39 vues

Transféré par ew1l

Jose Chabas Bergon, Bernard R. Goldstein Essays on Medieval Computational Astronomy

- George Tucker--A Voyage to the Moon
- SCIENCE POST TEST.docx
- nativity in the stars with improved graphics final
- science- project
- lib-smi-history-maps-34067-article and quiz
- Eclipses-libre.pdf
- ch16
- A forgotten Ptolemy: Harley Codex 3686 in the British Library
- Earth Not a Globe Review the Vol III No 4 5 July August 1896
- Ptolemy’s Map Projections and Coordinate Lists
- The Stony Brook Press - Volume 29, Issue 5
- Retrograde
- Sunny
- Ancient Egypt Astrology
- European Cartographers and the Ottoman World 1500 - 1750
- crew newsletter 10 28
- The Law of Vibration & the Archimedean Spiral
- Tides
- sc6_1998-live
- 10 Things You Didn’t Know Were Completely Misnamed

Vous êtes sur la page 1sur 422

texts and studies

Editors

Charles Burnett

Sacha Stern

Editorial Board

Leofranc Holford-Strevens Bernard R. Goldstein Alexander Jones

Daryn Lehoux Jrg Rpke Julio Sams Shlomo Sela John Steele

volume 5

Essays on Medieval

Computational Astronomy

By

Jos Chabs

Bernard R. Goldstein

leiden | boston

Cover illustration: Detail of the star chart in the Alfonsine Libro de las estrellas de la ochava espera in

Madrid, Real Academia de la Historia, ms 9/5707, f. 103. Courtesy Real Academia de la Historia .

Essays on Medieval computational astronomy / by Jos Chabs, Bernard R. Goldstein.

pages cm. (Time, astronomy, and calendars, ISSN 2211-632X ; volume 5)

Includes bibliographical references and index.

ISBN 978-90-04-28174-5 (hardback : alk. paper) ISBN 978-90-04-28175-2 (e-books) 1. Astronomy,

Medieval. 2. AstronomyTables. I. Goldstein, Bernard R. II. Title. III. Title: Medieval computational

astronomy.

QB26.C43 2015

521.01'51dc23

2014041332

This publication has been typeset in the multilingual Brill typeface. With over 5,100 characters covering

Latin, ipa, Greek, and Cyrillic, this typeface is especially suitable for use in the humanities. For more

information, please see www.brill.com/brill-typeface.

issn 2211-632X

isbn 978-90-04-28174-5 (hardback)

isbn 978-90-04-28175-2 (e-book)

Koninklijke Brill nv incorporates the imprints Brill, Brill Nijhoff and Hotei Publishing.

All rights reserved. No part of this publication may be reproduced, translated, stored in a retrieval system,

or transmitted in any form or by any means, electronic, mechanical, photocopying, recording or otherwise,

without prior written permission from the publisher.

Authorization to photocopy items for internal or personal use is granted by Koninklijke Brill nv provided

that the appropriate fees are paid directly to The Copyright Clearance Center, 222 Rosewood Drive,

Suite 910, Danvers, ma 01923, usa. Fees are subject to change.

Contents

Introduction 1

part 1

Conjunctions and Oppositions

Syzygy 40

Corpus: The Case of the Tables of Nicholaus de Heybech 57

part 2

Planetary Motions

Latitudes 73

5 Displaced Tables in Latin: The Tables for the Seven Planets for

1340 99

Corpus 150

part 3

Sets of Tables

(1320) 227

vi contents

Iberian Astronomical Tradition in the Late Middle Ages 338

part 4

Other Tables

Index 407

List of Figures

1.1 The mean conjunction ( s = m) at time t takes place after the true

conjunction at time t 11

1.2 The corresponding true conjunction (s = m) at time t takes place

before the mean conjunction at time t 12

1.3 Graphical representations of the entries in the five columns of Nicholaus

de Heybechs Table at intervals of 6 18

3.1 Facsimile of Nicholaus de Heybechs table (excerpt): Vienna,

Nationalbibliothek, ms 2440, f. 74v 59

3.2 Facsimile of tv 7 (excerpt): Madrid, Biblioteca Nacional, ms 3385,

f. 106r 62

3.3 Facsimile of tv 8 (excerpt): Madrid, Biblioteca Nacional, ms 3385,

f. 107r 63

4.1 The functions y =| cos x| and y = cos2 x in the interval (0, 180), where

the cosine function represents c5(x) in formulas (1) and (2) 91

5.a The solar equation displaced vertically 106

5.b Ptolemys second lunar model 110

5.c The model for Mars 120

5.d Mars, equation of center and equatio porcionis 123

5.e Mars, minutes of proportion 125

5.f First station of Mars 130

5.g Equation of anomaly near greatest distance for Mars 133

5.h Venus, equation of anomaly as a function of the true argument of

anomaly 140

6.1 Ptolemys model for Mars 152

6.2 The equation of center of Jupiter 170

7.1 The geometrical model underlying Ibn al-Kammds table for trepidation,

as reconstructed 209

8.1 A geometric interpretation of Vimonds tables for the mean motion for

Venus 253

8.2 Vimonds equation of center, col. 5, for Venus, Mars, Jupiter, and

Saturn 262

8.3 Vimonds equation of center, col. 5, for Mercury 262

8.4 Ptolemys model for the three superior planets and Venus 269

9.1 Correction for Saturn as a function of its mean argument of center 315

9.2 Correction for the Moon for a given age (20 days) 319

9.3 Correction for the Moon for a given value of mean lunar argument of

anomaly (4,0) 319

viii list of figures

12.2 Calendaric matters (Madrid, ms 3349, f. 3v) 394

12.3 Domiciles according to al-Brn 401

Introduction

During the Middle Ages and early modern times tables were a most success-

ful and economical way to present mathematical procedures and astronomical

models, facilitating computations based on them. One reason for depending

on these techniques was the absence of modern mathematical notation to

represent the algorithms that astronomers were using. Indeed, some major

sets of tables were the direct result of the development of new astronomical

approaches. In the second century ad Ptolemy composed the Almagest, a com-

prehensive handbook of astronomy and related mathematical procedures: in

it he presented a set of observations on the basis of which he determined the

parameters of his geometrical models for planetary motion. He then compiled

tables which were also included in this treatise. At a later date he revised these

tables, making them easier to use, in a work called the Handy Tables where, for

example, Ptolemy displayed entries at 1-intervals, rather than at 3- and 6-

intervals as he had done previously. Nicolaus Copernicus (14731543) provides

a somewhat different example of this pattern. In 1543 his magnum opus, De

revolutionibus, was published, in which he described a set of planetary mod-

els to replace those based on the Almagest, and included tables for computing

planetary positions with these new models. In 1551 Erasmus Reinhold (1511

1553) compiled the Prutenic Tables, based on the models of Copernicuss De

revolutionibus, presented in ways that facilitate computation. Most of the time,

however, the compilation of complete sets of tablesor of individual tables

just reflected partial changes in the parameters and the models underlying a

particular theory, or new methods to compute the positions of the celestial

bodies without changing the underlying models. This is particularly true in the

Middle Ages, when the Ptolemaic models were rarely challenged.

Astronomical tables are a basic component of astronomy, although they are

frequently neglected (or considered an unintelligible sequence of numbers)

because, more often than not, it is no simple matter to establish the structure

of such tables, and sometimes it is even difficult to identify the problem which

a particular table addresses. Yet, to understand astronomical tables and to

fully tackle the procedures used to compile them, various skills, both linguistic

and mathematical, are required at the same time. The papers included in this

volume give many examples where the meaning and purpose of such tables has

been determined by careful analysis.

As we have shown in our recent monograph, A Survey of European Astronom-

ical Tables in the Late Middle Ages (Brill, 2012), astronomical tables are a primary

source of historical information. Through their analysis it is possible to insert

2 introduction

pected links between authors. These analyses also allowed us to discover com-

putational techniques, interpolation methods, and approximation procedures,

as well as to identify changes in the standard parameters and the geometrical

models. It should be stressed that an analyst of a medieval table should only

appeal to modern mathematics very selectively and judiciously, given the risk

of falling into anachronisms that may lead to incorrect results. In particular,

modern mathematical functions are generally to be avoided although medieval

mathematical procedures expressed verbally may legitimately be translated

into algebraic notation. In recomputing a table consistency requires making

use only of computational techniques, concepts, and strategies that were avail-

able at the time. For example, in computing entries in a table for finding the

length of daylight on a given day, one needs to use trigonometrical tables that

can be found in the same set of tables or in a previous set; indeed, this is

how it was done by medieval astronomers. In other words, we have to evaluate

medieval scholars according to criteria consistent with their own time, which

may be different from those appropriate for other periods.

The pioneering work on this subject goes back to the beginning of the twen-

tieth century. In the years between 1899 and 1907 C.A. Nallino published an

edition, translation, and commentary on the zij (i.e., a set of astronomical

tables) of al-Battn (ca. 900). This work was extended in 1956 by E.S. Kennedy

in his survey of Islamic astronomical tables. Moreover, in 1962 O. Neugebauer

provided an extensive commentary on the zij of al-Khwrizm, originally com-

posed in Baghdad in the ninth century that is only extant in a Latin version

by Adelard of Bath (10751160), which was published by H. Suter in 1914. While

the publication and analysis of Islamic astronomical tables have continued to

this day, considerable attention has also been devoted to astronomical tables

in Latin, Byzantine Greek, Hebrew, and the vernacular languages in Europe,

including our own contributions to this field which have been built on those of

our distinguished predecessors. As a result of these scholarly efforts a critical

mass of studies is available which provides a good understanding of the overall

framework for the transmission of astronomical ideas and procedures through

tables in the period from Ptolemy to the early sixteenth century. The study of

many sets of tables in a variety of languages still to be examined in European

libraries (and others in different parts of the world) will surely lead to signif-

icant modifications of this framework in addition to an expansion of its con-

tents. In particular, there are at least three major topics concerning medieval

astronomy in the Latin West, where the analysis of tables is likely to play a crit-

ical role in enhancing our understanding of developments in this domain: the

early stages of astronomy in Latin Europe, based on Arabic materials in the

introduction 3

Iberian Peninsula and the Maghreb; the importance of several sets of tables in

Hebrew compiled in the Iberian Peninsula in the fourteenth and fifteenth cen-

turies; and the shaping of the Parisian Alfonsine Tables in the central decades

of the fourteenth century in Paris.

The twelve articles presented in this collection have been jointly written by

the two of us in a period covering about two decades. They are not the only

ones, for our collaboration has extended to many other papers and mono-

graphs, for the most part devoted to astronomical tables. This four-handed

collaboration, in the sense that all words and sentences as well as all numbers

and computations have been written and checked by both of us, has proved to

be fruitful and stimulating, an accomplishment that was achieved despite an

oceana few thousand kilometers wideseparating us.

The focus of attention in history of science is generally on scientists who

produced new theories, such as Kepler and Newton. In the Middle Ages, how-

ever, the innovations were made in the context of a long tradition. The essays

selected here address the major issues in medieval astronomy, and offer exam-

ples of several of the cleverest solutions in tabular form we have ever found.

The first group of essays concerns syzygies; specifically, the determination

of the time interval from mean to true syzygy, where mean syzygy refers to

a conjunction or opposition of the mean Sun with the mean Moon, and true

syzygy refers to a conjunction or opposition of the true Sun with the true Moon.

The concept of mean Sun and mean Moon (or the mean positions of the Sun

and the Moon) was used by Ptolemy in his Almagest, where the mean position

of the Sun, the Moon, or a planet is the place it would reach on the ecliptic if

its motion were uniform. This mean position is then correctedusing a table

derived from a geometrical modelto yield its true position at a given time.

Finding the time interval from mean to true syzygy is indeed a fundamental

problem in computational astronomy, for it is the first step in computing

the time and circumstances of a solar or lunar eclipse. Ptolemy had given

an approximate solution by means of an iterative procedure without a table,

and medieval astronomers devoted much effort to provide innovative, more

precise, and more user-friendly solutions. One impressive approach to this

problem is due to Nicholaus de Heybech (ca. 1400), who split the time from

mean to true syzygy into two terms, one for the Sun and the other for the Moon,

and managed to present his solution in the form of a single table in five columns

that makes computation easy to do in a few steps, for it requires only addition,

subtraction, and simple interpolation, rather than an iterative procedure with

many steps.

The second topic addressed here concerns the motion of the planets. The

history of tables for computing planetary latitude (that is, the angular distance

4 introduction

from a planet to the ecliptic) and using planetary equations (that is, the differ-

ence between the mean and true positions of a planet, where the Sun and the

Moon are included among the planets) is surveyed in two papers, and many

examples of their evolution are given. The other paper deals with the Tables for

the Seven Planets for 1340, compiled by an anonymous author working within

the framework of the Alfonsine Tables, who clearly had a profound understand-

ing of planetary astronomy. He succeeded in simplifying the astronomers task

by eliminating subtractions in the course of computing planetary positions.

This was very helpful at a time when negative numbers were not yet available,

for many complicated arithmetic rules were needed to express what can now

be represented algebraically using negative numbers. To accomplish this, the

compiler of these tables introduced what we call horizontal and vertical dis-

placements (following the terminology of E.S. Kennedy), such that the results

of the computations agree with those that depend on the standard Alfonsine

Tables, although the intermediate steps are all different.

In the third group we present different sets of tables by several authors,

beginning with the zij by Ibn al-Kammd (Crdoba, ca. 1100). In the Iberian

Peninsula astronomers were heirs to two traditions: a Greek tradition largely

based on the works of Ptolemy, and an Indian tradition introduced into the

Islamic world in the eighth and ninth centuries. The first is represented by the

zij of al-Battn, and the second by the zij of al-Khwrizm. Ibn al-Kammd

relied on both traditions, and we analyze each of his tables in terms of its

structure and sources. We then turn to the sets of tables by John Vimond

(ca. 1320) and John of Murs (ca. 1321), both of whom were active in Paris

and engaged in recasting the tradition of Alfonsine astronomy, specifically,

astronomical tables produced in Toledo in the thirteenth century under the

patronage of Alfonso x of Castile. The Parisian version of the Alfonsine Tables

is extant in many copies and it was the most influential set of tables in Latin

Europe during more than two centuries. Finally, we review the tables compiled

by Isaac Ibn al-adib (d. ca. 1426), who left Spain to live in Sicily in about 1396,

and whose tables, which concentrate on the motions of the two luminaries,

give innovative solutions. It is worth noting that the sets of tables by these four

authors were all available at the time in Latin manuscripts, even though the

work of Ibn al-Kammd was originally written in Arabic and that of Isaac Ibn

al-adib in Hebrew.

Finally, of the two papers listed under Other Tables, one is entirely devoted

to a list of 30 stars that depends on Ptolemys star catalogue. It was compiled

by Ibn al-Kammd, and it was surprisingly successful, for it is preserved in

texts in Arabic, Latin, and Hebrew. The other paper describes the contents of

a manuscript containing a mixture of tables of diverse origins and intended

introduction 5

for different purposes, some of which are strictly astrological. The tables, none

of them for planetary motion, reflect astronomical activity in Portugal during

the fourteenth century. Among the tables in this manuscript is one for the

daily lunar progress in an astrological month and another is for astrological

months, where 13 such months constitute a tropical year; this astrological

Moon (a fictitious celestial body) advances 390 in each astrological month.

We have found that these astrological tables appear in many medieval and

early modern sets of astronomical tables. For an explanation and historical

account of these astrological tables, see our article, Planetary velocities and

the astrological month, Journal for the History of Astronomy, 44 (2013), 465478.

Despite their great variety, underlying the tables reviewed in this collection

of essays are models for the Sun, the Moon, and the planets, which follow

essentially the Ptolemaic tradition in an ingenious and original way. Some of

them also provide elegant and intelligent solutions to standard astronomical

problems, within a general tendency to offer the practitioner user-friendly

tables that facilitate his task of computing planetary positions and celestial

events.

We thank the editors and publishers of the journals in which the chapters

of this book first appeared as independent essays for graciously granting us

permission to reprint them: bibliographic data for the original publication are

given in the first footnote of each chapter. Since it was decided to reset type for

these essays, we have taken the opportunity to make some minor changes in

them.

part 1

Conjunctions and Oppositions

chapter 1

His Table for Finding True Syzygy*

Introduction

The calculation of eclipses was a major task for medieval astronomers, and the

first step in this procedure was the determination of the time from mean to

true syzygy (where syzygy means either the conjunction or the opposition of

the Sun with the Moon). The basic approach given in Ptolemys Almagest was

refined by subsequent astronomers in various ways. In the Latin West of the

late Middle Ages the corpus of Alfonsine astronomy held a dominant position.

Of special interest in this tradition is the solution to this problem in Chap. 22

of John of Saxonys canons for the Alfonsine Tables (ca. 1327)1 as well as the

solution attributed to Nicholaus de Heybech of Erfurt (ca. 1400).2 The text

and table of Nicholaus de Heybech (see Appendix 1) have not been discussed

previously.3 lt will be shown that the method of Nicholaus de Heybech is much

1 Editio princeps, Ratdolt [1483]; cf. Poulle [1984, 17].

2 In a short note, Thorndike [1948] collected what little is known about Nicholaus de Heybech

of Erfurt. His principal text is the canon and table to find the time of true conjunction and

opposition of the Sun and the Moon. The date of composition of this text is uncertain but

some of the manuscripts of it were copied in the 1440s. Thorndike adds that most of the

manuscripts are of the fifteenth century but one manuscript, dated 1394, was in the Library of

Grenville Kane, Tuxedo Park, New York (the present location of that manuscript is in Prince-

ton). In addition to listing manuscripts of the text which will be discussed below, Thorndike

reports that ms Kln W* 178 (fol. 29v) contains a table of mean conjunctions for the years from

1384 to 1504 computed for the meridian of Paris and ascribed to a certain Nicholaus, sug-

gesting that it is the same author. In the manuscript we find the authors name in the phrase

given in the heading as composita per Nycolaum de Er* where the asterisk represents here

a balloon-shaped symbol. However, it is by no means certain that this author is to be iden-

tified with our Nicholaus. Moreover, Hartmann [1919, 1213] refers to a certain Nicholaus de

Heybech who was registered as a student at Erfurt University in 1421, and he suggests that

this might be our Nicholaus. According to Thorndike, in 1392 Nicholaus de Heybech com-

pleted a copy of Gerard of Cremonas Theorica Planetarum, preserved in ms Cues 213. From

this meager information we conclude that Nicholaus de Heybech was active circa 1400.

3 A translation of the text of Nicholaus de Heybech appears in Appendix ii.

10 chapter 1

simpler than that of John of Saxony. In the same period another way to solve

this problem by means of tables, introduced by Levi ben Gerson (d. 1344)

and followed by Jacob ben David Bonjorn (fl. ca. 1361), has recently received

extensive treatment.4 Levi ben Gersons method requires the use of a set of four

tables and calls for many more computations than the method of Nicholaus de

Heybech.

The Alfonsine tradition is identified with Alfonso x, king of Castile (Spain),

who reigned from 1252 to 1284. He sponsored much scientific activity, and a

number of astronomical works were written in, or translated into, Castilian

at that time. The Alfonsine Tables are a special case because they are only

extant in a Latin version produced in Paris in the 1320s, and derivatives from

it. lt has even been argued that there never was a Spanish version prior to the

Parisian one [see Poulle 1988]. The manuscript tradition of these tables is very

complex and has not been adequately examined. For this reason, it has been

customary to identify the Alfonsine Tables with those published by E. Ratdolt

in 1483 despite the difficulty in deciding which tables were intended to be so

designated by the Parisian group in the 1320s, let alone by the astronomers

serving under Alfonso x. One of the characteristic features of the Parisian

version (and the edition of 1483) is the division of the circle into physical

signs of 60 rather than natural signs of 30 that were prevalent in ancient

and medieval astronomy [Poulle 1988, 100]. We refer to the entire manuscript

tradition associated with these tables as the corpus of Alfonsine astronomy.

We shall first discuss Ptolemys method and then the method presented by

John of Saxony, neither of which is reduced to specific tables for this purpose.

Then we turn to the method of Nicholaus de Heybech, who successfully pre-

sented his solution to this problem in a single table. We begin by introducing

some definitions and notation.

Let s, s, m, and m be the true and mean longitudes of the Sun and the

Moon, respectively, at a mean syzygy that occurs at a given time t. The solar

correction (cs) and the lunar correction (cm) are defined as

cs() = s s

and

cm() = m m

4 On Levi ben Gersons method, see Goldstein [1974, 136144, 229241]; on Jacob ben David

Bonjorn see Chabs [1989, 2639, and 1991].

nicholaus de heybech and his table for finding true syzygy 11

after the true conjunction at time t (see Figure 1.2)

where is the mean solar anomaly and is the mean lunar anomaly. Note that

for 0 180, cs 0; and that in the simple lunar model, for 0 180,

cm 0. However, in Ptolemys complete lunar model cm depends on the mean

elongation = m s, as well as on . At mean syzygy 2 = 0, whereas at true

syzygy 2 = 0, where = m s.5

We add a prime symbol to all variables related to the true syzygy taking place

at a time t, corresponding to a mean syzygy that takes place at time t. Let t =

t t be the interval between the two events. When the true syzygy comes after

mean syzygy, t is positive and m < s. Figure 1.1 illustrates a mean conjunction

where m > s, and t is negative; Figure 1.2 illustrates the corresponding true

conjunction.

From the definition of t it follows that

(1)

5 For a discussion of Ptolemys solar and lunar models see, for example, Pedersen [1974, 122

202] and Neugebauer [1969, 191198].

12 chapter 1

t takes place before the mean conjunction at time t

(shown in Figure 1.1)

where the functions vm(t) and vs(t) are the time dependent velocities in longi-

tude of the Moon and the Sun. Since m = s at mean conjunction, Eq. (1) can

be written as

(2)

of t gave rise to a variety of approaches in the determination of the times of

true syzygies.

The functions vm(t) and vs(t) are not tabulated in Ptolemys Almagest. For the

calculation of t Ptolemy depended on the approximation vs(t)/vm(t) 1/13.6

Then Eq. (1) becomes

(3)

6 Almagest vi.4; Toomer [1984, 281]; Neugebauer [1975, 122]; Pedersen [1974, 221226].

nicholaus de heybech and his table for finding true syzygy 13

lunar anomaly by means of a relation which, expressed in modern notation, is

is the hourly mean lunar velocity in anomaly, and c = cm() cm( 1) is

the difference in the lunar correction corresponding to one degree of mean

anomaly at the time of mean syzygy. However, since Ptolemys table for the

lunar correction (Almagest iv.10) is arranged for intervals in the argument of

6 in the upper part and 3 in the lower part, the previous expression is stated

as c = (cm() cm( d))/d, where d is 6 or 3, as appropriate. At time t

+ t the true positions of the Sun and the Moon are to be computed anew.

If there is still a sensible difference between them, this procedure is to be

iterated.

If the time t at which a mean syzygy takes place is known, the values for all

the required quantities may be found and the time difference t from mean to

true syzygy may be derived by means of a successive approximation procedure

involving two principal steps.

1. First, it is necessary to compute a time interval which yields the time of true

syzygy, correct to the nearest hour. To do so, John of Saxony appeals to Ptolemys

complete lunar model in which Ptolemy introduces an equation of center q, a

function of . However, since q( ) = 0 at mean syzygy, the canons introduce an

equated lunar anomaly in order to determine a consistent value for vm. The

relation proposed is

(5)

where the mean lunar anomaly and the elongation are to be replaced by

their values at mean syzygy. In the text, the elongation = m s, is called

longitudo lune when > 0 and longitudo solis when < 0. The text offers

no justification for Eq. (5) but clearly its purpose is to account for the change

14 chapter 1

in lunar velocity in the time interval from mean syzygy to true syzygy. The

lunar velocity will now be treated as a function of rather than of , where

corresponds to the midpoint of the time interval from mean syzygy to true

syzygy.

To derive Eq. (5), let t be half the time interval from mean syzygy to true

syzygy and = v t, where v, the hourly mean lunar velocity in anomaly,

is 0;32,40 /h. In Eq. (5), the coefficient 13/24 (= 0;32,30) seems to represent an

approximation of v. We further assume that the velocity in elongation from

mean syzygy to true syzygy can be approximated by its mean value, v = 0;30,29

/h. It follows that the time, 2t, from mean syzygy to true syzygy equals /,

or 2; hence t = . Note that in Eq. (5) is therefore measured in hours. The

value of vm corresponding to is intended to represent the velocity of the

Moon at the midpoint of the time interval from mean syzygy to true syzygy.

The text assumes that the velocity of the Sun remains the same in this time

interval.

According to the canons, the time interval is then computed by means of

the relation

(6)

where = 0;0,1 < (|Int()| 1) and where Int() is the integer part of the

elongation at mean syzygy (in degrees). When = 0, Eq. (6) has the same

structure as Eq. (1); that is, the difference in lunar and solar longitudes is divided

by the difference in their velocities. Note, however, that in Eq. (6) lunar velocity

is a function of the equated anomaly defined in Eq. (5) rather than the mean

anomaly. In the text the difference between the lunar and solar velocities is

called superatio and the denominator of Eq. (6) is called superatio equata.

The solar velocity in the time from mean conjunction to true conjunction

is essentially constant; therefore, the term is intended to modify the lunar

velocity which changes noticeably in that same time interval. This term serves

the same function in the Toledan Tables and Azarquiels almanach, and John

of Saxony apparently followed the same tradition.8 This adjustment to lunar

8 Toomer [1968, 85] and Mills [19431950, 233] display a table of corrections to the hourly

lunar velocity corresponding to ; the argument varies from 1 to 7 and the entries from

0to 6. In alBattns zj a similar table is found [Nallino 1907, 88], but here the argument

varies from 1 to 7 and the corresponding entries from 1to 7. The tabulated function is thus

0;0,1 |Int()|. It would seem that for al-Battn this is a correction to the hourly velocity of

the Moon to account for the change in lunar velocity in the time interval from mean to true

nicholaus de heybech and his table for finding true syzygy 15

velocity is no longer needed because Eq. (5) satisfies the same purpose. Indeed,

has no sensible effect on the value of because its maximum of 0;0,6 is very

small compared with the least difference in lunar and solar velocities (0;27,45

according to the Toledan Tables, or 0;27,47 according to the edition of 1483 of

the Alfonsine Tables).9

The text presents rules for the algebraic sign of , but they are incomplete.

The text just considers the case where the longitude of the Moon is greater

than that of the Sun, i.e., > 0; this implies that true conjunction precedes

mean conjunction (t < 0), and that the lunar anomaly in that interval is always

smaller than its value at mean conjunction. Under this circumstance, the text

gives the correct rule for the algebraic sign of : if 180, then 0, and

vice versa. Unfortunately, the text uses the expression longitudo10 instead of

longitudo lune ( > 0) and omits any discussion of the case where < 0

(longitudo solis).

2. The second step requires the computation of a time interval * to yield the

time, correct to a minute of an hour, for the true syzygy: * + = t. In order to

do so, the canons propose a differential method that involves calculating and

comparing all the quantities at two different times 24 min apart (i.e., 1/60 of a

day): t + and t + + 0;1 d. For a given time dependent variable x, let

the time interval * is determined as

(7)

syzygy. Al-Battns maximum argument of 7 reflects the maximum value of which for

him is 7;0,10. The maximum entry in this table of 7is an average (where the range is from

0to 11) for the effect on lunar velocity in degrees per hour corresponding to a change in

lunar anomaly of about 3 (corresponding to half the time for the Moon to reach the

point of true conjunction).

9 Toomer [1968, 82]; Ratdolt [1483, fol. g6rg7r]. In his canons, John of Saxony mentions a

table by John of Lignres for solar and lunar velocity, but it is not clear which set of entries

is intended [cf. Poulle 1984, 82].

10 Ratdolt [1483, fol. b1r, line 5]; cf. Poulle [1984, 82, line 57].

16 chapter 1

where * is the elongation at time t + . The denominator of Eq. (7) can also be

written as

= (t + + 0;1) (t + )

= d.

computed according to Ptolemys complete lunar model rather than the simple

lunar model, both of which are represented in the Alfonsine Tables.11

The variation in elongation in a minute of a day, d, is very nearly the

instantaneous velocity in elongation at the time of true syzygy. Analogously,

dm represents the instantaneous velocity in lunar longitude in a minute of a

day.

To illustrate the method described in the canons to the Alfonsine Tables

by John of Saxony, we show in Table i some basic magnitudes (in degrees)

involved in the computation of the true conjunction corresponding to the

mean conjunction of 20 July 1327, 3;58,10 h after noon.12 True conjunction is

estimated to take place at time t + + * = 3;58,10 h 8;43,30 h + 0;9,49 h =

19;24,29 h (19 July 1327) in Toledo.

The Latin tradition of the Alfonsine Tables was mainly developed by the group

of astronomers who may be called the Johns (i.e., John of Murs, John of

Lignres, and John of Saxony), and also by such lesser known figures as John

of Genoa and John of Montfort who computed tables for lunar velocities.13

11 Ptolemys table for his complete lunar model appears in Almagest v.8 [Toomer 1984, 238].

The corresponding table (with modified parameters) appears in the Alfonsine Tables

[Ratdolt 1483, fol. e4re6v; cf. Poulle 1984, 148153].

12 Poulle [1984, 214218] presents a worked example according to John of Saxonys canons for

this conjunction. We have recomputed all the magnitudes involved (using the tables, as

he did, in the 1483 edition of the Alfonsine Tables) and our results differ somewhat from

his. For example, we compute the equation for the centrum lune according to John of

Saxony and the appropriate tables in [Poulle 1984] at time t + as 1;18,46 whereas he

found it to be 1;13,47. Despite the discrepancies, the resulting time of true conjunction

is essentially the same in both computations.

13 Nothing is known about John of Montfort except that, ca. 1332, he produced a table for

nicholaus de heybech and his table for finding true syzygy 17

according to the instructions by John of Saxony

t = 3;58,10 h t+ t + + *

(mean conj.) ( = 8;43,30 h) t + + 0;1 d (* = +0;9,49 h)

s 2, 3;40,35 2, 3;19,45 2, 3;20,42 2, 3;20, 8, 2

35;25, 4 35; 3,34 35; 4,34 35; 3,57

m 2, 4;53,30 2, 0; 6, 5 2, 0;19,16 2, 0;11,28,50

m 2, 8;26,14 2, 3;13,15 2, 3;28,37 2, 3;20, 7,57

3,42;26, 7 3,37;41, 8 3,37;54,12 3,37;46,28,42

3,42;26, 7 3,36;22,22 3,36;39, 6 3,36;29,12, 2

2 0 5,51; 8,10 5,51;32,33 5,51;18, 8,55

4;45,39 0; 5,29 0; 7,55 0; 0, 0, 5

Despite the differences of their tables in detail, all the Johns required two

kinds of tables for the determination of the time of true syzygy: tables of

correction and tables of velocity for each of the luminaries. However, some

manuscripts within the Alfonsine corpus contain a single table to solve this

problem where all entries are given in units of time. The author of that table,

according to its heading, is Nicholaus de Heybech of Erfurt, and this table is

usually accompanied by a short canon explaining its use. This canon adds that,

after the time of true syzygy is determined, a correction is required to take

into account the equation of time.14 Both the canon and the table survive in

solar and lunar velocities. John of Genoa was the author of a set of canons and tables for

eclipses (also dated 1332), and a computation for the solar eclipse of March 1337 according

to the Alfonsine Tables [cf. Thorndike and Kibre 1963, Cols. 51, 61, and 1690]. The tables for

lunar velocity by John of Montfort and by John of Genoa are analyzed in Goldstein [1992].

14 Nicholaus de Heybech refers to a table for the equation of time in the canon (see Appen-

dix i [18]). His allusion to a table for the equation of time is simply copied from John of

Saxonys canon (Chap. 22). This table appears in the 1483 edition of the Alfonsine Tables

(fol. k1rk2r) and has the heading Tabula elevationum signorum in circulo directo. It is

arranged for signs of 30, and each entry for the equation of time lies under each sign, as

it is described both by John of Saxony and Nicholaus de Heybech in Appendix 1: [18]. Note

that this table was taken from al-Battn [cf. Nallino 1907, 6164] and already appeared in

the Toledan Tables [cf. Toomer 1968, 3435], and that Poulle [1984, 6, 222] excluded it from

the Alfonsine Tables.

18 chapter 1

figure 1.3 Graphical representations of the entries in the five columns of Nicholaus de

Heybechs Table at intervals of 6

several Latin manuscripts, often together with the Alfonsine Tables and Jacob

ben David Bonjorns tables for syzygies.

Nicholaus de Heybechs canon and table are transcribed in Appendix i,

based on the following manuscripts:

D Dijon, bm 447, fol. 62rv (canon);

P Paris, bn Lat. 7287, fol. 72r73r (table), fol. 86va87ra (canon);

P Paris, bn Lat. 7290a, fol. 103r104r (table).

Thorndike and Kibre [1963, 1390, 1478, 1562] listed the first three as well as

three other manuscripts: Bern 454, Cues 211, and Vat. Pal. 1376. Most of the

information seems to have been taken from E. Zinner [1925]. This table seems

to have had a wide diffusion, for it is extant in quite a number of manuscripts

in addition to those already mentioned: Cracow, Jagiellonian Library, mss 609

(table and canon), 610 (table and canon), 613 (canon), 1852, and 1865; Princeton,

Library of Grenville Kane, ms 51; Vienna, Nationalbibliothek 2440 (table and

canon) and 5227 (table and canon).

The argument of the table is arranged at 1 intervals from 0 to 180 using

signs of 60. Column i has the heading equatio solis and the entries are

displayed in hours and minutes; column ii gives the diversitas equationis solis

in hours and minutes; column iii lists the minuta proportionalia; column iv

displays the equatio lune in hours and minutes; and column v gives the

diversitas equationis lune in minutes of an hour. Figure 1.3 illustrates the five

columns at 6 intervals.

nicholaus de heybech and his table for finding true syzygy 19

The structure of the table suggests that the time interval between mean and

true syzygy given by Eq. (2) is here distributed into two terms, each of which

accounts for the role of one luminary:

(8)

The calculation of the first, or solar term (ts), involves columns i, ii, and

iii, and that of the second, or lunar term (tm), involves columns iii, iv, and

v. The strategy consists in treating each term separately, and within each

term computing a set of minimum and maximum values and then to use an

interpolation scheme for intermediate values.

The solar term can be approximated as

(9)

where vs is the mean solar velocity (/h) and c3() is a coefficient of inter-

polation ranging from 0 to 1, depending on the mean lunar anomaly, , and

tabulated in col. iii in minutes. When = 0, the Moon reaches its minimum

velocity, min(vm), and c3() = 0. Hence Eq. (9) reduces to

(10)

Col. i. Analogously, when = 180, the Moon reaches its maximum velocity,

max(vm), and c3() = 1. In this case, Eq. (9) reduces to

(11)

The values computed by means of the term in brackets in Eq. (9) correspond

to those tabulated in Col. ii. Let c1 represent an entry in Col. i and c2 an entry

in Col. ii of Nicholaus de Heybechs table: c1 and c2 are both functions of the

mean solar anomaly, . Then

(13)

20 chapter 1

the mean solar anomaly, . Column iii serves as the interpolation scheme for

both the Sun and the Moon. When = 0, the Sun is at its minimum velocity,

min(vs), and c3() = 0. Hence Eq. (13) reduces to

(14)

Col. iv. Analogously, when = 180, the Sun is at its maximum velocity, max(vs),

and = 1. In this case, Eq. (13) reduces to

(15)

The values computed by means of the term in brackets in Eq. (13) correspond

to those tabulated in Col. v. Let c4 represent an entry in Col. iv and c5 an entry

in Col. v of Nicholaus de Heybechs table: c4 and c5 are both functions of the

mean lunar anomaly . Then

Finally, from Eqs. (12) and (16) we find that the time from mean to true syzygy

is

(17)

In this equation the algebraic signs of c1() and c2() are the same as the

corresponding sign of c3(), and those of c4() and c5() are the same as that of

cm().

In the canon by Nicholaus de Heybech we are first told (sentences [5][8])

to compute the six quantities in Eq. (17); note that the sign convention for c4()

and c5() differs from the one we have used. The rule for computing the sec-

ond term is found in sentence [9], and sentence [10] tells us to compute the

difference between the first and the second terms, that is, the solar term, ts. In

sentence [11] we are told to compute the fourth term, and sentence [12] tells us

to compute the lunar term, that is, tm (according to our convention). In sen-

tences [13] and [14] we are given the rules for adding the solar and lunar terms.

For the example given above (the mean conjunction of 20 July 1327), we find

c4 = +6;24 h c5 = +0;02 h c3() = 0;53,

nicholaus de heybech and his table for finding true syzygy 21

is 8;34 h, whereas the value according to John of Saxonys method is 8;33,41

h.

It is noteworthy that with the table of Nicholaus de Heybech essentially

the same result is obtained with much less work; we also note that his entire

procedure is in units of time restricted to two sexagesimal places. In contrast,

John of Saxonys procedure involves the computation of many intermediate

quantities, most of which are not in units of time, and he does not specify the

degree of precision of these quantities needed to assure the minutes of time in

the ultimate result.

Equation (17) incorporates the instructions in the canon of Nicholaus de

Heybechs table, and we have seen that, at least for the example presented

above, the entries are reasonable. As was frequently the custom among medie-

val astronomers, Nicholaus de Heybech does not describe his method for

computing the entries in his table. We now present our reconstruction of his

method for computing the table.

To recompute the entries in column i, we need to fix a set of values in Eq. (10).

For the solar correction we use the values in the 1483 edition of the Alfonsine

Tables; we consider min(vm) = 0;29,37,13 /h and vs = 0;2,27,51 /h.

The first parameter is the minimum lunar velocity found in John of Genoas

table, arranged at intervals of 6 and given to 3 sexagesimal places.15 The value

for the mean solar velocity vs is derived from the table for mean solar motion

in the editio princeps of the Alfonsine Tables.

To recompute the entries in Col. ii, we also have to fix the value for the

maximum lunar velocity in Eq. (11): max(vm) = 0;36,58,54 /h, which is found in

the same table by John de Genoa mentioned above.16 The values in Col. ii are

computed as the difference between those in Col. i and those found by means

of Eq. (11) with the same argument. These parameters yield the entries in Col. i

and Col. ii quite closely (see Table ii). It should be stressed that, restricting our

attention to attested parameters, we obtain best agreement with the entries

in column ii using John of Genoas velocity table which contains an isolated

15 The Alfonsine corpus has many variant tables for lunar velocity (see Goldstein [1992]) and

there is no reason to assume that Nicholaus de Heybech would have used the version in

the editio princeps of the Alfonsine Tables in which the minimum value is 0;30,21/h.

16 We use this value for the lunar velocity in John of Genoas table which appears in many

manuscripts, and hence it is textually secure. Nevertheless, it may have been miscomputed

according to an analysis of the entire table: see Goldstein [1992]. Other values for the max-

imum lunar velocity in the Alfonsine corpus are: 0;36,4/h [cf. Goldstein 1980], 0;36,25/h

[Ratdolt 1483, fol. g7r] and 0;36,53/h, ascribed to John of Lignres [cf. Goldstein 1992].

22 chapter 1

is consistent with the rest of the table). The fact that this value is so much

greater than any other strongly supports the claim that Nicholaus de Heybech

indeed depended on this table by John of Genoa. As we shall see, the entries

in column iv in Nicholaus de Heybechs table also agree with recomputation

based on this table by John of Genoa.

The recomputation of columns i and ii of Nicholaus de Heybechs table is

only displayed here for values of the argument in multiples of 6, for it is likely

that he derived the other entries by distributing them uniformly between the

values calculated at 6 intervals. For instance, in Col. i after finding 28 min for

= 6, one way to proceed would be to distribute 28 into six parts according

to the sequence 555544 (leading to the values 5, 10, 15, 20, 24, and 28,

that actually appear in Nicholaus de Heybechs table). After finding 57 min for

= 12, one can distribute 57 28 = 29 into 6 parts according to the sequence

555554 (leading to the values in Nicholaus de Heybechs table: 33, 38, 43,

48, 53, and 57). These sequences generate smooth curves within each interval

of 6, but they are no longer smooth when considered as a whole.

The results are shown in Table ii, where c is the recomputed value and tc

is its difference from the value given in the text.

We now argue that the coefficient c3, tabulated in column iii (minuta

proportionalia), has been recomputed according to the formula

(18)

at apogee), d = 60 e (its distance at perigee), d() is its distance for a true

anomaly, , and e is the eccentricity.17 In a simple eccentric model, this inter-

polation coefficient is a function of and e, but, as is shown in Table iii, c3 does

not depend strongly on the eccentricity, for similar results are obtained using

2;16 (the solar eccentricity corresponding to the maximum solar correction in

the Alfonsine Tables: 2;10) or 5;10 (the lunar eccentricity corresponding to the

maximum lunar correction in these tables: 4;56). This justifies the use of the

same interpolation coefficient in both the solar and lunar terms of t (Eqs. 9

and 13). Note also that while the entries in Col. iii have been computed with the

true anomaly as argument, the instructions by Nicholaus de Heybech indicate

17 The argument for the distance is here the true anomaly in an eccentric model, whereas

in the Almagest vi.8 [cf. Toomer 1984, 308] Ptolemy presents a similar interpolation table

based on the mean anomaly in an eccentric model [cf. Toomer 1984, 654].

nicholaus de heybech and his table for finding true syzygy 23

Nicholaus de Heybechs table

Arg. i c tc ii c tc

h h min h h min

6 0;28 0;28,28 0 0; 6 0; 6, 4 0

12 0;57 0;56,54 0 0;12 0;12, 9 0

18 1;25 1;25, 4 0 0;18 0;18, 8 0

24 1;53 1;52,50 0 0;24 0;24, 4 0

30 2;19 2;18,58 0 0;30 0;29,38 0

36 2;43 2;43,25 0 0;35 0;34,50 0

42 3; 6 3; 5,58 0 0;40 0;39,40 0

48 3;27 3;27,39 1 0;44 0;44,17 0

54 3;47 3;47,21 0 0;48 0;48,29 0

60 4; 4 4; 4,40 1 0;51 0;52,11 1

66 4;18 4;18,55 1 0;54 0;55,13 1

72 4;29 4;29,38 1 0;57 0;57,30 1

78 4;38 4;37,57 0 0;59 0;59,17 0

84 4;44 4;43,48 0 1; 1 1; 0,31 0

90 4;47 4;47, 7 0 1; 1 1; 1,14 0

96 4;47 4;46,54 0 1; 1 1; 1,11 0

102 4;43 4;43, 2 0 1; 0 1; 0,22 0

108 4;37 4;36,51 0 0;59 0;59, 3 0

114 4;27 4;27, 1 0 0;57 0;56,57 0

120 4;14 4;13,58 0 0;54 0;54, 9 0

126 3;58 3;57,53 0 0;50 0;50,44 1

132 3;39 3;38,37 0 0;46 0;46,37 1

138 3;18 3;18, 0 0 0;42 0;42,13 0

144 2;55 2;54,13 +1 0;38 0;37, 9 +1

150 2;29 2;28,17 +1 0;32 0;31,37 0

156 2; 1 2; 0,14 +1 0;26 0;25,39 0

162 1;32 1;31,52 0 0;20 0;19,35 0

168 1; 3 1; 2,34 0 0;14 0;13,21 +1

174 0;32 0;31,38 0 0; 7 0; 6,45 0

180 0; 0 0 0 0; 0 0 0

24 chapter 1

Nicholaus de Heybechs table

min min min

10 1 0 0

20 2 0 0

30 4 0 0

40 7 0 1

50 11 0 1

60 15 0 1

70 20 0 1

80 25 0 1

90 31 0 0

100 36 0 1

110 42 1 +1

120 47 2 +1

130 51 1 +1

140 54 1 0

150 56 0 0

160 58 0 0

170 60 0 0

180 60 0 0

that the argument for Col. iii is the mean anomaly. Indeed, these instructions

do not suggest that the true anomalies need be computed. In any event, the

differences between entries computed with true, rather that mean, anomaly

are small (less than 3 min in the worst case), and so there is little practical effect

on the final result.

For our recomputation of Cols. iv and v, we have taken the values of the lunar

correction cm() from the 1483 edition of the Alfonsine Tables, vm() from John

of Genoas lunar velocity table, and from his solar velocity table the following

values for the maximum and minimum velocities: max(vs) = 0;2,33,40 /h and

min(vs) = 0;2,22,30 /h.18

18 See, for example, Paris BnF Lat. 7282, fol. 129rv [cf. Poulle 1984, 210].

nicholaus de heybech and his table for finding true syzygy 25

These parameters are used to evaluate the time given by Eq. (14) which

corresponds to Col. iv and the difference between it and that resulting from

Eq. (15) with the same argument. It would seem that Nicholaus de Heybech

increased the argument of lunar anomaly by 6 in calculating the lunar velocity

in between mean syzygy and true syzygy. Hence Eq. (14) becomes

(19)

(20)

Again, the recomputation is only displayed for the values of the argument

in multiples of 6. The results are shown in Table iv. The recomputed value

according to Eq. (19) is c, and t c is its difference from the value given in the

text. The shift of 6 significantly diminishes the discrepancy between the text

and the recomputed values, with the result that a value in the text is always

within 1 min of time of the recomputed value. We note again that John of

Genoas table for lunar velocity is arranged at intervals of 6 so that a shift of 6

is a shift of a line in the table.

Column v is recomputed as the difference between the time found in col-

umn iv and the time found by means of Eq. (20) with the same argument.

Table v displays Col. v together with our recomputed values (the differences

between the times found by means of Eq. (19) and Eq. (20)). Since the entries

in Col. v are small and have a very limited range, the rounding procedure plays

a more significant role than in other columns. For this reason we have not dis-

played the differences between text and recomputation.

Conclusion

Our principal goal in studying the canon and the table of Nicholaus de Heybech

of Erfurt was to understand their use in finding the time from mean syzygy to

true syzygy, as well as to reconstruct his table. It turned out that it works very

well for its intended purpose and is quite easy to use.

When we review some of the different approaches prior to that of Heybech,

we see that Ptolemy had already solved the problem by means of an iterative

procedure without recourse to any tables. On the other hand, John of Saxonys

method took advantage of solar and lunar velocity tables, while also intro-

ducing what we have called a differential method for computing an instan-

26 chapter 1

Heybechs table

cm() +6 vm(+6) iv c tc

/h h h min

12 0;56,41 18 0;29,46,17 2; 5 2; 4,59 0

18 1;24,27 24 0;29,51,51 3; 5 3; 5,35 1

24 1;51,27 30 0;29,59,31 4; 3 4; 3,47 1

30 2;17,29 36 0;30, 8,34 4;59 4;59, 5 0

36 2;42,21 42 0;30,19,43 5;50 5;50,48 1

42 3; 5,46 48 0;30,32,15 6;38 6;38,24 0

48 3;27,30 54 0;30,46,53 7;21 7;21,10 0

54 3;47,20 60 0;31, 2,55 7;59 7;58,48 0

60 4; 5, 4 66 0;31,18,55 8;31 8;31,22 0

66 4;20,27 72 0;31,37,45 8;58 8;57,36 0

72 4;33,18 78 0;31,57,15 9;18 9;17,53 0

78 4;43,28 84 0;32,16,45 9;32 9;32,19 0

84 4;50,41 90 0;32,39,45 9;40 9;39,25 +1

90 4;54,54 96 0;33, 8,31 9;40 9;38,36 +1

96 4;55,56 102 0;33,30,36 9;34 9;33,37 0

102 4;53,38 108 0;33,49,21 9;23 9;23,34 1

108 4;48,10 114 0;34,10,19 9; 7 9; 6,58 0

114 4;39,15 120 0;34,35,24 8;44 8;43, 7 +1

120 4;27, 0 126 0;34,58,23 8;15 8;14,16 +1

126 4;11,23 132 0;35,18,25 7;41 7;40,39 0

132 3;52,47 138 0;35,38,47 7; 3 7; 2, 9 +1

138 3;31, 3 144 0;35,57,36 6;20 6;19, 9 +1

144 3; 6,35 150 0;36,13,37 5;33 5;32,32 0

150 2;39,35 156 0;36,28,15 4;42 4;42,22 0

156 2;10,26 162 0;36,38, 0 3;49 3;49,41 1

162 1;39,27 168 0;36,46,22 2;54 2;54,25 0

168 1; 7, 6 174 0;36,53,15 1;57 1;57,17 0

174 0;33,47 180 0;36,58,54 0;59 0;58,53 0

180 0 0; 0 0 0

nicholaus de heybech and his table for finding true syzygy 27

of Nicholaus de Heybechs

table

v c Eq.(19) Eq.(20)

min min h h

12 0 0;50 = 2; 4,59 2; 4, 9

18 1 1;15 = 3; 5,35 3; 4,20

24 2 1;39 = 4; 3,47 4; 2, 8

30 2 2; 1 = 4;59, 5 4;57, 4

36 3 2;20 = 5;50,48 5;48,28

42 3 2;38 = 6;38,24 6;35,46

48 3 2;53 = 7;21,10 7;18,17

54 3 3; 6 = 7;58,48 7;55,42

60 3 3;17 = 8;31,22 8;28, 5

66 4 3;25 = 8;57,36 8;54,11

72 4 3;29 = 9;17,53 9;14,24

78 4 3;34 = 9;32,19 9;28,45

84 4 3;34 = 9;39,25 9;35,51

90 4 3;30 = 9;38,36 9;35, 6

96 4 3;20 = 9;33,37 9;30,17

102 4 3;20 = 9;23,34 9;20,14

108 4 3;12 = 9; 6,58 9; 3,46

114 3 3; 1 = 8;43, 7 8;40, 6

120 3 2;50 = 8;14,16 8;11,26

126 3 2;35 = 7;40,39 7;38, 1

132 2 2;22 = 7; 2, 9 6;59,47

138 2 2; 6 = 6;19, 9 6;17, 3

144 2 1;50 = 5;32,32 5;30,42

150 2 1;33 = 4;42,22 4;40,49

156 2 1;15 = 3;49,41 3;48,26

162 1 0;57 = 2;54,25 2;53,28

168 1 0;38 = 1;57,17 1;56,39

174 0 0;19 = 0;58,53 0;58,34

180 0 0 =0 0

28 chapter 1

modern definition of a differential would require the evaluation of a function at

two moments separated by an infinitesimal time interval. A value for the veloc-

ity very close to that obtained by the modern definition can be computed by

taking a sufficiently small time interval, and John of Saxony set that interval as

0;1 days.

Less than a century after John of Saxony presented his method, Nicholaus de

Heybech offered a much simpler alternative without sacrificing the accuracy of

the previous methods. First, instead of requiring many steps of computation,

some of which called for tables while other did not, Heybech presented his

solution in the form of a single table whose entries are given in time and whose

arguments are the mean positions of the Sun and the Moon at mean syzygy,

the initial values for all methods of computation. Second, the use of the table

involves very simple arithmetic operations in contrast to the complexity of John

of Saxonys method. Nicholaus de Heybech has produced a user-friendly table

for which he deserves much credit. In this respect he is following Ptolemy who,

for example, established tables for finding planetary positions that replaced

complex trigonometric computations.

Both John of Saxony and Nicholaus de Heybech showed real insight into

the ways Ptolemys models work and, on this basis, they were able to facilitate

the computation of certain astronomical phenomena. This example shows that

astronomers in the Middle Ages could make significant contributions without

introducing any modification of the models for the motions of the celestial

bodies.

Finally, we wish to emphasize that it is inappropriate to define the corpus of

Alfonsine Tables by the tables as they appear in the editio princeps, or by the

canons of John of Saxony, or even to assign either of them any privileged status.

Indeed, it has long been known that there is much more in the vast number

of manuscripts within this corpus, but little attention has been paid to their

contents. In this case study we have indicated the significance of Nicholaus de

Heybechs table and its dependence on a lunar velocity table by John of Genoa,

neither of which is included in the restricted definition of the Alfonsine Tables.

It is reasonable to expect that similar studies of other parts of this unexplored

corpus would reveal hitherto unsuspected riches.

nicholaus de heybech and his table for finding true syzygy 29

Appendix i

a Text in Latin

The text presented here is a transcription of Dijon bm 447, fol. 62rv (D). Variant

readings from other manuscripts are noted: Basel f.11.7, fol. 37v (B) and Paris

BnF Lat. 7287, fol. 86va87ra (P). Sentence numbers and punctuation have been

added by the editors.

[1] Tempus uere coniunctionis et oppositionis solis et lune per tabulas a mag-

istro nicholao de heybech de erfordia compositas inuenire. [2] Scias tem-

pus medie coniunctionis si uolueris coniunctionem ueram, uel tempus medie

oppositionis si uolueris oppositionem ueram. [3] Scias eciam argumentum

solis et argumentum lune ad idem tempus. [4] Quibus scitis et habitis intra

primo cum argumento solis in tabulam equationis temporis uere coniunctio-

nis et oppositionis luminarium. [5] Et accipe equationem solis et scribe super

eam m si argumentum solis sit minus tribus signis phisicis, uel scribe a si sit

plus tribus signis. [6] Accipe eciam ibidem diuersitatem equationis solis atque

minuta proportionalia et scribe super ea solis. [7] Quibus habitis et seruatis

intra postea cum argumento lune in easdem tabulas in lineis numeri et accipe

equationem lune in directo existentem et scribe super am a si argumentum

sit minus tribus signis, uel m si sit plus. [8] Accipe eciam ibidem diuersi-

tatem equationis lune et minuta proportionalia et scribe super ea lune. [9]

Istis habitis et seruatis accipe partem proportionalem de diuersitate equatio-

nis solis secundum proportionem minutorum lune ad 60. [10] Quam partem

proportionalem minue ab equatione solis prius seruata et habebis eam bene

equatam. [11] Accipe similiter partem proportionalem de diuersitate equatio-

nis lune secundum proportionem minutorum proportionalium solis ad 60. [12]

Quam partem proportionalem adde super equationem lune prius inuentam

et habebis eam bene equatam. [13] Postea uide si super ambas equationes,

scilicet solis et lune, scriptum sita; tunc adde eas simul cum tempore medie

coniunctionis prius inuento et seruato; sed si super utramque equationem

scriptum fuerit m, tunc minue eas similiter ab eodem tempore. [14] Si uero

super unam scriptum fuerit a et super aliam m, tunc minue minorem equa-

tionem a maiori et residuum adde cum tempore predicto si super maiorem

scriptum fuerit a, uel minue si super maiorem scriptum fuerit m. [15] Et sic

habebis ueram coniunctionem solis et lune si operatus es ad coniunctionem,

et similiter ueram oppositionem si operatus es ad eam reperiendam, et hoc

diebus non equatis, ad meridianum loci ad quem predicta inquirebas. [16] Et

30 chapter 1

scire uolueris, oportet equare dies. [18] Intra ergo cum uero loco solis in tab-

ulam equationis dierum cum suis noctibus et inuenies in directo graduum

equationem dierum, scilicet sub signo illo in quo est sol. [19] Et si inueneris

ibi gradus et minuta accipe pro quolibet gradu quatuor minuta hore et pro

quolibet minuto quatuor 2a hore, que adde cum tempore predicto quodlibet

ad suum genus et proueniet tempus uere coniunctionis uel oppositionis diebus

equatis. [20] Et cum illo debet queri ascendens et reliqua que pertinent ad fig-

uram.

Title: canon heybech] om. B; post tabulas add. P coniunctionis solis et lune.

[1] per tabulas compositas] om. B.

[3] eciam] igitur P; idem] iddem B.

[5] sit] om. P.

[7] in easdem] easdem B D P.

[9] post minutorum add. B proportionalium. [13] inuento et] om. B; similiter] simul B.

[14] equationem] om. P; maiorem 1 ipsam B et add. maiorem s. l.; maiorem 2] ipsam B.

[17] si] si uero B.

[18] ergo] igitur B; graduum] graduum ascendentem P; equationem] equationum B.

[20] post figuram add. D et sic est finis secundum uasseun, add. P Et sic finis scripti Maruanti per

me Mertinj anno domini 1447 12 die Julij.

True Syzygy

The table presented here is a transcription of Basel f.ll.7, fol. 36r37r (B). Variant

readings from two other manuscripts are noted: Paris BnF Lat. 7287, fol. 72r73r

(P) and BnF Lat. 7290a, fol. 103r104r (P). Column iii is found before Col. ii in

P and P.

mss P and P share 28 errors (23 entries of Col. iv are shifted upward one

place); P exhibits 5 additional errors and P just one. Therefore, we suggest that

P was copied from P.

Title

solis et lune ordinata per Nycholaum dictum de Heybech de Erfordia

(f. 36r)

P Tabula equationum uere coniunctionis et oppositionis solis et lune (signs

are used for the last four words) secundum magistrum Nycholaum de

Heybech de Erfordia (f. 72v)

nicholaus de heybech and his table for finding true syzygy 31

cundum magistrum Nycholaum de Heybech de Erfordia (f. 103r)

Column Headings

i Equatio solis

ii Diversitas equationis solis

iii Minuta proportionalia

iv Equatio lune

v Diversitas equationis lune

Argument i ii iii iv v

s /s h min h min min h min min

0 1/5 59 0 5 0 1 0 0 11 0

0 2/5 58 0 10 0 2 0 0 22 0

0 3/5 57 0 15 0 3 0 0 33 0

0 4/5 56 0 20 0 4 0 0 43 0

0 5/5 55 0 24 0 5 0 0 53 0

0 6/5 54 0 28 0 6 0 1 3 0

0 7/5 53 0 33 0 7 0 1 14 0

0 8/5 52 0 38 0 8 0 1 24 0

0 9/5 51 0 43 0 9 1 1 34 0

0 10/5 50 0 48 0 10 1 1 45 0

0 11/5 49 0 53 0 11 1 1 55 0

0 12/5 48 0 57 0 12 1 2 5 0

0 13/5 47 1 2 0 13 1 2 15 0

0 14/5 46 1 7 0 14 1 2 25 1

0 15/5 45 1 12 0 15 2 2 35 1

0 16/5 44 1 17 0 16 2 2 45 1

0 17/5 43 1 21 0 17 2 2 55 1

0 18/5 42 1 25 0 18 2 3 5 1

0 19/5 41 1 30 0 19 2 3 14 1

0 20/5 40 1 35 0 20 2 3 24 1

0 21/5 39 1 40 0 21 3 3 34 1

0 22/5 38 1 44 0 22 3 3 43 2 P, P: 1

0 23/5 37 1 49 0 23 3 3 53 2

0 24/5 36 1 53 0 24 3 4 3 2

32 chapter 1

(cont.)

Argument i ii iii iv v

s /s h min h min min h min min

0 25/5 35 1 58 0 25 3 4 12 2

0 26/5 34 2 3 0 26 4 4 22 2

0 27/5 33 2 7 0 27 4 4 31 2

0 28/5 32 2 11 0 28 4 4 41 2

0 29/5 31 2 15 0 29 4 4 50 2

0 30/5 30 2 19 0 30 4 4 59 2

0 31/5 29 2 23 0 31 5 5 8 2

0 32/5 28 2 27 0 32 5 5 17 2

0 33/5 27 2 31 0 33 5 5 26 2

0 34/5 26 2 35 0 34 5 5 34 3

0 35/5 25 2 39 0 35 6 5 42 3

0 36/5 24 2 43 0 35 6 5 50 3

0 37/5 23 2 47 0 36 6 5 58 3

0 38/5 22 2 51 0 37 7 6 6 3

0 39/5 21 2 55 0 38 7 6 14 3

0 40/5 20 2 59 0 39 7 6 22 3

0 41/5 19 3 3 0 40 8 6 30 3

0 42/5 18 3 6 0 40 8 6 38 3

0 43/5 17 3 10 0 41 8 6 46 3

0 44/5 16 3 14 0 42 9 6 53 3

0 45/5 15 3 17 0 42 9 7 0 3

0 46/5 14 3 20 0 43 9 7 7 3

0 47/5 13 3 24 0 44 10 7 14 3

0 48/5 12 3 27 0 44 10 7 21 3

0 49/5 11 3 31 0 45 10 7 28 3

0 50/5 10 3 35 0 46 11 7 35 3

0 51/5 9 3 38 0 46 11 7 41 3

0 52/5 8 3 41 0 47 11 7 47 3

0 53/5 7 3 44 0 48 12 7 53 3

0 54/5 6 3 47 0 48 12 7 59 3

0 55/5 5 3 50 0 49 12 8 5 3

0 56/5 4 3 53 0 49 13 8 11 3

0 57/5 3 3 56 0 50 13 8 16 3

0 58/5 2 3 59 0 50 14 8 21 3

nicholaus de heybech and his table for finding true syzygy 33

Argument i ii iii iv v

s /s h min h min min h min min

0 59/5 1 4 2 0 51 14 8 26 3

1 0/5 0 4 4 0 51 15 8 31 3

1 1/4 59 4 6 0 52 15 8 36 3

1 2/4 58 4 9 0 52 16 8 41 3

1 3/4 57 4 11 0 52 16 8 46 4

1 4/4 56 4 14 0 53 17 8 50 4

1 5/4 55 4 16 0 53 17 8 54 4

1 6/4 54 4 18 0 54 18 8 58 4

1 7/4 53 4 20 0 54 18 9 2 4

1 8/4 52 4 22 0 55 19 9 6 4

1 9/4 51 4 24 0 56 19 9 9 4 P: 55

1 10/4 50 4 26 0 56 20 9 12 4

1 11/4 49 4 28 0 57 20 9 15 4 P: 56

1 12/4 48 4 29 0 57 21 9 18 4

1 13/4 47 4 31 0 57 21 9 21 4

1 14/4 46 4 32 0 58 22 9 24 4

1 15/4 45 4 34 0 58 22 9 26 4

1 16/4 44 4 35 0 58 23 9 28 4

1 17/4 43 4 37 0 59 23 9 30 4 P: 24

1 18/4 42 4 38 0 59 24 9 32 4

1 19/4 41 4 39 0 59 24 9 34 4

1 20/4 40 4 40 0 59 25 9 36 4

1 21/4 39 4 41 1 0 26 9 37 4

1 22/4 38 4 42 1 0 26 9 38 4

1 23/4 37 4 43 1 0 27 9 39 4

1 24/4 36 4 44 1 1 28 9 40 4

1 25/4 35 4 45 1 1 28 9 40 4

1 26/4 34 4 45 1 1 29 9 40 4

1 27/4 33 4 46 1 1 30 9 40 4

1 28/4 32 4 46 1 1 30 9 40 4

1 29/4 31 4 47 1 1 31 9 40 4

1 30/4 30 4 47 1 1 31 9 40 4

1 31/4 29 4 47 1 1 32 9 39 4

1 32/4 28 4 47 1 1 32 9 38 4

34 chapter 1

(cont.)

Argument i ii iii iv v

s /s h min h min min h min min

1 33/4 27 4 47 1 1 33 9 37 4

1 34/4 26 4 47 1 1 33 9 36 4

1 35/4 25 4 47 1 1 34 9 35 4

1 36/4 24 4 47 1 1 34 9 34 4

1 37/4 23 4 46 1 1 35 9 33 4

1 38/4 22 4 46 1 1 35 9 31 4

1 39/4 21 4 45 1 1 36 9 29 4

1 40/4 20 4 45 1 0 36 9 27 4

1 41/4 19 4 44 1 0 37 9 25 4

1 42/4 18 4 43 1 0 37 9 23 4

1 43/4 17 4 42 1 0 38 9 21 4

1 44/4 16 4 41 1 0 38 9 19 4

1 45/4 15 4 40 0 59 39 9 16 4

1 46/4 14 4 39 0 59 39 9 13 4

1 47/4 13 4 38 0 59 40 9 10 4

1 48/4 12 4 37 0 59 40 9 7 4

1 49/4 11 4 36 0 58 41 9 4 4

1 50/4 10 4 34 0 58 42 9 0 4

1 51/4 9 4 32 0 58 42 8 56 4

1 52/4 8 4 30 0 57 43 8 52 3

1 53/4 7 4 29 0 57 43 8 48 3

1 54/4 6 4 27 0 57 44 8 44 3

1 55/4 5 4 25 0 56 44 8 40 3

1 56/4 4 4 23 0 56 45 8 35 3

1 57/4 3 4 21 0 55 45 8 30 3

1 58/4 2 4 19 0 55 46 8 25 3

1 59/4 1 4 17 0 54 46 8 20 3

2 0/4 0 4 14 0 54 47 8 15 3 P, P: 16

2 1/3 59 4 12 0 53 47 8 10 3

2 2/3 58 4 10 0 52 48 8 4 3

2 3/3 57 4 7 0 52 48 7 59 3

2 4/3 56 4 4 0 51 49 7 53 3

2 5/3 55 4 1 0 50 49 7 47 3 P: 4 58, P: 3 58

2 6/3 54 3 58 0 50 49 7 41 3 P, P: 55

nicholaus de heybech and his table for finding true syzygy 35

Argument i ii iii iv v

s /s h min h min min h min min

2 7/3 53 3 55 0 49 50 7 35 3 P, P: 52

2 8/3 52 3 52 0 48 50 7 29 3 P, P: 49

2 9/3 51 3 49 0 48 50 7 23 3 P, P: 46

2 10/3 50 3 46 0 47 51 7 17 2 P, P: 43

2 11/3 49 3 43 0 46 51 7 10 2 P, P: 39

2 12/3 48 3 39 0 46 51 7 3 2 P, P: 36

2 13/3 47 3 36 0 45 52 6 56 2 P, P: 33

2 14/3 46 3 33 0 45 52 6 49 2 P, P: 29

2 15/3 45 3 29 0 44 52 6 42 2 P, P: 25, P: 0 47

2 16/3 44 3 25 0 43 53 6 35 2 P, P: 22

2 17/3 43 3 22 0 43 53 6 28 2 P, P: 18

2 18/3 42 3 18 0 42 53 6 20 2 P, P: 15

2 19/3 41 3 15 0 41 54 6 13 2 P, P: 11, P: 6 10

2 20/3 40 3 11 0 41 54 6 5 2 P, P: 7

2 21/3 39 3 7 0 40 54 5 57 2 P, P: 3

2 22/3 38 3 3 0 39 54 5 49 2 P, P: 2 59

2 23/3 37 2 59 0 39 55 5 41 2 P, P: 55

2 24/3 36 2 55 0 38 55 5 33 2 P, P: 51

2 25/3 35 2 51 0 37 55 5 25 2 P, P: 47

2 26/3 34 2 47 0 36 55 5 17 2 P, P: 43

2 27/3 33 2 43 0 35 56 5 9 2 P, P: 38

2 28/3 32 2 38 0 34 56 5 0 2 P, P: 34

2 29/3 31 2 34 0 33 56 4 51 2

2 30/3 30 2 29 0 32 56 4 42 2

2 31/3 29 2 24 0 31 56 4 34 2

2 32/3 28 2 19 0 30 57 4 25 2

2 33/3 27 2 15 0 29 57 4 16 2

2 34/3 26 2 10 0 28 57 4 7 2

2 35/3 25 2 6 0 27 57 3 58 2

2 36/3 24 2 1 0 26 57 3 49 2

2 37/3 23 1 57 0 25 58 3 40 2

2 38/3 22 1 52 0 24 58 3 31 1 P: 42

2 39/3 21 1 47 0 23 58 3 22 1

2 40/3 20 1 42 0 22 58 3 13 1

2 41/3 19 1 37 0 21 58 3 3 1

36 chapter 1

(cont.)

Argument i ii iii iv v

s /s h min h min min h min min

2 42/3 18 1 32 0 20 58 2 54 1

2 43/3 17 1 28 0 19 59 2 44 1

2 44/3 16 1 23 0 18 59 2 35 1

2 45/3 15 1 18 0 17 59 2 25 1

2 46/3 14 1 13 0 16 59 2 16 1

2 47/3 13 1 8 0 15 59 2 6 1

2 48/3 12 1 3 0 14 59 1 57 1 P, P: 0

2 49/3 11 0 58 0 12 59 1 47 1 P, P: 0

2 50/3 10 0 53 0 11 60 1 38 0

2 51/3 9 0 48 0 10 60 1 28 0

2 52/3 8 0 43 0 9 60 1 19 0

2 53/3 7 0 38 0 8 60 1 9 0

2 54/3 6 0 32 0 7 60 0 59 0

2 55/3 5 0 27 0 5 60 0 50 0

2 56/3 4 0 22 0 4 60 0 40 0

2 57/3 3 0 17 0 3 60 0 30 0

2 58/3 2 0 12 0 2 60 0 20 0

2 59/3 1 0 6 0 1 60 0 10 0

3 0/3 0 0 0 0 0 60 0 0 0

Appendix ii

Translation

The sentence numbers added to the Latin text have been retained in the

translation; words in square brackets have been added by the editors. Sentences

[16][20] are virtually identical with John of Saxonys canon to the Alfonsine

Tables, Chap. 22 [cf. Poulle 1984, 86, lines 127138].

[1] To find the time of true conjunction and opposition of the Sun and the

Moon by means of the tables compiled by Master Nicholaus de Heybech of

Erfurt, [2] you must know the time of mean conjunction if you seek [the time

nicholaus de heybech and his table for finding true syzygy 37

of] true conjunction, or the time of mean opposition if you seek [the time

of] true opposition. [3] You must also know the solar argument and the lunar

argument at that time; [4] and when these are known and obtained, first enter

the table for the equation of the time of true conjunction and opposition of

the luminaries with the solar argument. [5] Take the solar correction [in Col. i]

and write above it m if the solar argument is less than 3 physical signs [i.e.,

180], or write a if it is greater than 3 [physical] signs. [6] Take also in the same

place the difference of the solar correction [in Col. ii] as well as the minutes of

proportion [in Col. iii], and write Sun above them; [7] and when these are

obtained and recorded, then enter the same tables in the column [labeled]

argument [Lat.: linee numeri] with the lunar argument, and take the lunar

correction opposite it [in Col. iv], and write above it a if the [lunar] argument

is less than 3 [physical] signs, or m if it is more. [8] Take also in the same place

the difference of the lunar correction [in Col. v] and the minutes of proportion

[in Col. iii], and write Moon above them; [9] and when these are obtained and

recorded, take the proportional part from the difference of the solar correction

according to the ratio of the minutes of the Moon to 60. [10] Subtract this

proportional part from the solar correction which was recorded previously, and

you will have it well corrected. [11] Similarly, take the proportional part from

the difference of the lunar correction according to the ratio of the minutes

of proportion of the Sun to 60. [12] Add this proportional part to the lunar

correction previously found, and you will have it well corrected. [13] Then,

see if there is an a written above both corrections, the solar and the lunar:

then add together with them the time of mean conjunction previously found

and recorded; but if there has been an m written above both of them, then

subtract them from the same time in the same way. [14] However, if there is an

a written above one of them and an m above the other, then subtract the

smaller correction from the larger and add the remainder together with the

predicted time if there has been an a written above the larger, or subtract [it]

if there has been an m written above the larger one. [15] You will thus obtain

for the meridian of the place for which you were seeking the predictions the

true conjunction of the Sun and the Moon, if you were working to find the

[true] conjunction, and similarly, the true opposition, if you were working to

find it, but without the equation of time [Lat.: diebus non equatis]. [16] And

the positions of the planets must be sought for that particular time. [17] If you

wish to know the degree of the ascendant, it is necessary to equate the days. [18]

Therefore, enter the table for the equation of time [lit.: the equation of the days

with their nights] with the true solar position, and you will find the equation

of time opposite [the number] of degrees, that is, beneath the sign in which

the Sun is. [19] If you find degrees and minutes there, take four minutes of an

38 chapter 1

hour for each degree, and four seconds of an hour for each minute; and add

these together with the predicted time, each to its own rank, and there results

the time of true conjunction or opposition with the equation of time. [20] With

that [time] you must seek the ascendant and the other [magnitudes] relating

to the figure [of the sky for the horoscope].

Acknowledgments

We are grateful to Jos Luis Mancha (Sevilla) for establishing the Latin text of

the canon given in Appendix i, and we are indebted to him for many helpful

remarks. Alan C. Bowen (Princeton) read a draft translation of this canon, and

many of his suggestions have been incorporated in Appendix ii. We have also

benefited from comments by Emmanuel Poulle (Paris) on a previous draft of

this paper.

References

Alfonsine Tables. 1483. Tabulae astronomicae Alfontii regis castelle. Venice: E. Ratdolt.

Chabs, J. 1989. Anlisis del contenido astronmico de las tablas de Jacob ben David

Bonjorn. Barcelona: Ph.D. dissertation, University of Barcelona (u. of b. Microfilm

814).

Chabs, J. 1991. The astronomical tables of Jacob ben David Bonjom. Archive for History

of Exact Sciences 42, 279314.

Goldstein, B.R. 1974. The Astronomical Tables of Levi ben Gerson. Hamden, ct: Archon

Books (Transactions of the Connecticut Academy of Arts and Sciences, Vol. 45).

Goldstein, B.R. 1980. Solar and lunar velocities in the Alfonsine Tables. Historia Mathe-

matica 7, 134140.

Goldstein, B.R. 1992. Lunar velocity in the Ptolemaic tradition. In In the Prime of lnven-

tion: Essays in the History of Mathematics and the Exact Sciences, P. Harman &

A.E. Shapiro, Eds. Cambridge: Cambridge Univ. Press, pp. 317.

Hartmann, J. 1919. Die astronomischen Instrumente des Kardinals Nikolaus Cusanus.

Abhandlungen der kniglichen Gessellschaft der Wissenschaften zu Gttingen, Neue

Folge, 10 (6).

Mills, J.M. 19431950. Estudios sobre Azarquiel. Madrid/Granada: Consejo Superior de

Investigaciones Cientficas.

Nallino, C.A. 1907. Al-Battn sive Albatenii Opus Astronomicum, ii. Milan: Reale Osser-

vatorio di Brera in Milano.

Neugebauer, O. 1969. The Exact Sciences in Antiquity. New York: Dover.

nicholaus de heybech and his table for finding true syzygy 39

Springer Verlag.

Pedersen, O. 1974. A Survey of the Almagest. Odense: Odense University Press.

Poulle, E. 1984. Les tables alphonsines avec les canons de Jean de Saxe. Paris: Centre

national de la recherche scientifique.

Poulle, E. 1988. The Alfonsine Tables and Alfonso x of Castille. Journal for the History of

Astronomy 19, 97113.

Ratdolt, E. 1483. See Alfonsine Tables.

Thorndike, L. 1948. Nicholaus de Heybech of Erfurt. Isis 39, 5960.

Thorndike, L., and Kibre, P. 1963. A Catalogue of Incipits of Medieval Scientific Writings

in Latin. Cambridge, ma: Medieval Academy of America.

Toomer, G.J., 1968. A survey of the Toledan Tables. Osiris 15, 5174.

Toomer, G.J. 1984. Ptolemys Almagest. New York: Springer-Verlag.

Zinner, E. 1925. Verzeichnis der astronomischen Handschriften des deutschen Kulturge-

bietes. Munich: C.H. Beck.

chapter 2

Computational Astronomy:

Five Centuries of Finding True Syzygy*

1 Introduction

needed to follow the rules for determining astronomically significant events

and, from Ptolemy on, it was generally understood that numerical tables could

be constructed that represented the underlying geometrical models ade-

quately. Indeed, tables were already the result of many computations that

relieved the user of much work and thus reduced the possibility of making mis-

takes.1 Moreover, the use of a table could be described more simply than the

corresponding rules for calculating the same quantity. In short, tables were the

most successful and economical way to present complex mathematical proce-

dures in the Middle Ages. Many solutions to the same problem were generated,

usually without justification or reference to previous treatments of it. For this

reason, the task of the historian is to provide the astronomical significance of

the quantities present in the mathematical relationships, as well as to indicate

the dependence of later scholars on the works of their predecessors.

In this paper we explore the specific case of finding the time from mean

to true syzygy (conjunction or opposition of the Sun and the Moon).2 The

procedures introduced for this purpose rely entirely on arithmetic, and do not

involve any observations. Rather, they illustrate the methods and approaches

of medieval practitioners in computational astronomy, and yield insights into

1 See, for example, B.R. Goldstein, Descriptions of astronomical instruments in Hebrew, in

From deferent to equant: A volume of studies in the history of science in the ancient and medieval

Near East in honor of E.S. Kennedy, ed. by D.A. King and G. Saliba (New York Academy of

Sciences, New York, 1987), 105141 (espec. p. 128, where Ibn al-adibs view is cited).

2 In addition to the authors whose works are discussed here, we know of a few other treatments

of this question. The most important is probably that of John of Murs noted by E. Poulle, John

of Murs, in Dictionary of scientific biography (New York, 19701980), vii, 128133 (espec. p. 130),

and we have also located tables for this purpose in Hebrew astronomical manuscripts: Moses

Farissol Botarel (c. 1481) in Munich, ms heb. 343, ff. 96b97a, and Isaac ben Elia Kohen of

Syracuse (c. 1491) in London, ms British Library Or. 2806, ff. 43a44b.

computational astronomy: five centuries of finding true syzygy 41

the ways that various traditions in the Middle Ages affected early modern

astronomy. As far as we can tell, solutions to this problem that make use of

double argument tables first appeared in Spain, and later in other parts of

Europe. To be sure, medieval examples of double argument tables have been

noted previously, but they pertain to the calculation of lunar and planetary

equations (beginning with Ibn Ynus, d. 1009),3 and for problems concerning

the time of day.4 We will focus our attention on tabular solutions, while also

noting procedures that were not reduced to tables.

Simply stated, the problem is: given the time at which a mean syzygy takes

place (easily solved using available tables), to determine t, the time from mean

syzygy, t, to true syzygy, t, where t > 0 indicates that true syzygy takes place

after mean syzygy. There are four variables to be taken into consideration: s

and m, the true longitudes of the Sun and the Moon at mean syzygy; and vs

and vm, the velocities in longitude of the Sun and the Moon during the time

from mean to true syzygy. The most serious difficulty is that the lunar velocity

cannot properly be considered constant in this time interval, and there was no

simple way to approximate its average value.

In Almagest vi.4 Ptolemy (c. 150) presented an approximate solution (with-

out reducing it to a table) which may be expressed in modern notation as

(1) t = 13/12vm(t),

and = m s + 180 for opposition), together with a rule for computing

vm(t), the velocity of the Moon at the time of mean syzygy. This method makes

the simplifying assumptions that the solar and lunar velocities are constant

over the relevant time interval, and that the ratio of lunar to solar velocity

is 13 to 1. Though these approximations are crude, they can be refined by

3 C. Jensen, The lunar theories of al-Baghdadi, Archive for history of exact sciences, viii (1972),

321328; D.A. King, A double argument table for the lunar equation in Ibn Ynus, Centaurus,

xviii (1974), 129146; D.A. King, Some early Islamic tables for determining lunar crescent

visibility, in King and Saliba (eds), op. cit. (ref. 1), 185225; J.D. North, The Alfonsine Tables

in England, in Prismata, ed. by Y. Maeyama and W.G. Saltzer (Wiesbaden, 1977), 269301;

G. Saliba, The double argument lunar tables of Cyriacus, Journal for the history of astronomy,

vii (1976), 4146; M. Tichenor, Late medieval two-argument tables for planetary longitudes,

Journal of Near Eastern studies, xxvi (1967), 126128, reprinted in E.S. Kennedy, Studies in the

Islamic exact sciences (Beirut, 1983), 122124.

4 See, for example, B.R. Goldstein, A medieval table for reckoning time from solar altitude,

Scripta mathematica, xxvii (1964), 6166, reprinted in Kennedy, Studies (ref. 3), 293298.

42 chapter 2

for t, m and s may be recomputed from the solar and lunar correction tables

for t + t. If they are not equal, the procedure in Equation (1) is repeated,

using the new value for . The process converges rapidly, and it is rarely,

if ever, necessary to perform more than two iterations to achieve equality

of solar and lunar longitudes, to the precision of minutes of are. For most

medieval astronomers this approach (with or without refinement) was quite

satisfactory: it was adopted by al-Battn;5 it appears in the canons to the tables

of al-Khwrizm;6 and it is found in the principal astronomical tables composed

in medieval Spain, namely, the Toledan Tables and the Alfonsine Tables (as

attested in Chapter xxx of the Castilian canons).7

The earliest solution to the syzygy problem that we have found in the form of

a table appears in a work of Ibn al-Kammd (c. 1116) called al-Muqtabis.8 This

is one of his three known works (and the only one extant), preserved uniquely

in a Latin manuscript: Biblioteca Nacional de Madrid, ms 10023. The text was

translated into Latin in 1260 in Palermo by John of Dumpno. Ibn al-Kammds

solution consists of a double argument table (f. 52r): its entries, t(vm(t) vs(t),

), are given in hours and minutes, and are functions of the difference between

the hourly velocities of the Moon and the Sun (in minutes and seconds per

hour, from 0;27,30/h to 0;33,30/h at intervals of 0;0,30/h) and the elongation,

(in degrees and minutes, from 0;30 to 12;0 at intervals of 0;30). An excerpt

of this table is reproduced in Table 1. The entries were calculated by means of

the formula

5 O. Neugebauer, A history of ancient mathematical astronomy (New York and Berlin, 1975), 123,

points out that Ptolemy does not indicate whether iteration is necessary, whereas in works

of the Byzantine period it is explicitly stated that the procedure is to be iterated until no

elongation remains. See also C.A. Nallino, Al-Battn sive Albatenii opus astronomicum (2 vols,

Milan, 19031907), i, 94.

6 O. Neugebauer, The astronomical tables of al-Khwrizm (Copenhagen, 1962), 63.

7 On zijes (sets of astronomical tables), see E.S. Kennedy, A survey of Islamic astronomical

tables, Transactions of the American Philosophical Society (Philadelphia, 1956), xlvi/2; see

also F.S. Pedersen, Canones Azarchelis: Some versions, and a text, Cahiers de lInstitut du

MoyenAge grec et latin, liv (1987), 129218 (espec. p. 182); M. Rico y Sinobas, Libros del saber

de astronoma del Rey Alfonso x de Castilla (5 vols, Madrid, 1866), iv, 150151.

8 J. Chabs and B.R. Goldstein, Andalusian astronomy: al-Zj al-Muqtabis of Ibn al-Kammd,

Archive for history of exact sciences, xlviii (1994), 141.

computational astronomy: five centuries of finding true syzygy 43

[vm vs]

[] 27;30 28; 0 33; 0 33;30

1; 0 2;11 2; 8 1;49 1;47

1;30 3;17 4;12 2;43 2;41

6; 0 13; 5 12;51 10;45 * 10;45

11;30 25; 6 24;39 20;36 20;46 **

12; 0 26;11 25;43 21;30 21;50 **

* Read 10;54.

** These values seem to belong to the column

labelled 33;0, and the values there to this column.

This solution gives rather crude results (see Section 6, below), because the rela-

tive velocity of the luminaries is assumed to be constant during the time inter-

val from mean to true syzygy, but its simple presentation made it attractive,

and led to its adoption (with minor modifications) by a number of subsequent

astronomers.

Hebrew in London, Jews College, ms heb. 135, contains a table (ff. 92b93b)

with the same range, intervals, and precision for the horizontal argument

(vm vs) as that of Ibn al-Kammd. The vertical argument () has the

identical range and the same accuracy, but the interval is different, for here

it is 0;6 (rather than 0;30). The entries in this table are more accurately

computed from Equation (2) than those in Ibn al-Kammds table.

The Tables of Barcelona, extant in Catalan, Hebrew, and Latin manuscripts,

and probably completed in 1381,9 depend on the same solution as that

9 J.M. Mills, Las tablas astronmicas del Rey Don Pedro el Ceremonioso (Madrid and Barcelona,

1962); J. Chabs, Astronoma Andalus en Catalua: Las Tablas de Barcelona, in From Bagh-

dad to Barcelona: Studies in the Islamic exact sciences in honour of Prof Juan Vernet, ed. by

44 chapter 2

230, f. 78b)

[]

[vm] 0;10 0;20 1; 0 6; 0

12;10 0;21 0;42 2; 9 12;54

12;20 0;21 0;42 2; 7 12;40

13; 0 0;20 0;40 2; 0 12; 0

14;50 0;17 0;34 1;43 10;12

The horizontal argument has a minimum value of 0;28,0/h (rather than

0;27,30/h) and the same upper limit; the same interval (0;0,30/h) is also

used here. The range and interval of the vertical argument differ from those

in the tables of Ibn al-Kammd and Juan Gil: the elongation is given from

0;10 to 6;40 at intervals of 0;10.

Yosef lbn Waqr of Seville (mid-fourteenth century) presents a slightly dif-

ferent solution in the form of tables, uniquely preserved in Munich, ms heb.

230, f. 78b, where the headings are written in Arabic in Hebrew characters. He

treats the effect of the velocities of each luminary in Equation (2) separately

in two double argument tables. In the first table (Munich, ms heb. 230, f. 78b

top) only the lunar velocity varies while it is assumed that the solar veloc-

ity is constant. In this table the elongation, (in degrees and minutes, from

0;10 to 1;0, at intervals of 0;10, and from 1 to 6 at intervals of 1), is the

horizontal argument (rather than vertical, as before), and the lunar veloc-

ity is the vertical argument (in degrees and minutes per day, from 12;0/d

to 14;50/d, at intervals of 0; 10/d). Table 2 displays an excerpt of this table.

The entries, t(, vm), can be recomputed by means of a formula analogous

to Equation (2):

J. Casulleras and J. Sams (Barcelona, 1996), 477525; J. Chabs, Lastronomia de Jacob ben

David Bonjorn (Barcelona, 1992), 23.

computational astronomy: five centuries of finding true syzygy 45

230, f. 78b)

[t]

[vs] 0;10 0;20 1; 0 6; 0

0;58 0;25 0;49 2;25 14;30

0;59 0;25 0;49 2;27 14;42

1; 0 0;25 0;50 2;30 15; 0

1; 1 0;26 0;51 2;32 15;12

1; 2 0;26 0;52 2;35 15;30

Ibn Waqrs second double argument table is for determining the position of

the Sun at true syzygy (see the excerpt in Table 3): the vertical argument is vs,

the solar velocity (in degrees and minutes per day, from 0;57/d to 1;2/d, at

intervals of 0;1/d). The horizontal argument (not specified in the ms) is t, the

time from mean to true syzygy found in the previous table, and it is given in

hours, from 0;10h to 1h, at intervals of 0;10h, and from 1h to 6h, at intervals of

1h. The entries, (t, vs), represent the arc-distance (in minutes and seconds)

that the Sun travels during that time, and we have recomputed them by means

of the formula

(4) = (60/24)vs t.

In the canons to Ibn Waqrs tables (Munich, ms heb. 230, ff. 5a5b [Hebrew]

and ff. 14a14b [Arabic in Hebrew characters]) we are told that the true positions

of the Sun and Moon should be computed for noon of day 14 or day 29 of

the month (rather than for mean syzygy), and then the table should be used,

thus indicating that the table gives time from noon to true syzygy. Despite this

statement in the canons, the entries in the table display the time from mean to

true syzygy, rather than the time from noon to true syzygy (for an example, see

Section 6, below).

Two centuries later, the method presented by Ibn al-Kammd appears, with-

out ascribing it to any author, in the printed version of the Alfonsine Tables

edited by P. Du Hamel in Paris (1553). In this case we have exactly the same

46 chapter 2

pattern as in Ibn al-Kammd, but for minor details (p. 155): the entries,

t(vm vs, ), are given here to seconds; the relative difference in velocity

is given from 0;22/h to 0;34/h at intervals of 0;1/h; the elongation ranges

from 1 to 8 at intervals of 1. This table also includes columns for the dif-

ferences in the entries between two consecutive values of the elongation, to

assist in the cumbersome task of interpolation. The entries are computed by

means of Equation (2). On pp. 156157 there is another version of the same

table: its entries are sixtieths of the corresponding previous ones.

John of Saxony (c. 1330), one of the Parisian astronomers who adapted the

Alfonsine Tables and composed canons to it, offered a more sophisticated

solution using Ptolemys lunar models.10 He made allowances for the variation

in the lunar velocity in the time interval between mean and true syzygy, and

introduced a method of successive approximations to t, first to the nearest

hour, and then to the nearest minute of an hour.11 This yields an improvement

in accuracy, but involves a lot of computation that most practitioners in the late

Middle Ages were probably not prepared to follow. Unfortunately, this solution

cannot be displayed in tabular form.

Another attempt to give more accurate solutions was successfully devel-

oped by Levi ben Gerson (12881344) who lived in southern France and wrote

in Hebrew. He depended on his own lunar models rather than Ptolemys, but

remained within the framework defined by Ptolemy. He presented his original

solution in the form of four tables, and all of them contain additive coeffi-

cients to avoid calculations with negative terms.12 The equation of time is also

included in the tables, so that all coefficients add up to 24; 17h. Two different

10 E. Poulle, Les tables alphonsines avec les canons de Jean de Saxe (Paris, 1984), 80ff.

11 J. Chabs and B.R. Goldstein, Nicholaus de Heybech and his table for finding true syzygy,

Historia mathematica, xix (1992), 265289.

12 B.R. Goldstein, The astronomical tables of Levi ben Gerson (New Haven, 1974), 136146.

For examples of medieval planetary correction tables where negative terms are elimi-

nated by adding a constant. see H. Salam and E.S. Kennedy, Solar and lunar tables in

early Islamic astronomy, Journal of the American Oriental Society, lxxxvii (1968), 492

497; E.S. Kennedy, The astronomical tables of Ibn al-Aclam, Journal for the history of

Arabic science, i (1977), 1323 (espec. p. 14); and B.R. Goldstein, The survival of Arabic

astronomy in Hebrew, Journal for the history of Arabic science, iii (1979), 3139 (espec.

p. 37).

computational astronomy: five centuries of finding true syzygy 47

sets of Levis tables for this purpose are known: they share the same structure,

but they are based on different parameters.

The time from mean to true syzygy is dependent on the motions of both

luminaries in a way that renders it difficult to treat them separately. Neverthe-

less, Isaac Ibn aladib, an astronomer of Spanish origin who settled in Sicily

at the end of the fourteenth century,13 managed to treat the effect of each lumi-

nary separately. In his astronomical tables, entitled Ora selulah, he mentioned

the works of Bonfils (Tarascon, c. 1365), Bonjorn (Perpignan, c. 1361), Ibn al-

Kammd, and Ibn alRaqqm (Granada, d. 1315),14 and gave two corrections to

be applied to the time of mean syzygy in order to find that of true syzygy: one

due to the Sun and the other to the Moon, each one in separate tables (Paris, ms

heb. 1086, ff. 7a9b). The first (solar) correction has a maximum of 3;54,22h for

a value of the mean argument of the Sun of 91; the second (lunar) correction

reaches a maximum of 9;42,6h when the true argument of the Moon is 96. Ibn

al-adibs tables also give separate corrections for the positions of the luminar-

ies between mean and true syzygy, making use of Ptolemys values: 0;32,56/h

for the mean hourly lunar velocity in longitude, 0;32,40/h for the mean hourly

lunar velocity in anomaly, and 0;2,28/h for the mean hourly velocity of the Sun.

Within the Spanish astronomical tradition, Abraham Zacut (d. after 1515)

presented yet another approach to this problem.15 In his Almanach perpetuum

printed in Leiria (Portugal) in 1496 there is a double argument table for the

equation of syzygies. Its title is indeed surprising (pp. 65v66v): Tabula ad

verificandum horam aspectuum vel coniuntionis. The corresponding title for the

same table in Zacuts Ha-ibbur ha-gadol is Table for correcting the time of

conjunction and opposition and quarters of the month and all aspects of the

Moon with all the planets (Lyon, ms heb. 14, f. 142r). In this table the vertical

argument is the elongation (arcus distantie), , between the Sun and the Moon.

The values taken for the elongation are 0;5, 0;10, 0;20, , 1, and thereafter

for each half degree to 13, for a total of 31 values. The horizontal argument

goes from 10;36 to 16 at intervals of 0;12, and represents the daily increment

of elongation, . This table contains 868 entries, yielding not the time from

13 B.R. Goldstein, Scientific traditions in late medieval Jewish communities, in Les Juifs au

regard de l histoire: Mlanges en l honneur de M. Bernhard Blumenkranz, ed. by G. Dahan

(Paris, 1985), 235247.

14 On Ibn al-Raqqm, see J. Carandell, Risla fi ilm al-ill de Muammad Ibn al-Raqqm

al-Andalus (Barcelona, 1988).

15 On Zacut, see F. Cantera Burgos, El judo salmantino Abraham Zacut, Revista de la

Academia de Ciencias de Madrid, xxvii (1931), 63398; and F. Cantera Burgos, Abraham

Zacut (Madrid, 1935).

48 chapter 2

mean to true syzygy but the time, t, counted from noon, at which the true syzygy

occurs in Salamanca. The 842 non-zero entries in this table can be recomputed

by means of the following equation:

(5) t = 24 / ,

where and are deduced from the values of the true longitudes of the Sun

and the Moon, at Salamanca for noon, found in other tables of the Almanach

perpetuum (Tables 2 and 7 for the daily positions of the Sun and the Moon). The

results given by Zacut differ by more than 1 in only 26 cases, thus indicating

that the original table was satisfactorily calculated.

Nicholaus de Heybech of Erfurt (c. 1400) seems to have been the first astron-

omer to reach a solution that met the criteria of improved accuracy and a

user-friendly presentation in tabular form (see Table 4 for an excerpt).16 To

do this, he introduced two terms for the time from mean to true syzygy, to be

added algebraically, one for the Sun and one for the Moon, and treated them

separately. Each term can be considered as a combination of three functions,

some of which depend only on the mean solar anomaly, , and some only on

the mean lunar anomaly, . In calculating the entries in his table, Nicholaus de

Heybech made use of Ptolemys second lunar model for computing the under-

lying lunar velocities, in contrast to his predecessors who generally depended

on Ptolemys simple lunar model.17 The result is a single table in 5 columns and

180 rows that gives results as good as those of his predecessors (see Section 6,

below), but with much less work. The rule for computing t from the entries in

the table is as follows:

16 Heybechs entire table is published in Chabs and Goldstein, op. cit. (ref. 11), together with

an explanation of the way it was computed; see Essay 1.

17 For the use of Ptolemys second lunar model in computing lunar velocities, see B.R. Gold-

stein, Lunar velocity in the Ptolemaic tradition, in The investigation of difficult things:

Essays on Newton and the history of the exact sciences, ed. by P.M. Harman and A.E. Shapiro

(Cambridge, 1992), 317.

computational astronomy: five centuries of finding true syzygy 49

signs of 60. The columns are labelled: Argument:

linee numeri; i: equatio solis; ii: diversitas

equationis solis; iii: minuta proportionalia; iv:

equatio lune; and v: diversitas equationis lune.

Argument i ii iii iv v

s /s h min h min min h min min

0 1/5 59 0 5 0 1 0 0 11 0

0 2/5 58 0 10 0 2 0 0 22 0

1 0/5 0 4 4 0 51 15 8 31 3

2 0/4 0 4 14 0 54 47 8 15 3

3 0/3 0 0 0 0 0 60 0 0 0

Nicholaus de Heybechs method for finding the time from mean to true syzygy

requires a few, rather simple, computations. Not long after him, John of Gmun-

den (d. 1442), lecturer on astronomy at the University of Vienna, fully under-

standing Heybechs approach, presented it in an even simpler way by means

of a double argument table, and it is preserved in a holograph manuscript (ms

Vin. 5151). The table is entitled Tabula ostendens distantiam vere coniunctionis

et oppositionis a media (ff. 119v122r): the horizontal argument is , the mean

solar anomaly (in degrees, from 0s 0 to 11s 24, at intervals of 6), and its verti-

cal argument, , is the mean lunar anomaly (in degrees, from 0s 0 to 6s 0, at

intervals of 6). The table is preceded by a short canon (ff. 117v119r), at the end

of which we read: Iste canon editus et scriptus est Wienne per magistrum Johan-

nem de Gmunden die 20 mensis Maius anno domini 1440 currente. The table, an

excerpt of which appears in Table 5, displays 1,800 entries, t(, ), given in

hours and minutes, for 0 180. These entries can also be used for the

other values of because the following symmetry relation holds:

and, by Equation (7), the minimum value, t = 14;0h, corrresponds to = 90

and 96, and = 276.

50 chapter 2

table 5 John of Gmunden (ms Vin. 5151, ff. 119v122r). We have added a minus sign where

the text reads m(inue) and nothing where it reads a(dde).

[ ]

[] 0s 0 0s 6 3s 0 6s 0 9s 0 11s 18 11s 24

0s 6 1; 3 0;35 3;44 1; 3 5;49 1;59 1;31

0s 12 2; 5 1;37 2;41 2; 5 6;51 3; 1 2;33

3s 0 9;40 9;15 5;27 9;44 13;56 10;31 10; 5

6s 0 0; 0 0;22 3;46 0; 0 3;46 0;45 0;22

It is easy to derive this table from that of Nicholaus de Heybech, but note that

John of Gmunden uses signs of 30 whereas Nicholaus de Heybech uses signs of

60. For example, the maximum entry t(270, 84) (see Table 5) can be derived

directly from the entries in Nicholaus de Heybechs table: the result is 13;57h,

which differs from the entry in John of Gmundens table by 3 minutes. In other

cases we found even closer agreement.

In a fifteenth-century manuscript in Rome, ms Casanatense 1673, we have

found another copy of John of Gmundens double argument table (ff. 89v92r).

The table is entitled Tabule distantie vere coniunctionis vel oppositionis luminar-

ium a media: Composita Erfordie Duringie, and shares all the characteristics of

Gmundens except for two: signs of 60 rather than 30 are used, and the dif-

ferences between consecutive entries, whether in the same row or column, are

shown. There is no text in the manuscript explaining the use of this table, and

just after it there is an interpolation table (f. 92v) to be used in connection with

our table, with the indication that it was drawn per dominum nicolaum de

Reichenbach. This is the name of the copyist of the canons to the tables of

Iohannes Bianchinus, also found in this miscellaneous manuscript. The dou-

ble argument table bears no date, but in the manuscript it is found between

an almanac for the planets beginning in 1456 and a table for mean syzygies the

radices of which, as explicitly stated, are taken for the year 1452 (completed)

and the meridian of Vienna.

The same presentation as that given by John of Gmunden is also found in

an earlier work on eclipses, known as The book of six wings, that was written

computational astronomy: five centuries of finding true syzygy 51

Excerpt from the double

argument table for the time

from mean to true syzygy in

Wing 2 after subtracting

24; 16h.

[]

[ ] 0 90 180 270

0 0; 0 9;54 0; 0 9;54

90 4;10 6; 6 3;24 13;45

180 0; 0 9;57 0; 0 9;57

270 4;10 13;45 3;24 6; 6

fourteenth century. Originally written in Hebrew, this work was translated into

Latin in 1406 and into Greek in 1435.18 The table that he called Wing 2 is for

determining , the longitude from mean to true syzygy, and t, the time from

mean to true syzygy. Actually, the tabulated entries are + 3 and t + 24; 16h

(which will be called t* in what follows). The additive coefficient to t consists

of 0;16h, to account for the maximum equation of time, and 24h, to avoid cal-

culations with negative terms.19 In Bonfilss double argument table for finding

the time to true syzygies, the horizontal argument is , the mean lunar anomaly

(in degrees, from 0s 0 to 11s 24, at intervals of 6), and the vertical argument

is , the mean solar anomaly (in degrees, from 0s 0 to 12s 0, at intervals of 6).

The entries t*(, ) are given in hours and minutes. The table for t* is sim-

ilar to that of John of Gmunden except for interchanging the positions of the

arguments, and the fact that it has twice as many entries, 3,600.

Table 6 displays an excerpt of Bonfilss table, after subtracting 24;16h. Al-

though very close, the entries do not coincide with those given by John of

18 P.C. Solon, The Hexapterygon of Michael Chrysokokkes, Ph.D. dissertation, Brown Uni-

versity, 1968; P.C. Solon, The six wings of Immanuel Bonfils and Michael Chrysokokkes,

Centaurus, xv (1970), 120.

19 This device is reminiscent of a similar one used by Levi ben Gerson for the same purpose

relating to the same problem. Bonfils was aware of the work of his predecessor in southern

France, and probably depended on him here: cf. B.R. Goldstein, The astronomy of Levi ben

Gerson (12881344) (New York and Berlin, 1985), 9; see also ref. 12, above.

52 chapter 2

in Bonfilss table using the

values given by al-Battn

[]

[ ] 0 90 180 270

0 0; 0 9;51 0; 0 9;51

90 4;10 6; 4 3;26 13;38

180 0; 0 9;54 0; 0 9;54

270 4;10 13;38 3;26 6; 4

Alfonsine tradition, whereas Bonfilss table can be derived from al-Battns

tables, very nearly. Table 7 shows the recomputed values of some entries in Bon-

filss table using the values given by al-Battn for the solar and lunar equations

and velocities,20 and not taking into account any correction for the equation of

time.

It has not previously been noted in the scholarly literature that John of

Gmundens approach was retained by early modern astronomers. Georg Peur-

bach (14231461), one of the earliest advocates of humanism at the University

of Vienna, is the author of the Tabulae eclypsium, an extensive set of astro-

nomical tables first printed in 1514 in Vienna, edited by Georg Tannstetter, and

bound together with a work of Peurbachs associate and student, Regiomon-

tanus: Tabula primi mobilis Joannis de Monteregio. In Peurbachs work there is a

forty-eight-page table entitled Tabula distantie vere coniunctionis aut oppositio-

nis a media (ff. a3vd3r) and, in fact, it is a variant of John of Gmundens table

for finding the time from mean to true syzygy. In Peurbachs double argument

table, the mean solar anomaly ranges from 0s 0 to 11s 30, and the mean lunar

anomaly from 0s 0 to 6s 0, as was the case for John of Gmunden. But Peur-

bachs table is much more expanded, for the intervals are 2 for and 1 for ,

yielding an impressive total number of entries of 32,400. This table, together

with the rest of Peurbachs Tabulae eclypsium, was later printed in 1553 in Basel

in a volume entitled Luminarium atque planetarum motuum tabulae octoginta

quinque, omnium ex his quae Alphonsum sequuntur quam faciles together with

tables by Bianchini, Prugnerus, and Peurbach.

computational astronomy: five centuries of finding true syzygy 53

Nicholaus de Heybech, John of Gmunden, and Georg Peurbach, given at inter-

vals of 1, 6, and 1 of mean lunar anomaly, respectively, it is clear that Peur-

bachs table does not derive directly from Heybechs, but from Gmundens, and

that Peurbach used an interpolation scheme to fill in the intermediary values.

This puts Peurbach within a medieval tradition for computing the time from

mean to true syzygy rather than going back to classical sources, as might be

expected of a Renaissance scholar.

Copernicus seems to have had a method of his own to derive the time from

mean to true syzygy;21 it is described in De revolutionibus iv.29. His procedure

depends on Ptolemys second lunar model for computing lunar velocities at

syzygy, as did Nicholaus de Heybechs procedure, and considers the increment

in lunar anomaly in the interval from mean to true syzygy (without introduc-

ing any new tables), in a way that is reminiscent of John of Saxony. Although

Copernicus probably knew the solutions given by his immediate predecessors,

it is not clear that he ever used them. According to Swerdlow and Neugebauer,22

Copernicus had copied parts of Peurbachs Tabulae eclypsium and appended

them to a bound volume containing the 1492 edition of the Alfonsine Tables

and the 1490 edition of Regiomontanuss Tabulae directionum. This handwrit-

ten quire of astronomical tables known as the Uppsala Notebook seems to

come from Copernicuss studies at Cracow University (14911495?).23 Moreover,

the Prutenic tables by Erasmus Reinhold (15111553), based on the Copernican

models and first published in Tbingen in 1551, do not contain any table for the

time from mean to true syzygy.

In December 1590, the Danish astronomer, Tycho Brahe, observed a lunar

eclipse, and displayed the times computed for eclipse middle (which should

be very close to true syzygy) according to five procedures without indicat-

ing any of the intermediate values required by each procedure. The five pro-

cedures are labelled: Alfonso, Peurbach, Copernicus, Maestlin, and Brahes

ibus (New York and Berlin, 1984), 276.

22 Swerdlow and Neugebauer, op. cit. (ref. 21), 272.

23 P. Czartoryski, The Library of Copernicus, Studia Copernicana, xvi (1978), 355396 (espec.

p. 366).

54 chapter 2

own.24 Brahes times computed according to the Alfonsine Tables and those

of Peurbach differ from our recomputations, and the five values he presents

for the time for eclipse middle vary from each other by as much as 2 hours.

However, as can be seen from Table 8, below, the different methods we have

used yield results that are much closer together. Hence, we suspect that Brahe

made some mistakes in his calculations. Moreover, it has been claimed that

Brahe intended to test computed times against observed times, but the text is

too terse to draw any conclusions.25

To compute the time of true syzygy that was necessary for determining in

advance the circumstances of an eclipse, medieval astronomers had to go

through a complicated procedure. They had to use different tables for the

mean motions, equations, velocities, etc., of the two luminaries, and they had

to follow instructions, whenever available, to perform these calculations step

by step. Throughout the Middle Ages these tables and instructions evolved

in different ways, although they remained within the framework established

by Ptolemy. But they were adapted to different models and parameters, and

thus yielded different numerical results for a given problem. We know of no

case prior to the sixteenth century where the various methods for finding

time from mean to true syzygy proposed by the astronomers considered above

were tested against an observation. Nevertheless, we think it useful to compare

these methods, and so we have derived t for the same syzygy according to

the procedures of each author. To do this, we have used the same data in all

cases, rather than depending on the intermediate data that result from each

authors way of deriving them. As our test case, we have taken a syzygy, the

mean conjunction of 20 July 1327, occurring at 3;58,10h after noon in Toledo.26

The basic magnitudes for that event, as calculated according to the instructions

in the canons to the Alfonsine Tables by John of Saxony, are: = 35;25,4, =

222;26,7, and = 4;45,39. Our results, using the different methods and tables,

are displayed in Table 8.

24 Tychonis Brahe Opera omnia, ed. by J.L.E. Dreyer (15 vols, Copenhagen, 19131929), xii,

2025.

25 V.E. Thoren, Tycho Brahes discovery of the variation, Centaurus, xii (1967), 151166 (espec.

p. 158). Thoren also claimed that this observation led Brahe to the discovery of the lunar

variation, but we do not believe there is enough evidence to support his view.

26 Poulle, op. cit. (ref. 10), 214 ff.; Chabs and Goldstein, op. cit. (ref. 11), 271.

computational astronomy: five centuries of finding true syzygy 55

t

Table or method Date hours

lbn al-Kammd c. 1116 8;47

Yosef lbn Waqr mid-14th cent. 8;47

John of Saxonyb c. 1330 8;34

Levi ben Gersonc c. 1340 8;33

Immanuel Bonfils c. 1365 8;44

Nicholaus de Heybechd c. 1400 8;34

John of Gmunden 1440 8;33

8;51,30h (see ref. 5).

b. For details, see Goldstein and Chabs, op. cit. (ref. 11), 271.

c. With Levis second set of tables, the value obtained is 8;6h.

d. For details, see Goldstein and Chabs, op. cit. (ref. 11), 274.

7 Conclusion

The most important point is that observations were not the driving force for

innovations in the treatment of the time interval from mean to true syzygy.

Rather, computational methods were devised whose goal was to simplify com-

putations and to be user-friendly without sacrificing accuracy. We have seen

that a great deal of ingenuity went into these changes in procedure, and that

they were most successfully presented in the form of tables. The first step in

transforming the rules for finding the time from mean to true syzygy was a

rather simple table, and the earliest example of it is found in the tables of Ibn al-

Kammd, a Spanish Muslim astronomer. Other revisions of the rulesusually

in the form of tableswere made in Spain, notably by Ibn Waqr, and then

astronomers in France (and later in Germany) added new devices of their own

while ultimately depending on their Spanish predecessors. So far, we have not

found any examples of this type of table among astronomers from the eastern

Islamic countries. Moreover, as is the case for other scientific matters in Spain,

there is evidence to be gleaned from traditions in Arabic, Catalan, Castilian,

Hebrew, and Latin, reflecting the multicultural setting of the Iberian Penin-

sula. We have also seen that these medieval discussions continued to be copied,

56 chapter 2

printed, and cited in the sixteenth century, i.e., they were still a part of the living

astronomical tradition in the early modern period.

Acknowledgement

We thank R. Mercier for allowing us to consult his notes on Bonfilss table Wing

2 from which we have greatly benefited.

chapter 3

Transmission of Computational

Methods within the Alfonsine Corpus:

The Case of the Tables of Nicholaus de Heybech*

By the end of the fourteenth century the Alfonsine tradition, which origi-

nated in Castile a little more than 100 years earlier, had become the main

computational tool for European astronomers. A great variety of astronomi-

cal tables, often accompanied by texts, followed this tradition, using the same

models to describe the motions of the celestial bodies and the same under-

lying parameters, but differing in presentation. This Alfonsine corpus, as we

have recently named it, dominated the scene of Western astronomy for several

centuries.1

Within this corpus are the Parisian Alfonsine Tables of which hundreds of

copies in manuscript are extant as well as two editions that appeared before

1500. Of particular interest are various methods and tables for finding the time

from mean syzygy (i.e., conjunction or opposition of the Sun and the Moon)

to true syzygy, starting with the method described in the canons by John of

Saxony (1327).2 A specific approach to this problem, summarized below, was

introduced by an otherwise almost unknown Nicholaus de Heybech of Erfurt

(c. 1400).3 Heybechs table was modified as it was transmitted from Erfurt to

Poland, then to Salamanca, and finally to Jerusalem, and it is a remarkable

example of the variety within the Alfonsine corpus that did not involve any

1 See Jos Chabs and Bernard R. Goldstein, The Alfonsine Tables of Toledo (Dordrecht and

Boston, 2003).

2 Jos Chabs and Bernard R. Goldstein, Computational astronomy: Five centuries of finding

true syzygy, Journal for the history of astronomy, xxviii (1997), 93105.

3 Jos Chabs and Bernard R. Goldstein, Nicholaus de Heybech and his table for finding

true syzygy, Historia mathematica, xix (1992), 265289. We have seen a dozen manuscripts

of Nicholaus de Heybechs tables: Basel, Universittsbibliothek, f.ii.7; Dijon, Bibliothque

Municipale, 447; Paris, Bibliothque nationale de France, lat. 7287 and lat. 7290a; Cues, 211;

Vienna, Nationalbibliothek, 2440; Cracow, Biblioteka Jagielloska, 609, 610, 613, 1852, and 1865

(twice); and Princeton, University Library, Grenville Kane Collection 51. Several authors have

mentioned other manuscripts containing the same material: Bern, 454; Vatican, Pal. lat. 1376;

Vienna, Nationalbibliothek, 5245; and Munich, Clm 14111 and 26666.

58 chapter 3

the entries in his table, Heybech used a table for lunar velocity that first

appears in Paris in the 1330s; this is part of the Alfonsine corpus and differs

significantly from the tables for lunar velocity in the traditions of al-Battn

and of al-Khwrizm.4 We thus offer an illustration of the transmission of

a computational technique within the Alfonsine corpus that kept evolving

during its long journey through a considerable part of Europe and beyond.

Heybechs method was presented in the form of a single table in 5 columns

and consisted in taking the time interval between mean syzygy and true syzygy

as the difference between two independent terms, one for the Sun and one for

the Moon. Each term is calculated separately, and both require the computa-

tion of a set of minimum and maximum values and the use of an interpola-

tion scheme for intermediate values. In Heybechs table, this scheme is a list

of interpolation coefficients, ranging from 0 to 1 (column iii, headed minuta

proportionalia, where 60 minutes = 1), depending on the mean lunar anomaly,

when computing the solar term, and on the mean solar anomaly, in the case of

the lunar term. Besides column iii, the computation of the solar term requires

columns i (headed equatio solis) and ii (headed diversitas equationis solis), both

given in hours and minutes. As for the lunar term, besides column iii, its com-

putation requires columns iv (headed equatio lune) and v (headed diversitas

equationis lune), the former given in hours and minutes, and the latter in min-

utes of an hour.

The solar term (ts) can be obtained by means of the expression,

where is the mean solar anomaly and the mean lunar anomaly, and c1, c2,

and c3 represent entries in columns i, ii, and iii, respectively. The entries in

col. i depend on and assume that = 0; those in col. ii also depend on and

represent the differences between the values for = 0 and = 180, for a given

; and the entries in col. iii, ranging from 0 to 1, are for interpolation for other

values of between 0 and 180.

difficult things: Essays on Newton and the history of the exact sciences, ed. by P.M. Harman and

A.E. Shapiro (Cambridge, 1992), 317; idem, Lunar velocity in the Middle Ages: A comparative

study, in From Baghdad to Barcelona: Studies in the Islamic exact sciences in honour of Prof.

Juan Vernet, ed. by J. Casulleras and J. Sams (2 vols, Barcelona, 1996), i, 181194.

transmission of computational methods 59

ms 2440, f. 74v

60 chapter 3

Similarly, the lunar term (tm) can be obtained by means of the expression,

tm = c4 () c5 () c3 (),

in col. iv depend on and assume that = 0, and the entries in col. v also

depend on and represent the differences between the values for = 0 and

= 180, for a given . Again, the entries in col. iii are for interpolation. Thus,

according to Heybechs table, the time from mean syzygy to true syzygy is given

by

astronomers, for it simplified their computations without sacrificing accu-

racy.5 In the manuscripts that preserve this table there is usually a short canon

explaining its use, but not the method for computing the entries. In all cases

the tables are identical, but for copyists errors.

Madrid, Biblioteca Nacional, ms 3385, containing a set of tables in Latin

which we call the Tabule Verificate for Salamanca (ff. 104r113r), has much the

same material as in Heybechs table, but with a different presentation.6 In

these tables, fully part of the Alfonsine corpus, the epoch is 1 Jan. 1461. The

name of the author of the Tabule Verificate (henceforth tv) is unfortunately

not known, although we have identified Nicholaus Polonius as the most likely

candidate among the few astronomers working in the Castilian city of Sala-

manca. Polonius was a Polish scholar who came to Salamanca no later than

1460 and held the newly established chair in astronomy/ astrology at the uni-

versity there until 1464. It is reasonably clear that he brought the Tabule Res-

olute (a form of the Parisian Alfonsine Tables) with him from Poland.7 The

5 See Chabs and Goldstein, Computational astronomy (ref. 2). The conventions for the

algebraic signs in Eq. 1 are not well described in Heybechs canons, whereas the versions in

tv and Zacut are unambiguous because the headings tell the user when to add and when to

subtract.

6 Jos Chabs and Bernard R. Goldstein, Astronomy in the Iberian Peninsula: Abraham Zacut

and the transition from manuscript to print (Philadelphia, 2000), especially pp. 2336.

7 On this set of tables, see Jerzy Dobrzycki, The Tabulae Resolutae, in De astronomia Alphonsis

Regis, ed. by M. Comes, R. Puig, and J. Sams (Barcelona, 1987), 7177; Jos Chabs, Astronomy

at Salamanca in the mid-fifteenth century: The Tabulae Resolutae, Journal for the history of

astronomy, xxix (1999), 167175.

transmission of computational methods 61

tables we have labelled tv 7 (f. 106rv) and tv 8 (f. 107rv) in a previous pub-

lication represent Heybechs columns for the lunar correction and the solar

correction, respectively, but they are organized in a different way: column 2

in tv 7 displays c4(), that is, column iv in Heybechs table; column 4 in tv 7

is equivalent to 1 c3(), that is, the complement in 1 to column iii in Hey-

bechs table; column 2 in tv 8 represents c1() c2(), that is, the difference

between columns i and ii in Heybechs table. However, there is no longer a col-

umn equivalent to Heybechs column v. The rest of the columns in tv 7 and

tv 8 contain the arguments and line-by-line differences of the entries of other

columns.

With the entries in tv 7 and tv 8,

(3) t = c1 () c2 () c3 () c4 (),

and it agrees with the first three terms in Eq. (1). The suppression of the

fourth term, c5 () c3 (), is indeed an acceptable approximation because in

Eq. (1) its contribution is at most 0;04h, a small amount compared with the

maximum values of the first, second, and third terms (4;47h, 1;01h, and 9;40h,

respectively).

It should be noted that the changes introduced in these two tables by the

unknown author of the Tabule Verificate are not mere variations in the positions

of the columns, but imply a different approach from that in Heybechs table,

among other things because attention shifts from lunar apogee ( = 0), which

is assumed for col. i in Heybechs table, to lunar perigee ( = 180), which is

assumed for col. 2 in tv 8.

The next step in the transformation of this table took place in 1513 in Jerusa-

lem where Abraham Zacut (14521515) had recently arrived.8 Zacuts best

known astronomical work was composed in Hebrew in Salamanca (1478), and

entitled ha-ibbur ha-Gadol (The great composition). It was later published

(with a number of modifications) in Latin and Castilian in Leiria, Portugal

(1496), and entitled Almanach perpetuum. This work, in turn, was later trans-

8 For biographical details, see Chabs and Goldstein, Abraham Zacut (ref. 6), 615.

62 chapter 3

transmission of computational methods 63

64 chapter 3

lated into Arabic and diffused in the Islamic world.9 Zacuts tables are mainly

based on the Parisian Alfonsine Tables, but some depend on astronomical

traditions in Hebrew that began in Provence in the fourteenth century. For

his method of finding the time from mean to true syzygy in the Almanach

perpetuum Zacut depended on this Hebrew tradition that derived from the

work of Levi ben Gerson (12881344) and was transmitted to him through

Jacob ben David Bonjorns tables (c. 1360), and it was quite distinct from the

Alfonsine tradition. But in his new set of tables of 1513 for Jerusalem Zacut

included tables for finding t that represent a modified version of Heybechs

table. Ironically, in 1478 Zacut used the Christian calendar for mean motions,

whereas in 1513 he used the Jewish calendar for this purpose. Zacuts tables of

1513 in Hebrew for Jerusalem are extant only in fragments:10 we consulted New

York, Jewish Theological Seminary of America [jtsa], ms 2574 (not dated), 15

folios (containing only Zacuts canons and tables).11 This manuscript contains

three tables for computing t by a method that is similar to the one in the

Tabule Verificate.

Table 1 (f. 8b) is for the solar correction when the Moon is at perigee on

its epicycle. The argument is the solar longitude, that is, the solar anomaly

increased by 90 (under the assumption that the solar apogee is at 90, which

is an adequate value at the time), and it is given in degrees at intervals of 1.

The entries are displayed in hours and minutes. They derive from Heybechs

table, and correspond to the difference between the entries in columns i and

ii, c1() c2(), as was the case in the Tabule Verificate for Salamanca (tv 8,

col. 2). Note however that, contrary to the Tabule Verificate, Zacuts table for

the solar correction is presented as a single table, not as a column of a table, as

in tv 8.

Table 2 (f. 9a) is for the lunar correction when the Moon is at perigee on its

epicycle. The argument is the mean lunar anomaly and it is given in degrees

at intervals of 1. The entries are displayed in hours and minutes. The entries

9 For Zacuts tables in the Islamic world, see Julio Sams, Abraham Zacut and Joseph

Vizinhos Almanach perpetuum in Arabic, Centaurus, xlvi (2004), 8297; idem, In pursuit

of Zacuts Almanach perpetuum in the eastern Islamic world, Zeitschrift fr Geschichte der

Arabisch-Islamischen Wissenschaften, xv (20022003), 6793.

10 See Bernard R. Goldstein, The Hebrew astronomical tradition: New sources, Isis, lxxii

(1981), 237251, p. 248.

11 Another fragment containing Zacuts tables of 1513 is extant in NewYork, jtsa, ms 2567.

There is no hint in Zacuts canons that he was aware of Heybech or that he had direct

access to his table.

transmission of computational methods 65

York, jtsa, ms 2574, f. 8b). Table for the correction

of the Sun when the Moon is at the perigee of its

epicycle, in hours and minutes.

3s 4s 5s 6s 7s 8s

subtract

1 0; 4h 1;53h 3;16h 3;47h 3;18h 1;54h

2 0; 8 1;57 3;18 3;47 3;16 1;50

3 0;12 2; 0 3;19 3;47 3;14 1;47

4 0;16 2; 4 3;21 3;47 3;12 1;43

5 0;19 2; 6 3;22 3;47 3;10 1;39

add

in this table also derive from Heybechs table, and correspond to its column iv,

c4(), as was the case in the Tabule Verificate for Salamanca (tv 7, col. 2). Note

again that Zacuts table for the lunar correction is presented as a single table,

not as a column of a table, as in the Tabule Verificate.

Table 3 (f. 9b) is a double argument table where Zacut combined both

solar and lunar components. The rows were computed for values of the lunar

anomaly at intervals of 10, ranging from 0s 0 to 6s 0, whereas the columns

were computed for values of the solar longitude at intervals of 15, beginning

with 3s 0 (as in Table 1). The entries in this double argument table are given in

minutes of time.

We note that for lunar anomaly, the maximum correction takes place for

argument 0, and vanishes for argument 180. This should indeed be so, for

this correction is to be added to, or subtracted from, the value found in Table 1

when the Moon is at its epicyclic perigee; thus, the correction at argument 180

66 chapter 3

York, jtsa, ms 2574, f. 9a). Table for the correction of

lunar anomaly, in hours and minutes.

0s 1s 2s 3s 4s 5s

add

0 0; 0h 4;59h 8;32h 9;42h 8;15h 4;43h

1 0;11 5; 8 8;37 9;42 8; 9 4;34

2 0;22 5;17 8;42 9;42 7;58 4;25

3 0;33 5;26 8;42* 9;42 7;55 4;16

4 0;43 5;35 8;51 9;42 7;53 4; 7

30 4;59 8;32 9;42 8;15 4;43 0; 0

subtract

(i.e., lunar epicyclic perigee) is 0. The maximum effect of solar anomaly should

be at 90 from the solar apogee and this is represented by the column for 6s

0. Hence, this third correction is to improve the first correction displayed

in Table 1, where the only variable considered was solar longitude. Table 3

then takes into account the effect of the change in lunar anomaly in the time

interval due to the solar motion (where the heading is the solar longitude),

and represents the term [1 c3()] c2() in Eq. (2). This is certainly the case:

1 c3() appears as the column for a solar longitude of 6s 0, i.e., when the

Sun is 90 ahead of apogee and the correction reaches its minimum; it is the

complement in 1 of Heybechs column iii, as in tv 7, column 4. Moreover,

c2() appears as the row for a lunar anomaly of 0s 0, and it is column ii in

Heybechs table, as in tv 8, column 4. The product of the entries in this row

and this column generates the rest of the table. For example, consider the

entries in the row for 2s 0: 0, 12, 23, 32, , 47, , 24, 12, 0. Each one is found

transmission of computational methods 67

table 3 An excerpt of the table for the third correction (New York, jtsa, ms 2574, f. 9b). Table

for the correction of all values for the lunar anomaly to be added to its value at the

perigee of its epicycle: a double argument table [lua meubberet; lit.: a combined

table]

Solar

long. 3s 0 3s 15 4s 0 4s 15 6s 0 8s 0 8s 15 9s 0

Lunar

anom. subtract

0s 0 0 15 29 41 60 31 16 0

0s 10 0 15 29 40 59 30 15 0

0s 20 0 14 28 40 58 29 15 0

1s 0 0 14 28 39 56 29 15 0

1s 10 0 13 27 37 44* 28 14 0

1s 20 0 13 25 35 51 26 13 0

2s 0 0 12 23 32 47 24 12 0

3s 0 0 8 16 23 31 17 9 0

5s 20 0 0 0 1 1 1 1 0

6s 0 0 0 0 0 0 0 0 0

add

entry corresponding to a solar longitude of 6s 0). So, for 2s 0 of lunar anomaly

and

3s 15 of solar longitude: 0;15 0;47 = 0;11,45 = 0;12 (entry: 12 min);

4s 0 of solar longitude: 0;29 0;47 = 0;22,43 = 0;23 (entry: 23 min);

4s 15 of solar longitude: 0;41 0;47 = 0;32, 7 = 0;32 (entry: 32 min);

8s 15 of solar longitude: 0;16 0;47 = 0;12,32 = 0;13 (entry: 12 min);

9s 0 of solar longitude: 0; 0 0;47 = 0; 0 (entry: 0 min).

68 chapter 3

Thus, Zacut replaces two columns in two different tables in the Tabule Ver-

ificate by one double argument table where the multiplication is already done,

thus facilitating the task of the computer; all that is left is linear interpolation

in both the horizontal and vertical directions in the table. The use of a dou-

ble argument table is consistent with Zacuts preference for this kind of table,

which appears already in his ibbur, and this distinguishes him from the tra-

dition, always within the Alfonsine corpus, represented by Heybech and the

Tabule Verificate. On the other hand, the entries in these three tables are not

identical with those in the Tabule Verificate, but both sets are internally consis-

tent and differ very slightly.

Chapter 3 of the canons for these tables (jtsa, ms 2574, ff. 12b13a) concerns

the time interval from mean to true syzygy, with instructions on how to find this

interval from the three tables, but nothing is said about the origin of this table

or the way its entries were computed. Zacut adds a worked example (f. 12b) for

finding true conjunction for Tishri 5274am [= 30 Aug. 1513]: mean conjunction

took place 3(d) 18;6,30h after noon.12 According to the text, the Suns position

was then 5s 16;29 and the lunar anomaly was 8s 12;47. The first correction with

solar longitude 5s 16;29 as argument is 3;38,30h to be subtracted. The result is

14;28h [= 18;6,30h 3;38,30h]. With argument 8s 12 for the lunar anomaly, the

second correction is stated to be about 9;10h [ms: 10, written as a word] to be

subtracted. The result is then given in the text as 5;22h, although it should be

5;18h [= 14;28h 9;10h]. The third correction, with the two arguments, is 0;22h to

be subtracted. Thus, the final result, as given in the text, is 5;0h. This result can

be checked using the tables themselves. In Table 1 the entry for 5s 16 is 3;38h;

in Table 2 the entry for 8s 12 is 9;8h and for 8s 13 it is 9;11h. So 9;10h, the value

given in the text for the second correction, agrees with computation using the

table. In Table 3, with 5s 15 (rounded from 5s 16 for the solar longitude) and

8s 10 (rounded from 8s 12 for the lunar anomaly) as arguments, the entry is

0;22h. Hence, the total correction is 13;10,30h (= 3;38,30h 9;10h 0;22h),13

and the result should be: 18;6,30h 13;10,30h = 4;56h (text: 5;0h).

It is most likely that for this refinement of a technique in the Alfonsine cor-

pus Zacut depended on tables in Latin he had seen in Salamanca many years

before he arrived in Jerusalem. Zacut then transmitted his new method for

finding t in Hebrew, thus contributing to the enlargement of the Alfonsine

12 3(d) means weekday 3, i.e., Tuesday. And indeed 30 Aug. 1513 (jdn 2273923) was a Tuesday.

13 The absolute value of this amount for the total correction is close to its maximum: see

Richard L. Kremer, Wenzel Fabers tables for finding true syzygy, Centaurus, xlv (2003),

305329, p. 314 (table 2).

transmission of computational methods 69

fications in the models or the parameters that depended, directly or indirectly,

on the Alfonsine Tables compiled in Toledo in the 1270s and diffused from Paris

throughout Europe and beyond in the 1320s.

Epilogue

There is one known instance of a later text that depended on Zacuts tables of

1513 for finding t: a Geniza fragment in Hebrew, ms A 697-1, at the John Rylands

University Library, Manchester, England.14 In this brief fragment of an anony-

mous calendrical text for 5557am (= 17961797),15 the goal is to compute the

times of true opposition (full-moon) for each month in the year 5557 am as a

function of true solar longitude and mean lunar anomaly at mean opposition.

On the first line Zacut is credited for the method of determining true opposi-

tions, and the text includes values computed from his first two correction tables

(but there is no evidence of his third table). For example, the solar longitude in

this text for opposition in Tishri 5557 is 6s 23;47 whose correction is given as

3;33h, and the lunar anomaly at that time is 3s 20;26 whose correction is given

as 9;2h. These are exactly the values for these arguments in Zacuts Tables 1 and

2, where the arc minutes of the arguments have been ignored.

14 We are grateful to Y. Tzvi Langermann for bringing this manuscript to our attention.

15 The date given in the text is not easy to read but 5557am is confirmed by recomputing the

astronomical data with Zacuts tables for 1513.

part 2

Planetary Motions

chapter 4

Tables for Planetary Latitudes*

the models and tables for planetary longitudes, and considerably less attention

has been paid to Ptolemys models and tables for planetary latitudes. There are

good grounds for this imbalance both in medieval sources and in the modern

secondary literature, but it is not our goal here to examine the reasons for

this. Rather, we wish to focus on some special features of tables for planetary

latitude, particularly those of Giovanni Bianchini (d. ca. 1469) that are extant

in many manuscript copies,1 in a printed edition of 1526, and in a copy in

the hand of Copernicus.2 In Almagest xiii Ptolemy gives a full treatment of

planetary latitudes and, in the case of the inferior planets, he refers to three

components which we will call inclination (declinatio), slant (reflexio), and

deviation (deviatio).3 However, Ptolemys tables for the latitudes of Venus and

Mercury in Almagest xiii.5 display only the first two of these components,4

whereas Bianchini has columns for all three of them.

Our plan is first to give a brief survey of the history of tables for planetary

latitude, particularly those that include, for Venus and Mercury, columns for

the deviation. Then we will describe Bianchinis tables for planetary latitude

in detail. Finally, we will discuss Copernicuss copy of Bianchinis tables for

planetary latitude.

1 According to the catalogues we consulted, Bianchinis tables survive in a large number of

manuscript copies in many libraries: Naples, Nuremberg, Milan, Paris, Rome, Venice, Vatican,

among others (see, e.g., Boffito 1908; Thorndike 1950 and 1953; Zinner 1990). Of special interest

is Nuremberg, Stadtbibliothek, ms Cent v 57, copied in Vienna in 1460 by Regiomontanus.

2 Uppsala, University Library, ms Copernicana 4, ff. 276v281r. Since it is most likely that

Copernicus saw a manuscript of these tables while he was a student in Cracow between 1491

and 1495, we will refer to the catalogue of the scientific manuscripts there: Rosiska 1984a.

3 In Swerdlow and Neugebauer 1984, p. 523 et passim, deviatio is translated deflection.

4 O. Pedersen 1974, pp. 355386; Neugebauer 1975, pp. 216226; Riddell 1978; Toomer 1984,

pp. 632634; and Swerdlow 2005.

74 chapter 4

In the Handy Tables, composed after the Almagest, Ptolemy made some

changes in the theory of the planetary latitudes and introduced new tables

with a different presentation. The latitude tables in the Handy Tables had lit-

tle influence on subsequent astronomers, although they were the source for

parameters in the corresponding tables in the Mumtaan zij (9th century).5 In

the Iberian peninsula this tradition is represented by Ibn al-Kammd (ca. 1110)

whose zij (extant in a unique Latin manuscript) provides the only previously

known example of a set of astronomical tables where the planetary latitudes

follow the Handy Tables for the inferior planets.6 In the medieval astronom-

ical literature there is yet another tradition, not based on Ptolemaic models,

where the tables for the planetary latitudes differ substantially from those in

the Almagest and the Handy Tables. This tradition whose roots lie in Indian

astronomy appears in the zij of al-Khwrizm (9th century) and, later on, in

tables headed Tabula bipertalis numeri and Tabula quadripertalis numeri, in the

Toledan Tables (available in Latin in the 12th century, but manuscript copies of

it only begin to proliferate in the 13th century).7

5 For the Handy Tables, see Stahlman 1959, pp. 143155, 325334; Neugebauer 1975, pp. 1006

1016; and Swerdlow 2005. For the Mumtaan zij of Yay ibn Ab Manr, see Kennedy 1956,

pp. 145147, 173; and Vernet 1956. Kennedy 1956, p. 146, indicates that these tables for the lat-

itudes of Venus and Mercury have the same parameters as the corresponding tables in the

Handy Tables, but their structure is much more primitive. The extremal values for the lati-

tudes of Venus and Mercury in the Handy Tables are reported by al-Battn, but he does not

identify his source: see Nallino 19031907, 1:116.

6 For the zij of Ibn al-Kammd, see Chabs and Goldstein 1994, pp. 3132; and Madrid, Bib-

lioteca Nacional, ms 10023, f. 45rv. We have identified another copy of this table in the zij of

Juan Gil of Burgos (14th century) that survives in a single Hebrew manuscript: see f. 139a of

what was formerly London, Beth Din, ms 135 [olim, London, Jews College, ms 135; film no. 4796

at the Institute for Microfilmed Hebrew Manuscripts, Hebrew University (Jerusalem)]. The

same table is also found among the Tables of Barcelona (14th century): see Chabs 1996, p. 505.

7 For the zij of al-Khwrizm, see Suter 1914, pp. 138167; Neugebauer 1962, pp. 101103. For the

Toledan Tables, see Toomer 1968, pp. 6970; Richter-Bernburg 1987; and F.S. Pedersen 2002,

pp. 15, 13091321. On the rule for computing planetary latitudes in the zij of al-Khwrizm

see Kennedy and Ukashah 1969, espec. p. 89. A variant of this tradition for treating planetary

latitudes is found in the zij of Ibn Azzz (Fez, 14th century) where the entries in one column

for each planet are the same as in the zij of al-Khwrizm and those in the other are the

reciprocals of the corresponding entries in the earlier zij: see Sams 1999, pp. 114; and Sams

1997, p. 92.

ptolemy, bianchini, and copernicus 75

While these two traditions were represented in some texts, the mainstream

tradition found in most medieval sets of astronomical tables dealing with plan-

etary latitudes certainly derives from the Almagest and it was transmitted to

the West primarily via the zij of al-Battn (ca. 900) and the Toledan Tables.8

As was the case in the Almagest, neither of these tables displays columns for

the deviation for Venus and Mercury. Interest in the third component of lati-

tude is found in Maghribi sources, notably in a text written by an anonymous

author in Tunisia ca. 1280, a reworking of the canons to the zij of Ibn Isq al-

Tnis (ca. 1222), based on Ibn al-Kammd. Chapter 18 of this text (uniquely

extant in Hyderabad, Andra Pradesh State Library, ms 298) presents a worked

example for the latitude of Venus where the deviation is considered. Further

references to this method for reckoning the latitudes of the inferior planets

with three components are found in the zij of Ibn al-Bann (d. 1321) which,

in turn, depends on the zij of Ibn Isq.9 Written at much the same time,

the canons to the Castilian Alfonsine Tables include explanations for using

tables to compute the latitudes of the planets. These canons were composed no

later than 1272 and they are preserved in a unique manuscript now in Madrid;

the original tables, which are not extant, had an epoch of January 1, 1252.10

For the inferior planets, Chap. 22 specifically mentions columns for deviation,

here called third latitude, for Venus (22:12) and for Mercury (22:28), as fol-

lows:

[11] Mas quando quisieres saber la latitud de Venus. entra con su entro en

las tablas de su latitud y toma la que fuere en su derecho de los minutos

proporionales de la declinaion e de los minutos proporionales del decli-

namiento. et escrive cada uno dellos a su parte y escrive sobre cada uno lo

que hallares en somo de la regla donde lo tomas de alto o de baxo.

[12] E toma otrosi lo que fuere en aquel derecho del entro en la regla de la

latitud tercera et guardala otrosi et la parte que le hallares escripta de suso

es siempre septentrional.

()

8 For the zij of al-Battn, see Nallino 19031907, 1:115116, 2:140141. For the Toledan Tables,

see Toomer 1968, pp. 7172; and F.S. Pedersen 2002, pp. 13221326.

9 For the zij of Ibn Isq al-Tnis, see Mestres 1999, pp. 5659. For the zij of Ibn al-Bann,

see Vernet 1952, pp. 96100. For planetary latitude tables in the Islamic East, see van Dalen

1999, espec. p. 323.

10 Madrid, Biblioteca Nacional, ms 3306; see Chabs and Goldstein 2003a.

76 chapter 4

entro en las tablas de su ladeza et toma lo que fuere en su derecho de

los menudos proporionales de la su declinaion y de los menudos pro-

porionales del desviamiento.

[27] E escrivelos con sus titulos segun lo diximos en Venus.

[28] E toma otrosi con el entro su ladesa terera e escrivela con su parte

e siempre la hallaras meredional.

[11] But when you wish to know the latitude of Venus, enter the tables of its

latitude with the center and take the minutes of proportion for inclination

and the minutes of proportion for slant which are opposite it; and write

them down separately and note for each of them what you will find at the

top or at the bottom of the corresponding column.

[12] And also take what is in the column for the third latitude which is

in opposite it and keep it, and what is written is always northern.

()

[26] But when you wish to know the latitude of Mercury enter the

tables of its latitude with the center and take the minutes of proportion

for inclination and the minutes of proportion for slant which are opposite

it.

[27] And write them down with their headings, as we said for Venus.

[28] And also take with its center [as argument what is in the column

for] the third latitude and write it down with the rest, and it is always

southern.

The instructions in the text seem to refer to tables in the style of Almagest

xiii.5. However, in contrast to Ptolemys tables, the columns for the minutes

of proportion for the inclination and for the slant are not the same.

The Castilian Alfonsine Tables arrived in Paris in the early 14th century

and they began to spread in a modified form, in Latin, throughout Europe.

Among the Parisian astronomers, John Vimond, seems to be the first to have

constructed tables for the latitudes of each planet including, for Venus and

Mercury, a column for the deviation. His latitude tables are also in the style of

the Almagest with arguments at 12-intervals from 0s 12 to 12s 0.11 The inclusion

of columns for the deviation is exceedingly rare in the West and, to the best of

our knowledge, Vimonds tables are the earliest to display them. As a matter

11 Paris, Bibliothque nationale de France, ms lat. 7286c, ff. 1r8v; see Chabs and Goldstein

2003b, and Chabs and Goldstein 2004.

ptolemy, bianchini, and copernicus 77

of fact, the third component for Venus and Mercury is not tabulated in the

manuscripts of the Parisian Alfonsine Tables we have seen that date from the

14th century.

The Parisian Alfonsine Tables gave rise to a variety of tables that preserve the

basic parameters of Alfonsine astronomy but differ in presentation. In particu-

lar, the latitude tables were sometimes recast in the form of double argument

tables such that the entries display the latitude of the planet as a function of

both its anomaly and its center. Double argument tables for planetary latitudes

are found in several sets of tables: in the 14th century the Oxford Tables of 1348

ascribed to William Batecombe; and in the 15th century the tables of John of

Gmunden (Vienna), ha-ibbur ha-gadol by Abraham Zacut (Salamanca), and

the Almanach Perpetuum (based on Zacuts tables and printed in 1496 in Leiria,

Portugal).12 We have spot checked corresponding entries in both the Oxford

Tables and in the Almanach Perpetuum and it is clear that these entries all

derive from Ptolemys tables (or a minor variant of them). We can also say

that they were computed independently, based on small divergences between

corresponding entries. In particular, the entries for Venus and Mercury take

into account the deviation: when the anomaly and the center are 0, the incli-

nation and the slant are both 0 and the corresponding entries in the table

only account for the deviation. In the Almanach Perpetuum and in the Oxford

Tables, these entries are not 0, but +0;10 (Venus) and 0;45 (Mercury), which

are the standard extremal values for the deviation (see below). The deviation is

also embedded in the rest of the entries but some recomputations are required

to reveal it. Double argument tables for latitude were certainly an advance over

single argument tables such as those in Almagest xiii.5. The increase in size

offered more possibilities to the computer (Ptolemys table has 675 entries for

all 5 planets, whereas the Oxford Tables have 8,220 and those by Zacut 8,680,

of which 5,611 are for the inferior planets). But, above all, the latitude tables

with double arguments gained in user-friendliness, for with them the com-

puter could often find the latitude he sought (or, at least, a first approximation

to it) simply by inspecting the table. In this way he could avoid the tedious and

risky computations using the table where the entries in each column are a func-

tion of a single argument, looking up many entries (7 in the case of an inferior

12 For the Oxford Tables of 1348, see North 1977; for the tables of John of Gmunden, see

Porres 2003; for ha-ibbur ha-gadol and the Almanach Perpetuum, as well as a survey of

such double argument tables, see Chabs and Goldstein 2000, pp. 137143. According to

Tichenor 1967, p. 128, al-Ksh (d. 1429) also had double argument tables for planetary

latitudes. On al-Ksh, see now Kennedy 1998.

78 chapter 4

planet), each with its proper sign, and combining them (paying attention to the

algebraic signs) according to complicated rules.

John of Murs, who was also active in Paris in the early 14th century, compiled

various sets of tables, and among them are tables for the latitudes of Venus and

Mercury. For both planets we are given double argument tables for the first two

components of latitude, whereas the deviation is presented in a separate col-

umn headed, 3a latitudo.13 A contemporary of John de Lignres, also working

in Paris, John of Saxony, wrote canons to the Parisian Alfonsine Tables in 1327 in

which planetary latitudes are not even mentioned. His canons were frequently

copied in the 14th and 15th centuries, and published together with the edi-

tio princeps of the Alfonsine Tables (Ratdolt 1483). Despite their absence from

these canons, this edition of the Alfonsine Tables has tables for the planetary

latitudes but the deviation for Venus and Mercury is not taken into account.

The second edition of the Alfonsine Tables, edited by J.L. Santritter in 1492 with

a new set of canons, also has tables for the planetary latitudes which are the

same as those in the editio princeps (i.e., in both cases, the arguments are given

at intervals of 6, and the entries are the same), differing only in presentation.14

In his canons, chapters 25 and 26 (for Venus and Mercury, respectively), Santrit-

ter discussed planetary latitudes with instructions for computing the deviation

(que proveniet ex deviatione deferentis ab ecliptica) for the inferior planets. It

turns out that these instructions for computing the deviation were copied from

the canons by John of Lignres (1322), Priores astrologi motus corporum, chap-

ters 22 and 23, almost verbatim.15

13 Lisbon, Biblioteca de Ajuda, ms 52-xii-35, ff. 63r64r. The entries for the deviation for

Venus, given to minutes, are based on a maximum of +0;10, as explained in the text of

the Almagest. But those for Mercury have an extremal value of 0;23, rather than the

value in the Almagest of 0;45 (see below).

14 See Ratdolt (ed.) 1483, f. h1v; and Santritter (ed.) 1492, ff. e5r, f1r, f5r, g1r, and g5r. Rather than

keeping the components of latitude for all planets in columns in a single table, Santritter

collected columns for different phenomena (such as latitude, unequal daily motion, and

retrograde motion) in a separate table for each planet.

15 See Saby 1987, pp. 209, 211. For Venus, John of Lignres says: Postea accipe de minutis

proportionalibus in altero locorum servatis 6am partem que erit latitudo Veneris tertia

examinata que provenit ex deviatione deferentis ab ecliptica; et est semper hec tertia

latitudo septentrionalis, while Santritter (f. c2r) has: Postea accipe de minutis propor-

tionalibus in altero supra loco servatis sextam partem, que erit latitudo Veneris tertio

examinata que proveniet ex deviatione deferentis ab ecliptica, est quem [read: que] sem-

per ista latitudo septentrionalis.

ptolemy, bianchini, and copernicus 79

are not known.16 One of his earliest works, written in 1442, concerns the con-

struction and use of a surveying instrument, and it includes an explanation of

the use of the decimal point.17 In the 1460s he corresponded with Regiomon-

tanus on problems in astronomy in which Regiomontanus indicated difficulties

in Ptolemys treatment of various phenomena.18 Bianchinis tables were pub-

lished twice in Venice (ed. S. Beuilaqua 1495; and ed. L. Gaurico 1526),19 and

both editions include planetary latitudes along with planetary longitudes in

large double argument tables where the anomaly is given in days and the cen-

ter is given in degrees. The number of entries for the latitudes of all 5 planets

is 10,584, far greater than in the other double argument tables described above.

As we found in the Oxford Tables and the Almanach Perpetuum, Bianchinis

entries for the planetary latitudes are based on Ptolemys tables in Almagest

xiii.5 (with slightly different parameters), and take into account the deviation

for the inferior planets.

We now turn our attention to Bianchinis set of auxiliary tables for comput-

ing planetary latitudes in the ed. of 1526 that was not included in the editio

princeps of 1495, despite the fact that these tables appear in manuscript copies.

This set consists of two tables, one for the superior planets and one for the

inferior planets. These two tables have not been studied previously which is

somewhat surprising since they appear in a manuscript copied by Copernicus

that was printed by Curtze in 1875 and by Prowe in 1884.20 However, this copy

was not identified with the tables of Bianchini until 1984 by Rosiska.21 A brief

description of Bianchinis tables, with notes on the manuscripts in Cracow, was

also published by Rosiska, but her goal was to catalogue the tables in these

manuscripts, not to analyze their mathematical structure.22

Bianchinis first table is for the superior planets and has 10 columns (see

Table 1). The description that follows is based on Cracow, Biblioteka Jagiel-

17 Rosiska 1996, p. 57.

18 Swerdlow 1990.

19 There was a third edition, edited by N. Pruckner (Basel, 1553), but we have not seen it.

20 Curtze 1875, pp. 230238; Prowe 18831884, 2:231240. The transcription of Copernicuss

canon for the latitudes of Venus and Mercury appears in Curtze on p. 238, and in Prowe

on pp. 239240. Cf. Swerdlow and Neugebauer 1984, p. 524, n. 44.

21 Rosiska 1984b.

22 Rosiska 1984a, pp. 485486.

80 chapter 4

loska, ms 555 (ff. 237240), a manuscript dating from about 1453 and copied

in Perugia by a scholar from Cracow:23

col. 2: minutes of proportion for Saturn in minutes and seconds;

col. 3: northern latitude for Saturn in degrees and minutes, from 2;2

(for argument 360, or 0) to 3;4 (for argument 180);

col. 4: southern latitude for Saturn in degrees and minutes, from 2;1

(for argument 0) to 3;5 (for argument 180);

col. 5: minutes of proportion for Jupiter in minutes and seconds;

col. 6: northern latitude for Jupiter in degrees and minutes, from 1;6

(for argument 0) to 2;5 (for argument 180);

col. 7: southern latitude for Jupiter in degrees and minutes, from 1;4

(for argument 0) to 2;8 (for argument 180);

col. 8: minutes of proportion for Mars in minutes and seconds;

col. 9: northern latitude for Mars in degrees and minutes, from 0;5

(for argument 0) to 4;21 (for argument 180);

col. 10: southern latitude for Mars in degrees and minutes, from 0;2

(for argument 0) to 7;7 (for argument 180).

Note that the northern or southern latitudes are functions of the true anomaly,

whereas the minutes of proportion are functions of the true center. Table 1

displays an excerpt at 10-intervals from 0 to 180 of Bianchinis table for the

superior planets.

According to Rosiska, this table is found in ten manuscripts in Cracow.24 In

the printed edition of 1526, Table 1 is separated into three tables: the first is for

Saturn (324r326v) and displays, in this order, columns 1, 2, 3, and 4; the second

is for Jupiter (327r329v) and displays, in this order, columns 1, 5, 6, and 7; the

third is for Mars (330r332v) and displays, in this order, columns 1, 8, 9, and

10. The entries in columns 3, 4, 6, 7, 9, and 10 (for the northern and southern

latitudes of the three planets) follow very closely the corresponding values in

Almagest xiii.5. The three columns for the minutes of proportion (columns 2,

5, and 8) in fact are the same (but for shifts: see below), and they are based on

the column in Almagest xiii.5 for the minutes of proportion that applies to the

latitude for all planets.25 This particular column in Almagest xiii.5 is usually

24 Rosiska 1984a, p. 485.

25 See Toomer 1984, pp. 632634.

ptolemy, bianchini, and copernicus 81

table 1 Bianchinis table for the planetary latitudes of the superior planets

(1) (2) (3) (4) (5) (6) (7) (8) (9) (10)

10 30 0 2 4 2 3 58 56 1 8 1 6 58 56 0 9 0 4

20 20 24 2 6 2 4 60 0 1 8 1 6 56 12 0 12 0 6

30 10 24 2 8 2 6 58 54 1 10 1 8 52 0 0 14 0 8

40 0 0 2 11 2 8 56 12 1 12 1 10 45 44 0 17 0 11

50 10 24 2 13 2 11 52 0 1 13 1 12 38 24 0 22 0 16

60 20 24 2 16 2 15 45 44 1 16 1 16 30 0 0 28 0 22

70 30 0 2 20 2 20 38 24 1 20 1 20 20 24 0 35 0 28

80 38 24 2 25 2 25 30 0 1 25 1 25 10 24 0 42 0 38

90 45 44 2 30 2 30 20 24 1 30 1 30 0 0 0 52 0 49

100 52 0 2 35 2 35 10 24 1 35 1 35 10 24 1 3 1 2

110 56 12 2 40 2 40 0 0 1 40 1 40 20 24 1 17 1 16

120 58 56 2 45 2 45 10 24 1 45 1 45 30 0 1 34 1 37

130 60 0 2 49 2 50 20 24 1 49 1 50 38 24 1 56 2 3

140 58 56 2 54 2 54 30 0 1 53 1 55 45 44 2 22 2 41

150 56 12 2 57 2 58 38 24 1 58 2 0 52 0 2 55 3 29

160 52 0 3 0 3 1 45 44 2 1 2 4 56 12 3 30 4 39

170 45 44 3 1 3 3 52 0 2 3 2 6 58 56 4 6 6 0

180 38 24 3 4 3 5 56 12 2 5 2 8 60 0 4 21 7 7

referred to as c5(), for it is the fifth column in that table. Note that is the

argument of latitude counted from the northern limit on the deferent (i.e., 90

from the nodes where the deferent crosses the ecliptic) and that no algebraic

signs are associated with the coefficients c5(). In this paper c5() will be called

p(x), and Bianchini uses it for various purposes of interpolation. We note that

p(x) |cos x|.26

In the instructions to compute the planetary latitudes for the superior plan-

ets (Almagest xiii.6), Ptolemy indicates that the northern limits on the def-

erent differ in each case from the apogees of the superior planets by +50

(Saturn), 20 (Jupiter), and 0 (Mars). For example, according to Ptolemy, the

26 Neugebauer 1975, p. 219. The entries in Ptolemys col. 5 are very nearly those of a cosine

function; the divergence from the values of the cosine reaches a maximum of 0;0,11 (where

cos 0 is taken to be 1) at about 42.

82 chapter 4

ascending node for Saturn is at longitude 90, i.e., the northern limit (n) is at

longitude 180, and the apogee (a) of Saturns deferent is 230 (rounded from

233).27 Hence

a n = 50.

Bianchinis table takes these differences into account: the entries in column

c2 for Saturn are shifted 50 with respect to the corresponding entries in the

Almagest; column c5 for Jupiter +20; and column c8 for Mars has no shift. Thus,

if p(x) is an entry in the table in the Almagest for argument x, and ci(x) an entry

in col. i in Bianchinis table, then

Bianchinis inclusion of these shifts makes his table more user-friendly than

Ptolemys since there is one less step for the user who can enter the column for

interpolation with the true center in all cases. In other words, there is no longer

any need to compute an argument of latitude.

The second table is for the inferior planets and, as given in Cracow, Bib-

lioteka Jagielloska, ms 555 (ff. 241244), it has 11 columns (see Table 2):

col. 2: inclination (declinatio) for Venus in degrees and minutes with extremal

values 1;3 (for argument 360, or 0) and 7;22 (for argument 180);

col. 3: slant (reflexio) for Venus in degrees and minutes with extremal value

2;30 (for arguments 131139 and 221229);

col. 4: minutes of proportion for the inclination of Venus in minutes and

seconds;

col. 5: minutes of proportion for the slant of Venus in minutes and seconds;

col. 6: deviation for Venus in minutes and seconds, ranging from +0;10,0 (for

arguments 0 and 180) to 0;0 (for arguments 90 and 270);

col. 7: inclination (declinatio) for Mercury in degrees and minutes with ex-

tremal values 1;46 (for argument 0) and 4;4 (for argument 180);

col. 8: slant (reflexio) for Mercury in degrees and minutes with extremal value

2;30 (for arguments 112117 and 243247);

col. 9: minutes of proportion for the inclination of Mercury in minutes and

seconds;

ptolemy, bianchini, and copernicus 83

col. 10: minutes of proportion for the slant of Mercury in minutes and seconds;

col. 11: deviation for Mercury in minutes and seconds, ranging from 0;45,0

(for arguments 0 and 180) to 0;0 (for arguments 90 and 270).

The inclination and slant are functions of the true anomaly, whereas the min-

utes of proportion and the deviation are functions of the true center. Table 2

displays an excerpt at 10-intervals from 0 to 180 of Bianchinis table for the

inferior planets.

According to Rosiska, this table is found in eight manuscripts in Cracow.28

In the printed edition of 1526, Table 2 is separated into two tables. The first is

for Venus (333r338v) and displays, in this order, columns 1, 2, 3, 4, 5, and 6; the

second is for Mercury (339r344v) and displays, in this order, columns 1, 7, 8, 9,

10, and 11.

We note that there is a column for the minutes of proportion (col. 10),

not found in Almagest xiii.5, but the three other columns for the minutes of

proportion (columns 4, 5, and 9) are the same (but for shifts) as that for the

minutes of proportion in the planetary latitude tables of Almagest xiii.5, as

was the case for the superior planets. Nevertheless, entries in columns c4 and

c9 are shifted +90 with respect to the corresponding entries in the Almagest.

Thus, if p(x) is an entry in the table in the Almagest, then

are the same as the column for the minutes of proportion in Almagest xiii.5.

Again, the shifts are introduced to facilitate computation. In fact, this is also

the reason for repeating the same column (but for shifts) for each planet.

The entries in column 10, the minutes of proportion for the slant of Mercury,

decrease from 54;0 (at 0) to 0;0 (at 90), increase to 66;0 (at 180), decrease to

0;0 (at 270), and increase back to 54;0 (at 360). There is no indication of the

algebraic sign, and they may all be taken to be positive. The extremal values

for this correction are indeed 110 more and 110 less than 60 minutes, which

corresponds to the instructions given by Ptolemy in Almagest xiii.6 (but there

is no corresponding table in the Almagest). It is readily seen that the entries in

col. 10 can be computed from p(x), as follows:

9

c10(x) = 10 c5(x) for 0 x 90, 270 x 360

84 chapter 4

table 2 Bianchinis table for the planetary latitudes of the inferior planets

(1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11)

10 1 1 0 14 10 24 58 56 9 49 1 44 0 18 10 24 53 2 44 12

20 1 0 0 27 20 24 56 12 9 22 1 42 0 36 20 24 50 36 42 9

30 0 57 0 41 30 0 52 0 8 40 1 36 0 55 30 0 46 48 39 0

40 0 52 0 55 38 24 45 40 7 38 1 26 1 13 38 24 41 10 34 18

50 0 45 1 8 45 44 38 24 6 24 1 13 1 29 45 44 34 33 28 48

60 0 36 1 20 52 0 30 0 5 0 0 59 1 44 52 0 27 0 22 30

70 0 25 1 33 56 12 20 24 3 24 0 42 1 57 56 12 18 22 15 18

80 0 14 1 45 58 56 10 24 1 44 0 22 2 9 58 56 9 22 7 48

90 0 0 1 57 60 0 0 0 0 0 0 0 2 20 60 0 0 0 0 0

100 0 16 2 7 58 56 10 24 1 44 0 25 2 27 58 56 11 26 7 48

110 0 36 2 17 56 12 20 24 3 24 0 54 2 29 56 12 22 26 15 18

120 0 59 2 25 52 0 30 0 5 0 1 25 2 29 52 0 33 0 22 30

130 1 25 2 29 45 44 38 24 6 24 2 0 2 23 45 44 42 15 28 48

140 2 7 2 29 38 24 45 44 7 38 2 33 2 8 38 24 50 18 34 18

150 3 3 2 22 30 0 52 0 8 40 3 7 1 45 30 0 57 12 39 0

160 4 12 2 1 20 24 56 12 9 22 3 38 1 17 20 24 61 50 42 9

170 5 32 1 15 10 24 58 56 9 50 3 58 0 40 10 24 64 50 44 13

180 7 22 0 0 0 0 59 0 9 0 4 4 0 0 0 0 66 0 45 0

c5(180). bj, ms 555, reads 59;0 instead of 60;0 (as in ed. 1526).

c6(180). bj, ms 555, reads 9;0 instead of 10;0 (as in ed. 1526).

c5(40). bj, ms 555, reads 45;40 instead of 45;44 as in c4(50), c4(130), c5(140), c9(50),

and c9(130). This is just one example of miscopying in the table as it is found in this

manuscript; we have not indicated the others.

The number of errors in these two tables is low despite the fact that they contain

more than 7,500 entries (360 10 + 360 11), most of them with two sexagesimal

digits.

and

11

c10(x) = 10 c5(x) for 90 x 270.

As for the inclination and slant of the two inferior planets, the entries in

columns 2, 3, 7, and 8 follow very closely the corresponding values in Almagest

ptolemy, bianchini, and copernicus 85

xiii.5. Now, in Table 2, columns 6 and 11 give the deviation for Venus and

Mercury, respectively: the entries in col. 6 show an extremal value of +0;10,0

and those in col. 11 an extremal value of 0;45,0.29 It is readily seen that both

columns 6 and 11 can be computed from p(x), the fifth column in Almagest

xiii.5 or, equivalently, from c5(x) in Table 2, where the following holds:

and

c11(x) = 0; 45 c5(x).

In sum, all the entries in Bianchinis Tables 1 and 2 ultimately derive from

Almagest xiii.5, but Bianchini presented them differently. In particular, he

introduced one column for Venus (Table 2, col. 6) and two columns for Mercury

(Table 2, cols. 10 and 11) which are not found in Almagest xiii.5.

It is almost impossible to determine the specific source used by Bianchini

in compiling his tables for planetary latitudes because, globally, all such tables

in the tradition of the Almagest share the same entries (with minor variants),

except in some cases for the extremal values and entries near them. Table 3

displays the significant extremal entries in various sets of tables, all of them in

the style of the Almagest (e.g. Oxford, Bodleian Library, ms Can. Misc. 27, con-

taining tables by John of Lignres), except for Lisbon, ms Ajuda (mentioned

above), and the Oxford Tables of 1348,30 which are in the form of double argu-

ment tables. A glance at this table suggests that Bianchini may have depended

on a copy of the Toledan Tables.

The variation in extremal values from one set of tables to another suggests

more variation in the other entries than is the case. For example, in the column

for the inclination of Venus in different sets of tables, most of the entries are the

same. In Table 4 a sample of such entries is presented (Ptolemy tabulated the

inclination of Venus at 6-intervals from 6 to 90, and at 3 intervals from 93

to 180; the Alfonsine Tables [ed. 1483 and ed. 1492] at 6-intervals from 6 to

180; and Bianchini at 1-intervals from 1 to 360).

Both editions of Bianchinis tables are preceded by canons explaining their

use. Chapter 34 includes short comments on planetary latitudes and worked

examples for Jupiter, Venus, and Mercury. In the case of the inferior planets,

30 For these tables, we have consulted Vienna, Nationalbibliothek, ms Vin. 2440.

86 chapter 4

Almagest Tables Misc. 27 52-xii-35 Tables Bianchini 1483

Sat. n. 3; 2 3; 4 3; 2 3; 2 3; 2 3; 4 3; 3 c

s. 3; 5 3; 5 3; 5 3; 5 3; 4 3; 5 3; 5

Jup. n. 2; 4 2; 5 a 2; 5 2; 5 2; 5 2; 5 2; 8

s. 2; 8 2; 8 2; 8 2; 8 2; 8 2; 8 2; 8

Mars n. 4;21 4;21 4;21 4;21 4;31 4;21 4;21

s. 7; 7 7; 7 7;30 6;30 7; 7 7; 7 7;30

Ven. Incl. 6;22 6;22 a 7;22 7;22 7;22 7;22 b 7;12

Slant 2;30 2;30 2;30 2;25 2;30 2;30

Mer. Incl. 4; 5 4; 4 4; 5 4; 5 4; 5 4; 4 4; 5

Slant 2;30 2;30 2;30 2;44 2;30 2;30

a. Other mss read 2;4 for Jupiter and 7;22 for Venus (Pedersen 2002, p. 1326).

b. bj, ms 555: 7;22; ed. 1526: 7;12.

c. Santritters ed. (1492) reads 3;2.

Bianchini used the terms declinatio, reflexio, and deuiatio for the three com-

ponents of latitude, but we note that the edition of 1495 renders the first term

as declaratio, a mistake that was corrected in the edition of 1526. We present

Bianchinis worked examples for Venus and Mercury, without following them

word-for-word.

In the example for Venus the true anomaly = 258 and the true center

= 121. For this , Table 2 gives c2() = 0;20 (inclination) and c3() = 2;9

(slant). For this , Table 2 gives c4() = 51;24 (minutes of proportion for the

inclination), c5() = 30;52 (minutes of proportion for the slant), and c6() =

0;5,9 = 3 (deviation). The three components of latitude are 1 = c2() c4() =

0;17,8, 2 = c3() c5() = 1;6,22, and 3 = 0;5,9. The latitude of Venus is thus:

= 1 + 2 + 3 = 0;54,23, in agreement with the edition of 1495 (although the

edition of 1526 gives the result as 0;54,22).31

In the example for Mercury = 224 and = 189. Table 2 gives c7() = 2;20

(inclination), c8() = 2;15 (slant), c9() = 9;24 (minutes of proportion for the

31 If Bianchini had followed the instructions in the Almagest (see below), he would have

multiplied the entry in col. 6 by c5() and found 3 = 0;5,9 0;30,52 = 0;2,39; the final

result would have been = 0;50,53.

ptolemy, bianchini, and copernicus 87

6 1; 2 1; 2 1; 2

12 1; 1 1; 1 1; 1

18 1; 0 1; 0 1; 0

24 0;59 0;59 0;59

30 0;57 0;57 0;57

36 0;55 0;55 0;55

42 0;51 0;51 0;51

48 0;46 0;46 0;46

150 3; 3 3; 3 3; 3

153 3;23 3;24

156 3;44 3;44 3;43

159 4; 5 4; 5

162 4;26 4;26 4;26

165 4;49 4;49

168 5;13 5;13 5;24

171 5;36 5;42

174 5;52 6;12 6;24

177 6; 7 6;46

180 6;22 7;22 a 7;12

inclination), c10() = 65;1 (minutes of proportion for the slant), and c11() =

0;44,21 (deviation). The three components of latitude are 1 = c7() c9()

= 0;22, 2 = c8() c10() = 2;26, and 3 = 0;44. The latitude of Mercury is

thus: = 1 + 2 + 3 = 3;32, and this is the value computed by Bianchini.32

Bianchini (as well as Vimond, John of Lignres, and Batecombe before him,

Zacut at about the same time, and Santritter afterwards) assumed that the

formula for the deviation is

32 Analogously, if Bianchini had followed the instructions in the Almagest (see below), he

would have computed 3 = c5() c6() and found 3 = 0;44; in this case his final result

would have been the same.

88 chapter 4

(1) 3 = c5(x) d

where d = 0;10 for Venus and 0;45 for Mercury. However, Ptolemys model

requires the formula

and the instructions in Almagest xiii.6 indicate this (although Ptolemy ex-

pressed himself in a rather convoluted way):

Then we take these same sixtieths which were found by the second entry

with the longitude, calculate the amount which is the same fraction of

them as they are of 60, and, for Venus, take 16 of this and set it out

too, always with a northerly direction; but for Mercury we take of the

amount and set it out, always in a southerly direction.33

The critical phrase is calculate the amount which is the same fraction of them

as they are of 60. Ptolemy seems to indicate taking a fraction of a fraction which

is the product of the fractions; in this case, the fractions being the same, the

result is the square of the fraction used to compute 2 for Venus, i.e., the square

of the sixtieths (c5 c5). In the medieval Latin translation of the Almagest by

Gerard of Cremona (d. 1187), based on an Arabic version (rather than the Greek

original), this passage appears as follows:

Deinde post illud tendemus ad hec minuta eadem etiam que invenimus

mittendo longitudinem secundo, et accipiemus ex eis secundum quanti-

tatem partis qua ipsa sunt ex sexaginta partibus, et eius quod provenerit

in Venere accipiemus semper sextam et firmabimus in septentrione; et in

Mercurio semper accipiemus medietatem et quartam et firmabimus in meri-

die.34

After this we shall attend to the same minutes which we found by entering

with the longitude for the second time, and we shall take [a part] of them

according to the size of the part which they themselves are of 60 parts,

and of that which results in [the case of] Venus, we shall always take a

34 Liechtenstein (ed.) 1515, f. 149v. On Gerard of Cremonas translation, see Kunitzsch 1974,

pp. 83112.

ptolemy, bianchini, and copernicus 89

sixth and write it down35 as northern; and in [the case of] Mercury we

shall always take one half plus one quarter and put it down as southern.

the source for Bianchinis computation of the deviation according to formula

(1).

In contrast to the western tradition, eastern Islamic astronomers under-

stood Ptolemys instructions to require computations of the deviation for Venus

and Mercury according to formula (2).36

The western tradition for computing 3, accepted by Bianchini, seems to

have its origin in the zij of al-Battn.37 In the canons to the Toledan Tables,

which depend on al-Battn in this respect, the instruction is:

After this, if your computation is for Venus, take one sixth of the minutes

you wrote in the other place, and this [sixth] is always northern; and if it

is for Mercury, take one quarter plus a half [of them], and this is always

southern.38

depended on the Toledan Tables and their canons although he replaced the

rule given in words with a table.

For those using formula (1), the coefficient c5 was always taken to be positive

and, as we noted above, c5(x) |cos x|. But this function has a singularity at

35 Presumably the underlying Arabic word is a derived form of athbata which means to write

down although the root means to be firm; see Nallino 19031907, 2:325, and n. 38, below.

36 We have checked the zij al-Sanjar (ca. 1120) by al-Khzin (London, British Library, ms Or.

6669, ff. 148v, 152r; and Vatican, ms Arab. 761, ff. 179v, 186r), and the zij al-Khqn (ca. 1420)

by al-Ksh (Istanbul, Aya Sofia ms 2692, ff. 98v, 99v; and London, India Office, ms 430,

ff. 139v, 140v). In both zijes the deviation, called first latitude, was computed according to

formula (2).

37 Nallino 19031907, 1:116.

38 F.S. Pedersen 2002, pp. 268269: Post haec accipe ex ipsis minutis quae scripsisti in alio

loco, si fuerit numerus tuus Veneris, sextam partem eorum, et est semper septentrionalis;

et si fuerit Mercurii, quartam partem [eius] et dimidiam, et est semper meridiana; cf.

p. 515 [ 05]. Ibn al-Bann gives the same instructions: Take a sixth of the minutes of

proportion that you wrote down (athbata) in the second place, and it is the third latitude,

and it is always to the north [for Venus], and keep it (Vernet 1952, Arabic text, p. 38);

and these instructions are also found, almost verbatim, in the anonymous zij preserved

in Hyderabad, Andra Pradesh State Library, ms 298 (see Mestres 1999, Arabic text, p. 91:

chapter 18:12).

90 chapter 4

90 which has no physical meaning. On the other hand, the function y = cos2

x does not exhibit such a singularity. Figure 4.1 shows the behavior of the

functions represented by formulas (1) and (2).

Indeed, in De revolutionibus vi.8, Copernicus comments on this very point,

arguing that if c5 is the coefficient for the deviation (rather than the square of

c5), then for Venus and Mercury the angle between the deferent and the eclip-

tic would be constant (contrary to Ptolemys model for the inferior planets) and

this [component of] latitude () would suddenly leap back from the intersec-

tion [i.e., the nodal line] into the same latitude that it previously left.39 Clearly,

at the time when this passage was written Copernicus understood the behav-

ior of the deviation properly, but his approach had been different in previous

discussions of the latitude for the inferior planets.40

planetary latitudes in the hand of Copernicus: ff. 276v277v [superior planets];

278r279r, 280r [inferior planets]; 279v [all planets]; and 280v281r [canons

for the inferior planets only]. These tables are included in a quire bound

together with the 1492 edition of the Alfonsine Tables (edited by Santritter)

and Regiomontanuss Tabulae directionum (Augsburg, 1490), and it is almost

certain that they were copied at the time when Copernicus was a student in

Cracow.41 Comparison with the tables of Bianchini described above shows that

Copernicus copied and rearranged these tables that had been compiled by his

Italian predecessor.42 A brief description of the 10 tables follows, indicating the

corresponding table of Bianchini:

There are two sub-tables with a column for the argument at 1-intervals, from

1to 30, and six columns, one for each of the zodiacal signs. One sub-table

displays the entries in Bianchinis Table 1, col. 3 (northern latitude), and the

39 See Swerdlow and Neugebauer 1984, p. 523; Copernicus 1543, ff. 191v192r.

40 Cf. Swerdlow and Neugebauer 1984, pp. 508, 536.

41 For a description of this manuscript see Czartoryski 1978, p. 366.

42 Note that Copernicus has arguments at 1-intervals in agreement with Bianchini, in con-

trast to the 6-intervals in the corresponding tables edited by Ratdolt (1483) and by

Santritter (1492).

ptolemy, bianchini, and copernicus 91

figure 4.1 The functions y =| cos x| and y = cos2 x in the interval (0, 180),

where the cosine function represents c5(x) in formulas (1) and

(2). The maximum difference between these functions is 0;15 at

60 and 120

other, those in Table 1, col. 4 (southern latitude). The last few entries for

northern latitudes are not exactly the same as those in Bianchinis table: the

maximum value given by Copernicus is 3;3, and not 3;4, but this is probably

due to miscopying (or a faulty archetype).

There are two sub-tables presented in the same way as the corresponding

table for Saturn. One sub-table displays the entries in Bianchinis Table 1, col. 6

(northern latitude), and the other, those in Table 1, col. 7 (southern latitude).

There are two sub-tables presented in the same way as the corresponding tables

for Saturn and Jupiter. One sub-table displays the entries in Bianchinis Table 1,

col. 9 (northern latitude), and the other, those in Table 1, col. 10 (southern

latitude). In the latter, the entries for the minutes corresponding to arguments

151180 are all mistakenly shifted downwards 1, so that, at the end of the

column, Copernicus had to write the last number outside the frame of the table.

This table has a column for the argument at 1-intervals, from 1 to 30, and

six double columns, one for each of the first six zodiacal signs. In each double

92 chapter 4

and 3 (slant). The heading for the left side of the table is Prima pars tabule, and

that for the right side, Altera pars.

This table is headed Tabella latitudinis Mercurii. It has a column for the argu-

ment at 1-intervals, from 1 to 30, and six double columns, one for each of the

first six zodiacal signs. In each double column we are given the entries in Bian-

chinis Table 2, columns 7 (inclination) and 8 (slant). The heading for the left

side of the table is Superior pars circuli, and that for the right side, Inferior pars

circuli.

This table is headed northern deviation (deviacio borealis). It has a column for

the argument at 1-intervals, from 0 to 30, and three columns, one for each

of the first three zodiacal signs. The entries are taken from Bianchinis Table 2,

column 6 (deviation).

This table is headed southern deviation (deviacio australis). It has a column for

the argument at 1-intervals, from 0 to 30, and three columns, one for each

of the first three zodiacal signs. The entries are taken from Bianchinis Table 2,

column 11 (deviation).

This table is headed Minuta ad declinacionem. It has a column for the argument

at 1-intervals, from 0 to 30, and three columns, one for each of the first three

zodiacal signs. The entries are taken from Bianchinis Table 2, column 4, which

is identical to column 9.

This table is headed Tabella M(inutorum) proporcionabi(lium) 5 planeta(rum).

It displays the minutes of proportion, and it has a column for the argument at

1-intervals, from 0 to 30, and three columns, one for each of the first three

zodiacal signs. The entries are taken from Bianchinis Table 1, column 8, which

is identical to Table 2, column 5. Part of the column for 0s/6s and 5s/11s has

been crossed out, and the first 18 entries have been accurately copied in the left

margin. Note that Copernicus correctly gives the entry for 40 (45;44) rather

than 45;40 (see Table 2).

ptolemy, bianchini, and copernicus 93

This table is headed M(inuta) prop. ad reflectionem, and it has a column for the

argument at 1-intervals, from 0 to 30, and six columns, one for each of the

first six zodiacal signs. The entries are taken from Bianchinis Table 2, column 10.

The heading for the left side of the table is Prima pars ta(bule), and that for the

right side, Altera pars.

chinis Tables 1 and 2, avoiding all unnecessary repetitions by taking advantage

of the numerous symmetries they contain. This is not to say that he presented

his tables in a more compact way, for Copernicus split the two tables given by

Bianchini into ten, and treated the shifts differently. Moreover, the extremal

values in Copernicuss tables are the same as those of Bianchini in Table 1 (the

one exception for Saturn has already been noted above).

Copernicuss canons for Venus and Mercury are not very different from

those that we have discussed, beginning with the canons to the Toledan Tables.

In particular, he accepts formula (1) for computing 3. Hence, it is not easy

to decide whether he depended on Santritters canons in the edition of the

Alfonsine Tables (1492) with which these tables are bound,43 or on Bianchinis

canons, or on some other set of canons. Copernicus expresses himself in his

own way and so linguistic criteria are not sufficient to establish filiation. Santrit-

ters canons for 3 do not refer to a table; rather, they give instructions for taking

fractional parts of p(x) for Venus and for Mercury (with the directions north

and south, respectively), as in the canons to the Toledan Tables.44 Copernicus,

however, refers to tables of the kind he displays and does not give rules with

the fractions, a sixth or three-quarters. This suggests that Copernicus is closer

to Bianchini than to Santritter, although it is not impossible that he relied on a

different source.

44 For Mercury, Santritters instructions (again taken, almost verbatim, from John of Lign-

res) suggest computing with a quarter plus a half of a quarter (i.e., 38) rather than a half

and quarter (34). In this case, it would have been difficult for Copernicus (or any other

reader of Santritters canons) to compute the third latitude of Mercury correctly without

checking some other relevant text. See Santritter (ed.) 1492, f. c2v: Et accipe de postea

minutis proportionalibus in altero locorum servatis quartam partem et dimidium quarte,

que pars cum suo dimidio est latitudo Mercurii tertio examinata, que est semper merid-

ionalis. Where Santritter has que pars cum suo dimidio, John of Lignres has que 4a pars

cum suo dimidio (Saby 1987, p. 211). Note that John of Lignress instruction for using the

coefficient 38 (= 0;22,30) is consistent with the extremal value in his table for the deviation

of Mercury: 0;23 (see note 13, above).

94 chapter 4

planetary latitudes in his Commentariolus (ca. 1512), and modified it in his De

revolutionibus (1543). His aim was to transform Ptolemys geocentric models

into heliocentric models, but his success was far from complete.45 There are

no tables in the early work, but among the tables for planetary latitude in De

revolutionibus is a column that displays p(x)2 for the deviation.46

Despite the misunderstanding of the deviation for Venus and Mercury by

medieval and early modern astronomers, Georg Peurbach (d. 1461) correctly

described the models in the Almagest for the latitudes of both superior and

inferior planets, including the deviation: see his Theoricae novae planetarum

(first printed by Regiomontanus in Nuremberg in 1472). But this work does not

contain any tables.47

4 Conclusion

In the West the deviation for Venus and Mercury was generally not included

in tables for planetary latitudes and it was probably ignored by many users of

those tables as well. Bianchini was one of a very small number of astronomers

working in the Latin tradition of the Almagest who called attention to this third

component of latitude for Venus and Mercury and produced tables for it. Both

the zij of al-Battn and the canons to the Toledan Tables give instructions for

computing the deviation, but no tables accompany them. In the 13th century

the Castilian canons to the Alfonsine Tables mention the deviation, but the

tables associated with these canons are not extant. In the 14th century we

have identified John Vimond as having compiled tables for the deviation. John

of Lignres also compiled tables for the deviation but, as we noted, with an

idiosyncratic extremal value for Mercury; he also included instructions for

computing the deviation in his Priores astrologi but these instructions do not

refer to a table. In the 15th century, Peurbach described Ptolemys model for

the latitudes of Venus and Mercury correctly although he did not produce any

tables for this purpose. Moreover, Santritter gave instructions for computing

the deviation for Venus and Mercury, but did not display any table for it.

We have seen that prior to 1500 Ptolemys instructions for computing the

45 See Swerdlow 1973, pp. 482489, 494499, 509510; and Swerdlow and Neugebauer 1984,

pp. 483537, espec. pp. 535537.

46 Copernicus 1543, ff. 194v195r. See Swerdlow and Neugebauer 1984, pp. 530535.

47 Peurbach 1472, reprinted in Schmeidler (ed.) 1972, pp. 783787; Aiton 1987, pp. 3435.

ptolemy, bianchini, and copernicus 95

e.g., on Bianchinis worked examples would yield slightly incorrect results.

In his student days Copernicus accepted a table based on formula (1) for

the deviation, but in his magnum opus of 1543 he argued that the previous

interpretation of Ptolemys instructions (which he once shared) by earlier

astronomers was faulty, and that one needed to invoke formula (2) for the

deviation of Venus and Mercury.

Given the medieval tradition to which he adhered, Bianchini compiled a set

of user-friendly tables that simplified the computations required for using

the Alfonsine Tables. Many astronomers in the late Middle Ages (e.g., John

of Lignres, William Batecombe, John of Gmunden, and Abraham Zacut) also

had this as a goal (each interpreting it in his own way), and Bianchini fits

nicely in this group. Finally, there can be little doubt that early in his career

Copernicus depended on Bianchinis tables for planetary latitude which, in

turn, are based on Ptolemys models in the Almagest. Hence, Bianchinis tables

can be considered a source for Copernicuss knowledge of astronomy.

Acknowledgments

versitat de Barcelona, Spain) for providing us with a copy of the Tables of Bian-

chini (ed. 1526); to the Biblioteka Jagielloska (Cracow) for a partial copy of

ms 555; to Noel M. Swerdlow for a partial copy of Uppsala, University Library,

ms Copernicana 4; and to Benno van Dalen for partial copies of the Sanjar zij

by al-Khzin and the Khqn zij by al-Ksh. We also thank Fritz S. Pedersen

for supplying us with the text and a preliminary translation of the passage in

the Latin version of the Almagest (ed. 1515), Julio Sams for sending us a partial

copy of Ibn Azzzs zij, and Alan C. Bowen for comments on the instructions

for computing the deviation by John of Lignres and J.L. Santritter.

References

Aiton, E.J. 1987. Peurbachs Theoricae novae planetarum: A Translation with Commen-

tary, Osiris, 3:544.

Beuilaqua, S. (ed.) 1495. Tabularum Joannis Blanchini Venice: Augustinus Morauus.

Bianchini. See Beuilaqua 1495; Gaurico 1526; and Pruckner 1553.

Boffito, G. 1908. Le Tavole astronomiche di Giovanni Bianchini, La Bibliofilia, 9:378

388, 446460.

96 chapter 4

Casulleras, J. and J. Sams (eds.) 1996. From Baghdad to Barcelona: Studies in the Islamic

Exact Sciences in Honour of Prof. Juan Vernet. Barcelona.

Chabs, J. 1996. Astronoma andalus en Catalua: Las Tablas de Barcelona, in J. Ca-

sulleras and J. Sams (eds.) 1996, pp. 477525.

Chabs, J. and B.R. Goldstein 1994. Andalusian Astronomy: al-Zj al-Muqtabis of Ibn al-

Kammd, Archive for History of Exact Sciences, 48:141.

Chabs, J. and B.R. Goldstein 2000. Astronomy in the Iberian Peninsula: Abraham Zacut

and the Transition from Manuscript to Print. Transactions of the American Philo-

sophical Society, 90.2. Philadelphia.

Chabs, J. and B.R. Goldstein 2003a. The Alfonsine Table of Toledo. Dordrecht.

Chabs, J. and B.R. Goldstein 2003b. John Vimond and the Alfonsine Trepidation

Model, Journal for the History of Astronomy, 34:163170.

Chabs, J. and B.R. Goldstein 2004. Early Alfonsine Astronomy in Paris: The Astronom-

ical Tables of John Vimond (1320), Suhayl, 4:207294. [See essay 8, below]

Copernicus, N. 1543. De revolutionibus. Nuremberg.

Curtze, M. 1875. Reliquiae Copernicanae, Zeitschrift fr Mathematik und Physik, 20:

221248.

Czartoryski, P. 1978. The Library of Copernicus, Studia Copernicana, 16:355396.

Dalen, B. van 1999. Tables of Planetary Latitude in the Huihui li: Part ii, in Yung Sik

Kim and Francesca Bray (eds.), Current Perspectives in the History of Science in East

Asia. Seoul, pp. 316329.

Federici Vescovini, G. 1968. Bianchini, Giovanni, Dizionario Biografico degli Italiani,

10:194 196.

Gaurico, L. (ed.) 1526. Tabule Joa. Blanchini bononiensis Venice: Lucas Antonius

Giunta.

Kennedy, E.S. 1956. A Survey of Islamic Astronomical Tables. Transactions of the Ameri-

can Philosophical Society, 46.2. Philadelphia.

Kennedy, E.S. 1998. On the Contents and Significance of the Khqn Zj by Jamshd

Ghiyth al-Dn al-Ksh. Frankfurt.

Kennedy E.S. and W. Ukashah 1969. Al-Khwrizms Planetary Latitude Tables, Cen-

taurus, 14: 8696. Reprinted in King and Kennedy (eds.) 1983, pp. 125135.

King, D.A. and M.H. Kennedy (eds.) 1983. Studies in the Islamic Exact Sciences by E.S.

Kennedy, his colleagues, and former students. Beirut.

Kunitzsch, P. 1974. Der Almagest: Die Syntaxis Mathematica des Claudius Ptolemus in

arabisch-lateinischer berlieferung. Wiesbaden.

Liechtenstein, P. (ed.) 1515. Almagestu[m] Cl. Ptolemei Pheludiensis Alexandrini astrono-

mo[rum] principis. Venice.

Mestres, A. 1999. Materials Andalusins en el Zj dIbn Isq al-Tnis. University of

Barcelona (Ph.D. dissertation).

Nallino, C.A. 19031907. Al-Battn sive Albatenii Opus Astronomicum, 2 vols. Milan.

ptolemy, bianchini, and copernicus 97

Nauta, L. and A. Vanderjagt (eds.) 1999. Between Demonstration and Imagination: Essays

in the History of Science and Philosophy Presented to John D. North. Leiden.

Neugebauer, O. 1962. The Astronomical Tables of al-Khwrizm. Copenhagen.

Neugebauer, O. 1975. A History of Ancient Mathematical Astronomy. Berlin.

North, J.D. 1977. The Alfonsine Tables in England, in Y. Maeyama and W.G. Salzer,

(eds.), Prismata: Festschrift fr Willy Hartner. Wiesbaden, pp. 269301. Reprinted in

North 1989, pp. 327359.

North, J.D. 1989. Stars, minds and fate: Essays in ancient and medieval cosmology. London

and Ronceverte.

Pedersen, F.S. 2002. The Toledan Tables: A review of the manuscripts and the textual

versions with an edition. Copenhagen.

Pedersen, O. 1974. A Survey of the Almagest. Odense.

Peurbach, G. 1472. Theoricae novae planetarum. Nuremberg. Reprinted in Schmeidler

(ed.) 1972, pp. 755793.

Porres, B. 2003. Les tables astronomiques de Jean de Gmunden: dition et tude compara-

tive. Paris: cole pratique des hautes tudes (Ph.D. dissertation).

Prowe, L. 18831884. Nicolaus Coppernicus, 2 vols. (v. 1 is in two parts). Berlin. Reprinted

Osnabrck 1967.

Pruckner, N. (ed.) 1553. Luminarium atque planetarum motuum tabulae octoginta quin-

que Basel.

Ratdolt, E. (ed.). 1483. Tabulae astronomice illustrissimi Alfontij regis castelle. Venice.

Richter-Bernburg, L. 1987. id, the Toledan Tables, and Andalus Science, Annals of

the New York Academy of Sciences, 500:373401.

Riddell, R.C. 1978. The Latitudes of Venus and Mercury in the Almagest, Archive for

History of Exact Sciences, 19:95111.

Rosiska, G. 1984a. Scientific Writings and Astronomical Tables in Cracow: A Census of

Manuscript Sources (xivthxvith Centuries). Wrocaw.

Rosiska, G. 1984b. Identification of Copernican Tables of the Latitudes of the Planets,

Qwartalnik Historii Nauki i Techniki, 29:637644 [in Polish with an English sum-

mary].

Rosiska, G. 1996. The Fifteenth-Century Roots of Modern Mathematics, Qwartalnik

Historii Nauki i Techniki, 41:5370.

Saby, M.-M. 1987. Les canons de Jean de Lignres sur les tables astronomiques de 1321.

Unpublished thesis: cole Nationale des Chartes, Paris. A summary appeared as:

Les canons de Jean de Lignres sur les tables astronomiques de 1321, cole Nationale

des Chartes: Positions des thses, pp. 183190.

Sams, J. 1997. Andalusian Astronomy in 14th Century Fez: al-Zj al-Muwfiq of Ibn

Azzz al-Qusann, Zeitschrift fr Geschichte der Arabisch-Islamischen Wissen-

schaften, 11:73110.

Sams, J. 1999. Horoscopes and History: Ibn Azzz and His Retrospective Horoscopes

98 chapter 4

Related to the Battle of El Salado (1340), in L. Nauta and A. Vanderjagt (eds.) 1999,

pp. 101124.

Santritter, J.L. (ed.) 1492. Tabule Astronomice Alfonsi Regis. Venice.

Schmeidler, F. (ed.) 1972. Joannis Regiomontani Opera Collectanea, facsimiles of nine

works by Regiomontanus and one by Peurbach that was printed by Regiomontanus.

Osnabrck.

Stahlman, W.D. 1959. The Astronomical Tables of Codex Vaticanus Graecus 1291. Provi-

dence: Brown University (Ph.D. dissertation). University Microfilms, No. 625761.

Suter, H. 1914. Die astronomischen Tafeln des Muammad ibn Ms al-Khwrizm.

Copenhagen.

Swerdlow, N.M. 1973. The Derivation and First Draft of Copernicuss Planetary Theory:

A Translation of the Commentariolus with Commentary, Proceedings of the Ameri-

can Philosophical Society, 117:423512.

Swerdlow, N.M. 1990, Regiomontanus on the critical problems of astronomy, in T.H.

Levere and W.R. Shea (eds.), Essays on Galileo and the History of Science in Honour of

Stillman Drake. Dordrecht.

Swerdlow, N.M. 2005. Ptolemys Theories of the Latitude of the Planets in the Almagest,

Handy Tables, and Planetary Hypotheses, in J. Buchwald and A. Franklin (eds.),

Wrong for the Right Reasons. Dordrecht, pp. 4171.

Swerdlow, N.M. and O. Neugebauer 1984. Mathematical Astronomy in Copernicuss De

Revolutionibus. New York.

Thorndike, L. 1950. Giovanni Bianchini in Paris Manuscripts, Scripta Mathematica

16:512, 169180.

Thorndike, L. 1953. Giovanni Bianchini in Italian Manuscripts, Scripta Mathematica

19:517.

Tichenor, M.J. 1967. Late Medieval Two-Argument Tables for Planetary Longitudes,

Journal of Near Eastern Studies, 26:126128. Reprinted in King and Kennedy (eds.)

1983, pp. 122124.

Toomer, G.J. 1968. A Survey of the Toledan Tables, Osiris, 15:5174.

Toomer, G.J. 1984. Ptolemys Almagest. New York.

Vernet, J. 1952. Contribucin al estudio de la labor astronmica de Ibn al-Bann. Tetun.

Vernet, J. 1956. Las Tabulae Probatae, in Homenaje a Mills-Vallicrosa. Barcelona,

pp. 501522. Reprinted in Vernet 1979, pp. 191212.

Vernet 1979. Estudios sobre historia de la ciencia medieval. BarcelonaBellaterra.

Zinner, E. 1990. Regiomontanus, his life and work. Translated by E. Brown. Amsterdam.

chapter 5

The Tables for the Seven Planets for 1340*

The displaced tables are typical of a pervasive tendency in Islamic science

to provide extensive and elegant numerical tables for the convenience of

practitioners. The underlying astronomical theory is neither questioned

nor affected.1

edward s. kennedy

Introduction

dle Ages in which we discussed a wide range of tables, many of which were

previously known. In this monograph we mentioned an unusual set of tables,

whose significance had not been appreciated hitherto, that depends on a prin-

ciple of displacement to eliminate subtractions in the course of computing

planetary positions. This principle was employed in some zijes in the Islamic

world, but until now there was nothing comparable in the Latin West. As we

will see, this set of tables is a most ingenious reworking of the Parisian Alfon-

sine Tables, rather than a translation of an Arabic zij (where the term zij is

used in Arabic to refer to a set of astronomical tables with instructions for their

use). What makes this set of tables so unusual is that, for example, the mean

planetary motions are defined differently from those in the standard Alfon-

sine tables, and some of the functions for computing true planetary longitudes

from their mean motions also differ noticeably from those in the standard

* Archive for History of Exact Sciences 67 (2013), 142, communicated by George Saliba.

1 Kennedy 1977, p. 16.

100 chapter 5

the same results as computation with the standard Alfonsine tables. More-

over, the tables for first station are presented in a completely different way

from those we have seen in the Ptolemaic tradition, although the underlying

parameters are unchanged. Given the absence of instructions, it was not an

easy task to unravel the cleverness of the construction of these tables. We had

some helpful guidance from a paper by Kennedy (1977), but it was not suffi-

cient for uncovering various subtleties in these tables for which there are no

counterparts in the tables that Kennedy analyzed.

Paris, Bibliothque nationale de France, ms lat. 10262, is a 15th-manuscript

containing two sets of astronomical tables: an anonymous set for 1340 (ff. 2r

46v); and another set, called Tabule frequentine (ff. 47v71r), composed by Mel-

chion de Friquento of Naples in 1438 (f. 46v: 1437 completus).2 Neither set has

previously been examined. In this paper we focus on the first set, entirely com-

posed of tables with no accompanying text. At the bottom of the last table

(f. 46v), we read: Expliciunt tabule de septem planetis et de veris locis eius (Here

end the tables for the seven planets and their true positions). Therefore, we

shall refer to this set as the Tables for the Seven Planets (where the Sun and the

Moon are considered planets). The yearly tables begin in March, and we are

told that the tables are valid for 1340 completus, or 1340 (complete), meaning

that the epoch is noon of February 28, 1341, the last day of the year, counting

from March 1, 1340.3

The name of the author of these tables is not given in the manuscript, and no

locality is mentioned. However, on f. 9r, at the bottom of a table for the mean

motion of the Sun, we are told that the radix for the Sun, anno domini 1340

completo, is 8s 16;41,49 (= 256;41,49). In contrast to the use of physical signs

of 60 in the Parisian Alfonsine Tables, we note the use here of zodiacal signs

of 30. If we compute the mean motion of the Sun using the Parisian Alfonsine

Tables for Toledo, we find its mean longitude to be 346;20,53 and its mean

argument of center, 256;43,54.4 Therefore, the given radix refers to the mean

2 For a description of this manuscript, see Thorndike 1957, especially pp. 144145. On the Tabule

frequentine, see Kremer (forthcoming).

3 The last day of February belongs to 1340 according to the convention of this text (in fact, it is

the last day of that year), but to 1341 according to our modern convention (it is the 59th day

of this year). The same convention applies to all subsequent years in this text. Note also that

the common astronomical practice in the Middle Ages was to take noon as the beginning of

the day.

4 According to this computation, 89;36,59 is the longitude of the solar apogee, a, which is the

difference between the solar longitude and the solar mean argument of center at a given time.

displaced tables in latin 101

argument of center of the Sun. The difference in the mean argument of center,

0;2,5, is the amount traveled by the Sun in about 0;50h, which corresponds

to a difference in geographical longitude of about 12;40 east of Toledo. This

figure agrees, although not exactly, with the longitude of the meridian of Paris,

taken as 12;0 east of Toledo in many astronomical tables of the 14th and 15th

centuries.5

The Tables for the Seven Planets give multiple examples of displaced tables.

A table is said to be displaced with respect to another when its entries are the

same as those in the standard table after adding a constant to its argument

(horizontal displacement), or derive from the entries in the standard table

by adding a constant (vertical displacement). This can also be expressed in

algebraic terms: if y = f(x) is the function underlying a given table, then the

function embedded in the displaced table is y = f(x + kh), for a displacement

on the x-axis, or y = f(x) + kv, for a displacement on the y-axis. Of course, both

displacements can occur at the same time, leading to an equation such as

The purpose of displaced tables is to avoid subtractions, that is, the use of

complicated rules for handling negative numbers before they were available

to astronomers who computed planetary positions by means of astronomical

tables.6

This problem was already felt by astronomers in the 9th century, who called

the standard table al and the displaced table wa.7 Displaced tables were

clever computational devices, with implications for the method of computa-

tion of astronomical quantities, but did not challenge either the parameters

or the models on which the original tables were based. Kennedy (1977) pre-

sented an explanation of displaced planetary tables in the medieval Islamic

world, and demonstrated the equivalence of computations using one such set

with computations using Ptolemys tables. Kennedy focused on the tables of

6 It is important to distinguish between displaced tables and shifted tables. A shifted table

contains the same information as the standard table (the same columns and rows, and the

same entries) but the first row is not that for argument 0 or 1 but for some other convenient

number, to stress the fact that the entry reaches a specific value such as a maximum or

a minimum for that particular value of the argument. Shifts occur, for instance, for the

planetary latitudes of certain planets in John of Murss Tables of 1321 (Chabs and Goldstein

2009, see esp. pp. 308309).

7 Salam and Kennedy 1967, p. 497. See also Debarnot 1987, p. 43 (table 9); and Jensen 1971.

102 chapter 5

Ibn al-Aclam (d. 985) and called attention to the fundamental relation of three

displacements for each planet (see eq. 26, below).8 Byzantine astronomers

apparently depended on their Muslim predecessors for displaced tables: see

Tihon 19771981, espec. 68:76 and 110. As far as we can determine, none of the

displaced tables compiled by astronomers in the Islamic World and Byzantium

were known in the West. Two Jewish astronomers in Provence, Levi ben Gerson

(d. 1344) and Immanuel ben Jacob Bonfils of Tarascon (fl. 1350), composed zijes

in Hebrew in which they used the principle of displaced tables for the times and

longitudes of syzygies (but not for the motion of the planets); however, there is

no evidence to suggest that they depended on Islamic or Byzantine sources.9

In astronomy written in Latin the tables described here are the first to use dis-

placed tables systematically, although we know of an earlier use of this type

of table by John Vimond around 1320, limited to his table for trepidation (or

access and recess).10

As mentioned above, the Tables for the Seven Planets have no accompanying

text; therefore, the following comments are based entirely on the information

provided by the tables themselves. As will be seen, the terminology in these

tables often differs from that commonly used at the time, and many of the

tables have a presentation that diverges from other tables with the same param-

eters that are based on the same model.

There follows a table of contents, arranged by section number.

1. Multiplication table

2. Mean motion of the Sun

3. Solar equation

4. Length of daylight, diurnal seasonal hours, and the equation of time

5. Mean motions of the Moon

6. Lunar latitude

7. Lunar equations

8 Displaced tables in Islamic astronomy are also discussed in Jensen 1971, Saliba 1976, Saliba

1977, Mercier 1989, and Van Brummelen 1998. For displaced tables in the Maghrib, see

Sams and Mills 1998, and Sams 2003.

9 For Levi ben Gerson, see Goldstein 1974, pp. 136146, 229241; for Bonfils, see Solon 1970,

pp. 34, 11, and Kremer (forthcoming). It is noteworthy that the Tabule frequentine in this

very same manuscript contain an adaptation of Bonfilss tables, where displaced tables

occur.

10 On Vimond, see Chabs and Goldstein 2003, pp. 275277; and Chabs and Goldstein 2004,

pp. 265267.

displaced tables in latin 103

8. Lunar node

9. Precession/trepidation

10. Planetary mean motions

11. Planetary equations and stations

11.1. Equation of center and equatio porcionis

11.2. Equation of anomaly near greatest distance

11.3. Equation of anomaly near least distance

12. Latitudes of the superior planets

13. Planetary visibility

14. Possibility of an eclipse

15. Eclipsed fraction of the solar and lunar disks

16. Latitudes of Venus and Mercury.

1 Multiplication Table

The first table in this manuscript (ff. 2r7v) is a multiplication table for base

60 arithmetic, with entries for each integer from 1 to 60. This table, presented

as a 60 60 square matrix, is quite common in manuscripts containing sets

of astronomical tables (see, e. g., Chabs and Goldstein 2012, p. 227). The fact

that f. 8rv is blank makes it uncertain if this multiplication table belongs to the

set examined here, for it could very well have been inserted by the copyist to

facilitate computation.

Folio 9r contains five tables for the mean motion of the Sun. The first lists the

Radices ad 32 annos post annum 1340, i.e., the values of the mean motion of the

Sun at the beginning of each year for a period of 32 consecutive years. The entry

for year 1 is 8s 16;26,51, and it corresponds to the mean argument of center

of the Sun at noon of February 28, 1342; the entry for year 32 is 8s 16;36,12. It

should be emphasized that a period of 32 years for the motion of the Sun is

very unusual; in any case, we have not found it in the previous literature, where

periods of 20, 24, and 28 abound (see Chabs and Goldstein 2012, p. 54). We also

note that the usual phrase anni expansi (expanded years, that is, those within

a cycle, in contrast to anni collecti for the years at the interval of a cycle) does

not appear here or in the headings of the other tables.

104 chapter 5

The other four tables for the Sun have a general title, beginning Tabula

porcionis solis ad annos radicum , where porcio has to be understood here as

argument of center.11 The second table on this folio gives the mean motion

of the Sun for accumulated months; its first entry corresponds to March (1s

0;33,18) and the last one to February (11s 29;45,39). These entries agree exactly

with those given by John of Lignres in his Tabule magne for a year beginning

in January (see Chabs and Goldstein 2012, Table 5.1b, pp. 5556), who also

used zodiacal signs of 30. The third table is headed ad annos perpetuacionis,

from 1372 to 1852, at 32-year intervals, and again we see the use of non-standard

terminology, for the phrase anni collecti is totally absent from this set of tables.

The fourth and fifth tables are, respectively, for the mean motion of the Sun for

days from 1 to 31, and for hours and fractions of an hour from 1 to 60. The entry

for argument 1d is 0s 0;59,8.

As mentioned above, at the bottom of the folio we are told that the Radix

porcionis solis ad anno domini 1340 completo is 8s 16;41,49. From this value and

the entry for year 32 (8s 16;36,12) we derive a mean motion in the argument

of center of 0;59,8,13,32,49/d. As indicated previously, the entries in these

tables represent the mean argument of center of the Sun, , and this is another

peculiar characteristic of the Tables for the Seven Planets, because in other sets

the mean solar motion, that is, , is usually tabulated. The two quantities are

related by means of a, the longitude of the solar apogee, for = + a. A small

table, found later in the text (f. 17r), gives the motion of the solar apogee. It

is entitled Motus augium equatus and the entries are yearly values, probably

corrected for precession. The entry for 32 year is 0;19,42,35, which implies a

daily motion of 0;0,0,6,4/d. By adding this value to the one derived for the

mean motion in the argument of center of the Sun, the result is 0;59,8,19,37/d,

in good agreement both with the value used by Vimond (0;59,8,19,37,4/d) and

the value in the Parisian Alfonsine Tables (0;59,8,19,37,19,13,56/d) for the mean

11 We know of no prior usage of porcio for the argument of center of the Sun. There are,

however, examples where it means anomaly in the case of the planets. On porcio (or

portio) in the sense of anomaly, see Goldstein, Chabs, and Mancha 1994, pp. 6364

(see also Nallino 19031907, 2:328), where a reference is given to John of Lignress use

of this word in the canons to his tables, whose incipit is Priores astrologi motus corporum

celestium (1322). The Arabic for anomaly is ia, which means portion and one of the

Hebrew terms for anomaly, manah, also means portion. Both the Latin and Hebrew terms

apparently come from Arabic. The fact that in our text the argument of center of the Sun

is considered the anomaly indicates that the author had in mind an epicyclic model for

the Sun, which is equivalent to the eccentric model according to Apolloniuss theorem

(see Almagest iii.3; Toomer 1984, pp. 141153).

displaced tables in latin 105

solar motion in longitude.12 It would thus seem that the Tables for the Seven

Planets belong to this tradition. We will show that the rest of the tables provide

additional evidence of this relationship.

3 Solar Equation

The solar equation is given on ff. 9v10r. The title of the corresponding table

(Tabula equacionum solis cum auge eius equata) indicates that the value for

the solar apogee has already been taken into consideration. The argument,

which we call k , in parallel with the standard usage of ,13 begins at 0, not

at 1 as was the common practice at the time. The entry for 0 is 2s 29;37,9

(= 89;37,9), and corresponds to the longitude of the solar apogee at epoch,

89;36,59 (see n. 4). The minimum 2s 27;27,9 is reached at 9294, and the

maximum 3s 1;47,9 at 267269. Subtracting algebraically either of these two

values from the entry for 0, one gets 2;10,0, which is the standard Alfonsine

parameter for the maximum solar equation.14 Therefore, all entries for the solar

equation, c(k ), are displaced upwards by 89;37,9, compared to the entries in

the standard tables in the Alfonsine corpus, c(), to make all entries positive,

thus allowing the user to avoid dealing with a set of complicated rules for

adding and subtracting various terms:

This is the first example of a displaced table in this set, where the argument is

the mean argument of center, k , and the entries represent c(k ), which is the

sum of the Alfonsine solar equation (whether positive or negative), c(), and

the longitude of the solar apogee, a. In this case the displacement is vertical

(see Figure 5.a).

12 For Vimond, see Chabs and Goldstein 2004, p. 221; for the Parisian Alfonsine Tables, see

Ratdolt 1483, f. d5r. For a modern edition of these tables, based on Ratdolt 1483, see Poulle

1984.

13 Throughout this paper capital letters are used to represent variables and functions in the

Tables of the Seven Planets.

14 As in the tables of John Vimond (Chabs and Goldstein 2004, p. 223), the solar equation

with the same maximum of 2;10is not explicitly given. For the maximum solar equa-

tion, 2;10, in the Parisian Alfonsine Tables, see Ratdolt 1483, f. e3r.

106 chapter 5

The canons to this table, which are not extant, should have had an instruc-

tion indicating something like to find the true position of the Sun, enter the

table with the mean argument and add what you find to it or, in algebraic

terms,

(3) l = k + c(k ).

And, indeed, in Ptolemys model for the equation of the Sun, , the true longi-

tude of the Sun is obtained adding its mean longitude, , to the solar equation:

= + c(). Now the solar equation is also the difference between the true argu-

ment of center of the Sun, , and its mean argument of center, : it is negative

if 0 180, and positive if 180 360. The Tables of the Seven Planets

yield the same result as that deduced from Ptolemy, for

= + c()

= ( + a) + c()

= + (a + c())

= k + c(k ) = l,

provided that k = , which is the case here. The two terms in L are positive and

both are explicitly tabulated on ff. 9v10r.

displaced tables in latin 107

Time

On ff. 10v11v there is a table with a column for the argument from 1 to

30, and three other columns for each of the twelve zodiacal signs, begin-

ning with Aries. The headings are Hora equalis, tempus horarum, and equatio

dierum. The first entries for Aries 1 are 12;3h, 15;4, and 8;28 min, respec-

tively. The entries in the first column, under Hora equalis, display the length

of daylight (i.e., the time interval from sunrise to sunset) for a given locality;

it reaches a maximum, 15;10h, at Gem 28 Cnc 2, and a minimum, 8;50h, at

Sgr 28 Cap 2. If we consider the obliquity of the ecliptic, , to be 23;33,

the resulting geographical latitude, , for which the table is valid is = 42;44,

which agrees fairly well with the parallel through Toulouse (rather than with

that of Paris, where the longest daylight is 16;0h). The geographical longitude

(see above) and latitude that we derive from the tables do not yield good

agreement with any place where astronomy was practiced in the 14th cen-

tury, but a locality approximately fitting both computed coordinates is Perpi-

gnan.

The entries in the second column, under tempus horarum, display the length

of a diurnal seasonal hour (i.e., a twelfth of the length of daylight); it reaches a

maximum, 18;57h, at Gem 27 Cnc 3, and a minimum, 11;3h, at Sgr 27 Cap

3. These two columns are mutually consistent.

The third column displays the equation of time (i.e., the difference between

apparent and mean time where apparent time is counted from true noon,

that is, the moment that the true Sun crosses the meridian, and mean time

is counted from mean noon). The argument is the solar longitude expressed

in degrees and the entries are given in minutes and seconds of an hour. The

extremal values are 0;21,24h (at Tau 26), 0;11,16h (at Leo 5), 0;31,48h (at Sco 8),

and 0;0,0h (at Aqr 2023). When converting these values into time-degrees,

that is, multiplying each entry by 360/24h, we obtain 5;21, 2;49, 7;57, and

0,0, respectively, which are the characteristic values found in the equation

of time ascribed to Peter of Saint Omer and used by John of Lignres, among

others (see, e. g., Chabs and Goldstein 2012, pp. 3740).

The mean motions of the Moon are addressed in five tables on ff. 12r13r. In

addition to columns for the various arguments, each table has three columns

headed centrum lune (here meaning double elongation), porcio lune (argument

108 chapter 5

of lunar anomaly), and medius locus lune (mean longitude of the Moon). Again,

for the first two quantities this is not the standard terminology.

The first table lists the radices at the beginning of each consecutive year

from 1342 for a period of 32 years. The entries for year 1 are 6s 3;36,2 (double

elongation), 3s 17;36,22 (anomaly), and 8s 10;16,38 (longitude). The second

table gives the mean motion of the Moon for the months; the first entries,

for March, are 1s 5;49,36 (double elongation), 1s 15;0,53 (anomaly), and 1s

18;28,6 (longitude). The third and fourth tables are, respectively, from 1372 to

1852, at 32-year intervals, and for hours and parts of an hour, from 1 to 60. The

fifth table is for days from 1 to 31, and the first entries for argument 1d for the

three variables are 0s 24;22,54 (double elongation), 0s 13;3,54 (anomaly), and

0s 13;10,35 (longitude), confirming that the first column is indeed the double

elongation, for 0s 24;22,54 = 2 (0s 13;10,35 0s 0;59,8).

As was the case for the Sun, at the bottom of f. 12v we are given the radices for

the three variables, anno Christi 1340 perfecto: 9s 14;21,15,14 (double elongation),

0s 18;53,8,0 (anomaly), and 4s 0;53,34,57 (longitude). We note that the values

are given here to thirds and that the last digit for the anomaly is 0, indicating

that the author was sensitive to accuracy. From the values of the radices and

those for year 32 in the first table for the mean motions of the Moon, we

derive the following values for the mean motions: 24;22,53,23,16/d (double

elongation),15 13;3,53,57,30/d (anomaly),16 and 13;10,35,1,15/d (longitude).17 All

of them are typical parameters of the Parisian Alfonsine Tables (Ratdolt 1483,

ff. d5vd7r).

We have recomputed the radices using the Parisian Alfonsine Tables for

noon of February 28, 1341 for the meridian of Paris, 12 east of Toledo, that is,

at 0;2d before noon in Toledo. The results are presented in Table 1.

The agreement is good for the double elongation and the mean argument of

center of the Sun, but much better results are obtained when recomputing the

15 The value for year 32 given in the text is 4s 15;17,11, and the difference between this value

and that of the radix is 210;55,56. Thus the amount of the double elongation in 32 years

of 365;15 days is 210;55,56 and 791 complete revolutions, leading to a daily mean motion

of 24;22,53,23,16 /d.

16 The value for year 32 given in the text is 12s 22;28,24 (read 2s 22;28,24), and the difference

between this value and that of the radix is 63;35,6. Thus the amount traveled by the

Moon in the argument of anomaly in 32 years of 365;15 days is 63;35,6 and 424 complete

revolutions, leading to a daily mean motion of 13;3,53,57,30/d.

17 The value for year 32 given in the text is 1s 16;35,39, and the difference between this value

and that of the radix is 285;42,4. Thus the amount traveled by the Moon in longitude in

32 years of 365;15 days is 285;42,4 and 427 complete revolutions, leading to a daily mean

motion of 13;10,35,1,15/d.

displaced tables in latin 109

table 1 Recomputation for the Moon at epoch: noon, Feb. 28, 1341

Double elongation 9s 14;21,15,14 284;20,13,41 0; 1, 1,33

Mean argument of anomaly 0s 18;53, 8, 0 32; 1,34,53 13; 8,26,53

Lunar mean longitude 4s 0;53,34,57 128;29, 2, 0 7;35,27, 3

table 2 More detailed recomputation for the Moon at epoch: noon, Feb. 28, 1341

Double elongation 9s 14;21,15,14 284;21,14,39 +0; 0, 0,35

Mean argument of anomaly 0s 18;53, 8, 0 32; 2, 7,32 13; 8,59,32

Lunar mean longitude 4s 0;53,34,57 128;29,34,57 7;36, 0, 0

radices for 0;1,57,30d before noon in Toledo, that is, for a locality 11;47,45 east

of that Spanish city. The results are presented in Table 2.

From Table 2 it follows that the mean argument of anomaly of the Moon

tabulated here is diminished by 13;9 from that obtained from the Parisian

Alfonsine Tables for a locality near the meridian of Paris, and that the mean

lunar longitude is diminished by exactly 7;36. As will be seen in 7, the

displacements applied to the two quantities correspond, on the one hand, to

the maximum value of the lunar equation of center in the Parisian Alfonsine

Tables (13;9) and, on the other hand, to the sum (7;36) of two parameters, the

maximum values for the equation of anomaly (4;56) and for the increment

(2;40). So, if we call the mean argument of anomaly in the Tables of the Seven

Planets, and that in the Parisian Alfonsine Tables, we have

(4) = 13;9.

Similarly, if l is the mean lunar longitude in the Tables of the Seven Planets,

and that in the Parisian Alfonsine Tables, we have

(5) l = 7;36.

110 chapter 5

figure 5.b

observer, d is the center of the deferent circle

whose radius is dc, c is the center of the

epicycle, s is the direction to the mean Sun, l is

the true position of the Moon, v is the direction

to the vernal point, and d is 180 from d on a

circle about o with radius od. The angle is

the elongation of the mean longitude of the

Moon from the mean longitude of the Sun,

angle is the mean argument of anomaly

counted from the mean apogee e , whose

position is fixed by the direction dc, ae is the

true lunar apogee, angles c3 and c are the

corrections to the mean lunar longitude, and

angle is the true longitude of the Moon.

6 Lunar Latitude

The table for the lunar latitude (f. 13r) has a maximum of 5;0,0 at 90; it is

the standard table found in many zijes, including those used by Alfonsine

astronomers (see Chabs and Goldstein 2012, pp. 103104).

7 Lunar Equations

The lunar equations are presented in three separate tables. The author, well

aware of Ptolemys second lunar model (see Figure 5.b), has split the treatment

of the lunar equations, c3 and c, according to the two independent variables

involved: the double elongation, 2, and the true argument of anomaly, . The

true longitude of the Moon is the sum of four positive terms, all of which are

tabulated (see eq. 14, below), whereas in the standard Alfonsine tables the true

longitude of the Moon is the algebraic sum of three terms (see eq. 8, below).

Despite the differences in these procedures, we show that the results are the

same.

The longitude of the Moon, , is found by applying an equation, c, to its mean

longitude, , which is a linear function of time: = + c. Moreover, the mean

argument of anomaly, , is also a linear function of time. In Ptolemys second

lunar model (Almagest v.8), the lunar equation, c, depends on two indepen-

dent variables: the true argument of anomaly, (angle aecl), and the double

elongation, 2 (twice the angular distance between the mean longitude of the

Moon and the mean longitude of the Sun). In this context the lunar equation

displaced tables in latin 111

6th columns of Ptolemys table in Almagest v.8. The vast majority of later sets

of tables used Ptolemys approach, although some of them have slightly modi-

fied parameters or rearrangements of the order of the columns. In the Parisian

Alfonsine Tables (pat) these three columns correspond respectively to the 6th,

5th, and 4th columns; in the following, we will use the conventions of pat to

identify the columns ci:

where

(7) = + c3(2).

In eqs. 6 and 7, c3(2) is the equation of center, c4(2) represent the minutes of

proportion, c5() is called the increment, and c6() is the equation of anomaly.

Thus, the true longitude of the Moon is given by

The first table (ff. 13v14r) is for the equation of center; it has a column for

the argument, from 0 to 29, and then columns for each sign, from 0s to

11s. For each sign and each degree of the argument, we are given entries for

the equation of center and the minutes of proportion (c3 and c4, respectively,

in pats arrangement, both variables depending on the double elongation).

For the tabulated equation of center, the entry for 0s 0 is 13;9. It reaches

a maximum of 26;18 (twice 13;9) at 114115, and a minimum of 0;0 at

245246. So the equation of center of the Moon is tabulated with a vertical

displacement, for its entries are displaced upwards by 13;9 with respect to the

standard Alfonsine table (see Chabs and Goldstein 2012, Table 6.2a, p. 71). If

we call c3(2) the equation of center in the corresponding Alfonsine table, the

tabulated entries, c3(2), for the equation of center in the Tables for the Seven

Planets are given by

It is worth noting that eq. 7 and the definition of in eq. 4 imply that a = .18

112 chapter 5

For the minutes of proportion the entry for 0s 0 is 0. The entries increase

monotonically to 60 at 173187, and decrease to 0 at 349371. Except

for copyists errors, they agree with those in the Parisian Alfonsine Tables.

Therefore,

Let us now turn to the third table (ff. 14v15r), which is for the equation of

anomaly; it has a column for the argument, from 0 to 29, and then columns

for each zodiacal sign, from 0s to 11s. For each sign and each degree of the

argument, we are given entries for the equation of anomaly and the dyversitas

dyametri proporcionalis19 (corresponding to c6 and c5, respectively, in pats

arrangement, both variables depending on the true argument of anomaly). For

the equation of anomaly, the entry for 0s 0 is 4;56,0. It reaches a minimum

of 0;0,0 at 9199 and a maximum of 9;52,0 (twice 4;56,0) at 264, that is,

the vertical displacement is 4;56. This parameter was systematically used by

Parisian Alfonsine astronomers, but its origin goes back much earlier (Chabs

and Goldstein 2003, pp. 252253). Thus, if we call c6() the equation of anomaly

in the corresponding Parisian Alfonsine table, the tabulated entries, c6(a),

for the equation of anomaly in the Tables for the Seven Planets are given

by

The first entry in the column for the dyversitas dyametri proporcionalis, usu-

ally called increment is 2;40; it reaches a minimum of 0;0 at 113119 and

a maximum of 5;20 (twice 2;40) at 251257. Except for copyists errors, the

entries agree with those in the Parisian Alfonsine Tables, but for the vertical

displacement of 2;40. Again, if we call c5() the increment in the correspond-

ing Parisian Alfonsine table, the tabulated entries, c5(a), for the increment in

the Tables for the Seven Planets are given by

In the second table (f. 14v), headed dyversitas dyametri centralis, the argument is

given at intervals of 1, from 0 to 60, and the entries, d(m), decrease monotoni-

19 The word dyversitas, used in the headings of the lunar and the planetary equations, is often

spelled diversitas.

displaced tables in latin 113

centralis (excerpt)

() ()

0 2;40, 0

10 2;13,20

20 1;46,40

30 1;20, 0

40 0;53,20

50 0;26,40

60 0; 0, 0

cally from 2;40,0 (at 0) to 0;0,0 (at 60). Moreover, the entries show a constant

decrease of 0;2,40 for each degree of the argument (see Table a); hence,

The tables described above yield the same results as those obtained from the

Parisian Alfonsine Tables for the true position of the Moon, and were computed

with a similar expression (see eq. 8). If and l are the true lunar longitudes in

the Parisian Alfonsine Tables and the Tables for the Seven Planets, respectively,

then using eqs. 5, 8, 10, 11, and 12, can be written as

= (l + 7;36) + (c6(a) 4;56) + (c5(a) 2;40) c4(2)

= (l + c6(a) + 2;40) + c5(a) c4(2) 2;40 c4(2)

= (l + c6(a) + c5(a) c4(2)) + 2;40 2;40 c4(2) = l.

The term, 2;40 (1 c4(2)) = d(m), is found in the table for the dyversitas

dyametri centralis, and therefore the expression to compute the true longitude

of the Moon is

114 chapter 5

where all the terms are tabulated and none of them is to be subtracted. Note

that there is no counterpart to the function d(m) in the standard Alfonsine

tables.

The canons to this set of tables, which are not extant, should have had

an instruction more or less in the following terms: To determine the true

position of the Moon, find the mean longitude of the Moon, and keep it; find

the double elongation, and enter with it in the table for the equation of center

and the minutes of proportion, and keep them. Find the mean argument of

anomaly, and add to it the equation of center to obtain the true argument of

anomaly, and enter with it in the table for the equation of anomaly and the

increment, and keep them. Enter the table for the dyversitas dyametri centralis

with the minutes of proportion, and keep what you find there. Then add the

mean longitude to the true argument of anomaly, and add the result to what

is obtained from multiplying the increment by the minutes of proportion. To

the value obtained add what you found in the table for the dyversitas dyametri

centralis.

To illustrate eq. 14 consider the Moon at epoch (noon, February 28, 1341),

when the double elongation, 2, was 4,45;9,0 (= 285;9,0). The equation of

center in the standard Alfonsine tables is c3(2) = 10;26 and, because of

eq. 9, c3(2) = 2;43 (= 10;26 + 13;9). The corresponding minutes of pro-

portion are c4(2) = 19 and c4(2) = 19, because of eq. 10. In this case the

term d(m) = 1;49,20. Now, the argument of anomaly, , according to the stan-

dard Alfonsine tables, is 32;27,43, and thus = 22;1,43 (= 32;27,43 10;26).

From eq. 4, = 19;18,43, and thus a = 22;1,43 (= 19;18,43 + 2;43). The tab-

ulated values for the equation of anomaly, c6(a), and the dyversitas dyametri

proporcionalis, c5(a), are 3;13,19 and 1;48, respectively, whereas the values

for the equation of anomaly, c6(), and the increment, c5(), found in the

standard Alfonsine tables are 1;42,41 and 0;52, respectively. We note that

eqs. 11 and 12 hold. According to eq. 6, using the standard Alfonsine tables

the correction to be applied is c = (1;42,41) + (0;52) (19/60) = 1;59,9,

whereas using the Tables of The Seven Planets c = 3;13,19 + 1;48 (19/60) =

3;47,31.

With the standard Alfonsine tables the mean lunar longitude at epoch, ,

is computed to be 128;55,23; thus, with eq. 8, the true lunar longitude of the

Moon at epoch is

= 128;55,23 1;59,9 = 126;56,14.

displaced tables in latin 115

With the Tables for the Seven Planets we obtain l = 121;19,23 (see eq. 5); thus,

with eq. 14, the true longitude of the Moon at epoch is

= 121;19,23 + 3;47,31 + 1;49,20 = 126;56,14,

in agreement with .

8 Lunar Node

Folio 16r contains five tables for the mean motion of the lunar node. The first

lists the values of the mean motion of the node at the beginning of each

year, beginning with 1342, for a period of 32 years. The entry for year 1, i.e.,

1342, or more precisely, Feb. 28, 1341, is 3s 22;3;17. The second table gives the

mean motion of the Sun for the months; its first entry corresponds to March

(0s 1;38,30) and the last one to February (0s 19;19,42). The third table lists

values from 1372 to 1852, at 32-year intervals. The fourth and fifth tables are,

respectively, for the mean motion of the lunar node for days from 1 to 31, and

for hours and parts of an hour from 1 to 60. The entry for argument 1d is 0s

0;0,3,11. As was the case for other mean motions, at the bottom of the folio

we are told that the radix at anno Christi 1340 perfecto is 3s 2;43,34,43. From

this value and that for year 32 we derive 0;3,10,38,7/d as the mean motion of

the lunar node,20 in full agreement with the corresponding parameter in the

Parisian Alfonsine Tables.

9 Precession/Trepidation

framework of the Parisian Alfonsine Tables. In general, the approach found

in the Alfonsine corpus is to treat two terms separately: a linear term, usually

called motion of the apogees and the fixed stars, based on a mean motion

of 0;0,0,4,20,41,17,12/d, and presented in a single table; and a periodic term,

20 The value for year 32 given in the text is 11s 21;39,20, and the difference between this

value and that of the radix is 258;55,45,17. Thus the amount traveled by the lunar node in

32 years of 365;15 days is 285;42,4 and one complete revolution, leading to a daily mean

motion of 0;3,10,38,7/d.

116 chapter 5

usually called motion of access and recess of the 8th sphere, requiring the use

of two tables (one for the mean motion, based on a value of 0;0,0,30,24,49,0/d,

and another for its equation, found in a separate table with a maximum of

9;0,0): see Chabs and Goldstein 2012, pp. 4852.

The author of the Tables for the Seven Planets follows the same approach,

uses the same parameters, but gives a different presentation. Folio 16v has four

tables under the general title Tabula motus augium et stellarum fixarum atque

8 spere, displaying the two mean motions of the components of Alfonsine

variable precession. Each table has two columns, in addition to that for the

argument: the mean motion of the apogees and the fixed stars on the one hand,

and the mean motion of the 8th sphere on the other. The motion of trepidation

is to be applied to the positions of the fixed stars as well as to the positions of

the planetary apogees.

The first table gives entries for both quantities at the beginning of each

year for a period of 32 consecutive years. The entries for year 1 are 0s 0;51,12

(apogees) and 2s 8;10;51 (8th sphere), and those for year 32 are 0s 1;4,52

(apogees) and 2s 9;46;31 (8th sphere). The second table gives the mean motion

of both quantities for accumulated months. The third table, headed ad annos

perpetuacionis, has entries for years 1372 to 1852, at 32-year intervals, of both

quantities, whereas the fourth table is for their mean motions for days from

1 to 31. As was the case for other quantities, at the bottom of the folio we are

given the radices for both quantities, anno Christi 1340 perfecto: 0s 9;50,45,59,27

(apogees), and 2s 8;7,45,20,10 (8th sphere).21 From these radices and the en-

tries for year 32 we derive the following mean motions: 0;0,26,28 /y (apogees)

and 0;3,5,11/y (8th sphere), which correspond to about 0;0,0,4,20,51/d (apo-

gees) and 0;0,0,30,25,12 /d (8th sphere), that is, 1 revolution in 49,000 years

and 7,000 years, respectively. These are the standard values found in Alfonsine

astronomy for precession/trepidation. Combining the mean motion of the 8th

sphere and its radix, we find that the argument of the 8th sphere was 0 about

1324 years and 9 months before the epoch of these tables (February 28, 1341),

that is, in May 16ad. This date is explicitly given by Giovanni Bianchini (15th

century) as the epoch of the Alfonsine model; see Chabs and Goldstein 2009,

p. 32, and the Tables for the Seven Planets offer a justification of it.

It remains to compute the periodic term of Alfonsine trepidation, that is,

the equation of the 8th sphere, which is tabulated on f. 17r. Indeed, the entries

given in the table are based on the standard Alfonsine equation of the 8th

21 We note that the radix for the apogees has the wrong number of degrees (probably a copy-

ists error); to be consistent with the 32 values in the table, it should be 0s 0;50,45,59,27.

displaced tables in latin 117

and vanishes at 0 and 180 (see also Ratdolt 1483, f. d3v). The corresponding

entries on f. 17r are 9;0,0 at 0, 18;0,0 (maximum) at 90, 9;0,0 at 180,

0;0,0 (minimum) at 270, and 0;0,0 at 360; thus, the entries are displaced

upwards by 9 from the standard table, and can be represented by the modern

expression (where y is the entry and x is the argument):

Interestingly, the only example in Latin of a displaced table previously known

to us is that of John Vimond (see n. 10), whose table for the equation of

the 8th sphere presents both a vertical displacement (8;17) and a horizon-

tal displacement (113). However, Vimonds purpose was not to avoid nega-

tive numbers in his table (otherwise he would have used a vertical displace-

ment of at least 9;0, as here); rather, he wished to set the origin of the equa-

tion of the 8th sphere at the value (8;17) it had at the time he composed

his tables (1320). Apparently, the author of the Tables for the Seven Planets

did not have the same goal as Vimond, although he used the same princi-

ple.

As mentioned above, on folio 17r there is also a small table entitled Motus

augium equatus ad 32 annos post annum 1340, listing 32 values. The first entry is

0;0,37,11 and the last entry, for 32 years, is 0;19,42,35. The entries show a steady

increase but the line-by-line differences do not reveal a clear pattern. A com-

plement to this table is found on f. 18r among the tables for the mean motions,

where we find one headed Motus augium equatus ad annos perpetuacionis, list-

ing entries from 1372 to 1852 at intervals of 32 years. The first entry is 0;19,42,35

but, as in the previous case, the line-by-line differences do not vary in a smooth

way. With all these cautions, from the entry for 32 years it is possible to derive

a motion of the apogee of 0;0,36,57/y, or 0;0,0,6,4/d, which is the amount to

be added to the mean motion in solar anomaly to obtain the mean motion in

the argument of solar longitude (see 2). The value we deduce from the table is

certainly close to Ptolemys value for precession of 1 in 100 years, that is 0;0,36

in 1 Egyptian year of 365 days.

Next come tables for the mean motions, equations, and stations of the five

planets, on ff. 17v22r (Saturn), ff. 22r26v (Jupiter), ff. 27r31r (Mars), ff. 31r35v

118 chapter 5

Centrum Porcio

Jupiter 0s 1;38,17,48 4s 29;28,59,11

Mars 1s 25; 6,26, 4 2s 24;48,42,[..]

Venus 6s 25;41,48,36 7s 17; 6,11

Mercury 3s 19;27,49 3s 23;27,56

(Venus), and ff. 35r40r (Mercury). In this section we review the mean motions

of the planets, for each of which we are given five tables. As we will see, all the

mean motions are displaced with respect to those in the standard Alfonsine

tables.

The first of these five tables lists the radices of two variables, centrum (argu-

ment of center) and porcio (here meaning the displaced argument of anomaly,

as will be seen below), for the beginning of each year after 1340 for a period

of 32 consecutive years. The entries for both variables for year 32 are, respec-

tively: 1s 6;7,52 and 1s 14;30,11 (Saturn); 8s 12;53,57 and 8s 18;7,43 (Jupiter);

2s 0;5,0 and 2s 19;44,28 (Mars); 6s 25;36,12 and 7s 22;56,20 (Venus); and 3s

19;22,12 and 2s 4;36,10(Mercury). The second of these tables gives the mean

motion of the planets for accumulated months, beginning in March, for the

two variables. The third of these tables is headed ad annos perpetuacionis and

displays entries for the two variables from 1372 to 1852, at 32-year intervals.

The fourth and fifth of these tables are, respectively, for the mean motion of

the planets for days from 1 to 31, and for hours and parts of an hour from 1 to

60.

From the values of the radices and those for year 32 in the first table for the

mean motions of the planets, we derive the following values for the daily mean

motions: 0;2,0,29,13 /d (Saturn, argument of center); 0;4,59,9,23/d (Jupiter,

argument of center); and 0;31,26,32,35/d (Mars, argument of center). These

three parameters differ by 0;0,0,6,4/d from the standard Alfonsine mean

motions, as was the case for the argument of solar anomaly, indicating that the

precession of 1 in 100 years was applied to the planets as well as to the Sun.

We have also derived the daily mean motions of anomaly for the inferior plan-

ets: 0;36,59,27,24/d (Venus) and 3;6,24,7,43/d (Mercury), which agree exactly

with the corresponding parameters in the Parisian Alfonsine Tables.

As was the case for the Sun and the Moon, we are also given the radices anno

christi 1340 completo, which we display in Table 3.

displaced tables in latin 119

table 4 Recomputation for the planets at epoch: noon, Feb. 28, 1341

center TextComp. value anomaly TextComp. value

Jupiter 19;18,21 17;40, 3 * 18 155;28,57 5;59,58 6

Mars 116; 6,24 60;59,58 61 96;48,44 12; 0, 2 12

Venus 256;41,56 51; 0, 7 51 230; 6, 9 2;59,58 3

Mercury 137;27,46 27;59,57 28 117;27,52 3;59,56 4

* The difference is 18;0,3 if we take the radix for the argument of center in Table 3 to

be 1;18,17,48, which is the result one would obtain from a set of standard Alfonsine

tables.

We have recomputed the mean positions of the planets using the Parisian

Alfonsine Tables for noon of February 28, 1341 for the meridian of Paris, 12 east

of Toledo. The results are presented in Table 4.

From Table 4 it follows that the arguments of center of the five planets given

in these tables are diminished, but for a few seconds, by 14, 18, 61, 51, and

28 from those obtained from the Parisian Alfonsine Tables for a locality near

the meridian of Paris. Table 4 also indicates that the entries for the argument

of anomaly are diminished, but for a few seconds, by 7, 6, 12, 3, and 4 from

those derived from the Parisian Alfonsine Tables. As will be shown in 11, all

these integer values correspond to the horizontal and vertical displacements

(kh3 and kv3, respectively) used in the tables for the equations of the planets. If

we call and the planetary arguments of center and anomaly in the Parisian

Alfonsine Tables, the corresponding quantities in the Tables for the Seven

Planets are defined by:

(16) k = kh3,

and

(17) = kv3.

120 chapter 5

figure 5.c

observer, d is the center of the deferent

circle acp, a is the apogee, p is the

perigee, e is the equant point, ov is the

direction to the vernal point on the

ecliptic, s is the direction to the mean

Sun, e is the mean apogee on the

epicycle whose center is c, ae is the true

apogee on the epicycle, and are the

mean arguments of center and anomaly,

respectively, m is the mean position of the

planet, and m is its true position (where,

for an outer planet, cm is parallel to os).

In this section we review the equations and the stations of the five planets, for

each of which we are also given five tables, in which the equation of center

(one table; see 11.1) and the equation of anomaly (four tables; see 11.2 and

11.3) are treated separately. Figure 5.c displays the model for Mars. As we will

see, the true longitude of a planet is found by adding four positive terms, all of

which are tabulated (see eqs. 43 and 45, below). These terms are all different

from those in the standard Alfonsine tables, where the true longitude is found

as the algebraic sum of four terms. The author of these tables has considered

two cases: (1) the epicyle is near the apogee of the deferent, and (2) the epicycle

is near the perigee of the deferent. In case 1 he has the distance of the center of

the epicycle vary from its maximum to its mean value, whereas in case 2 he has

the distance of the center of the epicycle vary from its mean to its minimum

value.

In the first table the argument ranges from 0 to 29 for each zodiacal sign, and

for each degree we are given three entries: equatio centri, equatio porcionis, and

minutes of proportion. The first two are given in degrees and minutes, and the

third entry is given in minutes (see Table b).

Table 5 displays significant values of the equation of center for the five

planets. To determine the relation between the equation of center presented

here and that in the Parisian Alfonsine Tables, let us consider the case of

Mars. The entry for 0 is 2;29; the minimum (0;36) is reached at 3135

displaced tables in latin 121

and minutes of proportion for

Saturn (excerpt)

centri porcionis of proportion

() () () ()

0 5;30 8;30 0

1 5;24 8;36 0

2 5;17 8;43 1

60 0;48 13;12 37

79 0;29 13;31 58

80 0;29 13;31 59

81 0;30 13;30 1

82 0;30 13;30 2

255 13;31 0;29 0

265 13;23 0;37 1

266 13;22 0;38 59

267 13;21 0;39 58

and the maximum (23;24) at 203207. The difference between the two

extremal values is 22;48, that is, twice 11;24, the standard parameter used

for Mars in the Parisian Alfonsine Tables. The entries in the table are dis-

placed exactly 12 upwards as compared with the corresponding entries in

the Parisian Alfonsine Tables, and 12 is the minimum integer for a vertical

displacement ensuring that all entries for the equation of center are positive.

But the entries for the equation of center have a second displacement, a hori-

zontal displacement, of 61 with respect to the table in the Parisian Alfonsine

122 chapter 5

Entry equation equation lying displace- displace-

for 0 of center of center parameter ment (kv3) ment (kh3)

(7680) (252256)

Jupiter 4;15 0; 3 11;57 5;57 6 18

(7278) (246252)

Mars 2;29 0;36 23;24 11;24 12 61

(3135) (203207)

Venus 1;21 0;50 5;10 2;10 3 51

(3747) (211221)

Mercury 2;47 0;58 7; 2 3; 2 4 28

(6569) (235239)

Tables. With the notation used above, k = 61. If we call c3() the equa-

tion of center in the corresponding Parisian Alfonsine table,22 the tabulated

entries, c3(k ), for the equation of center in the Tables for the Seven Planets

are given by c3(k ) = c3(k + 61) + 12. This is the first example of a horizon-

tal displacement in these tables. For each planet the corresponding displace-

ment applied to (see Table 5) accounts for the difference between the entries

given in the Tables for the Seven Planets and the Parisian Alfonsine Tables.

Thus, the general rule for the equation of center of the planets can be written

as

where kh3 and kv3 are the horizontal and vertical displacements associated

with c3(), the equation of center in the Parisian Alfonsine Tables. Once c3(k )

is known, the true argument of center, k, is defined as:

(19) k = k + c3(k ),

22 In the Parisian Alfonsine Tables the equation of center is displayed in the third column in

the tables for planetary equations, hence the notation used here.

displaced tables in latin 123

on analogy with the expression for the true argument of center, , used in the

Alfonsine tables,

(20) = + c3(),

All the underlying parameters displayed in Table 5 are strictly in the tradition

of the Alfonsine corpus. Interestingly, the function for Venus is presented with

displacements that differ from that for the Sun (see 3, above), although they

share the same maximum equation, 2;10.

As for the column labeled equatio porcionis, note that for a given argument

its entries and those corresponding to them in the equation of center always

add up to an integer number of degrees: 14 (Saturn), 12 (Jupiter), 24 (Mars),

6 (Venus), and 8 (Mercury). This is exactly twice the vertical displacement

applied to each planet, that is, 2 kv3. Thus, if we call the equatio porcionis

e(k ), it follows that the author of the Tables of the Seven Planets defined it

as

below: there is no counterpart to the function e(k ) in the standard Alfonsine

tables. Figure 5.d shows the equation of center and the equatio porcionis in the

case of Mars. The other planets follow the same pattern (see Table 5).

When e(k ) is known, the true anomaly, a, is defined as:

124 chapter 5

(22) a = + e(k ),

on analogy with the expression for the true argument of anomaly, , used in the

Alfonsine tables,

(23) = c3()

In the third column of this table the entries for the minutes of proportion

are also dependent upon those in the Parisian Alfonsine Tables (see Figure

5.e for the case Mars) except for two features: a horizontal displacement, dif-

ferent for each planet, and a vertical displacement of 60, which is limited to

the positions when the planet is near its apogee. Note that in the Parisian

Alfonsine Tables the minutes of proportion for these positions are subtrac-

tive, whereas they are additive for the positions near perigee. Consequently,

in order to avoid subtractions, the author of the Tables for the Seven Plan-

ets only applied a vertical displacement of 60 to half of the column, that

is, where the center of the epicycle is near apogee, and left unchanged the

additive part (for details see Table 6). With this approach, if an entry, c4(),

gives the minutes of proportion in the Parisian Alfonsine table,23 the cor-

responding entry in the Tables for the Seven Planets in the case of Mars is

c4(k) = c4(k + 49) + 60, near apogee, and c4(k) = c4(k + 49), near perigee.

Thus, the general rule for the equation of center of the planets can be written

as

near perigee, where kh4 and kv4 (= 60) are the horizontal and vertical displace-

ments, respectively, associated with c4(), the minutes of proportion. Figure 5.e

23 In the Parisian Alfonsine Tables the minutes of proportion are displayed in the fourth

column in the tables for planetary equations; hence the notation used here. Note, however,

that in the corresponding table in Ptolemys Almagest they are found in the eighth column

and depend on the mean argument of center, whereas in the Parisian Alfonsine Tables

they are a function of the true argument of center. For this function in the Almagest, see

Neugebauer 1975, pp. 185186.

displaced tables in latin 125

displays the minutes of proportion for Mars; the other planets follow the same

pattern (see Table 6).

As in previous cases, the vertical displacements are intended to avoid cum-

bersome rules for addition and subtraction corresponding to the simple rules

we now give by means of algebraic signs. The horizontal displacements are

intended to counterbalance another displacement in a different column or

table. That this is indeed the case can be seen from the fact that the horizontal

displacements applied to the argument of center (the column on the right side

in Table 5) and to the minutes of proportion (the column on the right side in

Table 6) add up algebraically to the vertical displacement applied to the equa-

tion of center (see Table 5):

(27) k = kh4.

of center to obtain the true argument of center in the Tables of the Seven

Planets.24

24 Indeed, k = k + c3(k ) = kh3 + c3() + kv3 = + c3() kh3+ kv3 = (kh3+ kv3) = kh4.

126 chapter 5

for 0 (kv4) (kh4)

Saturn 0 60 7

(265266 to 360 and 0 to 8081)

Jupiter 0 60 12

(259260 to 360 and 0 to 7778)

Mars 20 60 49

(223224 to 360 and 0 to 3839)

Venus 19 60 48

(223224 to 360 and 0 to 4041)

Mercury 10 60 24

(271272 to 360 and 0 to 4142)

For each planet the equation of anomaly is displayed in four tables: two, here

called (i) and (ii), associated with greatest distance (i.e., between greatest and

mean distance) and two, with exactly the same presentation, associated with

least distance (i.e., between mean and least distance: see 11.3). In the Almagest

and sets of tables related to it the equation of anomaly is treated separately

when the center of the epicycle is near apogee in contrast to the case when the

center of the epicycle near perigee. Ptolemy first assumed that the center of

the epicycle is at mean distance and computed the equation of anomaly under

this condition (Ptolemys column 6). He then considered the case where the

center of the epicycle lies between maximum distance (apogee), where the

equation of anomaly is least, and mean distance. Ptolemys column 5 gives

the subtractive difference to be applied to the equation of anomaly at mean

distance. This is because at maximum distance the equation of anomaly is

least. To interpolate between mean distance and maximum distance Ptolemy

applied the minutes of proportion in his column 8 to the subtractive difference;

thus, in this case, the total equation is

emy used different formulas for c8() near apogee and near perigee, but put the

values in a single column. Between mean distance and perigee Ptolemy com-

displaced tables in latin 127

puted the entries in his column 7 which is the amount to the added to the value

for mean distance to get the equation of anomaly at least distance (perigee)

where the equation of anomaly is greatest; thus, in this case, the total equation

is

where is near perigee. So in some cases one has to subtract and in other cases

one has to add. The author of the Tables for the Seven Planets wished to avoid

subtractions and he took a different approach without changing the model

or the parameters. Instead of first tabulating the values for mean distance

with a subtractive difference, he tabulated the equation of anomaly at greatest

distance, where the equation is least. Hence, all corrections are positive, and

the total equation is

emys c5(), and c4(k) (defined in eq. 24) corresponds to Ptolemys c8(); all

terms are positive. For values of near perigee, the author of the Tables for

the Seven Planets adhered more closely to Ptolemys formula since it did not

involve subtraction. Thus, between mean distance and minimum distance, our

author used

c7(), and c4(k) (defined in eq. 25) corresponds to Ptolemys c8(). We will see

that the entry for mean distance is the same whether one uses the formula near

apogee or the formula near perigee.

(i) The argument is the minutes of proportion, as indicated in the heading, dis-

played here as integers from 0 to 60. There are two other columns, one labeled

diversitas dyametri centralis and another for the first station near greatest dis-

tance (see Table c). The entries corresponding to 0 refer to greatest distance,

and the entries corresponding to 60 refer to mean distance.

This table is unprecedented in the astronomical literature and, in the ab-

sence of canons explaining its use, it is difficult to interpret the meaning of the

expression, diversitas dyametri centralis. Table 7 shows the extremal values for

all planets.

128 chapter 5

and first station near greatest

distance for Jupiter (excerpt)

of prop. dyametri centralis station

() (s, ) (s, )

0 5 22;37,40 4 4; 5

1 5 22;37,10 4 4; 7

15 5 22;30,10 4 4;30

30 5 22;22,40 4 4;53

45 5 22;14,40 4 5;17

60 5 22; 7,40 4 5;39

A close inspection of the entries for the diversitas dyametri centralis indicates

that they are strongly dependent on the apogees of the planets at epoch (noon,

February 28, 1341, Toledo). A comparison of our computation of the planetary

apogees for the city of Toledo at that time and the entries for 0 for each planet

is displayed in Table 8.

From Table 8 it follows that the entries for 0 tabulated here for the diversitas

dyametri centralis are obtained by adding a constant to the longitude of the

apogee of each planet at epoch, 0. This constant is specific to each planet

and results in turn from the displacements applied to its variables, kh4 max

(c6 c5): see Table 10. Therefore, these entries play a crucial role in computing

the true planetary longitudes but have no direct astronomical meaning. On

the other hand, for each planet the range for the entries is 0;20,42 (= 8s

12;42,18 8s 12;21,36); or 0;21 (if we consider 8s 12;21,18 as the intended entry

for 60) for Saturn, 0;30 for Jupiter, 5;38 for Mars, 1;42 for Venus, and 3;12

for Mercury. It turns out that these values agree with the vertical displacement

applied to the difference at greatest distance, c5, associated with the argument

of anomaly. We have called this quantity kv5 (see Table 11). In the case of Mars,

its value is 5;38, and the diversitas dyametri centralis near greatest distance,

d5, which is a function of the minutes of proportion, m, can be written as

displaced tables in latin 129

table 7 Diversitas dyametri centralis and first station near greatest distance for all planets

Entry for 0 Entry for 60 Entry for 0 Entry for 60

Jupiter 5s 22;37,40 5s 22; 7,40 4s 4; 5 4s 5;39

Mars 4s 25;38,49 4s 20; 0,49 5s 7;28 5s 13;11

Venus 3s 2;48, 9 3s 1; 6, 9 5s 15;51 5s 17; 9

Mercury 7s 3;50, 9 7s 0;38, 9 4s 27;14 4s 25; 8

* Probably an error for 8s 12;21,18. For justification, see below. One should also note

that the number of seconds in the entries for 0 and 60 in columns 1 and 2 agree for

each planet (but for Saturn here), both in this table and in the equivalent, Table 12.

Jupiter 171;31,37 1; 6, 3 1; 6 (= 12 10;54)

Mars 133;23,50 12;14,59 12;15 (= 49 36;45)

Venus 89;37, 0 3;11, 9 3;11 (= 48 44;49)

Mercury 208;51,10 4;58,59 4;59 (= 24 19; 1)

* As will be seen below, the rounded value for each planet should result from subtract-

ing the maximum of c6 c5 (see Table 10) from kv3 (see Table 5). In the case of Jupiter,

this gives 1;26 (= 12 10;34), whereas we obtain 1;6 when using the entry for 0 in the

table, 5s 22;37,40 (see Table 7), and a set of standard Alfonsine tables. If the author

had taken 5s 22;57,40 as the entry for 0, the rounded value would have been fully

consistent with those for the other planets.

d5(m) = d5(0) 5;38 c4(k). Thus, the general rule for the diversitas dyametri

centralis near greatest distance is given by

that is,

130 chapter 5

The third column in this table displays the first station as a function of the

minutes of proportion. The entries for the first station represent the true argu-

ment of anomaly of the first stationary point, and those for 0 agree with the

entries in the Almagest (see, e.g., Toomer 1984, p. 588) and sets of tables related

to it. They correspond to the greatest distance of the center of the epicycle,

when the mean argument of center, , is zero. It should be recalled that the

table in Almagest xii.8 gives the positions of the stationary points on the epicy-

cle as a function of (see Neugebauer 1975, pp. 202205), not as a function of the

minutes of proportion, as is the case here. The entries increase (except for Mer-

cury) monotonically and vary in a range different for each planet: 1;23 (Saturn),

1;34 (Jupiter), 5;43 (Mars), 1;18 (Venus), and 2;6 (Mercury). Figure 5.f dis-

plays the first station of Mars, both near apogee (see Table 7) and near perigee

(Table 12).

If we call s5(m) the tabulated argument of anomaly of the first station near

apogee, the entries are can be computed by means of a linear function of m, the

minutes of proportion:

(ii) In the second table the argument ranges from 0 to 29 for each zodiacal

sign. There are two columns, one for the equation of anomaly (the heading

has argumentum for anomaly) and another for the diversitas dyametri. The

displaced tables in latin 131

Diversitas Diversitas

Equation of dyametri (directus Equation of dyametri (directus

anomaly or retrogradus) anomaly or retrogradus)

() () () () () ()

1 37; 7 5;40

191 21;19 1;34

90 67;40 8; 5 statio

192 20; 3 1;20

125 73;29 9;48

126 73;30 9;52 203 9;15 0; 1

127 73;30 9;55 secunda

128 73;29 9;59 directus

204 8;31 0; 0

154 66;22 11;16 205 7;48 0; 0

155 65;42 11;16 206 7; 8 0; 0

156 64;59 11;16

157 64;15 11;15 232 0; 1 1;17

statio 233 0; 0 1;21

158 63;29 11;14 234 0; 0 1;24

235 0; 0 1;28

169 52;11 9;42

prima 270 5;50 3;11

retro.

170 50;55 9;26 358 36; 0 5;35

359 36;23 5;36

title indicates that the entries are given for greatest distance of the center of

the epicycle (ad longitudinem longiorem), that is, near apogee. The heading of

the entries for the diversitas dyametri is generally diversitas dyametri directus,

and sometimes diversitas dyametri retrogradus, a change which is indicated by

the insertion of the words statio, prima, and secunda, among the entries. See

Table d.

The relevant information for the equation of anomaly is summarized in

Table 9, where the underlying parameter was derived by taking half the

difference between maximum and minimum equation of anomaly.

132 chapter 5

distance for 0 of anomaly of anomaly parameter

(261266) (9297)

Jupiter 10;34 0;0 21; 8 10;34

(258263) (97102)

Mars 36;45 0;0 73;30 36;45

(233234) (126127)

Venus 44;49 0;0 89;38 44;49

(225) (135)

Mercury 19; 1 0;0 38; 2 19; 1

(250252) (108)

The underlying parameters in this table are the maximum values of the dif-

ference c6 c5, that is, the maximum difference between the equation of

anomaly at mean distance and the subtractive difference in the Parisian Alfon-

sine Tables.25 All these values are shown in Table 10.

Figure 5.g displays the tabulated values of the equation of anomaly for Mars

near greatest distance, here called d(a) and the difference between c6 and c5 in

the Parisian Alfonsine Tables. It is readily seen that d(a) is displaced upwards

with respect to c6 c5. The other planets follow the same pattern.

In the case of Mars the expression for the equation of anomaly at greatest

distance can be written as

25 In the Parisian Alfonsine Tables the fifth and sixth columns in the tables for planetary

equations display, respectively, the subtractive difference and the equation of anomaly at

mean distance; hence the notation used here.

displaced tables in latin 133

of c6 of c5 of c6c5

(at 9499) (at 100106) (at 9499)

Jupiter 11; 3 0;30 10;34

(at 99102) (at 107117) (at 97102)

Mars 41;10 5;38 36;45

(at 131132) (at 153156) (at 126127)

Venus 45;59 1;42 44;49

(at 135136) (at 161162) (at 135)

Mercury 22; 2 3;12 19; 1

(at 111112) (at 129131) (at 108)

The diversitas dyametri tabulated for each planet is the difference near

greatest distance. The entries in this column agree with those for the same

quantity, c5(), in the Parisian Alfonsine Tables, but for a vertical displacement,

which differs from one planet to another (see Table 11).

In the case of Mars this displacement amounts to 5;38. In general, if c5() is

the difference near greatest distance in the corresponding Parisian Alfonsine

table, that in the Tables for the Seven Planets is given by

134 chapter 5

distance for 0 at greatest distance at greatest distance placement (kv5)

(253261) (99104)

Jupiter 0;30 0;0 1; 0 0;30

(242252) (108118)

Mars 5;38 0;0 11;16 5;38

(204207) (153156)

Venus 1;42 0;0 3;24 1;42

(198199) (161162)

Mercury 3;12 0;0 6;24 3;12

(229231) (129131)

We note that a = , which is a consequence of equations 17, 18, 21, and 22.26

Therefore, the displacements applied by the author of the Tables for the Seven

Planets to the standard quantities used in Alfonsine planetary astronomy are

such that they keep invariant the argument of anomaly.

As in eq. 28, the total correction when the planet is near apogee in Ptolemys

notation is

In the Tables for the Seven Planets the corresponding expression near apogee

is

displaced tables in latin 135

near least distance for Venus (excerpt)

of prop. dyametri centralis station

() (s, ) (s, )

0 3 1;38, 9 5 17; 9

1 3 1;36,17 5 17;10

15 3 1;10, 9 5 17;25

30 3 0;42, 9 5 17;43

45 3 0;14, 9 5 18; 4

59 2 29;48, 1 5 18;20

60 2 29;46, 9 4 18;21

an expression only involving positive terms and leading to the same true lon-

gitude of the planets as the standard Alfonsine procedure.

The following two tables, for least distance (i.e., between mean and least dis-

tance), have the same presentation as those for greatest distance (i.e., between

greatest and mean distance), reviewed in 11.2 (i) and (ii).

and there are two other columns, one labeled diversitas dyametri centralis and

another for the first station near least distance (see Table e). Here the entries

for 0 correspond to mean distance, and the entries for 60 correspond to least

distance.

The extremal values of the two columns (diversitas dyametri centralis and

first station) for all planets are listed in Table 12.

As was the case for greatest distance, the entries for 0 tabulated here for the

diversitas dyametri centralis are obtained by adding a constant to the longitude

of the apogee of each planet at epoch, 0. This constant is specific to each planet

and results in turn from the displacements applied to its variables: kh4 kv6

(see Table 13).

136 chapter 5

table 12 Diversitas dyametri centralis and first station near least distance for all planets

Entry for 0 Entry for 60 Entry for 0 Entry for 60

Jupiter 5s 22; 8,40 5s 21;35,40 4s 5;40 4s 7;11

Mars 4s 21;13,49 4s 13;10,49 5s 13; 6 5s 19;15

Venus 3s 1;38, 9 2s 29;46, 9 5s 17; 9 5s 18;21

Mercury 7s 0;49, 9 6s 28;48, 9 4s 25; 8 4s 24;28

Jupiter 171;31,37 0;37, 3 0;37 (= 12 11;23)

Mars 133;23,50 7;49,59 7;50 (= 49 41;10)

Venus 89;37, 0 2; 1, 9 2; 1 (= 48 45;59)

Mercury 208;51,10 1;57,59 1;58 (= 24 22; 2)

* As will be seen below, the rounded value for each planet should result from subtract-

ing kv6 (see Table 14) from kv3 (see Table 5). In the case of Jupiter, this gives 0;57

(= 12 11;3), whereas we obtain 0;37 when using the entry for 0 in the table, 5s

22;8,40 (see Table 12) and a set of standard Alfonsine tables. If the author had taken

5s 22;28,40 as the entry for 0, the rounded value would have been fully consistent with

those for the other planets.27

On the other hand, the range for the entries is: 0;25 for Saturn, 0;33 for Jupiter,

8;3 for Mars, 1;52 for Venus, and 2;1 for Mercury. The entries for least distance

are not exactly a continuation of those for greatest distance (Table 7), because

they were computed by means of different expressions and different ranges

27 This is the same situation we noted in Table 8; thus, it would seem that the author either

miscopied an entry for Jupiter on which he built up his two tables for the diversitas

dyametri centralis, or that he had at his disposal a table for Jupiter generating a difference

of 0;20.

displaced tables in latin 137

were used. In the case of Mars the vertical displacement applied here, kv7,

is 8;3 rather than kv5 = 5;38, as was the case at greatest distance. Then the

diversitas dyametri centralis near least distance, d7, can be written as d7(m)

= d7(0) 8;3 c4(k), where m is a value for the minutes of proportion. Thus,

the general rule for the diversitas dyametri centralis near least distance is given

by

that is,

The third column in this table displays the argument of anomaly of the first

station near least distance as a function of the minutes of proportion (Table 12).

The entries follow the same pattern as those near greatest distance, but we note

that the entries for 60 at greatest distance (see Table 7) do not coincide with

the entries for 0 near least distance (see Table 12) in three cases: Saturn, Jupiter,

and Mars. The entries for 60 essentially agree with those in the Almagest and

sets of tables related to it, corresponding to the least distance of the center of

the epicycle, when the mean argument of center, , is 180. The entries, graphed

in Figure 5.f, increase (except for Mercury) monotonically and vary in a range

which is different for each planet: 1;19 (Saturn), 1;31 (Jupiter), 6;9 (Mars), 1;12

(Venus), and 0;40 (Mercury).

If we call s7(m) the tabulated argument of anomaly of the first station near

perigee, the entries are can be computed by means of a linear function of m,

the minutes of proportion,

(ii) In the second table the argument ranges from 0 to 29 for each zodiacal

sign. There are two columns (equation of anomaly and diversitas dyametri). The

title indicates that the entries are given for least distance of the center of the

epicycle (ad longitudinem propiorem), that is, near perigee (see Table f). As was

the case for Table d, the words statio, prima, and secunda are inserted among

various entries.

138 chapter 5

Diversitas Diversitas

Equation of dyametri (directus Equation of dyametri (directus

anomaly or retrogradus) anomaly or retrogradus)

() () () () () ()

1 22;19 2; 2

213 5;27 0; 8

90 42;35 3;30 secunda

214 5; 7 0; 6

110 44; 3 9;48 215 4;48 0; 4

111 44; 4 9;52 statio

112 44; 4 9;55 directus

113 44; 3 9;59 216 4;30 0; 3

130 42;46 4; 1 223 2;38 0; 1

131 42;37 4; 2 224 2;25 0; 0

136 41;39 4; 2 229 1;27 0; 0

137 41;26 4; 1 230 1;18 0; 1

144 39;34 3;59 247 0; 1 0; 9

prima 248 0; 0 0;10

145 39;16 3;58 249 0; 0 0;11

146 38;57 3;56 250 0; 1 0;12

147 38;37 3;54

statio 270 1;29 0;32

retro.

148 38;16 3;52 358 21;29 1;59

359 21;45 2; 0

Table 14, where the underlying parameter was derived by taking half the

difference between maximum and minimum equation of anomaly.

For each planet the underlying parameter agrees with that in the Parisian

Alfonsine Tables for mean distance, and so do the entries for the equation of

anomaly, but for a vertical displacement. If c6() represents the equation of

displaced tables in latin 139

Least Entry equation of equation of Underlying displacement

distance for 0 anomaly anomaly parameter (kv6)

(261266) (9499)

Jupiter 11; 3 0;0 22; 6 11; 3 11; 3

(258261) (99102)

Mars 41;10 0;0 82;20 41;10 41;10

(229) (131)

Venus 45;59 0;0 91;58 45;59 45;59

(224225) (135136)

Mercury 22; 2 0;0 44; 4 22; 2 22; 2

(248249) (111112)

anomaly in the corresponding Parisian Alfonsine table, that in the Tables for

the Seven Planets is given by

where kv6 = 45;59 in the case of Venus.28 Figure 5.h displays the tabulated

values of the equation of anomaly for Venus near least distance in the Tables

for the Seven Planets and at mean distance in the Parisian Alfonsine Tables.

The terminology in the Tables of the Seven Planets may be confusing, but

the author sought to maintain a parallel structure in his treatment of the

equation of anomaly near apogee and near perigee. The entries near apogee

range from greatest distance to mean distance, and those near perigee range

from mean distance to least distance. Hence, what is here called the equa-

tion of anomaly near least distance actually refers to the equation of anomaly

at mean distance, to be applied to distances between mean and least dis-

tance.

The diversitas dyametri tabulated for each planet is the difference near

least distance to be added to the corresponding equation of anomaly at mean

distance (see the extremal values in Table 15). The entries in this column agree

140 chapter 5

figure 5.h Venus, equation of anomaly as a function of the true argument of anomaly

distance for 0 at least distance at least distance placement (kv7)

Jupiter 0;33 0;0 (240253) 1; 6 (107120) 0;33

Mars 8; 3 0;0 (201) 16; 6 (159) 8; 3

Venus 1;52 0;0 (197199) 3;44 (161163) 1;52

Mercury 2; 1 0;0 (224229) 4; 2 (131136) 2; 1

with those for the same quantity, c7(), in the Parisian Alfonsine Tables, but

for a displacement, which differs from one planet to another. If c7() is the

difference at least distance in the corresponding Parisian Alfonsine table, the

difference at least distance in the Tables for the Seven Planets is given by

Note that at least distance of the center of the epicycle, i.e., perigee, the equa-

tion of anomaly is greatest.29

29 These tables bear some similarity with the tables of John Vimond, who tabulated c6

displaced tables in latin 141

when the planet is near perigee is

In the Tables for the Seven Planets, the corresponding expression near perigee

is

This is an expression involving only positive terms and leads to the same true

longitude of the planets as the standard Alfonsine procedure.

To find the true longitude, , of the planet at that time according to the

Parisian Alfonsine Tables one has to add the correction c() to its mean lon-

gitude, that is, to the sum of the longitude of the apogee, 0, and the true

argument of center of the planet, :

= 0 + + c(),

Now, the author of the Tables of the Seven Planets introduced an analogous

expression to find the true longitude of the planet:

where c(a) = c6(a) + c7(a) c4(k): see eq. 41. The term d7(m) is tabulated under

diversitas dyametri centralis.

The general expression for the longitude of a planet near perigee is therefore

given as

c5, but he did not apply any displacements to the planets. Vimond also tabulated c5 + c7,

which has no counterpart in these tables (see Chabs and Goldstein 2004, pp. 248256).

142 chapter 5

To be sure,

= 0 + (k + kh4) + (c6() kv6) + (c7() kv7) c4(k)

= 0 + kh4 kv6 kv7 c4(k) + k + c6() + c7() c4(k) = l,

provided that d7(m) = 0 + kh4 kv6 kv7 c4(k), which is the case: see eq. 37. All

terms appearing in eq. 43 are positive and are found directly in the tables. Note

that d5(m) and d7(m) have no counterparts in the standard Alfonsine tables.

To illustrate eq. 43 consider the position of Mars at epoch (noon, February 28,

1341), when the planet is near perigee. The mean argument of center, k , and the

mean argument of anomaly, , are given as 55;6,26 and 84;48,42, respectively,

rounded to the seconds (see Table 3). From eqs. 16 and 17 it follows that

and in the standard Alfonsine Tables are 116;6,26 and 96;48,42, respectively

(where kh3 = 61 and kv3 = 12: see Table 5). The corresponding values of the

equation of center in both sets of tables are c3(k ) = 1;23 and c3() = 10;37, and

eq. 18 holds. Thus, the true arguments of center are k =56;29 and = 105;29,

and eq. 27 holds, where kh4 = 49 (see Table 6). The tabulated equatio porcionis

for k = 55;6,26 is e(k ) = 22;37, and thus a = 84;48,42 + 22;37 = 107;26. The

argument of anomaly, , according to the Parisian Alfonsine Tables, is 107;26

(= 96;48,42 + 10;37), and we note that a = . The minutes of proportion

corresponding to k = 56;29 and = 105;29 are c4(k) = 17/60 and c4() = 17/60,

and we note that c4(k) = c4(), as indicated in eq. 25, valid near perigee. The

tabulated values for the equation of anomaly, c6(a), and the diversitas, c7(a), are

79;11 and 11;48, respectively, whereas the values for the equation of anomaly,

c6(), and the additive difference, c7(), found in the Parisian Alfonsine Tables

are 38;1 and 3;45, respectively. We note that eqs. 39 and 40 hold, where kv6 =

41;10 and kv7 = 8;3 (see Tables 14 and 15). From eq. 41, c(a) = 79;11 + 11;48

(17/60) = 82;32, whereas the Parisian Alfonsine Tables yield c() = 38;1 + 3;45

(17/60) = 39;5.

Now, the true longitude derived with the standard Alfonsine Tables is =

133;23,50 + 105;29 + 39;5 = 277;57,50 (see Table 8). On the other hand, from

eq. 37, d7(17) = 141;13,49 8;3 17/60 = 138;56,58 (see Table 12). Thus, l =

138;56,58 + 56;29 + 82;32 = 277;57,58, in agreement with the previous result.

Similarly, the general expression for the longitude of a planet near apogee is

given as

displaced tables in latin 143

To be sure,

= 0 + (k + kh4) + c6() c5() + c5() + c5() c4()

= 0 + (k + kh4) + d(a) max (c6 c5) + c5() + c5() c4()

= 0 + kh4 max (c6 c5) + k + d(a) + c5() (1 + c4())

= 0 + kh4 max (c6 c5) + k + d(a) + (c5() kv5) c4(k)

= 0 + kh4 max (c6 c5) + k + d(a) kv5 c4(k) + c5(a) c4(k)

= 0 + kh4 max (c6 c5) kv5 c4(k) + k + d(a) + c5(a) c4(k) = l,

provided that d5(m) = 0 + kh4 max (c6 c5) kv5 c4(k), which is the case:

see Table 7 and eq. 32. All terms appearing in eq. 44 are positive and are found

in the tables.

To illustrate eq. 44, consider the position of Mars at noon, October 1, 1340,

when = 37;31,6, and thus the planet is near apogee. From eq. 16 it follows that

k = 336;31,6 (where kh3 = 61; see Table 5). In the Parisian Alfonsine Tables, the

corresponding equation of center is c3() = 6;26,40, whereas the tabulated

equatio centri and equatio porcionis for k = 336;31,6 are c3(k ) = 5;33,20 and

e(k ) = 18;26,40, respectively. We note that eq. 18 holds, where kv3 =12 (see

Table 5). We also note that eq. 27 holds, where kh4 = 49 (see Table 6). Then,

= 31;4,26 because of eq. 18, and k = 342;4,26 because of eq. 19. In the Parisian

Alfonsine Tables, the corresponding minutes of proportion is c4() = 51/60,

whereas the minutes of proportion tabulated here is c4(k) = 9/60. Therefore eq.

24 holds, where kv4 = 60. The mean argument of anomaly of Mars for that date

is = 27;35,24, and thus the true argument of anomaly is = 34;2,4, because

of eq. 23, and = 15;35,24, because of eq. 17, where kv3 =12. Therefore, a =

34;2,4, because of eq. 22. We note that a = . In the Parisian Alfonsine Tables,

the corresponding entries for the equation of anomaly and the difference at

greatest distance are c6() = 13;26 and c5() = 0;48, respectively. Thus, the

total correction is c() = 13;26 + 0;48 (51/60) = 12;45,12. On the other hand,

the tabulated values for the diversitas dyametri and the equation of anomaly

are c5(a) = 6;26 (we note that eq. 35 holds, where kv5 = 5;38: see Table 11)

and d(a) = 49;23. From eq. 36 it follows that c(a) = 49;23 + 6;26 9/60 =

50;20,54.

144 chapter 5

Jupiter 6 18 60 12 0;30 11; 3 0;33

Mars 12 61 60 49 5;38 41;10 8; 3

Venus 3 51 60 48 1;42 45;59 1;52

Mercury 4 28 60 24 3;12 22; 2 2; 1

Now, the true longitude derived with the standard Alfonsine Tables is =

133;23,50 + 31;4,26 + 12;45,12 = 177;13,28 (see Table 8). On the other hand,

from eq. 32, d5(9) = 145;38,49 5;38 9/60 = 144;48,7 (see Table 7). Thus, l

= 144;48,7 + 342;4,26 + 50;20,54 = 177;13,27, in agreement with the previous

result.

In Table 16 we present a summary of the values for the displacements, both

vertical and horizontal, applied by the anonymous author of the Tables of the

Seven Planets to the Sun, the Moon, the 8th sphere, and the planets.

Folio 40r displays a table for the latitudes of the three superior planets (the

inferior planets are addressed in a very different way on ff. 41r46v: see 16,

below). This table is in the Almagest tradition, and is found in many other sets

of tables such as the zij of al-Battn and the Toledan Tables (see Chabs and

Goldstein 2012, Table 9.2b, p. 109). For the zij of al-Battn, see Nallino 19031907;

and for the Toledan Tables, see Toomer 1968 and Pedersen 2002.

13 Planetary Visibility

On f. 40v there is a table entitled Tabula visionis et occultationis for the three

superior planets and the two inferior planets. It is a table for the visibility of

the planets, also called a table of planetary phases, which is already found in

displaced tables in latin 145

(s, ) ()

0 0 6 0 12 0 6 0 0; 0

0 1 5 29 11 29 6 1 5;13

0 12 5 18 11 18 6 12 62;16

0 13 5 17 11 17 6 13 67;23

the Handy Tables, the zij of al-Battn, and the Toledan Tables (see Chabs and

Goldstein 2012, Table 11.2, p. 125).

14 Possibility of an Eclipse

On f. 40v there is a small table entitled Tabula latitudinis lune in principio medio

et fine eclipsis, an excerpt of which we reproduce above (see Table g). The

argument, which is the argument of lunar latitude, is presented in four columns

and, as indicated in the title, it is restricted to the values for which an eclipse

is possible, that is, 13 from the lunar nodes. The purpose of the table is to

show the correspondence between the argument of latitude and the latitude

of the Moon, and the entries can be recomputed by means of the modern

formula

the lunar orb to the ecliptic, taken here as 5;0 (the same parameter as in the

table for lunar latitude on f. 13r).

A table for the eclipsed fraction of the solar and lunar disks is also found

on f. 40v. The argument ranges from 1 to 12 linear digits (where the diameter

of the eclipsed body is 12 digits), and the entries are the corresponding area

146 chapter 5

Anomaly

0s6/11s24 3s0/9s0 5s24/6s6 6s0/6s0

Center 1st 2nd 1st 2nd 1st 2nd 1st 2nd 3rd

(s, ) (s, ) (s, ) (s, ) () () () () () () () () ()

0 12 6 12 5 18 11 18 0;13 0; 2 0;0 0;23 1;15 0;10 1;30 0;0 2

0 18 6 18 5 12 11 12 0;19 0; 2 0;0 0;35 1;52 0;14 2;14 0;0 3

2 24 8 24 3 6 9 6 1; 1 0; 8 0;0 1;55 6; 8 0;48 7;16 0;0 10

3 0 9 0 3 0 9 0 1; 2 0; 8 0;0 1;57 6;12 0;48 7;22 0;0 10

digits (where the area of the eclipsed body is 12). It is very common in sets of

astronomical tables, such as the zij of al-Battn, the Toledan Tables, and it is

already found in Almagest vi.8. However, this was not one of the tables included

in the editio princeps of the Alfonsine Tables (see Chabs and Goldstein 2012,

Table 15.4, p. 175).

The tables for the latitude of Venus and Mercury are presented as six sub-tables

for each planet, on ff. 41r43v and ff. 44r46v, respectively. Both are double

argument tables. The vertical argument (center) is given in four columns at

6-intervals and altogether there are 30 columns for the horizontal argument

(anomaly), also at 6-intervals. Each of these columns contains entries for the

inclination (latitudo prima) and the slant (latitudo secunda). The deviation

(latitudo tercia) is presented in another column, which remains invariant in

all sub-tables for each planet. Table h displays an excerpt of the table for the

latitude of Venus.

The maximum values for Venus and Mercury are, respectively, 7;22 and

4;5 for the inclination, 2;30 and 2;30 for the slant, and +10 and 13 for the

displaced tables in latin 147

deviation. This table differs, strongly in presentation and slightly in the basic

parameters, from tables by Parisian astronomers who included the third com-

ponent of latitude, John Vimond and John of Murs.30 For the maximum val-

ues of deviation, Vimond had +10and 45 (Ptolemys values in the Almagest),

for Venus and Mercury, respectively, whereas John of Murs had +10and 23,

in contrast to +10and 13 in the tables reviewed here. John of Lignres was

aware of the deviation, for he mentions it in the chapters on the latitude

of Venus and Mercury in the canons of his Priores astrologi motus corporum

celestium,31 but we do not know of any tables by him similar to those pre-

sented here. All in all, the tables for planetary latitudes in the Tables for the

Seven Planets, although certainly in the same tradition, are not simply related

to those by John Vimond and John of Murs, or by any other known table-

maker.

Conclusion

Throughout the time from the reception of the Alfonsine Tables in Paris (no

later than 1320) to the publication of the editio princeps in 1483 in Venice, an

intense effort was made to adapt tables in the Alfonsine framework to the needs

of practitioners. These adaptations focused on presentation rather than on the

parameters underlying the tables of the Alfonsine corpus of tables. The set of

tables which we have called the Tables for the Seven Planets for 1340 is an early

example of this kind of work, with zodiacal signs of 30 (rather than physical

signs of 60), vacillation in the beginning of the year, and cyclical radices of

32 years, with special characteristics such as displacements, different param-

eter for the latitude of Mercury, and different terminology. The use of double

argument tables and the extensive use of displacements, both vertical and hor-

izontal, show a deep insight into planetary astronomy and great skill in pro-

ducing astronomical tables. This set of tables is also the first example that has

been discovered in the Latin world of a systematic use of displaced tables, of

which only a few examples were previously known in the medieval astronomi-

cal literature. Unfortunately, no name is associated with this set, but the author

was an astronomer working around 1340, probably in Southern France (judg-

ing by the geographical coordinates underlying these tables), who deserves the

highest praise for his skill in providing clever and complex solutions to many

30 Chabs and Goldstein 2004, pp. 257258; Chabs and Goldstein 2009, p. 309.

31 An edition of chapters 22 and 23 is found in Saby 1987, pp. 207211.

148 chapter 5

problems, and for constructing a compact and consistent set of tables for the

planets, building upon the work done by the Parisian group of astronomers in

the early 1320s.

References

Chabs, J. and B.R. Goldstein 2003. The Alfonsine Tables of Toledo. Archimedes: New

Studies in the History and Philosophy of Science and Technology, 8. Dordrecht and

Boston.

Chabs, J. and B.R. Goldstein 2004. Early Alfonsine Astronomy in Paris: The Tables of

John Vimond (1320), Suhayl 4:207294.

Chabs, J. and B.R. Goldstein 2009. John of Murss Tables of 1321, Journal for the History

of Astronomy 40:297320.

Chabs, J. and B.R. Goldstein 2012. A Survey of European Astronomical Tables in the Late

Middle Ages. Boston.

Debarnot, M.-T. 1987. The Zj of abash al-sib: A Survey of ms Istanbul Yeni Cami

784/2, in King and Saliba 1987, pp. 3569.

Goldstein, B.R. 1974. The Astronomical Tables of Levi ben Gerson. Transactions of the

Connecticut Academy of Arts and Sciences, 45. New Haven.

Goldstein, B.R., J. Chabs, and J.L. Mancha 1994. Planetary and Lunar Velocities in

the Castilian Alfonsine Tables, Proceedings of the American Philosophical Society

138:6195.

Hogendijk, J. and A.I. Sabra 2003. The Enterprise of Science in Islam. Cambridge, ma.

Jensen, C. 1971. The Lunar Theory of al-Baghdd, Archive for History of Exact Sciences

8:321328.

Kennedy, E.S. 1977. The Astronomical Tables of Ibn al-Aclam, Journal for the History of

Arabic Science 1:1323.

King, D.A. and M.H. Kennedy (eds.) 1983. Studies in the Islamic Exact Sciences by E.S.

Kennedy, Colleagues and Former Students. Beirut.

King, D.A. and G. Saliba (eds.) 1987. From Deferent to Equant: A volume of studies in the of

history science in the ancient and medieval Near East in honor of E.S. Kennedy. Annals

of the New York Academy of Sciences, 500.

Kremer, R.L. (forthcoming). Melchion de Friquentos eclipse tables of 1437, a revised

Latin version of Immanuel Bonfilss Six Wings.

Kremer, R.L. and J. Dobrzycki 1998. Alfonsine meridians: Tradition versus experience

in astronomical practice c. 1500, Journal for the History of Astronomy 29:187199.

Mercier, R. 1989. The parameters of the Zj of Ibn al-Aclam, Archives Internationales

dHistoire des Sciences 39:2250.

Nallino, C.A. 19031907. Al-Battn sive Albatenii Opus Astronomicum, 2 vols. Milan.

displaced tables in latin 149

Pedersen, F.S. 2002. The Toledan Tables: A review of the manuscripts and textual versions

with an edition. Copenhagen.

Poulle, E. 1984. Les tables alphonsines avec les canons de Jean de Saxe. Paris.

Ratdolt, E. (ed.) 1483. Tabule astronomice illustrissimi Alfontij regis castelle. Venice.

Saby, M.-M. 1987. Les canons de Jean de Lignres sur les tables astronomiques de 1321.

Unpublished thesis: cole Nationale des Chartes, Paris. A summary appeared as:

Les canons de Jean de Lignres sur les tables astronomiques de 1321, cole Nationale

des Chartes: Positions des thses, pp. 183190.

Salam, H. and E.S. Kennedy 1967. Solar and Lunar Tables in Early Islamic Astronomy,

Journal of the American Oriental Society 87:493497. Reprinted in King and Kennedy

1983, pp. 108113.

Saliba, G. 1976. The Double-Argument Lunar Tables of Cyriacus, Journal for the History

of Astronomy 7:4146.

Saliba, G. 1977. Computational Techniques in a Set of Late Medieval Astronomical

Tables, Journal for the History of Arabic Science 1:2432.

Sams, J. 2003. On the Lunar Tables in Sanjaq Drs Zj al-Sharf , in Hogendijk and

Sabra 2003.

Sams, J. and E. Mills 1998. The computation of planetary longitudes in the zj of Ibn

al-Bann, Arabic Sciences and Philosophy 8: 259286.

Solon, P. 1970. The Six Wings of Immanuel Bonfils and Michael Chrysokokkes, Centau-

rus 15:120.

Thorndike, L. 1957. Notes on some Astronomical, Astrological and Mathematical

Manuscripts of the Bibliothque Nationale, Paris, Journal of the Warburg and Cour-

tauld Institutes 20:112172.

Tihon, A. 19771981. Un trait astronomique chypriote du xive sicle, Janus 64:281

308; 66:4981; 68:65127. Reprinted in Tihon 1994, Essay vii.

Tihon, A. 1994. tudes dastronomie byzantine. Aldershot.

Toomer, G.J. 1968. A survey of the Toledan Tables, Osiris 15:5174.

Toomer, G.J. 1984. Ptolemys Almagest. New York.

Van Brummelen, G. 1998. Mathematical Methods in the Tables of Planetary Motion in

Kshyr ibn Labbns Jmic Zj, Historia Mathematica 25:265280.

chapter 6

User-Friendliness and the Alfonsine Corpus*

Astronomical tables are ways to turn the treatment of complex problems into

elementary arithmetic. Since Antiquity astronomers have addressed many

problems by means of tables; among them stands out the treatment of plan-

etary motion as well as that for the motions of the Sun and the Moon. It was

customary to assign to the planets constant mean velocities to compute their

mean longitudes at any given time in the past or the future, and to add to

these mean longitudes corrections, called equations, to determine their true

longitudes. In this paper we restrict our attention to the five planets,1 with an

emphasis on their equations. Part 1 deals with what we call the standard tra-

dition, beginning with Ptolemys Handy Tables, and Part 2 deals with the new

presentations that proliferated in Latin Europe in the fourteenth and fifteenth

centuries, some of which reflect a high level of competence in mathematical

astronomy.2

By the middle of the second century ad Ptolemy displayed tables for the

equations of the five planets with specific layouts and based on specific models,

algorithms, and parameters. We argue that this category of tables, as is the case

for many others, provides a clear example of user-friendliness, the driving force

that prevailed in the history of table-making.

In Almagest xi.11 Ptolemy presented tables for the planetary equations, one

for each of the five planets.3 Each table has eight columns, of which the first two

1 Unless otherwise specified, by planets we mean the five visible planets of Antiquity, although

we are well aware that at the time the Sun and the Moon were also considered planets.

2 We do not treat tables in Islamic zijes systematically but, occasionally, we refer to some

of them. For a survey of these zijes, see D.A. King and J. Sams, with a contribution by

B.R. Goldstein, Astronomical Handbooks and Tables from the Islamic World (7501900): An

Interim Report, Suhayl, ii (2001), 9105.

3 G.J. Toomer, Ptolemys Almagest (New York and Berlin, 1984), 549553.

computing planetary positions 151

are for the argument (one from 6 to 180 and the other for its complement in

360). The argument is given at intervals of 6, from 6 to 90 (and for 270 to

354), and at intervals of 3, from 90 to 180 (and for 180 to 270). According to

Toomer, Ptolemy computed the entries at 6-intervals, even where the function

is tabulated at 3-intervals.4 Columns 3 and 4 are for the equation in longitude

and the difference in equation, respectively. Column 3 assumes an eccentric

model, which Ptolemy rejected in favor of an equant model. Column 4 displays

the difference between the equation for an equant model and the equation for

an eccentric model. The sum of corresponding entries in these two columns is

the equation of center, which replaced columns 3 and 4 that appear in Almagest

xi.11 (see Table a, col. 3).5 Columns 5 and 7 give the subtractive and additive

differences to be applied to the equation of anomaly (displayed in col. 6), when

the planet is at greatest and least distance, respectively. Column 8 is for the

minutes of proportion, to seconds, used for interpolation purposes. We note

that, in the case of Venus, the entries for the equation in longitude (col. 3) are

exactly the same as those for the solar equation, although Ptolemy does not call

attention to this fact.6 We display Ptolemys model for Mars to illustrate how a

planets position can be computed directly from the model: see Figure 6.1. To

do this, one must solve plane triangles by means of trigonometric procedures

that were already available in Ptolemys time. The solution is as follows. Given

, we wish to compute the correction angle, c3, by solving triangle eco. But,

before we can do this, we have to find the length of ec, where dc, the radius of

the deferent, is 60. So first we must solve triangle edc to find ec, where angle

ced is the supplement to angle and ed is the eccentricity (a given parameter

in the model). With , ec and eo (twice the eccentricity), we can solve triangle

eco, which yields the values for c3 and co. We then have to solve triangle mco

to find c(). In this triangle two sides and an angle are known: angle mco is

equal to 180 ( c3), cm is the radius of the epicycle (a given parameter in

the model), and co has already been determined. Then

= (a) + + c3 + c(),

5 O. Neugebauer, A History of Ancient Mathematical Astronomy (Berlin, 1975), 183.

6 Although the equation of center for the Sun and the equation in longitude for Venus are

the same in the Almagest (Toomer, Almagest (ref. 3), 167, 552), their apogees differ: the solar

apogee is 65;30 and tropically fixed (Toomer, Almagest (ref. 3), 155), whereas the apogee of

Venus is 55 in Ptolemys time and sidereally fixed, and thus subject to precession (Toomer,

Almagest (ref. 3), 470; cf. Neugebauer, Ancient Astronomy (ref. 5), 58, 154, 182).

152 chapter 6

figure 6.1

o; the center of the deferent, ac, is at d; the

equant is at e such that ed = do; the

direction to the vernal equinox is ov; a is

the apogee, c is the center of the epicycle

such that dc = 60. The radius of the epicycle

is cm; the mean apogee of the epicycle is at

e and the true apogee of the epicycle is at

ae ; the mean longitude of Mars is in the

direction om , and its true position is at m.

The mean argument of center is and the

mean argument of anomaly is ; the

equation of center is c3, and the equation of

anomaly is c(). The true longitude of Mars,

, is angle vom. For an outer planet the

direction cm is always parallel to os, where

s is the direction to the mean Sun.

where (a), the longitude of the apogee, is a given parameter in the model.

Using the planetary equation tables takes trigonometric functions out of the

computational scheme.

In the Handy Tables Ptolemy did not modify the models or the param-

eters for the planetary equations, but he introduced a series of changes to

make the tables more suitable for calculation. Firstly, the arguments are now

given at intervals of 1, rather than at intervals of 3 or 6, as was the case

in the Almagest.7 This certainly simplifies interpolation. Secondly, he merged

columns 3 and 4 in the Almagest into a single column representing the equa-

tion of center, thus reducing the number of operations required for using these

tables. This also reduced the number of columns, from 8 to 7. Thirdly, the col-

umn for the minutes of proportion was also modified by avoiding unnecessary

precision (the entries are given to seconds in the Almagest but only to minutes

in the Handy Tables) and by changing the argument (mean argument of center

in the Almagest and true argument of center in the Handy Tables).8 This new

7 W.D. Stahlman, The Astronomical Tables of Codex Vaticanus Graecus 1291 (unpublished Ph.D.

thesis, Brown University, 1959; University Microfilms no. 625761), 295324. This dissertation

includes an edition of Ptolemys Handy Tables. A. Tihon and R. Mercier are currently editing

the Handy Tables; only two volumes have been published so far, but they do not deal with

planetary equations.

8 See Neugebauer, Ancient Astronomy (ref. 5), 183186 (Almagest) and 10021003 (Handy

Tables).

computing planetary positions 153

presentation (see Table a for Mars) set the standard for most tables dealing

with planetary equations for about 14 centuries.

Equation Min. Subtractive Equation Additive

Argument of center prop. difference of anom. difference

() () () () () () ()

2 358 0;22 60 0; 3 0;48 0; 3

3 357 0;33 60 0; 4 1;12 0; 4

86 274 2

87 273 1

88 272 1

89 271 2

90 270 2;28 33;22 2;49

92 268 11;24

93 267 11;25

96 264 11;25

97 263 11;24

131 229 41;10

132 228 41; 9

153 207 5;38

157 203 5;37

158 202 8; 2

159 201 8; 3

154 chapter 6

Equation Min. Subtractive Equation Additive

Argument of center prop. difference of anom. difference

() () () () () () ()

160 200 8; 2

179 181 0;14 60 0;26 1;57 0;48

180 180 0; 0 60 0; 0 0; 0 0; 0

2) for the equation of center (col. 3), c3; the true argument of center, , can then

be computed, for

= + c3(),

also as argument for the minutes of proportion (col. 4), which are necessary

to compute the true position of the planet when not found at maximum or

minimum distance of the epicycle from the observer. Now the true argument

of anomaly, , serves as argument (columns 1 and 2) for the equation of anomaly

(col. 6), and is obtained from the mean argument of anomaly:

= c3().

that the center of the epicycle is at mean distance. When the epicycle lies

between maximum distance (apogee) and mean distance, a subtractive differ-

ence (col. 5) must be applied. Similarly, when the epicycle is between minimum

distance (perigee) and mean distance, an additive difference (col. 7) must be

applied. The true argument of anomaly serves as argument for both subtractive

and additive differences. Then, the total equation of anomaly is

when ranges from 270 to 90, that is, when the planet is near apogee, and

computing planetary positions 155

when ranges from 90 to 270, that is, when the planet is near perigee. The

combined effect of the equation of center and the total equation of anomaly is

thus c3() + c(, ), and the true position of the planet, , at a given time is:

= + + c3() + c(, ),

where , the mean longitude of that planet at a given time t since epoch, is

defined as:

= 0 + t,

0 being the planets mean longitude at epoch, and the planets mean

motion in longitude.

In the early Islamic world, the Zj al-Sindhind of al-Khwrizm (fl. 830) fol-

lowed the Indo-Iranian tradition, which was not based on Ptolemaic models

and parameters, and made no use of equants.10 This tradition was represented

by the Zj al-Shh, a work composed in Sasanian Persia and translated into Ara-

bic c. 790, where the maximum value for the equation of Venus is set equal to

that of the solar equation (2;13 or 2;14); the identity of these parameters is

also found in the Almagest.11 Accordingly, the tables for the planetary equa-

tions are quite different, both with respect to the entries and the presentation,

from those in the Almagest or the Handy Tables.

The Greek tradition was represented in the eastern Islamic world by the Zj

al-bi of al-Battn (d. 929) which is strongly Ptolemaic; indeed, the tables in

it for the planetary equations followed exactly those in the Handy Tables, but

for the equation of center of Venus.12 Both the Almagest and the Handy Tables

have 2;24 as the maximum value for Venuss equation of center, whereas it

is 1;59 in the zij of al-Battn. This change in the equation of center of Venus

was not due to new observations of Venus; rather, it was the result of a new

value found from observations for the eccentricity of the solar model that

11 In this tradition the apogees of Venus and the Sun are the same and both are sidereally

fixed: see B.R. Goldstein and F.W. Sawyer, Remarks on Ptolemys Equant Model in Islamic

Astronomy, in Prismata: Festschrift fr Willy Hartner, ed. by Y. Maeyama and W.G. Salzer

(Wiesbaden, 1977), 165181, p. 168.

12 C.A. Nallino, Al-Battn sive Albatenii Opus astronomicum (2 vols, Milan, 19031907), ii,

108137.

156 chapter 6

implied a maximum solar equation of 1;59,10. This new solar parameter was

simply applied to the equation of center for Venus, where it was rounded to

1;59. In this modification al-Battn followed other Islamic zijes, such as that

of abash al-sib (fl. 850).13 Toomer pointed out that modifying the entries

for the equation of Venus was inconsistent with leaving unchanged the entries

for the subtractive and additive differences (at greatest and least distances,

respectively), because they also depend on eccentricity.14 In any case, in the

zij of al-Battn only the entries for the equation of center of Venus differ from

those in the Handy Tables whereas all the rest remain unchanged.

The Toledan Tables were compiled in the second half of the eleventh century,

but the original Arabic version is not extant. In the Latin versions of the Toledan

Tables the presentation and the numerical entries agree with those in the zij

of al-Battn, but for (in most cases) an added column for the first station of

each of the planets.15 In Almagest xii.8 Ptolemy displayed the first and second

stations of the five planets in a single table, using the mean center as argument,

with entries at intervals of 6.16 In the Handy Tables, Ptolemy gave more entries,

at 3-intervals, and presented a table for the two stations for each planet. He

also introduced a change in the argument (true argument of center, instead of

mean argument of center), thus making the entries slightly different from those

in the Almagest.17 In his zij al-Battani reproduced in a separate table the entries

for the first and second stations in the Handy Tables, and only displayed them at

6-intervals. The compilers of the Toledan Tables probably realized that it was

unnecessary to give entries for both the first and the second stations (because

corresponding entries add up to 360) and just included a specific column for

the first station of each of the planets. Thus, in the tables for the planetary

equations, ultimately derived from the Handy Tables, the number of entries

13 See Goldstein and Sawyer, Equant (ref. 11), 168. abash identified both the eccentricities

and the apogees of Venus and the Sun, despite the lack of justification based on obser-

vations or based on anything said by Ptolemy in the Almagest (or elsewhere). In modern

terms, this would mean that the solar orb serves as the deferent for Venus; but this claim

was not made by any medieval scholar. Nevertheless, the medieval tradition was to keep

the apogee and eccentricity of Venus equal to those of the Sun, such that whenever the

parameters for the Sun were changed, the same changes were applied to Venus.

14 See G.J. Toomer, A survey of the Toledan Tables, Osiris, xv (1968) 5174, p. 67.

15 Toomer, Toledan Tables (ref. 14), 6068; and F.S. Pedersen, The Toledan Tables: A review

of the manuscripts and the textual versions with an edition (Copenhagen, 2002), 12651306.

16 Toomer, Almagest (ref. 3), 588.

17 See Neugebauer, Ancient Astronomy (ref. 5), 10051006; and J. Chabs and B.R. Goldstein,

A Survey of European Astronomical Tables in the Late Middle Ages (Leiden, 2012), 118.

computing planetary positions 157

increased, for they are given here at intervals of one degree, and gained one

column which was eliminated as a separate table.18 The Toledan Tables were

by far the most popular tables in Latin Europe, and the presentation in them

for tables of planetary equations can be considered standard.

The Maghrib astronomers Ibn Isq al-Tnis (c. 11931222), Ibn al-Bann

of Marrakesh (12651321), and Ibn al-Raqqm (Tunis and Granada, d. 1315) used

new parameters for the equations of center of Saturn, Jupiter, and Venus. In

contrast, the values given to the equations of anomaly agreed precisely with

those in the standard tradition, namely that of the Handy Tables, the zij of

al-Battn, and the Toledan Tables, but for the fact that in this tradition the

columns displayed are combinations of cols. 5, 6, and 7.19 We further note that

the tables for the planetary equations of Ibn Isq and his followers depart

from the standard tradition not only in the three basic parameters already

mentioned, but also in presentation. Indeed, for each of the planets there are

two tables of equations: one for quantities that depend on the argument of

center and one for those that depend on the argument of anomaly.20

The Castilian Alfonsine Tables were produced in Castile by two astronomers

working under the patronage of Alfonso x (reigned: 12521284), Judah ben

Moses ha-Cohen and Isaac ben Sid. We do not know how the tables for the

planetary equations were presented in these tables, because the tables them-

selves are not extant. However, the canons have been preserved in Castilian,

and chapter 18 (De la equaion de los v planetas) describes the way to com-

pute planetary longitudes by means of tables. Although no numerical values are

given, the description agrees perfectly with the layout of tables in the standard

tradition of the Handy Tables, the zij of al-Battn, and the Toledan Tables.21

This tradition was transmitted from the Iberian Peninsula to the rest of

Europe. The earliest astronomer to depend on this Iberian tradition was John

Vimond, who was active in Paris c. 1320. He compiled a set of tables which

18 The only known example of this kind of table where the entries are given at intervals of

half a degree is preserved in a double folio now in the General Archive of Navarre: see

J. Chabs, The Toledan Tables in Castilian: Excerpts of the planetary equations, Suhayl,

xi (2012), 179188.

19 See J. Sams and E. Mills, The computation of planetary longtiudes in the zj of Ibn

al-Bann, Arabic Sciences and Philosophy, viii (1998), 259286; reprinted in J. Sams,

Astronomy and Astrology in al-Andalus and the Maghrib (Aldershot, 2007), Essay viii.

20 A. Mestres, Materials Andalusins en el Zj dIbn Isq al-Tnis (unpublished Ph.D. thesis,

Universitat de Barcelona, 1999), 5051 and 234235.

21 J. Chabs and B.R. Goldstein, The Alfonsine Tables of Toledo (Dordrecht, 2003), 3839 and

157160.

158 chapter 6

Maghrib and developed in Castile in the late thirteenth century on the one

hand, and the activity of the astronomers working in Paris in the 1320s and

the 1330s that resulted in the Parisian Alfonsine Tables on the other.22 In many

ways Vimonds tables follow a tradition unattested in Latin prior to 1320; for

example, his tables for the planetary equations are also split into two tables for

each planet, much as the Maghrib-Andalusian astronomers did.

In addition to changes in structure, which will be examined later, the main

modification in Vimonds tables is found in the entries for the equations of cen-

ter of Jupiter and Venus, with maximum values of 5;57 and 2;10, respectively.

Not much can be said about the value 5;57 other than it does not appear in any

text or table prior to Vimonds tables, and no medieval discussion of its origin

has been found. However, the value 2;10, also used by Vimond as the maximum

solar equation, appears in previous texts: implicitly in a table for the daily solar

positions for 1278 contained in the Libro del astrolabio llano composed by the

astronomers in the service of King Alfonso x of Castile,23 and explicitly in an

account in John of Murss Expositio of two observations of autumnal equinox,

one by Ptolemy in 132 and the other attributed to Alfonso in 1252, where John

explains that he has seen this observational report in what he calls the Tables

of Alfonso.24 We thus think it likely that these two new values for Jupiter and

Venus/Sun were taken from an earlier work, and the most reasonable candidate

is the Alfonsine Tables in the original Castilian version.25

The Parisian Alfonsine Tables, produced in the 1320s by a group of astrono-

mers working in Paris, were built on material coming from the Iberian Penin-

sula. They are best known today from the editio princeps that appeared in

Venice in 1483. While each part of this printed edition has a complicated his-

tory, the planetary equation tables in it are faithful to the Parisian Alfonsine

Tables as they were presented in the 1320s. The layout of the tables for the

planetary equations conforms to the standard tradition, although they have no

additional column for the first station. We will refer to the presentation and

parameters of this version of the Alfonsine Tables as standard. The entries

22 J. Chabs and B.R. Goldstein, Early Alfonsine Astronomy in Paris: The Tables of John

Vimond (1320), Suhayl, iv (2004), 207294, pp. 236256.

23 J. Chabs, Were the Alfonsine Tables of Toledo first used by their authors?, Centaurus, 45

(2003), 142150.

24 E. Poulle, Jean de Murs et les tables alphonsines, Archives dhistoire doctrinale et littraire

du moyen ge, xlvii (1980), 241271, p. 253.

25 Chabs and Goldstein, Toledo (ref. 21), 251254.

computing planetary positions 159

table b Maximum values of the equations of center and anomaly in various sets of tables

(new values are shown in boldface type)

Eq. of Eq. of Eq. of Eq. of Eq. of Eq. of Eq. of Eq. of Eq. of Eq. of

center anom. center anom. center anom. center anom. center anom.

Almagest * 6;31 6;13 5;15 11; 3 11;25 41; 9 2;24 45;59 3; 2 22; 2

Handy Tables 6;31 6;13 5;15 11; 3 11;25 41;10 2;24 45;59 3; 2 22; 2

al-Khwrizm 8;36 5;44 5; 6 10;52 11;13 40;31 2;14 47;11 4; 2 21;30

al-Battn 6;31 6;13 5;15 11; 3 11;25 41; 9 1;59 45;59 3; 2 22; 2

Toledan Tab. 6;31 6;13 5;15 11; 3 11;24 41; 9 1;59 45;59 3; 2 22; 2

Maghrib astr. 5;48 6;13 5;41 11; 3 11;25 41; 9 1;51 45;59 3; 2 22; 2

Vimond ** 6;31 6;13 5;57 11; 3 11;24 41; 9 2;10 45;59 3; 2 22; 2

Parisian Alf. 6;31 6;13 5;57 11; 3 11;24 41;10 2;10 45;59 3; 2 22; 2

* The values for the equation of center shown here are found by adding algebraically the equation

in longitude and the difference in equation in Alm. xi.11 (cols. 3 and 4).

** The values for the equation of center shown here result from subtracting the motus completus

(col. 2) from the mean argument of center (col. 1). The values for the equation of anomaly shown

here result from adding the motus completus (col. 2) to the correction for maximum distance (col. 5

in the standard tradition); see below.

are given at intervals of one degree, as was already established in the Handy

Tables.26 Moreover, out of ten basic parameters for the five planets, only two

differ from those defined by Ptolemy, namely, the equations of center of Jupiter

and Venus, and both of these parameters are already found in John Vimonds

tables. It is difficult to find other examples of such great stability in the trans-

mission of astronomical tables for more than 13 centuries. Table b provides a

summary of the main parameters for the equations of center and anomaly used

by different authors.

26 Characteristic of the Parisian Alfonsine Tables is the consistent use of sexagesimal days

and angles. Angles are given in physical signs of 60 (in contrast to zodiacal signs of 30):

an angle a,b means a 60 + b degrees (where a and b are integers such that 0 a 5 and

0 b 59; and 6,0 = 360). In our notation 10s 25 means 10 30 + 25 = 325, that is, s

signifies a zodiacal sign of 30. Sexagesimal fractions of a degree are used in the same way

with both physical signs and zodiacal signs.

160 chapter 6

Prior to the first edition of the Parisian Alfonsine Tables in 1483, a variety of

original approaches for presenting tables for the planetary equations were

undertaken within the Alfonsine corpus. They coexisted with the standard

tradition, which is preserved in a number of manuscripts dating from the

fourteenth and fifteenth centuries.27 The goal of these new approaches that

depart from the standard tradition was, once again, to facilitate computa-

tion.

Let us return to about 1320, the date of John Vimonds tables, in which the two

equations for each planet are displayed in different tables. In those where all

the tabulated functions depend on the mean argument of center (see Table c),

the entries are given at 6-intervals. Vimond displayed the true argument of

center (col. 2: motus completus) and added columns for the increment of the

true argument per degree of the argument (col. 3: motus gradus), planetary

velocity (col. 4: motus diei), minutes of proportion (col. 5: diametri), and first

station (col. 6). Moreover, the equation of center incorporates a displacement

which is the difference between the apogee of each of the planets and that

of the Sun (no displacement is therefore needed in the case of Venus, for its

apogee is assumed to be the same as that for the Sun). Analysis of Vimonds

tables shows that the motion of the solar apogee was included in the motions

of the planetary apogees, thus following a theory for which there was no

previous evidence outside al-Andalus and the Maghrib.28 With this particular

arrangement Vimond intended to present a more user-friendly table than the

standard table for the equation of center.

Now, in the tables where all the tabulated functions depend on the argu-

ment of anomaly (given at 6-intervals for Saturn, Jupiter, and Mercury, and at

3-intervalsand at 2-intervals in the vicinity of 180for the other two plan-

ets), Vimond also added columns for planetary velocities and other corrections,

such as col. 5 (see Table d), which results from adding the correction for max-

imum distance to the correction for minimum distance (cols. 5 and 7, respec-

tively, in the Handy Tables, the zij of al-Battn, and the Toledan Tables). As

was the case for the equation of center, the entries for the equation of anomaly

are not explicitly displayed. Rather, we are given entries for the motus comple-

tus (col. 2), which is the difference between the equation of anomaly and the

27 For example, for a list of manuscripts extant in Spain or of Spanish origin, see Chabs and

Goldstein, Toledo (ref. 21), 292303.

28 Sams and Mills, Ibn al-Bann (ref. 19), 268270.

computing planetary positions 161

(excerpt)

Motus Motus Motus First

Argument completus gradus diei Diametri station

s () s () min min min s ()

0 12 0 17;36 50; 0 26 4 5 8;21

1 18 1 17;16 49;10 25 0 5 7;31

7 18 7 18;54 73;10 38 59 5 19;13

12 0 12 7;24 51;10 26 8 5 9; 6

cols. 6 and 5 in the standard tradition, we find agreement with Vimonds motus

completus, indicating that he kept all the basic parameters for the equation

of anomaly in this tradition. It is noteworthy that, as indicated by North, this

implies that Ptolemys eccentricities underlie these tables even though, in

the case of Venus and Jupiter, the eccentricities were modified for computing

the equation of center.29 The only text of which we are aware that treats

the equation of anomaly in this way is of Maghrib origin: the Minhj of Ibn

al-Bann, dependent on the zij of Ibn Isq. In the Minhj the tables for the

equations of anomaly of Saturn and Jupiter give entries for al-mufrad (c6 c5

in the standard terminology for columns) and al-bucd (c5 + c7).30 These are

30 Sams and Mills, Ibn al-Bann (ref. 19), 278285.

162 chapter 6

Motus Motus Motus Motus Motus

Argument compl. gradus diei Diametri grad. diei

s () s () () min min min sec sec

0 3 11 27 1; 8 23 11 0; 8 3 1

0 6 11 24 2;16 23 11 0;17 3 1

4 3 7 27 36;40 1 1 8;53 9 4

4 6 7 24 36;44 0 0 9;19 9 4

4 9 7 21 36;43 3 1 9;46 9 4

5 6 6 24 28;15 46 21 13;30 0 0

5 9 6 21 25;56 53 25 13;37 6 2

5 12 6 18 23;17 62 29 13;19 13 6

5 28 6 2 3; 1 90 42 2;29 74 35

6 0 6 0 0; 0 90 42 0; 0 74 35

precisely two of the columns found in Vimonds tables (cols. 2 and 5). This

particular choice of columns was intended to facilitate the computation of the

planetary equations of anomaly.

In addition to the Expositio, already mentioned, John of Murs, a key fig-

ure in the Parisian milieu for the transmission of Alfonsine astronomy, was

responsible for a set of tables, called the Tables of 1321, devoted exclusively

to the planets and the two luminaries. With these tables the computation of

true planetary positions is entirely different from that described in any other

text of which we are aware.31 The most significant feature of the Tables of

1321 is a new organizational principle, which does not require the equations

of the planets to be displayed explicitly. To be sure, the mean motions of

the planets are here presented in tables for the mean conjunctions of each

planet with the Sun (tabula principalis), and the corrections to be applied for

times between consecutive conjunctions are given in double argument tables

31 J. Chabs and B.R. Goldstein, John of Murss Tables of 1321, Journal for the History of

Astronomy, xl (2009), 297320.

computing planetary positions 163

the typically cumbersome computations for determining true planetary posi-

tions, compared with using tables previously available in Latin. One unusual

feature of these double argument tables is that the horizontal argument is the

mean argument of center of the planet (at intervals of 12) and the vertical

argument is the age of the planet, that is, the time after a mean conjunc-

tion with the Sun, expressed as a number of days. It is also noteworthy that

for each planet, besides the tabula and contratabula, we are given a table for

its equation of center and first station. The values of the maximum equation

of center agree in all cases with those used by Vimond, and so do the rest of

the entries (given at 6-intervals in both sets of tables, but presented differ-

ently).

The tables of Vimond and those of John of Murs for 1321 certainly made

the computation of the true positions of the planets much easier, but their

approaches do not seem to have been very popular. The main improvement

in that direction came from double argument tables, which greatly simplified

computations and only required linear interpolation.32 John of Lignres (also

active in Paris) was probably the first astronomer in Latin Europe to draw

up a double argument table combining the equations of center and anomaly

32 There were a few double argument tables in Islamic zijes prior to 1320, e.g., Ibn Ynus

(c. 990) had such a table for the lunar equations as did al-Baghdd (c. 1285): see D.A. King,

A Double-Argument Table for the Lunar Equation Attributed to Ibn Ynus, Centaurus,

xviii (1974), 129146 and C. Jensen, The Lunar Theories of al-Baghdd, Archive for History

of Exact Sciences, viii (1972), 321328. Double argument tables for planetary latitudes

attributed to Ibn al-Bayr, who is otherwise unknown, are preserved in Hyderabad, Andra

Pradesh State Library, ms 298, Tables 6677, where the horizontal headings are the true

arguments of center and the vertical headings are the true arguments of anomaly (both at

intervals of 6): see A. Mestres, Maghrib Astronomy in the 13th Century: a Description of

Manuscript Hyderabad Andra Pradesh State Library 298, in From Baghdad to Barcelona:

Studies in the Islamic Exact Sciences in Honour of Prof. Juan Vernet, ed. by J. Casulleras and

J. Sams (Barcelona, 1996), 383443, p. 428. These tables are also cited in B. van Dalen,

Tables of Planetary Latitude in the Huihui li (ii), in Current Perspectives in the History

of Science in East Asia, ed. by Y.S. Kim and F. Bray (Seoul, 1999), 316329, p. 327. This

manuscript contains the zij of Ibn Isq (early thirteenth century), and Ibn al-Bayr is

mentioned in chapter 18 of the canons to this zij. This implies that Ibn al-Bayr was active

no later than the time of Ibn Isq. A summary of chapter 18 appears in Mestres 1996,

pp. 396397. We are most grateful to van Dalen for sharing with us his translation of the

Arabic text of chapter 18, the Arabic text of chapter 18 and of Tables 6677, as well as his

notes on this material.

164 chapter 6

in a single table for each planet in his Tabule magne (c. 1325).33 The vertical

argument is the mean argument of anomaly (at 6-intervals in all planets, and

also at 3-intervals from 150 to 180 in the case of Mars and Venus),34 and the

horizontal argument is the mean argument of center (at 6-intervals). In Table

e we reproduce an excerpt from the table for the combined equation of Venus

in John of Lignress Tabule magne, as found in Lisbon, ms Ajuda 52-xii-35,

ff. 83r87v, with the title Tabula equationum ultimarum veneris. We note the

use of physical signs of 60, and the inclusion of columns for the differences,

to minutes, of successive entries for a fixed value of the argument of anomaly

(not displayed here).35

For each planet there is a total of at least 1860 entries (2160 in the case of Mars

and Venus) presented as 31 60 or 36 60 matrices, not taking into account the

columns and rows that display the successive differences. None of the entries

explicitly corresponds to the maximum values of the equations of center or

anomaly, which could lead to the identification of the tradition to which it

33 E. Poulle, John of Lignres, Dictionary of Scientific Biography (16 vols, New York, 1970

1980), vii (1973), 122128, pp. 123124; J.D. North The Alfonsine Tables in England, in Pris-

mata: Festschrift fr Willy Hartner, ed. by Y. Maeyama and W.G. Salzer (Wiesbaden, 1977),

269301, pp. 273274, 278; reprinted in J.D. North, The universal frame: historical essays in

astronomy, natural philosophy, and scientific method (London, 1989), 327359.

34 The reason is that in the range 150210 the entries for these two planets vary quite

rapidly, and thus the accuracy of interpolation is increased by doubling the number of

entries. Vimond was already aware of this rapid variation and, in his tables for the equation

of anomaly for Mars and Venus given at 3-intervals, he displayed entries in the range

168192 at 2-intervals.

35 Only three manuscripts containing theses tables are known. The other two use zodiacal

signs of 30 (Erfurt, Biblioteca Amploniana, ms ca 2 388, and London, British Library, ms

add. 24070). All three have a column at the far right for the mean argument of anomaly

from 180 to 360, but the numbers in this column, and only in this one, are inverted in

different ways in two manuscripts; for example, 234 is written as 54 3 (meaning 3,54) in

ms Lisbon; 24 7s in ms Erfurt; and 7s 24 in ms London. It is possible that this reflects

an archetype in Arabic or Hebrew, but see C. Burnett, Why we read Arabic numerals

backwards, in Ancient and Medieval Traditions in the Exact Sciences, ed. by P. Suppes et

al. (Stanford, 2000), reprinted in C. Burnett, Numerals and arithmetic in the Middle Ages

(Aldershot, 2010), Essay vii, for some examples of this inversion in medieval Latin texts.

Moreover, the three manuscripts differ in another respect: ms Lisbon has a column for

successive differences, ms Erfurt has no such column, and ms London has columns and

rows for successive differences. This is definitely a good example of the intervention of

copyists when transmitting the very same table, without altering its presentation or any

of its essential features.

computing planetary positions 165

table e John of Lignress double argument table for the combined equation of Venus

(excerpt)

0,6 0,12 3,54 4,0 4,6 5,54 0,0

() () () () () () () ()

m* m a a a a a

0, 0 0; 8 0;16 1; 0 ** 1; 4 1; 8 0; 8 0; 0

a a

0, 6 2;22 2;14 3;35 3;39 3;42 2;38 2;30

0,12 4;50 4;43 6; 6 6;10 6;13 5; 7 4;58

2,12 44;32 44;20 48;10 48;10 48; 8 44;57 44;44

2,18 44;29 40;18 48;33 48;33 48;30 44;59 44;43

2,24 43;41 41;44 48;13 48;16 48;14 44;13 43;57

2,57 6;23 5;38 27;14 19;14 15;14 7;44 7; 7

m m

3, 0 0;44 1;28 31; 6 24; 7 19; 7 0;39 0; 0

* m stands for minue (to be subtracted) and a for adde (to be added).

** ms Erfurt: 1;1.

belongs (see Table b), but a few entries are easy to track. Let us consider the

case when = 0 or 180. Then c3() = 0 and = , and the entries for = 90

reduce to c6(90) c5(90) and c6(90) + c7(90), respectively, in the usual termi-

nology for columns. We find agreement in all cases, except for the equation

for Mercury at greatest distance (the entry reads 21;32, whereas computation

with the tables in the standard tradition give 22;2).36 In all other cases there

is good agreement, but it is not always perfect because columns 5, 6, and 7,

which depend exclusively on the argument of anomaly sometimes vary in the

36 In the case of Mars, for instance, c5(90) = 2;28, c6(90) = 33;22, and c7(90) = 2;49 in the

standard tradition beginning in the Handy Tables (see Table a). Thus, c6(90) c5(90) =

30;54 and c6(90) + c7(90) = 36;11. The entry in John of Lignress table of Mars for = 0

and = 90 is 30;54 and that for = 180 and = 90 is 36;11.

166 chapter 6

minutes. To show that the entries in John of Lignress table are specifically

based on the values used by John Vimond and, in particular, on those maximum

values for the equation of center found for the first time in Vimonds tables,

we have recomputed a few critical entries.37 The maximum entries in John of

Lignress tables could not have been computed with values as low as those

in the tradition represented by the Toledan Tables, and we conclude that they

were calculated with Vimonds tables, or that both astronomers had a common

source.

It should be noted that in his tables of 1322 John of Lignres had used the

parameters 1;59 (Venus) and 5;15 (Jupiter) that are found in the Toledan

Tables for the maximum equations of center, replacing them with 2;10 (Venus)

and 5;57 (Jupiter) in his double argument tables for the planetary equations

in 1325. As a matter of fact, John of Lignress tables for planetary equations for

1322, as presented in Bibliothque nationale de France, ms 7286c (ff. 33r47v),38

share the same entries and layout, including a column for first station, with

the Toledan Tables. This change was much the same as John of Murs had done

a few years previously, given that in his earliest astronomical work of 1317,

beginning Auctores calendarii , he had praised the Tables of Toulouse and

seemed unaware of Alfonsine material.39

37 For Venus, the maximum entry in John of Lignress table of 1325 is 48;33 (at = 3,54 and

4,0; and = 2,18) as displayed in Table e. If = 4,0 = 240, then c3(240) = 1;55, using

an equation of center with a maximum of 2;10 (Vimonds value), and c4(240) = 31. Thus,

= 138 1;55 = 136;5. Therefore, c5(136;5) = 1;11, c6(136;5) = 45;59, and c7(136;5) = 1;16.

Finally, the combined equation is 1;55 + 45;59 + (1;16 31/60) = 48;32, in agreement with

the entry. However, when performing the same calculation using an equation of center

with a maximum of 1;59 (as in the Toledan Tables), one finds a combined equation of

48;19. As we shall see in the computation that follows, the results are also unambiguous

in the case of Jupiter. The maximum entry in John of Lignress table is 17;1 (at = 4,24

and = 1,48). If = 4,24 = 264, then c3(264) = 5;57, which is the maximum equation

of center in Vimonds tables, and c4(264) = 7. Thus, = 108 5;57 = 102;3. Therefore,

c5(102;3) = 0;29, c6(102;3) = 11;3, and c7(102;3) = 0;32. Finally, the combined equation is

5;57 + 11;3 + (0;32 7/60) = 17;4, very close to the entry in John of Lignress Tabule magne.

However, if we use a table with a maximum equation of center of 5;15 (as in the Toledan

Tables), one finds a combined equation of 16;21.

38 For a description of this manuscript, see M.-M. Saby, Les canons de Jean de Lignres sur

les tables astronomiques de 1321, (Unpublished thesis: cole Nationale des Chartes, Paris,

1987), 516520. A summary appeared as Les canons de Jean de Lignres sur les table

astronomiques de 1321, cole Nationale des Chartes: Positions des thses (1987), 183190.

39 Chabs and Goldstein, Toledo (ref. 21), 278. The Tables of Toulouse are an adaptation of

the Toledan Tables for the Christian calendar (instead of the Muslim calendar): see Poulle,

computing planetary positions 167

displayed in a compact and clever way intermediate calculations needed to

obtain a final numerical result.40 This kind of presentation was not an inven-

tion of the Parisian astronomers, for it is already found in Arabic sources, e.g.,

it was used by Ibn al-Kammd (Crdoba, c. 1100) in his tables for the time from

mean to true syzygy as a function of the difference between the hourly veloc-

ities of the Moon and the Sun and the elongation.41 Double argument tables

proliferated in fourteenth-century Europe and were not restricted to the plan-

etary equations: they were also used to display true planetary positions (the

Tabule anglicane, also called the Oxford Tables of 1348, associated with William

Batecombe); planetary conjunctions (John of Murss Tables of 1321); planetary

latitudes (John of Murss Tables of 1321, and the Oxford Tables); syzygies (John

of Murs and Firmin of Beauval in their Tabulae permanentes, Immanuel ben

Jacob Bonfils of Tarascon, Levi ben Gerson, Juan Gil of Burgos, Joseph Ibn

Waqr of Seville, and the Tables of Barcelona); lunar motion (Levi ben Gerson);

and lunar and planetary velocities (Judah ben Asher ii of Burgos).42

matriser la nature au moyen ge: Mlanges dhistoire des sciences offerts Guy Beaujouan

(Geneva and Paris, 1994), 5581.

40 For interpolation in double argument tables, see M. Husson, Ways to read a table: reading

and interpolation techniques of early fourteenth-century double-argument tables, Jour-

nal for the History of Astronomy, xliii (2012), 299319.

41 These tables are only extant in Latin and Hebrew versions; the Latin version was com-

posed by John of Dumpno in 1260 in Palermo and survives uniquely in Madrid, Biblioteca

Nacional de Espaa, ms 10023. See J. Chabs and B.R. Goldstein, Andalusian Astronomy:

al-Zj al-Muqtabis of Ibn al-Kammd, Archive for History of Exact Sciences, xlviii (1994),

141; and B.R. Goldstein, Solomon Franco on the Zero Point for Trepidation, Suhayl, x

(2011), 7783.

42 For the Tabule anglicane see J.D. North, England (ref. 33); for the Tables of 1321, see

J. Chabs and B.R. Goldstein, John of Murs (ref. 31); for the Tabulae permanentes, see

B. Porres and J. Chabs, John of Murss Tabulae permanentes for finding true syzygies,

Journal for the History of Astronomy, xxxii (2001), 6372; for Immanuel ben Jacob Bonfils

of Tarascon, see P. Solon, The Hexapterygon of Michael Chrysokokkes (Brown University,

Ph.D. thesis (unpublished), 1968; Proquest, umi, aat 6910019), and P. Solon, The Six Wings

of Immanuel Bonfils and Michael Chrysokokkes, Centaurus, xv (1970), 120; for Levi ben

Gerson, see B.R. Goldstein, The Astronomical Tables of Levi ben Gerson (Hamden, ct, 1974);

for Juan Gil of Burgos and Joseph Ibn Waqr of Seville, see J. Chabs and B.R. Goldstein,

Computational Astronomy: Five Centuries of Finding True Syzygy, Journal for the History

of Astronomy, xxviii (1997), 93105, pp. 9496; for the Tables of Barcelona, see J.M. Mills,

Las Tablas Astronmicas del Rey Don Pedro el Ceremonioso (Madrid and Barcelona, 1962),

and J. Chabs, Astronomia andalus en Catalua: Las Tablas de Barcelona, in From

168 chapter 6

Another set of tables in the Alfonsine corpus that adheres strictly to its

parameters and models is the set we call the Tables for the Seven Planets for

1340; they are a most ingenious reworking of the Parisian Alfonsine Tables and

include several displaced tables. The purpose of displaced tables is to elimi-

nate all subtractions in the derivation of planetary positions, thus facilitating

computations.43 This anonymous set of tables, most likely of French origin, is

uniquely preserved in Paris, Bibliothque nationale de France, ms 10262 (ff. 2r

46v). The two planetary equations are given in separate tables for each planet,

and are not explicitly displayed. Rather, for the equation of center we are given

entries which are displaced both vertically and horizontally with respect to

those in the standard Parisian Alfonsine Tables, whereas the entries for the

equation of anomaly are only displaced vertically. In modern algebraic terms,

the vertical and horizontal displacements of a function underlying a displaced

table are such that y = f(x + kh) + kv, where y = f(x) is the original function to

which the displaced table is compared, kh is the displacement on the x-axis,

and kv is the displacement on the y-axis. Tables f and g display excerpts of

the equation of center of Jupiter in the Tables for the Seven Planets and in the

Parisian Alfonsine Tables, respectively (where kh = 18 and kv = 6).

Figure 6.2 illustrates the situation for Jupiter. The graph labeled ms 10262

displays the entries in the Tables for the Seven Planets for 1340, and that labeled

pat corresponds to those in the Parisian Alfonsine Tables.

In general the vertical displacements are intended to avoid complicated

rules for addition and subtraction corresponding to the simple rules we now

give by means of algebraic signs. The horizontal displacements are intended to

counterbalance other displacements, such as those applied to the minutes of

proportion. It is easy to recognize that the vertical displacements of the entries

for the equation of anomaly agree with the parameters found in the Parisian

Alfonsine Tables: see Table b. The Tables for the Seven Planets use a total of 40

different displacements for the planets (including the Sun and the Moon): see

Table h.

Baghdad to Barcelona: Studies in the Islamic Exact Sciences in Honour of Prof. Juan Vernet,

ed. by J. Casulleras and J. Sams (Barcelona, 1996), 477525; for Judah ben Asher ii

of Burgos, see B.R. Goldstein, Abraham Zacut and the Medieval Hebrew Astronomical

Tradition, Journal for the History of Astronomy, xxix (1998), 177186, pp. 179181.

43 For details, see J. Chabs and B.R. Goldstein, Displaced Tables in Latin: The Tables for the

Seven Planets for 1340, Archive for History of Exact Sciences, lxvii (2013), 142. Note that the

term displaced applied to tables was coined by E.S. Kennedy in 1977 as a translation of

the Arabic wa (see his The Astronomical Tables of Ibn al-Aclam, Journal for the History

of Arabic Science, i (1977), 1323, espec. p. 16).

computing planetary positions 169

Jupiter in the Tables for the in the Parisian Alfonsine

Seven Planets44 Tables45

centri of proportion of center of proportion

() () () () () ()

0 4;15 0 1 0; 6 60

1 4; 9 1 2 0;12 60

2 4; 3 1 3 0;12 60

71 0; 4 53 88 5;56 1

72 0; 3 54 89 5;56 1

90 5;57 2

77 0; 3 59

78 0; 3 2 96 5;57 7

79 0; 4 3 97 5;56 8

168 6;39 60 180 0; 0 60

245 11;56 14 263 5;56 8

246 11;57 13 264 5;57 7

252 11;57 7 270 5;57 2

253 11;56 6 271 5;56 1

272 5;56 1

259 11;52 1 273 5;55 2

260 11;51 58

358 0;12 60

359 4;20 0 359 0; 6 60

44 Note that the equatio centri is always positive; it reaches a minimum of 0;3 at 7278,

and a maximum of 11;57 at 246252. For the minutes of proportion there are two

discontinuities (from 60 to 0 between 77 and 78, and from 0 to 60 between 259 and

260), to keep them positive in all cases.

45 Note that the equation of center is negative between 0 and 180, and positive between

180 and 360; it reaches a minimum of 5;57 at 9096, and a maximum of 5;57 at

264270. The minutes of proportion are positive between 89 and 271, and negative for

the rest of the arguments.

170 chapter 6

figure 6.2 The equation of center of Jupiter. The graph labeled ms 10262 is displaced

vertically by 6 and horizontally by 18 with respect to the graph labeled

pat. The maximum of the upper graph is 11;57 and takes place at

arguments 246252; the maximum of the lower graph is 5;57 and takes

place at arguments 264270.

In any case, computation with this compact and consistent set of tables gives

the same results as those obtained with the Parisian Alfonsine Tables, while

avoiding subtractions at any stage in the computation.

In the fifteenth century the Paduan astronomer, Prosdocimo de Beldomandi

(d. 1428), compiled a new set of tables that belong to the Alfonsine corpus.46 His

tables for the planetary equations follow the Parisian Alfonsine Tables, includ-

ing the 2;10 and 5;57 used by Vimond for Venus and Jupiter, in agreement with

those that were printed in 1483 in the editio princeps.

Giovanni Bianchini (d. after 1469) spent most of his life in Ferrara where he

served as administrator for the estate of the prominent dEste family. About

1442 he compiled an extensive set of astronomical tables which depend on the

Alfonsine Tables, but have a completely different presentation.47 Bianchinis

tables offer a whole new approach for computing the true positions of the

Journal for the History of Astronomy, xxxviii (2007), 269281.

47 J. Chabs and B.R. Goldstein, The Astronomical Tables of Giovanni Bianchini (Leiden, 2009).

computing planetary positions 171

for the Seven Planets for 1340

Vert. displac. Horiz. displac. Vert. displac.

Saturn 7 14 6;13

Jupiter 6 18 11; 3

Mars 12 61 41;10

Venus 3 51 45;59

Mercury 4 28 22; 2

planets. Although tables for the planetary equations are not explicitly given, the

true positions of the planets are computed by means of double argument tables

where the vertical argument is the mean anomaly, represented here by the time

within an anomalistic period for each planet, and the horizontal argument is

the mean center. These tables were first published in 1495 in Venice under the

title Tabulae astronomiae, and again in 1526 and 1553.48

The Tabulae resolutae were compiled in central Europe, and circulated

widely in manuscripts during the fifteenth century and in print during the

sixteenth century.49 One of their characteristics is that the mean motions are

arranged according a system of cyclical radices at intervals of 20 year. The Tabu-

lae resolutae are also strictly based in the Parisian Alfonsine Tables; in fact, they

are a particular form of presenting them. The tables for the planetary equations

display the same parameters as the Parisian Alfonsine Tables but, unlike them,

add a column for first station, thus following the layout of most versions of the

Toledan Tables.

In Vienna John of Gmunden (c. 13801442) collected a great variety of tables

within the framework of the Parisian Alfonsine Tables. He displayed them in

48 For a list of manuscripts that contain Bianchinis tables, see Chabs and Goldstein, Bian-

chini (ref. 47), 14. The owners of manuscript copies of these tables include Johannes

Regiomontanus (Nuremberg, Stadtbibliothek, Cent v 57) and Johannes Virdung (Vatican,

Biblioteca Apostolica, ms Pal. lat. 1375).

49 J. Dobrzycki, The Tabulae Resolutae, in De Astronomia Alphonsi Regis, ed. by M. Comes,

R. Puig, and J. Sams (Barcelona, 1987), 7177; J. Chabs, Astronomy in Salamanca in

the Mid-Fifteenth Century: The Tabulae Resolutae, Journal for the History of Astronomy,

xxix (1998),167175; and J. Chabs, The Diffusion of the Alfonsine Tables: The Case of the

Tabulae resolutae, Perspectives on Science, x (2002), 168178.

172 chapter 6

various sets, called First Version, Tabulae maiores, and Tabulae breviores.50 He

presented his tables for the equations of the planets in three different ways:

at 1-intervals following the standard tradition; at 3-intervals in an abridged

form of the latter; and as double argument tables, reproducing those by John of

Lignres. Therefore, with respect to the planetary equations, John of Gmunden

was not an innovator; rather, he offered table users several possibilities that

were already known in Latin Europe.

In the early sixteenth century Johannes Angelus, a follower of Peurbach and

Regiomontanus, claimed that these two authors had compiled a new table of

planetary equations giving better results than the standard Alfonsine Tables,

but this new table has not been found in any manuscript or printed edition.51

As already mentioned, the Alfonsine Tables were first printed in 1483 by

Erhard Ratdolt in Venice. A few years later (1492) and in the same town, a sec-

ond edition appeared, edited by Johannes Lucilius Santritter. The entries for

the planetary equations are the same in both sets of tables but, in the second

edition, the planets were inexplicably presented in the order Venus, Mercury,

Mars, Jupiter, and Saturn (rather than in the usual order where Mercury pre-

cedes Venus). Also in the 1492 edition, the second column for the argument,

displaying the complement in 360, was eliminated; this left enough space on

the page to include five extra columns for the differences between successive

entries in the remaining columns.

In 1503 Petrus Liechtenstein printed another set of tables in Venice, the Tab-

ule Astronomice Elisabeth Regine. It was much less popular than the standard

version of the Parisian Alfonsine Tables, but it is historically significant because

in his Commentariolus Copernicus cited its author, Alfonso de Crdoba, who

was in the service of Pope Alexander vi in Rome.52 The tables for the planetary

(unpublished Ph.D. thesis, cole pratique des hautes tudes, Section iv, Paris, 2003).

51 J. Dobrzycki and R.L. Kremer, Peurbach and Margha Astronomy? The Ephemerides of

Johannes Angelus and Their Implications, Journal for the History of Astronomy, xxvii

(1996), 187237, pp. 187188.

52 Alfonso de Crdoba, known as Hispalensis, came from Seville (Latin: Hispalis) and his

origin is well attested in various printed texts of the early sixteenth century. His place of

origin is explicitly given as patria hispalensis and he is cited as Alfonso hispalensi de Cor-

duba. Hence there is no reason to emend the text to hispaniensis, that is, from Spain, as

has been suggested: see N.M. Swerdlow, The Derivation and First Draft of the Copernicuss

Planetary Theory: A Translation of the Commentariolus with Commentary, Proceedings

of the American Philosophical Society, cxvii (1973), 423512, pp. 451452. Copernicuss Com-

mentariolus is undated, but it is usually taken to be from about 1514. Alfonso dedicated his

computing planetary positions 173

equation of center of

Mars (excerpt)

Mars

Longitude () min

Leo 15 Leo 15 0; 0 60

Leo 20 Leo 10 0;55 60

Leo 25 Leo 5 1;49 59

Sco 15 Tau 15 11;23 3

Sco 20 Tau 10 11;24 3

Sco 25 Tau 5 11;21 8

Aqr 5 Aqr 25 2;13 58

Aqr 10 Aqr 20 1; 7 59

Aqr 15 Aqr 15 0; 0 60

equations, as well as all the others in this set, depend on the Parisian Alfonsine

Tables, both for models and parameters. However, this is not true for the

presentation. First, for each planet the equation of center is given in a different

table from that for the equation of anomaly. As we have seen, John Vimond

had also used this two-fold presentation, which was most uncommon in Latin

astronomy,53 but the two sets differ in several important aspects (see Table i).

Second, the argument in the tables for the equation of center is given at

5-intervals (as is the case in the tables for the equation of anomaly), in contrast

to the tables in the standard tradition (1-intervals). But most important of all

is the fact that the argument in the table for the equation of center represents

work to Queen Isabella of Castile and Aragon (14511504), whose name in Latin was Elis-

abeth. On this set of tables, see J. Chabs, Astronomy for the Court in the Early Sixteenth

Century: Alfonso de Crdoba and his Tabule Astronomice Elisabeth Regine, Archive for

History of Exact Sciences, lviii (2004), 183217.

53 This separation was intended to distinguish clearly between the columns that depend

on one variable from those that depend on the other. We know of another example of

this two-fold presentation in Erfurt, Biblioteca Amploniana, ms q 362 (ff. 28r36r), also in

the Alfonsine corpus, whose layout seems unrelated to those by either John Vimond or

Alfonso de Crdoba.

174 chapter 6

the mean longitude of the planet, , that is, the mean argument of center plus

the longitude of the planets apogee. Thus, the argument is shifted by a quantity

that, in each case, corresponds to the longitude of the apogee (Leo 15 in the

case of Mars). Again, the purpose is to facilitate calculation. In turn, the tables

for the equation of anomaly display the usual columns of the tables in the

standard tradition (cols. 1, 2, 5, 6, and 7).54

The tabular innovations developed to facilitate computation of the true

longitude of the planets paved the way to a substantial increase in the number

of almanacs in the fourteenth and fifteenth centuries. Although there are some

earlier examples of this genre, the various new presentations of the tables for

planetary equations (such as double argument tables) made almanacs much

easier to compile. In turn, since almanacs and ephemerides display directly

the true positions of the planets at successive times, the user did not have the

difficult task of computing planetary equations; hence, they were very popular,

for they could be used even by those who had not mastered all the subtleties of

astronomy.55

Perhaps the most elaborate and influential almanac in the late Middle Ages

was the Almanach perpetuum.56 Its tables, together with a short explanatory

text, were first printed in two editions (one in Latin and the other in Castilian)

in Leiria, Portugal, in 1496. The tables were derived from a set of astronomical

tables in Hebrew called ha-ibbur ha-gadol (The Great Composition) compiled

by Abraham Zacut of Salamanca (14521515).57 As regards the positions of the

planets, Zacuts work was compiled in the framework of the Parisian Alfonsine

Tables with 1473 as epoch. For each planet it gives the true longitude, the true

argument of center, and the true argument of anomaly for several days in each

54 The fact that the argument in the table for the equation of center was chosen to be the

mean longitude of the planet makes the table less useful in the long term, for it does not

take into account the slow motion of the apogee due to precession.

55 An almanac lists successive true positions of each planet at intervals of one day or a few

days, that is, each planet is listed separately; hence, the information for a given day is

scattered among various tables. An ephemeris displays the true positions of all planets

in a single row at intervals of one day or a few days for a certain number of years. That is,

the difference between an almanac and an ephemeris is only one of presentation.

56 J. Chabs and B.R. Goldstein, Astronomy in the Iberian Peninsula: Abraham Zacut and the

Transition from Manuscript to Print (Transactions of the American Philosophical Society,

90.2, Philadelphia, 2000).

57 On the various editions of the Almanach perpetuum and its impact on the Jewish com-

munity as well as on Christian and Muslim scholars, see Chabs and Goldstein, Zacut (ref.

56), 161171.

computing planetary positions 175

month (sometimes daily) for periods as long as 125 years in the case of the

longitude of Mercury. For these three quantities there is a total of more than

42,100 entries and, in each case, the sign, the degrees, and the minutes are

specified. We have certainly come a long way from the tables for the planetary

equations in the Handy Tables!

Conclusion

Handy Tables, survived at least until the first printed editions of the Parisian

Alfonsine Tables. Ptolemys underlying models and most of the parameters

involved were rarely challenged from about the middle of the second century

to the end of the fifteenth century. Only two parameters appearing in the

tables were changed in that period, and John Vimond seems to have been the

first astronomer to have used them in Latin Europe.58 Vimond depended on

material from the Iberian Peninsula, most likely of Arabic origin.

Astronomers in Latin Europe in the fourteenth and fifteenth centuries were

actively engaged with this well-defined tradition, but they did not simply repro-

duce the tables and texts of their predecessors, and many of them devel-

oped innovative approaches to facilitate computational tasks, such as double

argument tables, displaced tables, separated tables, or shifted variables. User-

friendliness, rather than improvement of the models or enhancement of preci-

sion, was the driving force for most of the efforts developed by table-makers

in the computation of planetary positions. Nevertheless, the editio princeps

of the Parisian Alfonsine Tables did not incorporate any of the various new

presentations. These innovations in presentation have only been recognized

in recent years and, taken together, they indicate that astronomers in Latin

Europe reached a high level of mathematical competence in the late Middle

Ages.

58 To be sure, the apogee of Venus was also changed, but this does not modify the table for

the equations for this planet.

part 3

Sets of Tables

chapter 7

Andalusian Astronomy:

al-Zj al-Muqtabis of Ibn al-Kammd*1

Introduction

which he described, briefly in most cases, over 100 sets of tables, called zijes

(after the Arabic: zj), from the 8th to the 15th centuries and from all parts of

the medieval Islamic world. At that time only two of them had been published,

and it was clear that our understanding of scientific activity in the Middle Ages

would be greatly enhanced by detailed treatment of the others. Indeed, such

has proven to be the case, as we learn from the many studies that have followed

this pioneering essay.

Astronomers in Islamic Spain, al-Andalus, composed zijes and, beginning

in the 12th century, they were adapted and translated into Hebrew, Latin,

Castilian, and Catalan, the most famous examples being the Toledan Tables

(see Toomer (1968)) and the Alfonsine Tables. In Spain, as elsewhere in the

Islamic world, these zijes were largely based on the work of predecessors going

back to Ptolemy on the one hand, and Hindu astronomers on the other. More

often than not, a table comes with instructions for using it, rather than the

method used to construct it. For this reason, much scholarly energy has been

devoted to describing the methods underlying these tables, as well as their

lines of descent. By such analysis, guided by textual material, one can now

distinguish tables that are based on entirely new models, tables that are merely

copies of tables in previous zijes (at the two extremes), from tables based on

previous models but with new parameters and tables composed by modified

or new mathematical methods.

The work of Ibn al-Kammd, an Andalusian astronomer of the 12th century,

illustrates most of these characteristics. He composed 3 zijes, none of which

survives in the original Arabic, but a Latin manuscript contains a translation

of what appears to be one complete zij with references to the others. Ibn

1 The authors thank Professor J. Sams (Barcelona) for his valuable comments.

180 chapter 7

Hindu astronomers as they were known in zijes prior to his time, and his

influence was felt by later astronomers writing in Arabic, Latin, and Hebrew. As

is often the case, this work preserves parts of otherwise lost texts: in particular,

we gain valuable information on the solar theory of the Andalusian astronomer

Azarquiel (or Ibn al-Zarqlluh), who lived in the 11th century (see Toomer

(1969)). In this article our primary intention is to describe the astronomical

work of Ibn al-Kammd as it is preserved in Latin in ms Madrid 10023.

Ibn al-Kammd is cited or criticized by a number of his successors: Ibn

al-Him writing in Arabic ca. 1205 (see Sams (1992), pp. 321 ff.); Ab l-asan

cAl al-Marrkush (ca. 1262) (see Mills (1950), p. 347); Juan Gil (ca. 1350) whose

astronomical tables are preserved in a Hebrew version (see Goldstein (1985),

p. 237); Ibn al-adib, a 14th-century astronomer from Spain who went to Sicily

where he wrote astronomical works in Hebrew (see Goldstein (1985), p. 239);

Joseph Ibn Waqr, a Spanish astronomer of the mid-14th century who wrote in

Arabic and Hebrew (see Goldstein (1985), p. 237); and most importantly, as we

shall see, Ibn al-Kammd had a profound influence on the Tables of Barcelona

(see Mills (1962)).

In the discussions that follow new information of particular interest for the

history of astronomy in Spain is presented: see, e.g., the solar equation table

(Section ii); the preservation of material that goes back to the zij al-Mumtaan

of Yay ibn Ab Manr who lived in the 9th century (Section iii, c, f and j);

the table for trepidation (Section iv, b); and the tables for planetary latitude

(Section iv, e).

In sum, Ibn al-Kammd was a major player in medieval Spanish astronomy;

his achievements and the extent of his legacy have not yet been sufficiently

appreciated.

This is the name by which the astronomer Ab Jacfar Amad ben Ysuf Ibn

al-Kammd is known, although a variety of names has been associated with

him (Vernet (1949), pp. 7273). He was probably from Sevilla, but active in

Crdoba in the 12th century. In a remarkable study, Mills (1942), pp. 231247,

called attention to a 14th-century Latin manuscript at the Biblioteca Nacional

de Madrid, ms 10023, containing one of Ibn al-Kammds works; he described it

and gathered all the information then available on this astronomer. Not much

more is known now.

Ibn al-Kammd is the author of 3 zijes:

andalusian astronomy 181

1. al-Kawr al al-dawr (the periodic rotations [?]) (in 60 chapters). This trea-

tise has been partially preserved in Arabic in ms Escorial Ar. 939,4 (not seen

by us). There is also a short text in Castilian in Segovia, Biblioteca de la Cat-

edral, ms 115, ff. 218vb220vb, attributed to Yuaf Benacomed, and entitled

Libro sobre ircunferencia de moto.

2. al-Amad al al-abad (the eternally valid [tables])

3. al-Muqtabis (the compilation [of the two previous works]). ms Madrid 10023

contains the Latin translation of al-Muqtabis. In the explicit it is clearly

stated that the translation was done by John of Dumpno in 1260 in Palermo.

In Arabic only chapter 28 has survived: see ms Alger 1454,2, ff. 6263.

The dates for Ibn al-Kammd are uncertain. Mills (1950), p. 346, considered

the period towards the end of the 12th century, and suggested the year 1195

as that of his death, probably following Ahlwardt (1893), p. 219, where this date

is given with no specification of his source. More recently, Ibn al-Kammd has

been taken to be a direct disciple of Azarquiel. This claim is based on a note in

the margin of f. 30r of ms lat. 7281, a 15th-century manuscript at the Bibliothque

nationale de France, and it has been argued that this claim is supported by

the date ah480 (10871088ad) that appears in ms Madrid 10023, f. 65v: see

Section v, m, below. The marginal note, already transcribed in Mills (1950),

p. 14, reads: Post uenit Alcamet discipulus , referring to Azarquiel. According

to Mills, the same hand, or a similar one, has added: Similiter discipulus

Messalle. It does not seem at all warranted to deduce from this expression

that Ibn al-Kammd was a direct pupil of Azarquiel. Instead, we understand

this to mean only that Ibn al-Kammd was a follower of Azarquiels methods.

On the other hand, the date in ms Madrid 10023, f. 65v, is not the only one

mentioned in the last section of this manuscript (e.g., on f. 66r there is a table

for ah550: see Section v, r, below); as we shall see, the last section of this ms

contains a variety of tabular material not directly related to al-Muqtabis. Sams

(1992), p. 322, noted that Ibn al Him (fl. 1205) criticized Ibn al-Kammd. Since

the available evidence suggests that Ibn al-Kammd lived after Azarquiel and

before Ibn al-Him, we conclude that Ibn al-Kammd lived in the 12th century,

without offering any greater precision.

a al-Muqtabis

Text in Latin, presented in two columns.

Index (ff. 1va2rb). Transcribed in Mills (1942), pp. 234235.

182 chapter 7

Canons, 1 to 30, each of them called porta (ff. 2rb18vb). Three chapters

have been published so far: canon 1 is transcribed in Mills (1942), pp. 235

236, canon 28 in Vernet (1949), pp. 7478, and canon 30 in Mills (1942),

pp. 237238. Note that canon 30 explicitly mentions the other two works of

Ibn al-Kammd: al-Kawr al al-dawr and al-Amad al al-abad.

Explicit (f. 18vb). Transcribed in Mills (1942), p. 238.

b Other Texts

From f. 18vb to f. 24rb there is a set of chapters, in some disorder, that are dis-

tinct from those of al-Muqtabis. Some of them are associated with al-Kawr

al al-dawr, and were also translated by John of Dumpno in 1262 in Palermo;

their incipits and explicits were published in Mills (1942), pp. 238242. Mil-

ls also transcribed some of the texts therein, and Toomer (1969), pp. 323324,

transcribed and translated a text concerning the variation of solar eccentric-

ity.

Two sets of tables can easily be distinguished in the manuscript:

The tables are mentioned, or their use is explained, in the text in 30 canons

called al-Muqtabis. The tables are calculated for the meridian of Crdoba.

Folio 54v contains the last table of al-Muqtabis (a geographical table), as we

learn from canon 10 (f. 6va): tabulam longitudinum terrarum positam in

ultimo huius canonis. We will discuss all the tables in al-Muqtabis as follows:

the solar equation in Section ii, eclipse theory in Section iii, and the remaining

tables in Section iv.

These tables do not form a homogeneous set. They are not mentioned in the 30

canons of al-Muqtabis, and are presumably distinct from that work. Some of the

tables are related to al-Kawr al al-dawr, some are attributed to astronomers

other than Ibn al-Kammd, and still others are calculated for places other than

Crdoba. These tables will be discussed in Section v.

seductionis (f. 35r)

andalusian astronomy 183

Above the heading, the longitude of the solar apogee, presumably for the

Hijra, is given: Aux 2s 16;45,21. This table displays the solar equation in

degrees, minutes, and seconds as a function of mean solar anomaly. The max-

imum solar equation, which occurs at 92, is 1;52,44, thus differing from the

more common values: 1;59 (abash al-sib, Yay ibn Ab Manr), 1;59,10

(Toledan Tables, al-Battn), 2;14 (al-Khwrizm), 2;23 (Ptolemy).

The use of this table is explained in canon 13 (f. 7vb).

In his Tractatus super totam astrologiam, Bernardus de Virduno (ca. 1300)

attributes an eccentricity of 1;58 to Azarquiel, which yields a maximum equa-

tion of 1;52,42 (see Toomer (1987), pp. 515517 and Sams (1992), p. 213). How-

ever, this is not the only parameter used by Azarquiel for, in the Alfonsine trans-

lation of his treatise on the construction of the equatorium, 1;52,30 is explicitly

called a rounded parameter for the maximum solar equation (see Sams (1987),

p. 468). It is therefore likely that Ibn al-Kammd accepted a parameter from

Azarquiel.

The columns in Table 1, The Solar Equation in al-Muqtabis, are arranged as

follows:

(1) Mean solar anomaly, , for each integer degree from 1 to 180.

(2) Entries in the text (f. 35r): solar equation, c(), in degrees, minutes, and

seconds.

(3) Line-by-line differences in (2): c( + 1) c(). This column allows us to

recognize quite a number of errors in the entries for the solar equation

in the text, for these should increase monotonically from anomaly 0

to reach a maximum at anomaly 92, and then decrease monotonically

to anomaly 180. Consequently, the set of line-by-line differences should

follow a smooth decreasing pattern. In a great number of cases, pairs of

successive line-by-line differences in (3), whether erroneous or not, are

identical, thus strongly suggesting that the table for the solar equation

was originally computed for every other degree, and that interpolation

was used to derive the rest of the entries.

(4) Reconstructed line-by-line differences. Not all the line-by-line differences

which do not fit smoothly in (3) have been reconstructed; we have only

made suggestions in those cases where the tabulated values for the solar

anomaly can be easily derived from the reconstructed values. A question

mark indicates that, although there is a peculiar value in (3), we do not

offer any alternative value.

(5) Reconstructed values for the solar equation. These values result from an

easy explanation of the way the copying, or the calculational error was

made. There follows a list of various kinds of such errors:

184 chapter 7

easily confused in Arabic); c(28) = 0;50,27 instead of 0;51,27;

inversion of the order of the entries: in the text, the seconds for c(175)

and c(176) are apparently inverted;

incorrect interpolation between computed entries: c(30) and c(32)

were computed correctly, but c(31) seems to result from an incorrect

interpolation between them;

displacement of columns; from c(145) to c(150) the entries for the

minutes have been shifted one line downwards.

We have left unchanged those erroneous entries of the solar equation

for which we do not have an easy explanation; some of them may be due

to incorrect computation by the author of the table rather than to copyist

errors.

(6) Recomputed values: for the recomputation of the entries, we used a

simple eccentric model, and the following expression:

(7) Differences in seconds: T(ext)C(omp.), i.e., col. (2) col. (6); in those

cases where we have confidence in our reconstructed values, this column

displays the differences between the reconstructed values in col. (5), and

computation in col. (6).

2 3,47 1,54 3,49 2

3 5,40 1,53 5,43 3

4 7,34 1,54 7,37 3

5 9,25 1,51 1,56 9,30 9,31 1

6 11,25 2, 0 1,55 11,24 +1

7 13,18 1,53 13,18 0

8 15,11 1,53 15,12 1

9 17, 4 1,53 17, 5 1

10 0;18,57 0; 1,53 0;18,58 1

andalusian astronomy 185

13 24,39 1,57 ? 24,34 +5

14 26, 9 1,30 ? 26,26 15

15 28,13 2, 4 ? 28,16 3

16 30,17 2, 4 ? 30, 7 30, 7 0

17 32, 1 1,44 ? 31,57 4

18 33,45 1,44 ? 33,46 1

19 35,35 1,50 35,35 0

20 0;37,25 0; 1,50 0;37,24 +1

22 40,19 1, 7 1,47 40,59 40,58 +1

23 42,45 2,26 1,46 42,45 0

24 44,30 1,45 44,30 0

25 46, 1,46 46,16 46,15 +1

26 48,16 1,44 48, 0 47,59 1

27 49,44 1,28 1,44 49,43 +1

28 50,27 0,44 1,44 51,27 51,25 +2

29 53, 7 2,40 1,40 53, 7 0

30 0;54,48 0; 1,41 0;54,48 0

32 58, 7 1, 9 1,39 58, 7 0

33 0;59,45 1,38 0;59,44 +1

34 1; 1,23 1,38 1; 1,21 +2

35 2,38 1,15 1,35 2,58 2,57 +1

36 4,33 1,55 1,35 4,32 +1

37 6, 7 1,34 6, 6 +1

38 7,37 1,30 7,38 1

39 9, 9 1,32 9,10 1

40 1;10,41 0; 1,32 1;10,40 +1

42 13,39 1,29 13,37 +2

43 15, 7 1,28 15, 4 +3

44 16,35 1,28 16,29 +6

45 17,56 1,21 17,53 +3

186 chapter 7

47 20,33 1,16 1,21 20,38 20,37 +1

48 21,58 1,25 1,20 21,57 +1

49 23,45 1,47 1,17 23,15 23,15 0

50 1;24,32 0; 0,47 0; 1,17 1;24,33 1

52 27, 2 1,15 27, 3 1

53 28,19 1,17 1,13 28,15 28,15 0

54 29,27 1, 6 1,12 29,26 +1

55 30,37 1,10 30,36 +1

56 31,46 1, 9 31,44 +2

57 32,50 1, 4 32.51 1

58 33,54 1, 4 33,56 2

59 34, 7 0,13 1, 3 34,57 34,59 2

60 1;36, 0 0;1,53 0; 1, 3 1;36, 1 1

62 37,59 1, 1 ? 37,59 0

63 38,53 0,54 38,56 3

64 39,49 0,56 39,51 2

65 40,48 0,59 40,44 +4

66 41,37 0,49 41,35 +2

67 42,25 0,48 42,25 0

68 43,13 0,48 43,12 +1

69 43,59 0,47 43,58 +1

70 1;44,43 0; 0,44 1;44,43 0

72 46, 6 0,41 46, 5 +1

73 46,44 0,38 46,44 0

74 47,21 0,37 47,21 0

75 47,57 0,36 47,55 +2

76 48,33 0,36 48,28 +5

77 49, 1 0,28 48,59 +2

78 49,29 0,28 49,28 +1

andalusian astronomy 187

80 1;50,19 0; 0,25 1;50,20 1

82 51, 3 0,12 0,22 51, 4 1

83 51,22 0,19 51,23 1

84 51,41 0,19 51,40 +1

85 51,58 0,17 51,56 +2

86 52,15 0,17 0,12 52,10 52, 9 +1

87 52,14 0, 1 0, 9 52,19 52,20 1

88 52,19 0, 5 0,10 52,29 52,29 0

89 52,29 0,10 0, 6 52,35 52,35 0

90 1;52,40 0; 0,11 0; 0, 5 1;52,40 0

91 52,42 0, 2 52,43 1

92 52,44 0, 2 52,44 0

93 52,42 0, 2 52,43 1

94 52,39 0, 3 52,39 0

95 52,33 0, 6 52,34 1

96 52,27 0, 6 52,26 +1

97 52,16 0,11 52,17 1

98 52, 6 0,10 52, 6 0

99 51,50 0,16 51,52 2

100 1;51,34 0; 0,16 1;51,36 2

102 50,59 0,23 ? 50,58 +1

103 50,41 0,18 50,36 +5

104 50,22 0,19 50,12 50,12 0

105 49,48 0,34 0,26 49,46 +2

106 49,18 0,30 49,18 0

107 48,50 0,28 ? 48,48 +2

108 48,16 0,34 48,16 0

109 47,31 0,45 0,35 47,41 47,41 0

110 1;47, 5 0; 0,26 0,36 1;47, 5 0

188 chapter 7

113 45,33 1,12 0,42 45, 3 45, 4 1

114 44,20 1,13 0,43 44,20 0

115 43,12 1, 8 0,48 43,32 43,33 1

116 42,22 0,50 42,42 42,45 3

117 42,22 0, 0 0,50 41,52 41,55 3

118 41, 1 1,21 0,51 3 41, 3 2

119 40,56 0, 5 0,55 40, 6 40, 9 3

120 1;39,52 0; 1, 4 0; 0,54 1;39,12 1;39,13 1

122 37,53 0,59 37,13 37,15 2

123 36,51 1, 2 36,11 36,13 2

124 35,51 1, 0 35,11 35,10 +1

125 33,25 1,26 1, 6 34, 5 34, 5 0

126 32,58 0,27 1, 7 32,57 +1

127 31,50 1, 8 31,48 +2

128 30,39 1,11 30,38 +1

129 29,25 1,14 29,25 0

130 1;28,11 0; 1,14 1;28,11 0

132 26,40 0,45 1,15 25,40 25,37 +3

133 24,50 1,50 1,20 24,20 24,20 0

134 23, 0 1,50 1,20 22,57 +3

135 21,54 1, 6 1,26 21,34 21,35 1

136 20,11 1,43 1,23 20,10 +1

137 18,44 1,27 18,45 1

138 17,18 1,26 17,17 +1

139 15,49 1,29 15,49 0

140 1;14,20 1,29 1;14,18 +2

142 11,13 1,34 11,13 0

143 9,39 1,34 9,39 0

144 8, 6 1,33 8, 3 +3

andalusian astronomy 189

146 6,49 0,37 1,37 4,49 4,47 +2

147 5, 8 1,41 3, 8 3, 7 +1

148 3,27 1,41 1; 1,27 1; 1,26 +1

149 1; 1,10 2,17 1,37 0;59,50 0;59,43 +7

150 0;59,19 1,51 0;57,59 0;58, 0 1

152 54,31 2,50 1,50 54,29 +2

153 52,28 2, 3 ? 52,42 14

154 50,37 1,51 ? 50,54 17

155 48,47 1,50 ? 49, 6 19

156 47,17 1,30 ? 47,15 +2

157 44,55 2,22 1,52 45,25 45,24 +1

158 43,32 1,23 1,53 43,33 +1

159 41,39 1,53 41,40 1

160 0;39,46 1,53 0;39,46 0

162 35,57 1,54 35,57 0

163 34, 2 1,55 34, 1 +1

164 32, 6 1,56 32, 5 +1

165 30,14 1,52 ? 30, 8 +6

166 28,19 1,55 28,10 +9

167 26,16 2, 3 ? 26,11 +5

168 24,13 2, 3 ? 24,13 0

169 22,14 1,59 22,13 +1

170 0;20,15 1,59 0;20,14 +1

172 16,43 1,31 2, 1 16,13 16,13 0

173 14,12 2,31 2, 1 14,12 0

174 12,11 2, 1 12,11 0

175 10, 8 2, 3 2, 1 10,10 10,10 0

176 8,10 1,58 2, 2 8, 8 8, 8 0

177 6, 6 2, 4 2, 2 6, 6 0

178 4, 4 2, 2 4, 4 0

190 chapter 7

179 2, 2 2, 2 2, 2 0

180 0; 0, 0 2, 2 0; 0, 0 0

All the tables related to eclipse theory are mentioned in canons 25 (ff. 13rb14rb)

and 27 (ff. 15ra16va). Some specific terminology in Latin is used in canon 25:

longitudo for elongation, preuentio for opposition, peruenencia solis for

the fraction of the elongation that belongs to the Sun, precessio for the

hourly relative velocity of the luminaries.

Tabula diuersi motus solis in una hora quod est respectus (f. 51v)

Tabula diuersi motus lune in una hora quod est respectus (f. 51v)

The use of these two tables is explained in canon 25 (f. 13va). Both tables

coincide, except for copying errors, with those in al-Khwrizm (Suter (1914),

pp. 175180, tables 6166). The extremal values are:

The same two tables are also found in the tables of Juan Gil (London, Jews

College, ms Heb. 135, f. 91r), and the Tables of Barcelona (Mills (1962), table

43), as well as in a manuscript containing the Tables of Toulouse (Paris, BnF,

ms Lat. 16658, ff. 90v93r). The Toledan Tables (cf. Toomer (1968), p. 82) and

the Almanac of Azarquiel (Mills (1950), p. 174) have tables for solar and lunar

velocities that agree with those in al-Battn but differ from the present ones.

andalusian astronomy 191

For the solar velocity table, the formula used in recomputing the entries is:

[2] v = v + v .

of the mean anomaly, , taken from al-Khwrizms tables 2126, col. 2. The

recomputed values (which are not displayed here) are in good agreement with

the entries in the table. Equation [2] has been used by analogy with equation

[3], used for computing the lunar velocity table (cf. Goldstein (1992); Goldstein

et al.):

, taken from al-Khwrizms tables 2126, col. 3. The results are shown in

Table 2, together with a comparison between the entries in the tables of Ibn

al-Kammd and al-Khwrizm. The original table seems to be better preserved

by al-Khwrizm than by Ibn al-Kammd. lt should be noticed that previously

this table had not been recomputed successfully (cf. as-Saleh (1970), p. 162, and

Neugebauer (1962), p. 106).

(2) Entries in the text (f. 51v): lunar velocity in minutes and seconds of are per

hour.

(3) Variant readings in al-Khwrizms lunar velocity table (Suter (1914), tables

6166, col. 3); only the seconds are displayed here, except for 72.

(4) Recomputed values at multiples of 5 of anomaly; only the seconds are

displayed here, except for 135.

1 30;12 12 91 32;58

2 30;12 92 33; 0

3 30;12 93 33; 2

4 30;13 94 33; 5

5 30;13 13 95 33; 7 15

192 chapter 7

6 30;13 96 33;10

7 30;14 13 97 33;13

8 30;14 98 33;16

9 30;14 99 33;20

10 30;15 14 14 100 33;23 26

12 30;15 102 33;29

13 30;16 103 33;32

14 30;17 104 33;36

15 30;18 17 17 105 33;39 44

16 30,19 18 106 33;42

17 30;20 19 107 33;45

18 30;21 20 108 33;48

19 30;22 21 109 33;51

20 30;23 22 22 110 33;54 54

22 30;25 24 112 34; 0

23 30;26 25 113 34; 3

24 30;27 114 34; 5 6

25 30;28 26 115 34; 9 12

26 30;29 116 34;12

27 30;30 117 34;15

28 30;32 118 34;17

29 30;33 119 34;20

30 30;34 34 120 34;22 24

32 30;37 122 34;27

33 30;38 123 34;30

34 30;40 124 34;32

35 30;41 39 125 34;36 34 36

36 30;43 126 34;37

37 30;44 127 34;40

38 30;46 128 34;45 42

andalusian astronomy 193

40 30;49 47 130 34;48 49

42 30;52 132 34;53

43 30;54 133 34;55

44 30;56 134 34;58

45 30;58 56 135 35; 0 34,57

46 31; 0 136 35; 1

47 31; 2 137 35; 3

48 31; 4 138 35; 4

49 31; 6 139 35; 5

50 31; 8 7 140 35; 6 7

52 31;12 142 35; 9

53 31;14 143 35;10

54 31;16 144 35;12

55 31;21 18 17 145 35;13 17

56 31;23 21 146 35;14

57 31;25 23 147 35;16

58 31;27 25 148 35;17

59 31;29 27 149 35;18

60 31;29 29 150 35;20 22

62 31;35 152 35;22

63 31;37 153 35;24

64 31;40 154 35;25

65 31;42 42 155 35;26 49

66 31;45 156 35;27

67 31;47 157 35;28

68 31;49 158 35;29

69 31;51 52 159 35;30

70 31;54 55 160 35;31 31

194 chapter 7

73 32; 5 163 35;33

74 32; 8 164 35;34

75 32;12 12 165 35;35 36

76 32;17 16 166 35;35

77 32;23 20 167 35;36

78 32;27 24 168 35;37

79 32;27 169 35;37

80 32;31 32 170 35;38 38

82 32;37 172 35;38

83 32;39 173 35;39

84 32;42 174 35;39

85 32;46 45 55 175 35;39 39

86 32;47 176 35;39

87 32;49 177 35;39

88 32;52 178 35;40

89 32;54 179 35;40

90 32;56 57 180 35;40 40

preuentionis (f. 52r)

The use of this table is explained in canon 25 (f. 13va). It is a double argument

table: the vertical argument is the elongation (e), given in degrees and minutes,

from 0;30 to 12;0 at intervals of 0;30. The horizontal argument is the velocity

of the Moon relative to that of the Sun (vm vs), in minutes and seconds of arc

per hour, from 0;27,30 /h to 0;33,30 /h, at intervals of 0;0,30/h.

Each entry (t) can be computed from the following equation:

andalusian astronomy 195

where t is the time, given in hours and minutes, that the Moon takes to travel

the longitudinal arc between the Sun and the Moon at mean syzygy, i.e., the

time interval from mean to true syzygy. For a discussion of tables for finding

the time from mean to true syzygy, see Chabs and Goldstein.

The entries in the column for 0;32,0 /h are also copied, erroneously, in the

column for 0;32,30/h. A similar, but not quite identical, table is found in the

Tables of Barcelona (Mills (1962), table 42).

c Solar Eclipses

This table in 7 columns (see table 3) is mentioned in canon 27 (f. 15rb), and

gives the declination of midheaven, i.e., the intersection of the ecliptic and the

local meridian, in degrees and minutes, as a function of the longitude of the

ascendant. The manuscript uses septentrionalis for north, and meridionalis

for south, which we have transcribed by assigning positive or negative signs,

respectively, to the tabulated entries.

The declination reaches its maximum of 23;50 at Libra 1 and its minimum

of 23;51 at Aries 0. Note that 23;51 is the value used by Ptolemy in the

Handy Tables for the obliquity of the ecliptic. This table is closely related to

a table found in an Arabic manuscript (ms Escorial Ar. 927, ff. 9v, 12r, 12v: see

Kennedy (1986)), where it is called the Table of samt for determining solar

eclipses ( jadwal al-samt li-cilm kusf al-shams). Kennedy and Faris (pp. 2124)

have describedand presented a graphof this table, based on the copy

in the Escorial where it appears among tables attributed to Yay ibn Ab

Manr (9th century). The entries in this table in ms Madrid 10023, f. 52v, are

displayed in table 3 and, for comparison, those in the Arabic manuscript in the

Escorial are displayed in a separate table (see table 4), where the underlining

of entries indicates that they differ from those in table 3. Both tables are clearly

variants of the same archetype despite many discrepancies that can be ascribed

to copyist errors. The Arabic copy displays 180 additional entries because it

fails to recognize the symmetries (see Kennedy and Faris, p. 24). In addition

to the value adopted for the obliquity of the ecliptic, the table depends on

the latitude of the place for which it is intended. In this case, the latitude

used seems to be 35;55,48 (cf. ms Escorial Ar. 927, f. 8v), even though the

use of seconds for geographical latitude was not meaningful at the time; this

latitude corresponds to Yays native abaristan (cf. Kennedy and Faris, p. 24).

However, it is quite likely that Ibn al-Kammd, who is associated with Crdoba

(latitude = 38;30 as it appears in the geographical table on f. 54v), did not know

196 chapter 7

table 3 Table of samt for determining solar eclipses (ms Madrid 10023, f. 52v)

Degree of the

ascendant Lib Sco Sgr Cap Aqr Psc

2 28 23;50 19; 5 6;23 8;32 18;50 22;49

3 27 23;48 18;50 6; 0 9;27 19; 5 22;52

4 26 23;47 18;35 5;53 9;50 19;19 22;57

5 25 23;45 18;33 5;50 10;12 19;20 22;59

6 24 23;42 18; 2 4; 1 10;32 19;35 23; 0

7 23 23;39 17;46 3;37 10;53 19;50 23; 5

8 22 23;36 17;29 3;33 11;18 20; 3 23;12

9 21 23;32 17;12 2;50 11;59 20;16 23;17

10 20 23;23 16;55 2; 1 12; 1 20;33 23;18

11 19 23;16 16;36 1;36 12;22 20;35 23;23

12 18 23;12 16;18 0;48 12;48 20;39 23;24

13 17 23; 5 15;41 +0;24 13; 4 20;47 23;26

14 16 22;59 15; 3 0; 0 13;23 20;52 23;27

15 15 22;52 14;24 0;24 13;43 20;55 23;32

16 14 22;44 14; 5 0;48 14; 5 20;58 23;36

17 13 22;37 13;24 1;36 14;24 21; 0 23;37

18 12 22;23 13; 4 1; 1 14;44 21; 4 23;39

19 11 22;18 12;22 2;20 15; 3 21; 7 23;39

20 10 22;11 12;39 3;53 15;23 21;18 23;42

21 9 21;51 11;39 3;37 15;41 21;29 23;42

22 8 21;40 11;33 4; 1 16; 0 21;40 23;45

23 7 21;38 10;36 4;25 16;18 21;41 23;45

24 6 21;18 10;12 5;13 16;36 21;51 23;47

25 5 21; 7 9;52 5;36 16;44 21; 1 23;48

26 4 20;55 9;44 6; 0 16;56 22;10 23;48

27 3 20;42 8;42 6;24 17;12 22;10 23;50

28 2 20;29 8;19 7;10 17;27 22;19 23;50

29 1 20;16 7;23 7;18 17;46 22;28 23;51

30 0 +20; 3 +7;10 7;16 18; 2 22;35 23;51

andalusian astronomy 197

table 4 Table of samt for determining solar eclipses (ms Escorial Ar. 927, ff. 9v, 12r, and 12v)*

Degree of the

ascendant Lib Sco Sgr Cap Aqr Psc

2 28 19; 5 6;23 8;42 18;50 22;45

3 27 23;48 18;50 6; 0 9;27 19; 5 22;51

4 26 23;46 18;35 5;53 9;50 19;15 22;56

5 25 23;45 18;33 5; 7 10;12 19;25 23; 0

6 24 23;42 18; 2 5; 1 10;32 19;35 23; 4

7 23 23;39 17;56 4;37 10;56 19;50 23; 8

8 22 23;36 17;29 4;13 11;18 20; 3 23;11

9 21 23;32 17;12 3;50 11;39 20;18 23;14

10 20 23;23 16;55 3; 1 12; 1 20;33 23;17

11 19 23;16 16;36 2;36 12;22 20;35 23;20

12 18 23;12 16;18 1;48 12;48 20;39 23;23

13 17 23; 5 15;41 +0;24 13; 5 20;42 23;27

14 16 22;59 15; 3 0; 0 13;23 20;47 23;28

15 15 22;52 14;24 0;24 13;43 20;55 23;30

16 14 22;44 14; 5 0;48 14; 5 20;58 23;32

17 13 22;37 13;24 1;36 14;24 21; 0 23;34

18 12 22;27 13; 4 2; 1 14;44 21; 5 23;36

19 11 22;18 12;22 2;25 15; 3 21; 7 23;38

20 10 22;11 12;30 3;53 15;23 21;18 23;40

21 9 21;51 11;39 3;36 15;41 21;29 23;41

22 8 21;40 11;33 4; 1 16; 0 21;40 23;42

23 7 21;30 10;36 4;24 16;18 21;41 23;43

24 6 21;18 10;12 5;13 16;36 21;51 23;44

25 5 21; 7 9;50 5;36 16;44 21; 1 23;45

26 4 20;55 9; 5 6; 0 16;55 22;10 23;46

27 3 20;42 8;42 6;24 17;12 22;14 23;47

28 2 20;29 8;19 7;10 17;26 22;19 23;48

29 1 20;16 7;13 7;13 17;46 22;28 23;49

30 0 +20; 3 +7;10 7;16 18;10 22;32 23;50

198 chapter 7

* The columns for Libra, Scorpio, and Sagittarius are transcribed from f. 9v; the columns

for Capricorn and Aquarius from f. 12v; and the column for Pisces from f. 12r, for only

one entry appears on f. 12v while the rest of the column on that page is blank. In the

following list we display all entries on f. 12r, that differ from those of f. 12v. Variant

readings from f. 12r:

Cap 14 = 13;25 Cap 15 = 13;45 Cap 18 = 14;43 Aqu 2 = 18;47

Aqu 3 = 19; 1 Aqu 4 = 19;15 Aqu 5 = 19;26 Aqu 6 = 19;38

Aqu 8 = 20; 4 Aqu 9 = 20;18 Aqu 11 = 20;36 Aqu 13 = 20;42

Aqu 14 = 20;47 Aqu 15 = 20;52 Aqu 17 = 21; 5 Aqu 18 = 21;12

Aqu 19 = 21;19 Aqu 20 = 21;26 Aqu 21 = 21;33 Aqu 23 = 21;47

Aqu 24 = 21;54 Aqu 26 = 22; 7 Aqu 27 = 22;18 Aqu 29 = 22;25

the geographical latitude for which the table was computed. It is worth noting

that the table for solar declination (f. 35v) is based on a different value for the

obliquity of the ecliptic: 23;33.

Clearly, in the Arabic copy f. 12v has the better readings; the copyist appar-

ently realized that f. 12r had many errors, and so he tried again on f. 12v. The

table of the samt discussed here is also found in the Tables of Barcelona (Mills

(1962), table 47), and suffers from the same errors as the table of al-Muqtabis in

ms Madrid 10023.

A method for recomputing the entries of the table has been suggested by

Neugebauer (cf. Kennedy and Faris, p. 24), and it consists of 3 steps:

(i) for each integer value of the argument (the longitude of the ascendant)

find its oblique ascension by means of a table for the appropriate latitude

(Neugebauer suggested using a table for 36 since no such table is known

for 35;55,48);

(ii) in a table for normed right ascensions (see ff. 48v49r; cf. al-Khwrizms

table in Suter (1914), pp. 171173), find the longitude for which its normed

right ascension equals the value obtained previously;

(iii) the declination corresponding to that longitude in a table of declinations

is the entry sought.

andalusian astronomy 199

f. 52v f. 57v f. 52v f. 57v f. 52v f. 57v

12 0 55 0 54 0 6 0 0 6 0 0 0 0 0 0 0

11 1 35 1 35 0 14 0 0 54 5 0 0 0 0 0 0

10 3 50 3 50 0 50 40 1 50 40 0 0 0 0 0 0

9 4 16 4 56 1 6 50 2 6 50 0 0 0 0 0 0

8 6 47 6 47 1 43 9 2 2 9 0 0 0 0 0 0

7 8 19 8 39 1 57 50 2 36 50 0 0 0 0 0 0

6 10 29 10 29 2 0 9 3 0 20 0 0 0 0 0 0

5 12 0 11 0 2 0 50 3 0 50 0 0 48 0 0 40

4 12 0 12 0 2 0 0 3 0 0 1 2 40 0 2 40

3 12 0 12 0 2 20 0 4 20 0 1 6 48 1 6 48

2 12 0 12 0 2 25 0 4 40 0 1 8 48 1 8 43

1 12 0 12 0 2 30 0 4 50 0 1 10 12 1 10 42

d Lunar Eclipses

The use of this table (see table 5) is described in canon 27 (f. 15vb). Columns 24

display the lunar eclipse magnitude (digits), the duration of the eclipse (hours)

and the duration of totality (hours), as functions of the argument of latitude

of the Moon (degrees). The headings are prima porta, secunda porta, tercia

porta, respectively, while that for the argument is longitudo a capite et cauda.

Among the material appearing at the end of this manuscript after the tables

associated with al-Muqtabis, folio 57v has the same table under the title: Hec

tabula est quam extraxit et composuit Alkemed, in eclipsibus lunaris in canone

suo que est extracta a canone Ebi Iusufi cognoscitur Byn Tarach, que est ualde

uerax. This author is probably to be identified with the late 8th-century astron-

omer Yacqb ibn riq, a collaborator of al-Fazr at Baghdad, and whose zij

was called the Sindhind (cf. Pingree (1968b) and (1970)). Note that Ebi Iusufi, or

Ab Ysuf, means the father of Joseph, and, in Arabic nomenclature, this nick-

name can be substituted for Jacob, who was the father of the Biblical Joseph.

Previously, Mills (1942), p. 245, had suggested that the name here was a cor-

rupt form of the name of the 11th-century astronomer Muammad ben Ysuf

ben Amad ibn Mucdh, from Jan.

200 chapter 7

The two versions of this table are displayed in table 5. The entries in both

versions of this table seem to be quite corrupt, and do not allow us to derive

the parameters underlying them. Another version of this table is found in

the Tables of Barcelona (Mills (1962), table 50), and the entries in it are also

corrupt.

Note that there is a single table for lunar eclipses, which is quite uncom-

mon, for almost all zijes have two such tables (one for minimum, and one for

maximum, lunar distance). The astronomical work of Jacob ben David Bon-

jorn, an astronomer of the 14th century from Perpignan, has only one table for

lunar eclipses, but the entries in it are unrelated to those here (cf. Chabs (1991),

p. 309).

e Color of Eclipses

This table (see table 6) is arranged in 6 columns: column 1 displays the argu-

ment of lunar latitude; col. 2 gives the color of solar eclipses as a function of the

argument of lunar latitude, in degrees; col. 3 displays lunar latitude in minutes;

col. 4 gives the color of lunar eclipses in terms of lunar latitude; cols. 5 and 6 (not

shown here) display the magnitudes of solar eclipses in area digits as a function

of the magnitude of the eclipse in linear digits. Canon 27 (f. 16va) refers to the

first four columns, whereas the magnitudes of eclipses are treated on f. 15vb. All

six columns are also found in the Tables of Barcelona (Mills (1962): cols. 14

appear in table 51, and cols. 56 in table 48).

Kennedy (1956a), p. 159, has noted that Treatise viii of al-Brns Qnn

al-Masc d has a chapter on the colors of solar and lunar eclipses. On the colors

of lunar eclipses, see Goldstein (1967), pp. 234235. For Ibn al-Muthann (10th

century), the color changes during the eclipse, whereas for Ibn Ezra (Mills

(1947), p. 167) color is a function of latitude, as is the case here. Chapter 151

of Kitb alcAmal bil-Asturlb, by the Persian al-f (903986), contains a

similar list for the colors of lunar eclipses, but with different entries from those

presented here (see Kennedy and Destombes, p. 413).

Chapter 35 of the Libro de las Taulas Alfonsies, De qu color sera ell eclipsy,

concerns lunar eclipses, and gives two different rules for their colors. These

rules show similarities with those in the above-mentioned tables, but do not

fully agree with them.

andalusian astronomy 201

Arg. Solar Lat. Lunar

Lat. eclipse eclipse

2 niger clarus in nigredine

3 turbatus rubeus 20 niger

4 turbatus croceus cum rubedine

5 turbatus clarus 30 niger

6 turbatus cinereus cum rubedine

7 cinereus 40 niger

8 cinereus cum croceo

9 cinereus 50 turbatus

10 cinereus

11 croceus 60 cinereus

12 rubeus albus

The columns concerning the areas of eclipses (cols. 56) appear in a number

of earlier tables: the Toledan Tables (Toomer (1968), p. 113, table 76), the zij of

al-Battn (Nallino, ii, p. 89), the Almanac of Azarquiel (Mills (1950), p. 233),

and in al-Khwrizms zij (Suter (1914), p. 190, table 76, columns 67). In fact,

this table is already found in Ptolemys Almagest (vi, 8) and Handy Tables

(Stahlman (1959), p. 258). The entries in all these tables agree, except in the

Toledan Tables, where the entries for 3 and 5 (linear) digits are, respectively,

1;50 d and 3;20 d instead of 1;45 d and 3;40 d.

f Parallax in Latitude

Tabula latitudinis solis iudicate que est diuersitas respectus lune in lati-

tudine specialiter (f. 53r)

The use of this table (see table 7, below) is described in canon 27 (f. 15rb). The

table displays the adjusted parallax in latitude, p, in minutes and seconds of

arc, where adjusted parallax means the difference between the lunar and the

solar parallax (see Kennedy (1956b), p. 35). Its maximum value is 0;48,32 at 90.

The same table appears in al-Khwrizms zij (Suter (1914), pp. 191192, tables 77

and 77a, col. Diversitas respectus in latitudine), but the values tabulated there

202 chapter 7

1 0;50 46 34;30

2 1;41 47 35;56 34;56

3 2;32 2;33 48 35; 0

4 3;23 49 35;52

5 4;13 4;14 50 36; 8 37;11

6 5; 3 5; 4

7 5;55 51 36;44

8 6;45 52 37;21

9 7;37 53 38; 7

10 8;23 8;27 8;26 54 38;59 38;39

55 39;28 38;48 39;45

11 9;16 56 39;22 39;21

12 10; 5 57 39;36

13 10;54 58 39;50

14 11;44 59 40;56 40;16

15 12;33 12;34 60 40;44 42; 2

16 13,23

17 14;12 61 40;55

18 15; 1 62 41;43

19 15;51 63 41;54 41;53

20 16;41 16;36 64 42;54 42;34 43;37

65 42;54 43;59

21 17;32 17;38 66 43;57 43;17 44;20

22 18;22 18;26 18;11 67 43;59 43;39

23 18;12 19;14 18;58 68 44; 0

24 18;58 19;58 19;44 69 44;22

25 19;44 20;44 20;31 70 44;43 45;36

26 20;31 21;17

27 21;17 22; 2 71 45; 5

28 22; 3 22;47 72 45;27

29 22;49 23;32 73 45;48

30 23;35 24;16 74 46; 2 46;39

75 46;16 46;53

31 23;35 24;22 25; 0 76 46;29 46;39 47; 6

32 24;22 25; 4 25;43 77 46;42

33 25;46 26;26 78 46;55

andalusian astronomy 203

35 27;10 27;50 80 47;21 47;48

36 28;32 28;32

37 29;34 28;44 81 47;34

38 29;56 82 47;40

39 30;37 83 47; 0 48; 0

40 31;38 31;22 31;12 84 47; 5 48; 5

85 47; 9 48; 9

41 31;22 31;42 86 47;14 48;14

42 32;48 87 47;18 48;18

43 33;52 33;12 88 47;22 48;22

44 33;58 89 48;27

45 34; 4 34;19 90 48;32 48;34 48;32

differ in all cases, e.g., the entry for 90 is 0;48,45. This specific table was

discussed by Neugebauer (1962), pp. 121123.

Among the eclipse tables attributed to Yay ibn Ab Manr, and anal-

ysed by Kennedy and Faris (pp. 2038), there is a table entitled table for

the solar latitude ( jadwal ard al-shams) which coincides with this one (ms

Escorial Ar. 927, ff. 10v and 71v). Note the absurdity of the title for this table

which, in fact, deals with the latitudinal component of the adjusted parallax.

Kennedy and Faris (p. 25) give the function that underlies the entries of the

table:

where is the solar zenith distance. They also state that the author of the

calculations necessary for this table did an extraordinarily bad job, as the

results almost never agree with the recomputed ones. These irregularities,

which also occur in this table by Ibn al-Kammd, allow us to relate it with

confidence to that of Yay.

The same table is also found in the Tables of Barcelona (Mills (1962),

table 46), and it exhibits the same inconsistencies as our text. Comparison

between the entries in al-Muqtabis and the recomputed values seem to indicate

that some shifts of entries occurred in the copying process (entries for 2330

have been shifted one place downwards, entries for 3335 have also been

204 chapter 7

shifted one place, and entries around 67 may have been shifted two places

downwards). In any case, we are far from the smoothness of the analogous table

in al-Khwrizm where the maximum value is 0;48,45.

The columns in table 7 for the adjusted parallax in latitude are the following:

(1) Solar zenith distance for each integer degree from 1 to 90.

(2) Entries in the text: adjusted parallax in latitude (p), in minutes and

seconds of arc.

(3) Variant readings in the Tables of Barcelona (ms Ripoll 21).

(4) Recomputed values by means of equation [5].

g Lunar Latitude

This table is mentioned in canon 27 (f. 15rb), and it gives the lunar latitude ()

as a function of the argument of lunar latitude (). It is also found, with minor

variant readings, in al-Khwrizms zj (Suter (1914), pp. 132134, tables 2126; cf.

Neugebauer (1962), pp. 9598), where the maximum latitude is 4;30. Kennedy

and Ukashah, pp. 9596, have shown that the entries in this table were com-

puted according to the method of sines given by the formula:

Kennedy (1956a), p. 146, mentions that a similar table, with the same maximum

value, appears in the zij of Yay ibn Ab Manr.

The maximum value in the table on f. 53r is 4;29 instead of 4;30. This does

not suggest a different parameter; rather, it should be interpreted as a variant

reading of 4;30. The same table is also found in the Tables of Barcelona (Mills

(1962), table 44), and exhibits the same characteristics.

The table for the lunar latitude on f. 35v of this ms displays a different

maximum, 5;0, which is the value used by Ptolemy, al-Battn, Azarquiel,

among many others.

andalusian astronomy 205

h Elongation

Canon 27 (f. 13rb) explains the use of this table. Two sets of entries are tabulated,

the lunar longitude (lm) and the solar longitude (ls); both sets of entries are

functions of the elongation (e) between the Moon and the Sun, given in degrees

and minutes, from 0;30 to 12;0, at intervals of 0;30. The entries lm and ls are

such that e= lm ls, where lm = 13e/12 and ls = e/12.

A similar, but more extensive table, is found in the Tables of Barcelona

(Mills (1962), table 41): the elongation is given at intervals of 0;6 instead of

0;30, and it ranges from 0 to 13;12 instead of from 0 to 12;0.

i Parallax in Longitude

This table, mentioned in canon 25 (f. 14ra), gives the longitudinal component

of adjusted parallax (p), in hours and minutes, as a function of the argument,

given in time from 0;15 h to 9 h, at intervals of 0;15 h.

The entries reach a maximum of 1;36 h, and can be easily derived from the

column with the heading Horae diversorum/diversitatis respectuum lunae

[in longitudine] in al-Khwrizms zij (Suter (1914), pp. 191192, tables 77 and

77a; cf. Neugebauer (1962), pp. 121126). However, in al-Khwrizms zij, (i) the

argument is not given in time, but in degrees, and (ii) the parallax in longitude

is given to seconds. Kennedy (1956b), pp. 4950, has shown that al-Khwrizms

table can be computed by means of the following formula:

where

such that , the argument (in degrees), meets the condition that 0 150.

The coefficient 1;36 = 0;4 24 = 24/15 contains the standard Hindu value for the

obliquity of the ecliptic, 24. In the case of Ibn al-Kammds table, the entries

can be recomputed by means of equation [7], where

206 chapter 7

for all 0 135; now, 135 = 9 h 15/h, and 9 h is the maximum value of the

argument, expressed in time.

Only two such adjusted longitudinal parallax tables of this kind are known:

that of al-Khwrizm and the one discovered by Kennedy (1956b, p. 48) in the

zij of Ibn al-Shir (ca. 1350), explicitly paraphrasing an early Islamic source

that has not been identified (Kennedy and Faris, pp. 3338). We can now add

to that short list the parallax table of Ibn al-Kammd and that in the Tables of

Barcelona (Mills (1962), table 45).

Column 2 of this table gives the magnitude of the eclipse, in (linear) digits and

minutes, as a function of the adjusted latitude (at conjunction) of the Moon

displayed in column 1, in minutes and seconds of arc, from 0;34,13 to 0. The

explanation of this table in canon 27 (f. 15va) confirms the above value for the

eclipse limit: si fuerit minus 34 minutis et 13 secundis erit eclipsis.

Columns 39 form a double argument table. The vertical argument is the

eclipse magnitude in digits (col. 2); the horizontal argument is the relative

velocity of the Moon with respect to the Sun (vm vs) in minutes and seconds

of arc per hour, from 0;27,30 /h to 0;33,30/h, at intervals of 0;1 /h.

This table, as some previous ones, seems to derive from Yay ibn Ab

Mansr (ms Escorial Ar. 927, f. 13r). Kennedy and Faris (pp. 2730), once again,

have transcribed and explained this table.

The adjusted latitude of the Moon is a linear function of the magnitude of

the eclipse, so that the graph of the function relating columns 1 and 2 should be

a straight line. However, this is not the case: there are jumps at 4;20 and 9;20

digits; and one should consider the first three entries in col. 2 and 0 d, 20 d, and

40 d, instead of 0;15 d, 0;30 d, and 0;45 d (these errors appear in this table in the

Arabic ms as well as in our Latin text).

The second part of the table displays the half-duration of the eclipse, in hours

and minutes, as a function of its magnitude and the relative velocity of the

Moon with respect to the Sun. The recomputations made by Kennedy and Faris

(p. 29) give good results, but fail to reproduce the entries of the table precisely.

andalusian astronomy 207

The purpose of these tables is to convert dates from the Arabic calendar to

Roman (i.e., Julian) and Egyptian calendars. The epoch of the radix given

is the Hijra: noon of July 14 ad 622. Canon 9 (f. 6ra) states that the epoch

of the radix is in the day of Mercury (Wednesday). The tables display two

correspondences:

ah0 = 932 Julian years 9 months 17;0 days from the beginning of the Seleucid

era = 9 Egyptian years 11 months 9 days from the beginning of the Yazdijird era.

These tables entirely or partially reproduce calendaric tables in the zij of

al-Khwrizm/Maslama (Suter (1914), p. 110, table 2; p. 111, table 2a; and p. 113,

table 3), and/or in pseudo-Battn (Maslama) (Nallino, ii, pp. 301, 304305).

The same radix and equivalent entries for the mean motion of the vernal

point are found in Azarquiels Treatise on the motion of the fixed stars (Paris,

BnF, ms Heb. 1036; see Mills (1950), pp. 266, 324); the Liber de motu octave

sphere, attributed to Thbit ibn Qurra displays a similar table (Mills (1950),

p. 507); cf. Morelon (1987), p. xix. On the theory of trepidation see Goldstein

(1964), Dobrzycki (1965), North (1967), North (1976), vol. 3, pp. 155158, Mercier

(19761977) and Sams (1992).

The entries (see table 8) display very nearly a sine function whose maximum

is 9;59 at 90; we have not succeeded in explaining the deviations from the

sine function, e.g., the entry for 30 is not half the entry for 90. Toomer (1968),

118, gives two tables, which are also sine functions, associated with the Toledan

Tables, and in fact they already appear in the Liber de motu octave sphere (Mills

(1950), pp. 507508). The radix and the mean motion of the first point in Aries

(see f. 28v and the comments on it, above) are taken from Azarquiels Treatise on

the motion of the fixed stars. The use of this table is briefly outlined in canon 12

(f. 7v), where another work by Ibn al-Kammd is explicitly mentioned: al-Amad

al al-abad. This table is also found in the Tables of Barcelona (Mills (1962),

table 20), a treatise by al-Marrkush (Sdillot (1834), p. 131), and in the tables

of Juan Gil (London, Jews College, ms Heb. 135, f. 78b).

208 chapter 7

We have tried to recompute this table in many ways, and by far the best fit

comes from formula [10] which is intended to represent Azarquiels somewhat

vaguely defined second model (Mills (1950), pp. 287289, 317318; cf. Sams

(1992), p. 230):

where e is the entry in the table, r = 10;24, and i is the argument. The maximum

entry in the table, 9;59, is indeed a rounded value for arcsin (10;24/60) =

9;58,54.

point in Aries

2 28 0;20 5;25 8;56

3 27 0;31 5;34 9; 1

4 26 0;41 5;43 9; 5

5 25 0;53 5;52 9;10

6 24 1; 3 6; 4 9;14

7 23 1;54 6;16 9;17

8 22 1;25 6;29 9;21

9 21 1;35 6;41 9;24

10 20 1;45 7;53 9;28

11 19 1;56 7;57 9;31

12 18 2; 8 7; 2 9;35

13 17 2;19 7; 6 9;38

14 16 2;30 7;10 9;41

15 15 2;41 7;14 9;45

16 14 2;50 7;21 9;47

17 13 2;18 7;28 9;49

18 12 3; 6 7;35 9;51

19 11 3;15 7;42 9;52

20 10 3;25 7;48 9;55

21 9 3;36 7;54 9;56

22 8 3;47 8; 0 9;56

andalusian astronomy 209

23 7 3;57 8; 7 9;57

24 6 4; 8 8;13 9;58

25 5 4;19 8;20 9;59

26 4 4;29 8;25 9;59

27 3 4;39 8;31 9;59

28 2 4;49 8;37 9;59

29 1 4;38 8;41 9;59

30 0 5; 7 8;47 9;59

5/11 4/10 3/9

Ripoll 21, f. 137r):

e (7) = 1;14 e(16) = 2;51 e(17) = 2;58 e(29) = 4;58

e(40) = 6;43 e(41) = 6;57 e(50) = 7;54 e(51) = 7;58

e(53) = 8; 6 e(55) = 8;19

Ibn al-Kammds table for

trepidation, as reconstructed

ure 7.1: a small circle or epicycle, bce, whose center is a, lies in the plane of the

ecliptic, circle ag, and the center of the sphere is o. Note that the small circle

bce is partly inside, and partly outside, the sphere. The angle i is equal to arc

bc; radius ab = 10;24, and oa = 60. Through c, draw a line parallel to ba, reach-

ing the large circle ag at f. Draw of, and let aof be the angle e, which we seek.

To compute angle e, we drop a perpendicular fh from f to ao; then fh = cd,

210 chapter 7

al-Kammd

10 1;45 1;43 2

20 3;25 3;24 1

30 5; 7 4;58 9

35 5;52 5;42 10

40 6;43 * 6;24 19

45 7;14 7; 2 12

50 7;48 7;38 10

60 8;47 8;38 9

70 9;28 9;22 6

80 9;55 9;50 5

90 9;59 9;59 0

entry (see Table 8); hence we display

surrounding values.

and cd = r sin (i). In right triangle fho, fo = 60, and sin (e) = fh/fo; equa-

tion [10] follows. We cannot account for the remaining differences between

text and computation. The columns in table 9 for trepidation according to Ibn

al-Kammd are as follows:

(1) i: argument

(2) e: text of Ibn al-Kammd

(3) e: computed, using equation [10]

(4) the difference: T(ext)C(omp.), in minutes

The value 10;24 for r in equation [10] is made plausible by the two other

models for trepidation associated with Azarquiel and his followers. Sams

(1992), pp. 235236, describes the third model of Azarquiel and shows how

it yields a maximum value, pmax very close to 10;24 (10;23,29), based on the

equations:

andalusian astronomy 211

al-Bann and Ibn al-Raqqm

20 3;32 3;32;23 3;32,24

30 5;11 5;10,43 5;10,40

40 6;40 6;39,48 6;39,40

50 7;57 7;56,55 7;56,42

60 8;59 8;59,39 8;59,20

70 9;46 9;45,59 9;45,35

80 10;14 10;14,25 10;13,56

90 10;24 10;24 10;23,29

and 23;33 is the value for the obliquity. Moreover, Neugebauer (1962), p. 184,

presents a formula based on Azarquiels first model:

If we substitute 10;24 for Pmax in equation [13], we find a set of values that

agree very well with those preserved by the late 13th-century astronomers Ibn

al-Bann (ms Escorial Ar. 909, f. 22v) and Ibn al-Raqqm (ms Kandilli 249, f. 66v)

who only tabulated these entries to degrees and minutes. However, since the

agreement is equally good using equations [11] and [12] on the one hand, and

equation [13] on the other, we cannot decide which procedure was used for

computing this table. Table 10 displays these computations, where the columns

are the following:

(1) i: argument

(2) e: text of Ibn al-Bann and Ibn al-Raqqm

(3) e: computed, using equation [13]

(4) e: computed, using equations [11] and [12]

The headings of the tables indicate that they were intended for the meridian

of Crdoba and calculated for Arabic years, months, etc. We have computed all

mean motions from the corresponding tabulated values for ah 720, except for

212 chapter 7

that of the lunar anomaly. The method is simply to subtract the corresponding

radix from each value for ah720, taking into account the full rotations, and then

to divide the result by the number of days elapsed from the epoch of the Hijra

calendar.

The tabulated values for the mean motions are given to seconds. They do

not agree with either al-Khwrizm (Suter (1914), p. 115, table 4), or al-Battn

(Nallino, ii, p. 20). The entry for the radix is 1s 6;35,9, and the tabulated

value for ah720 is 7s 16;13,19. The difference between these two values is 6s

9;38,10 = 189;38,10. Now, in 720 Arabic years, the Sun has completed 698 full

rotations; therefore, the are length traveled by the Sun in 720 Ar. y. (= 255,144

days) is 189;38,10 + (698 360) degrees, and the daily mean motion of the

Sun resulting is 0;59,8,9,21,15, /d, which yields a year-length of 365;15,36,34,

d. The value deduced from the table is sidereal. It differs from the daily

mean motion embedded in the Toledan Tables, and attributed to Azarquiel

(0;59,8,11,28,27, /d) by an amount which is exactly equal to Azarquiels value

for the daily motion of the solar apogee (0;0,0,2,7,10,39, /d, cf. Toomer (1969),

p. 319).

The radix given here (1s 6;35,9) corresponds to the solar centrum (the

distance from a sidereally fixed apogee). To obtain the longitude of the Sun at

epoch, add the given radix to the longitude of the apogee (2s 16;45,21, f. 35r);

the result is 3s 23;20,30, a value which is close to, but not identical with, those

of Azarquiel or al-Khwrizm/Maslama (see Toomer (1968), p. 44).

The daily mean motion of the solar apogee resulting from the tabulated value

for ah720 (0s 2;24,24) is 0;0,0,2,2,14,46, /d. This implies a progress of 1o in

about 299 Arabic years or in about 290 Julian years, a value which differs from

the daily motion of the apogee used by Azarquiel (0;0,0,2,7,10,39, /d), which

corresponds to a progress of 1 in about 279 Julian years. A parameter very

similar to that of Azarquiel is also found in the works of Ibn Isq, Ibn al-Bann

and Ab l-asan cAl b. Ab cAl al-Qusanayn (Mills (1950), pp. 352353;

Sams (1992), p. 212).

The recomputations show that Ibn al-Kammd used Azarquiels value for

the mean motion of the Sun, but that he incorporated a different parameter for

the mean motion of the apogee which does not appear in any known text prior

to this one.

andalusian astronomy 213

The daily mean motion of the Moon in longitude resulting from the tabu-

lated value for 720ah (10s 8;37,1) is 13;10,34,52,46, /d. This is exactly al-

Khwrizms value (cf. Neugebauer (1962), pp. 42, 92), and very nearly that in

the Toledan Tables (cf. Toomer (1968), p. 44).

The daily mean motion of the Moon in anomaly resulting from the tabulated

value for 900 a.r. (9s 13;12) is 13;3,53,56,19, /d. This is very nearly the value

in the Toledan Tables (cf. Toomer (1968), p. 44), which differs from that of

al-Khwrizm (cf. Neugebauer (1962), p. 92).

The tabulated values for the mean motion in longitude agree almost exactly

with those in al-Khwrizm (Suter (1914), pp. 117119, table 68), and exhibit very

small differences with those in the Toledan Tables (Toomer (1968), p. 48). How-

ever, the tabulated values for the mean motion in anomaly are in agreement

with those of al-Battn (Nallino, ii, p. 20), and in the Toledan Tables (Toomer

(1968), p. 49), except for the fact that the motion in anomaly is given to seconds

there.

For Saturn, Jupiter, Mars and Mercury, the tables give entries for the mean

motion of the centrum (i.e., the distance from a fixed apogee) and anomaly.

For Venus, the entries display the mean motion in anomaly only, and for the

lunar node, the entries are the complement in 360 of the mean motion of

the ascending node in longitude. All entries are given to minutes. Those for

the yearly progress of the centrum of Mercury agree with those for the Sun

(f. 28r), except for the fact that the latter are given to seconds. The values for

the radices are for the Hijra (see table 11, below), and differ from those in the zij

of al-Khwrizm/Maslama and those in the Toledan Tables.

The mean motions in anomaly of the superior planets are not generally tab-

ulated in other sets of astronomical tables (and this is also the case for almost

all copies of the Toledan Tables). The recomputed values obtained here can

be compared, however, with those in one copy of the Toledan Tables (Oxford,

Bodleian Library, ms Laud. Misc. 644, cf. Toomer (1968), p. 45): 0;57,8/d (Sat-

urn), 0;54,9/d (Jupiter) and 0;27,41/d (Mars). The same values are explicitly

found in the work of abash (Debarnot (1987), p. 44), where the value for Mars

is 0;27,42/d.

All parameters computed from the tabulated entries for the mean motions

show close, although not perfect, agreement with those derived from the Tole-

dan Tables, the differences never being greater than 0;0,0,1.

214 chapter 7

Solar apogee 0;0,0,2,2,14,46, 2s 16;45,21 **

Vernal point 0;0,0,54,56,57, 0s 3;51,11

Lunar longitude 13;10,34,52,46 4s 0;34,42 ***

Lunar anomaly 13;3,53,56,19, 3s 18;11

Double elongation 0s 14;33

Lunar node 0;3,10,46,41, 4s 6;30

Saturn (longitude) 0;2,0,25,36, 7s 26;52 *

Saturn (anomaly) 0;57, 7,44,57, 11s 27;48

Jupiter (longitude) 0;4,59,6,43, 5s 21;58 *

Jupiter (anomaly) 0;54,9,3,37, 4s 23; 2

Mars (longitude) 0;31,26,31,40, 3s 1;46 *

Mars (anomaly) 0;27,41,40,34, 8s 22;14

Venus (anomaly) 0;36,59,29,21, 1s 15;21

Mercury (longitude) 0;59,8, 11,23, 9s 4;58 *

Mercury (anomaly) 3;6,24, 7, 19, 2s 14; 1

* The values of the radices for the Sun and the planets correspond to

their centrum, i.e., their distance from apogee. Note that the ms does

not provide this information for Venus.

** This value for the longitude of the solar apogee is given on f. 35r.

However, in the table for the mean motion of the solar apogee (f. 28v)

one finds 0s 0;0,0 opposite Radix.

*** The radix for the Moon is called its centrum, here meaning

longitude.

d Spherical Astronomy

This is a table giving the declination of the Sun for each integer degree; the max-

imum entry is 23;33. This value for the obliquity of the ecliptic is associated

with the zij al-Mumtaan (Vernet (1956), p. 515). The entries in this table agree,

except for minor differences, with those in the Almanac of Azarquiel (Mills

(1950), p. 174), where the argument is given only at intervals of 3. The value

found in the Toledan Tables, and attributed to Azarquiel, is 23;33,30 (Toomer

(1968), p. 30).

andalusian astronomy 215

The heading of the table mentions Crdoba, but it does not specify a value

for its latitude. The geographical table on f. 54v gives its latitude, , as 38;30,

the most prevalent value for it at the time. This table gives the half-length of

daylight as a function of the solar longitude. The maximum entry represents

half of the longest daylight (m/2), and it is 7;21 h for Cancer 0. This value follows

from the formula:

The same table of right ascension, beginning with Capricorn 0, is found in al-

Battn (Nallino, ii, pp. 6164), for an obliquity () of 23;35. It is also found

in the Toledan Tables (Toomer (1968), p. 34, table 17) and, with copying errors

in an abridged version, in the Almanac of Azarquiel (Mills (1950), pp. 220

221, tables 6970). The table differs from that in the Handy Tables (Stahlman

(1959), pp. 206209, table 1) and in al-Khwrizms zij (Suter (1914), pp. 171173,

tables 5959b), both calculated for higher values of the obliquity.

The mean values between the rising times of Aries and Virgo, of Taurus and

Leo, and of Gemini and Cancer, are respectively 27;50, 29;54,30 and 32;15,30.

They are almost identical with those derived from the Toledan Tables for the

seven climates: 27;50, 29;54 and 32;16, which are the Ptolemaic values for the

right ascensions (cf. Toomer (1968), p. 42). They are also embedded in the zij of

al-Khwrizm, but differ from those in the zij of al-Battn.

When recomputing the entries, close, although not exact, agreement is

obtained with = 38;30 and = 23;51. With the two other values of the

obliquity found in our text (23;33 and 23;35), the agreement is worse.

Another column in this table represents the length of the seasonal hours.

The entry for Cancer 0 is 18;24. Now 18;24 12/15 = 14;43 h, which gives a half

daylight of 7;21,30 h, and this is quite close to the value 7;21 h for Cancer 0 (see

the table for the half-length of daylight as a function of the solar longitude on

f. 47v).

A similar table for Sal, a place in North Africa near Rabat, whose latitude is

given here as 33, is found on ff. 59v61r.

216 chapter 7

e Latitude

Moon (f. 35v)

The maximum value in the table for the latitude of the Moon is 5;0. The same

value for the inclination of the lunar orb is found in the Almagest v, 8, and in

many other texts, including the Toledan Tables and the Almanac of Azarquiel

(cf. Mills (1950), p. 173, where the argument is given only at intervals of 6). For

another table for the lunar latitude in al-Muqtabis, see Section iii. g.

This table for the latitude of the superior planets is the same as the one in the

Almagest xiii, 5, and in al-Battn (Nallino, ii, p. 140 (columns 13) and p. 141

(column 4)). The pattern of this table differs greatly from the corresponding

one in the Handy Tables. Toomer listed some mss associated with the Toledan

Tables that contain such a table, but concluded that it is not part of the original

Toledan Tables (Toomer (1968), p. 72).

In contrast to the superior planets, the latitude table for the inferior planets

does not conform to the pattern of the Almagest, the zij of al-Battn, or the

tables associated with the Toledan Tables. Rather, this table reproduces, with

variant readings, the entries which are multiples of 6 in the Handy Tables

(Stahlman (1959), pp. 331334, tables 4950, where the entries are given at

3-degree intervals). In particular, the maximum values for the mean latitude

of Mercury (3;52) agree in both sets of tables, but those for the mean latitude

of Venus differ (8;35 in our text and 8;51 in the Handy Tables). However,

canon 16 (f. 10va) gives 8;36 and 4; 18 as the values for the maximum latitude

of Venus and Mercury. Kennedy (1956a), p. 173, reports maximum values for

Venus (8;56) and Mercury (4;18), and associates the zij al-Mumtaan and Ibn

Hibint with them.

Nevertheless, the outstanding feature here is the juxtaposition of different

Ptolemaic tabular material: the Almagest for the superior planets, and the

Handy Tables for the inferior planets. The source for such a mixed approach

has not been determined.

Note that in the tables associated with al-Muqtabis no values are given for

the longitudes of the planetary nodes, although ms 10023 (f. 66r) has a list of

them.

f Equations

Moon (ff. 36r37r)

The same table is found in the Almagest v, 8, in the Handy Tables, as well

as in many medieval tables, such as the zij of Yay ben Ab Manr (Salam

and Kennedy, pp. 495496), the zij of al-Battn (Nallino, ii, pp. 7883) and

andalusian astronomy 217

the Toledan Tables (Toomer (1968), pp. 5859). The table lists columns for the

equation for mean to true apogee, an interpolation function, the increment in

the equation of center, and the equation of center. There is no column here for

the lunar latitude, which is tabulated separately (see f. 35v). The entries for the

equation of anomaly agree with those in al-Battns zij, but differ slightly from

those in the Toledan Tables (e.g., in our table the maximum of 5; 1,0 is reached

at 95, while in the Toledan Tables the value is 5;0,59, and it occurs at 9495).

On the other hand, in our table the order of the columns is the same as in the

Toledan Tables, and differs from that in al-Battns zij.

The tables for the equations of the five planets are essentially those found in the

Almagest xi, 11, in the Handy Tables, as well as in the zij of al-Battn (Nallino, ii,

pp. 108137) and the Toledan Tables (Toomer (1968), pp. 6068), among many

others. Although in most respects the zij of al-Battn and the Toledan Tables

agree for the planetary equations, there are some differences between them:

for instance, there is a column for the planetary stations in the Toledan Tables,

which is neither in al-Battns zij nor in our table. All we can deduce from these

tables is that Ibn al-Kammd accepted the Ptolemaic tradition, as displayed in

the Handy Tables, followed by most Muslim astronomers.

It is worth noting that in the tables of Ibn al-Kammd the maximum solar

equation (1;52,44: f. 35r) differs from the maximum equation of center for

Venus (1;59). This value for Venus is not that of the Almagest, but follows

al-Battn, etc. (cf. Goldstein and Sawyer). This indicates that Ibn al-Kammds

contribution was restricted to solar theory and that he did not introduce any

changes in planetary theory.

Values for the sidereally fixed apogees appear above the tables for the equa-

tions on ff. 37v38v (Saturn), 39r40r (Jupiter), 40v41v (Mars), 42r43r (Venus),

and 43v44v (Mercury). The apogee for Venus is that ascribed to the Sun on

f. 35r.

Saturn 238;38,30

Jupiter 158;21, 0

Mars 119;41, 0

Venus 76;45,21

Mercury 198;21, 0

g Stations ( f. 46r)

Only four values are given for each planet: the positions of the first and the sec-

ond stationary points for arguments of 0 and 180 (see table 12). The tabulated

218 chapter 7

2nd st. at apogee 8s 7;16 7s25;55 6s22;32 6s14; 9 7s 2;46

1st st. at perigee 3s25;30 4s 7;11 5s19;15 5s18;21 4s24;42

2ndt st at perigee 8s 4;30 7s22;49 6s10;45 6s11;39 7s 5;18

values for the same argument add up correctly to 360. In all cases, they coin-

cide with those found in the Toledan Tables (Toomer (1968), pp. 6068), and the

zij of al-Khwrizm (Suter (1914), pp. 138167, tables 2756), both zijes display-

ing tables for each integer degree of the argument. Nearly the same values are

also found in the zij of al-Battn (Nallino, ii, pp. 138139), which have a Ptole-

maic origin. However, Almagest xii.8 and the Handy Tables (Stahlman (1959),

pp. 335339, tables 5155) have slightly different tables for the stations: in the

latter the argument was modified, as well as the interval for the calculation of

the entries (6 in the Almagest, 3 in the Handy Tables).

Beneath the table we read: Mediatus solis in radice posita ad directionem

dierum cum noctibus: 10.23.24.50. a puncto capitis arietis. This value agrees

with that appearing in canon 11 (f. 7rb), and seems to correspond to the argu-

ment (in signs and degrees) for the minimum entry in the table. Hence it is to be

understood as 10s 23;24,50. The entries in this table, in time-degrees, rounded

to the nearest integer, may have been taken from the more precise values given

in the zij of al-Battn (Nallino, ii, pp. 6164) or in the Toledan Tables (Toomer

(1968), pp. 3435).

i Trigonometry

Functions Related to the Sine (f. 46v)

We adopt the convention that Sin = 60 sin , and similarly other capitalized

trigonometric functions are normed for r = 60 (cf. Kennedy (1956a), p. 139).

Three functions are given for each integer degree: Sin , Cos , and Vers =

r Cos . The Almanac of Azarquiel (Mills (1950), p. 229) has a table with

three functions (sine, cosine, and versine), but given at intervals of 3. Except

for copying errors, the entries agree in both tables.

andalusian astronomy 219

The entries in this table represent the length of a shadow (s) projected by a

gnomon of 12 units as a function of the altitude of the Sun (h):

in the tradition of al-Khwrizm (Suter (1914), p. 174, table 60) and al-Battn

(Nallino, ii, p. 60).

This is a list of 30 stars and it displays the following information for each

star: magnitude, name, ecliptic coordinates (longitude and latitude), and the

planets associated with it (for astrological purposes).

Kunitzsch (1966), pp. 99102, described this list under his type xv: the eclip-

tic longitude of each star is derived from that in Ptolemys star catalogue by

adding 6;38, thus indicating that the epoch of this star list is the Hijra. The

type defined by Kunitzsch only includes two versions of this list: the other ver-

sion is uniquely represented by ms Vienna 5311, f. 129v. Although the stars in

both versions are the same, the coordinates for them do not always agree, thus

suggesting a common Arabic ancestor. Kunitzsch also reports close similarity

with the star list of Ab l-asan cAl al-Marrkush (ca. 1262). We have found

additional copies of this list in some copies of the Tables of Barcelona, e.g., ms

Vatican Heb. 356, f. 65b, where the names of stars are given in Hebrew.

k Excess of Revolution

In the first table on f. 54v, the excess of revolution is given in degrees and

minutes, for 1, 2, 3, , 10, 20, , 100 years. The entry for 1 year is 92;36, which

seems to be an isolated error. To be coherent with all other entries in the table,

one should read 93;36, leading to a year-length of 365;15,36 days = 365d 6;14,24

h. The entry for 100 years is 0;5, and the resulting length of the solar year is

365;15,36,0,30 days.

In the second table on f. 54v, the excess of revolution is given in hours and

minutes of time, for 1, 2, 3, , 10, 20, , 100 years. The entry for 1 year is 6;14 h,

which corroborates the emendation above. The entry for 100 years is 23;58 h,

and the resulting length of the solar year is 365;15,35,58,30 days, a value close,

but not equal, to the parameter derived from the previous table.

These values for the solar year are to be compared with that appearing in

the canons (f. 2va): 365;15,36,19,34,12 days (cf. Mills (1942), p. 236), as well as

with the information given on f. 65v of this manuscript: Length of the solar

year according to Ibn al-Kammd:

220 chapter 7

365;15,36,19,35,32 days.

The daily mean motion of the Sun resulting from the first value is 0;59,8,9,23,44,

53 /d (the second value yields a very similar parameter: 0;59,8,9,23,44,40 /d),

which agrees quite well with the value derived above from the table for the

mean solar anomaly on f. 28r.

On f. 57v of this manuscript there is an analogous table attributed to Azar-

quiel, and for a sidereal year of 365;15,24d.

There follows a list of the various lengths of the solar year associated with

Ibn al-Kammd:

365;15,35,58,30d (computed from the second table: f. 54v)

365;15,36,19,34,12d (mentioned in canon 1: f. 2va)

365;15,36,19,35,32d (attributed to Ibn al-Kammd: f. 65v)

This is a list of 30 places: for each of them we are given its longitude and

latitude, in degrees and minutes. The prime meridian used here is located west

of the shore of the Western Ocean. It thus differs from that in the Toledan

Tables, where the shore of the Western Ocean seems to have been used for

most longitudes (cf. Toomer (1968), p. 136). For a general discussion of the prime

meridian in Islamic sources, see Kennedy and Kennedy, p. xi.

The entry for the latitude of Crdoba is 38;30, and that for its longitude

27;0, which is the same value given in canon 9 (f. 6ra): longitudo a circulo

occidentis ex centro Erin est gradus 27. The same values for the longitude and

latitude of Crdoba are found in some other Islamic sources, notably in a work

by Ab l-asan cAl al-Marrkush (see Kennedy and Kennedy, p. 95). We note

that in the Toledan Tables, the longitude of Crdoba is given as 9;20 and its

latitude as 38;30 (Toomer (1968), p. 134).

m Miscellaneous

This table has 3 columns: (1) gradus longitudinis, (2) directio arcus lumi-

nis, and (3) minuta diuersitatis transitus. Col. 1 lists degrees at 3 intervals

from 3 to 90 (with some copying errors); col. 2 has entries in degrees and

minutes from 0; 14 to 1507;0; and col. 3 has entries in minutes and seconds

from 0;1,23 to 0;23,33. The structure of this table is similar to that of some

andalusian astronomy 221

tables in the Almanac of Azarquiel (Mills (1950), p. 226), and most of the

entries are the same (the entries in col. 2 agree with those in the correspond-

ing column in the Almanac of Azarquiel for arguments from 3 to 84, but

for roundings and copying errors: note that the column in the Almanac dis-

plays degrees, minutes, and seconds, rather than degrees and minutes). But

the headings for the columns in the Almanac of Azarquiel are different from

those here, and we have not succeeded in determining the purpose for which

our table (or the corresponding table in the Almanac of Azarquiel) was com-

puted.

a. Just after the last table associated with al-Muqtabis, there are two tables

related to Azarquiels solar theory: (f. 55r) Tabula motus centri circuli exeun-

tis centrum in longitudine longiori et propinquiori a centro terre; (ff. 55v

56r) Tabula directionis composite centri circuli solis exeuntis centrum de

diuersitate centri eiusdem in longitudine propinquiori et longiori a circulo

diuersitatis motus centri morantis tempus. Toomer (1969), p. 325, has repro-

duced an excerpt and has explained the two tables.

b. (f. 56v) Tabula uisuum lunarium post occasum solis in climatibus septem

c. (f. 57r) Tabula eclipsis lune et quot digiti eclipsantur ex ea et hore dimidii

temporis eclipsis

d. One of the tables on f. 57v has already been analysed above (see Section iv,

k), in connection with a table in al-Muqtabis (f. 54v) giving the length of the

solar year. For another table on f. 57v, an eclipse table, see the comments

in Section iii, d, concerning lunar eclipses in al-Muqtabis (f. 52v). Folio 57v

has still another table for the maximum values for latitudes in the seven

climates.

e. There follow two tables that are clearly related to table 1 in the zij of al-

Khwrizm: (f. 58r) Tabula cuius est inter annos gentium et alios annos

preter illos ad inuicem; (f. 58v) Numeri dimissi per 28, 28 secundum annos

romanorum et egyptiacumi. According to Mills (1942), p. 245, they derive

from Maslama rather than from al-Khwrizm.

f. (f. 58v) Circuitus planetarum magni in sectis et divinationibus. These are

the Mighty Years of the planets: see table 13, where the entries of this table

are compared with those of Ab Macshar as preserved by al-Sijz (Pingree

(1968a), p. 64).

g. (f. 58v) Tabula dierum prouenientium in retrogradationibus planetarum et

directionibus eorum

222 chapter 7

et planetarum

i. Folios 59v to 61v present tables for the city of Sal: see the comments in

Section iv, d, on the table in al-Muqtabis for the oblique ascension for

Crdoba (ff. 49v51r).

j. The following table is certainly a part of al-Kawr al al-Dawr: (ff. 62v

64r) Tabula extractionis annorum quantitatis durationis creature in uentre

matris per longitudinem lune a gradu occidentis. It deals with astrological

obstetrics and has been discussed by Vernet (1949), pp. 273300.

k. (f. 64v) Tabula circuituum annorum planetarum in natiuitatibus

l. There follows information on houses, exaltations, triplicities, and signs, pre-

sented in tabular form (f. 65r), but every other entry has been left blank. The

information given agrees with that in al-Khwrizm/Maslama (Suter (1914),

p. 231, table 116). The table with the heading, Tabula terminorum egyptio-

rum (f. 65r) is the same as the fourth sub-table in al-Khwrizm/Maslama

(Suter (1914), p. 231, table 116), but some of the entries are blank here.

m. (f. 65v) Mediatus cursus solis in descensu eius ad quartas circuli secundum

probationem huius canonis. The data are for ah 480 (10871088 ad).

n. (f. 65v) Tabula terminorum ciuium Babillonie ueteris qui sunt magistri

ymaginum. Two parameters are given in this table: the length of the solar

year according to Ibn al-Kammd (365;15,36,19,35,32 days, cf. our com-

ments in Section iv, k, on the table in al-Muqtabis on f. 54v) and his length

of the lunar month (29;31,50,5,1 days).

o. (f. 66r) Tabula extracta per misilme de eo quod confirmatum extitit per

ciues huius artis yspanenses super diuisionem Yspanie per signa duodecim

12 12 12 12. Mills (1942), p. 256, suggested that misilme stands for Mas-

lama.

p. (f. 66r) Residuum ascensionum ad reuoluciones annorum solarium secun-

dum Muhad Arcadius. Mills (1942), p. 256, identified Muhad Arcadius

with Ab cAbd Allh Muammad ben Ysuf ben Amad Ibn Mucdh al-

Jayyn, from Jan. The value given for one year, 93;2,15, corresponds to

365;15,30,22 days. The same value is found in al-Khwrizm/Maslama (cf.

Neugebauer (1962), p. 132, and Goldstein (1967), pp. 143, 242). In fact, this

table attributed to Ibn Mucdh reproduces, with some scribal errors, the

table of conversion for the years of Nativity given in units of time-degrees in

al-Khwrizm/Maslama (Suter (1914), p. 230, table 115; cf. Neugebauer (1962),

p. 131).

q. (f. 66r) Tabula uisuum planetarum et absconsionum eorum sub radiis solis

r. (f. 66r) Capita draconum planetarum in anno quingentesimo et quinqua-

gesimo ab annis seductionis. Note that ah550 corresponds to 11551156ad.

andalusian astronomy 223

(f. 58v)

Venus 1151 1151

Mercury 480 480

Moon 420 * 520

Saturn 625 * 265

Jupiter 567 * 427

Mars 684 * 284

See our comments in Section iv, e, on the table of al-Muqtabis for the lati-

tude of the planets (ff. 45rv).

We are grateful to ngel Mestres (University of Barcelona) for calling our atten-

tion to the following passage in ms Hyderabad, Andra Pradesh State Library 298

(no foliation): chap. 35 Ab l-cAbbs al-Kammd said in a horoscope he drew

in Crdoba in the year 510 Hijra (= 11161117ad).

References

vol. v. Berlin.

Chabs, J. 1991, The Astronomical Tables of Jacob ben David Bonjorn, Archive for the

History of Exact Sciences 42: 279314.

Chabs, J., and Goldstein, B.R. 1992, Nicholaus de Heybech and His Table for Finding

True Syzygy, Historia Mathematica 19, pp. 265289.

Debarnot, M.-T. 1987, The zj of abash al-sib: A Survey of ms Istanbul Yeni Cami

784/2, in David A. King and George Saliba (eds.), From Deferent to Equant: A Volume

of Studies in the History of Science in the Ancient and Medieval Near East in Honor of

E.S. Kennedy, Annals of the New York Academy of Sciences 500: 3569.

Dobrzycki, J. 1965, Teoria precesji w astronomii sredniowiecznej, Studia i Materialy

Dziejow Nauki Polskiej, Seria c 11: 347.

224 chapter 7

Goldstein, B.R. 1967, Ibn al-Muthanns Commentary on the Astronomical Tables of

al-Khwrizm, New Haven and London.

Goldstein, B.R. 1985, Scientific Traditions in Late Medieval Jewish Communities, in

G. Dahan (ed.), Les Juifs au regard de lhistoire, Paris, pp. 235247.

Goldstein, B.R. 1992, Lunar Velocity in the Ptolemaic Tradition, in P.M. Harman and

A.E. Shapiro (eds.), An Investigation of Difficult Things: Essays on Newton and the

History of Exact Sciences, Cambridge, pp. 317.

Goldstein, B.R., and Sawyer, F.W. iii, 1977, Remarks on Ptolemys Equant Model in

Islamic Astronomy in Y. Maeyama and W.G. Saltzer (eds.) Prismata, Wiesbaden,

pp. 165181.

Goldstein, B.R., Chabs, J., and Mancha, J.L. 1994, Planetary and Lunar Velocities in the

Castilian Alfonsine Tables, Proceedings of the American Philosophical Society, 138:

6195.

Kennedy, E.S. 1956a, A Survey of Islamic Astronomical Tables, Transactions of the

American Philosophical Society ns 46.

Kennedy, E.S. 1956b, Parallax Theory in Islamic Astronomy, Isis 57: 3353; reprinted in

Kennedy, E.S., Studies in the Islamic Exact Sciences, Beirut (1983), pp. 164184.

Kennedy, E.S. 1986, The Verified Astronomical Tables for the Caliph al-Mamn by Yay

b. Ab Manr. A photographic reproduction of ms Escorial Ar. 927, with an intro-

duction. Frankfurt.

Kennedy, E.S., and Destombes, M. 1966, lntroduction to Kitb al-cAmal bil-Asturlb,

Osmania Oriental Pub., Hyderabad-Dn. (1966); reprinted in Kennedy, E.S., Studies in

the Islamic Exact Sciences, Beirut (1983), pp. 405447.

Kennedy, E.S., and Faris, N. 1970, The Solar Eclipse Technique of Yay b. Ab Manr,

Journal for the History of Astronomy 1: 2038; reprinted in Kennedy, E.S., Studies in

the Islamic Exact Sciences, Beirut (1983), pp. 185203.

Kennedy, E.S., and Kennedy, M.H. 1987, Geographical Coordinates of Localities from

Islamic Sources, Frankfurt.

Kennedy, E.S., and Ukashah, W. 1969, Al-Khwrizms Planetary Latitude Tables, Cen-

taurus 14: 8696; reprinted in Kennedy, E.S., Studies in the Islamic Exact Sciences,

Beirut (1983), pp. 125135.

Kunitzsch, P. 1966, Typen van Sternverzeichnissen in astronomischen Handschriften des

zehnten bis vierzehnten Jahrhunderts, Wiesbaden.

Mercier, R. 19761977, Studies in the Medieval Conception of Precession, Archives

internationales dhistoire des seiences 26: 197220, and 27: 3371.

Mills, J.M. 1942, Las traducciones orientales en los manuscritos de la Biblioteca Catedral

de Toledo, Madrid.

Mills, J.M. 1947, El libro de los fundamentos de las Tablas astronmicas, MadridBarce-

lona.

andalusian astronomy 225

Mills, J.M. 1962, Las Tablas Astronmicas del Rey Don Pedro el Ceremonioso, Madrid

Barcelona.

Morelon, R. 1987, Thbit ibn Qurra. Oeuvres dastronomie, Paris.

Nallino, C.A. 18991907, Al-Battn sive Albatenii Opus Astronomicum, 3 Vols. Milan.

Neugebauer, O. 1962, The Astronomical Tables of al-Khwrizm. Copenhagen.

North, J.D. 1967, Medieval Star Catalogues and the Movement of the Eighth Sphere,

Archives internationales dhistoire des sciences 17:7383.

North, J.D. 1976, Richard of Wallingford: An edition of his writings with introductions,

English translation and commentary, 3 Vols. Oxford.

Pingree, D. 1968a, The Thousands of Ab Macshar. London.

Pingree, D. 1968b, The Fragments of the Works of Yacqb ibn riq, Journal of Near

Eastern Studies 27: 97125.

Pingree, D. 1970, The Fragments of the Works of al-Fazr, Journal of Near Eastern

Studies 29: 103123.

Salam, H., and Kennedy, E.S. 1967, Solar and Lunar Tables in Early Islamic Astronomy,

Journal of the American Oriental Society 87: 492497.

As-Saleh, J.A. 1970, Solar and Lunar Distances and Apparent Velocities in the Astro-

nomical Tables of abash al-sib, Al-Abhath 23: 129177; reprinted in Kennedy,

E.S., Studies in the Islamic Exact Sciences, Beirut (1983), pp. 204252.

Sams, J. 1987, Al-Zarql, Alfonso x and Peter of Aragon on the Solar Equation, in David

A. King and George Saliba (eds.), From Deferent to Equant: A Volume of Studies in the

History of Science in the Ancient and Medieval Near East in Honor of E.S. Kennedy,

Annals of the New York Academy of Sciences 500: 467476.

Sams, J. 1992, Las Ciencias de los Antiguos en al-Andalus. Madrid.

Sdillot, J.-J., and Sdillot, L.-A. 1834, Trait des instruments astronomiques des Arabes.

Paris; reprinted Frankfurt 1984.

Stahlman, W.D. 1959, The Astronomical Tables of Codex Vaticanus Graecus 1291. Brown

Univ. Ph.D. Thesis.

Suter, H. 1914, Die astronomischen Tafeln des Muammad ibn Ms al-Khwrizm.

Copenhagen.

Toomer, G.J. 1968, A Survey of the Toledan Tables, Osiris 15: 5174.

Toomer, G.J. 1969, A History of Errors, Centaurus 14: 306336.

Toomer, G.J. 1987, The Solar Theory of Az-Zarql: An Epilogue in David A. King and

George Saliba (eds.), From Deferent to Equant: A Volume of Studies in the History of

Science in the Ancient and Medieval Near East in Honor of E.S. Kennedy, Annals of the

New York Academy of Sciences 500: 513519.

Vernet, J. 1949, Un tractact dobstetricia astrolgica, Boletn de la Real Academia de

Buenas Letras de Barcelona 22: 6996; reprinted in Vernet, J., Estudios sobre Historia

de la Ciencia Medieval, BarcelonaBellaterra (1979), pp. 273300.

226 chapter 7

Vernet, J. 1956, Las Tabulae Probatae, Homenaje a Mills Vallicrosa, Barcelona, ii: 501

522; reprinted in Vernet, J., Estudios sobre Historia de la Ciencia Medieval, Barcelona

Bellaterra (1979), pp. 191212.

chapter 8

The Tables of John Vimond (1320)*

It has been clear for many years that medieval European astronomy in Latin

was heavily dependent on sources from the Iberian peninsula, primarily in

Arabic, but also in Hebrew, Castilian, and Catalan. The Castilian Alfonsine

Tables, compiled by Judah ben Moses ha-Cohen and Isaac ben Sid under the

patronage of Alfonso x (d. 1284), were an important vehicle for the transmission

of this body of knowledge to astronomers north of the Pyrenees, but the details

of this transmission remain elusive, in part because only the canons to these

tables survive (see Chabs and Goldstein 2003a). In this paper we build on

our preliminary studies of a figure who previously had barely been mentioned

in the recent literature on medieval astronomy (Chabs and Goldstein 2003a,

pp. 267277, and 2003b). John Vimond was active in Paris ca. 1320 and, as we

shall see, his tables have much in common with the Parisian Alfonsine Tables

(produced by a group in Paris, notably John of Murs and John of Lignres), but

differ from them in many significant ways. As far as we can tell, there is no

evidence for any interaction between Vimond and his better known Parisian

contemporaries and in our view the best hypothesis is that they all depended

on Castilian sources. As a result of our analysis, we are persuaded that Vimonds

tables are an intelligent reworking of previous astronomical material in the

Iberian peninsula to a greater extent than is the case for the Toledan Tables

(compiled in Toledo about two centuries before the Castilian Alfonsine Tables).

It is most likely that Vimonds principal source was the Castilian version of the

Alfonsine Tables.

Paris, Bibliothque nationale de France, ms lat. 7286c is a 14th-century

manuscript containing an unusual set of tables (ff. 1r8v) as well as the canons

and tables of 1322 by John of Lignres (ff. 9r58r). In a brief text at the end of

the first set of tables they are attributed to John Vimond (Iohannes Vimundus),

an astronomer who compiled them for the use of students at the University of

Paris (f. 8v):

228 chapter 8

posicionibus planetarum et stellarum fixarum; et cum istis sequitur de hiis

que per ipsum ordinantur ad conversionem temporum verorum et equalium

sociatorum, et de disposicionibus eclipsalibus solis et lune sibi pertinentibus,

et de aliis disposicionibus ipsorum et aliorum corporum celestium, ad utili-

tatem scolarium universitatis parisiensis et omnium aliorum.

Here ends the work by John Vimond of the diocese of Bayeux on the

dispositions of the planets and the fixed stars; () and on the dispositions

of solar and lunar eclipses and [other syzygies] corresponding to them,

and on the other dispositions of these and other celestial bodies, for the

use of students at the University of Paris and all others.

The complete set of Vimonds tables are uniquely extant in this manuscript,

and no canons for them have been identified. They are a coherent set of tables

with all the elements needed to compute the positions of the celestial bodies,

much in the tradition of the Arabic zijes and their derivatives. The exact date of

composition of Vimonds tables is not given in the text, but they were probably

produced shortly before 1320. In the paragraph preceding his tables, Vimond

tells us that they were compiled for Paris with 1320 as epoch (f. 1r: see below) and

this date is confirmed by recomputation. These tables also include a calendar

with the dates of syzygies: this strongly suggests that they were constructed

prior to the year of the calendar because the astronomical information would

no longer be of any use after the year had passed. However, the calendar poses

special problems which will be discussed below.

Vimonds only other known work is a short treatise on the construction

of an astronomical instrument, extant in Erfurt, ms ca 2 377 (ff. 21r22r),

beginning Planicelium vero componitur ex eis que sunt diversorum operum ,

and ending Explicit tractatus johannis vimundi in a manuscript containing

various works by other Parisian astronomers such as John of Murs and John of

Lignres (Thorndike and Kibre 1963, col. 1050; Saby 1987, pp. 471, 474).

John Vimond and his works were seldom mentioned by his contempo-

raries. However, in Vatican, Biblioteca Apostolica, ms Ottob. lat. 1826, we are

told that John of Spira (14th century), the author of a commentary on John

of Lignress canons (Thorndike and Kibre 1963, col. 204), composed his own

canons to several of Vimonds tables (for a description of this manuscript, espe-

cially ff. 148153, see F.S. Pedersen 2002, p. 177). This manuscript includes a

text that begins on f. 148ra ascribed to a certain M.J.C., Canon tabulae sequen-

tis quae intitulatur tabula motus diversi solis et lunae in una hora et semidi-

ametrorum secundum tabulas Alfonsi, at the end of which John Vimond is

early alfonsine astronomy in paris 229

lioteca Nacional, ms 4238, a manuscript containing a few tables that can be

attributed to him, as well as a copy of the Parisian Alfonsine Tables computed

for Morella (in the province of Valencia) for the years 1396 and 1400 (Chabs

2000).

As far as we can tell, John of Murs and John of Lignres do not refer to

Vimond at all in any of their numerous works, but it seems implausible that

they did not know him or his work which was addressed to the students at the

University of Paris. Indeed, there were not so many competent astronomers

working in Paris around 1320 and both Vimond and Murs came from the

same region, Normandy, from places about 70km apart, Bayeux and Lisieux,

respectively.

We would expect Vimond to be well known and frequently cited by practi-

tioners of astronomy, for he is named as one of the outstanding astronomers

of his time by Simon de Phares in his Recueil des plus celebres astrologues

(14941498), a chronologically ordered list with comments, edited by Boudet

(19971999, 1:467). In fact, Vimond is mentioned before John of Lignres, John

of Saxony, John of Janua, and John of Murs:

Maistre Jehan Vymond fut a Paris, homme moult singulier et grant astrolo-

gien, lequel eut en ce temps grant cours pour la science des estoilles. Entre

ses euvres, fist une verifficacion de la conjunction des lu[mi]naires, aussi des

eclipses et estoilles fixes pour plusieurs ans. Cestui predist les grans vens qui

furent en son temps et fist plusieurs beaulx jugemens, dont il acquist grant

loz et renommee en France et fut moult devost en Nostre Seigneur.

Master John Vimond lived in Paris, a most singular man and a great

astrologer, who had at that time much prestige because of (his knowledge

of) the science of the stars. Among his works is a verification of the

conjunction of the luminaries, as well as eclipses and the fixed stars, for

many years. He predicted the great winds which took place in his time

and made many fine judgments for which he acquired great praise and

renown in France and he was most devoted to our Lord.

These tables are arranged very differently from those of his Parisian contempo-

raries and are based, in part, on parameters that probably came from the Castil-

ian Alfonsine Tables or a tradition closely associated with them. Of special

interest is the proper motion of the solar and planetary apogees, a feature pre-

viously unknown in medieval tables produced outside Spain and North Africa.

230 chapter 8

Paris of new astronomical material coming from Castile, in the sense that they

propose new approaches to replace those based on the Toledan Tables and

developed at the end of the 13th century by astronomers working in Paris such

as Peter Nightingale, Geoffreoy of Meaux, and William of St.-Cloud. Further,

we believe that Vimonds tables are prior to, and independent of, the tabular

work developed in the early 14th century, which we call the Parisian Alfon-

sine Tables, by the group of Parisian astronomers that included John of Murs

and John of Lignres, which were also based on Castilian sources. Vimonds

tables and the Parisian Alfonsine Tables have many parameters in common

both for mean motions and equations. In principle, it is possible that one set of

tables depended on the other, but the differences between them suggest to us

that it is far more likely that they depended on a common source. Moreover, if

Vimond composed his tables prior to 1320, he did so before any datable text of

the Parisian Alfonsine Tables.

A description and analysis of Vimonds tables follow.

f. 1r The first numerical information given in this set of tables is the radix

for mean conjunctions of the Sun and the Moon: 13;54,54d. In modern ter-

minology, the initial time for a set of tables is called its epoch whereas its

radices are the positions of the Sun, Moon, planets, etc., at that time. The

medieval convention, however, is to use radix for both the time and the posi-

tion.

We are convinced that the author refers to the time, in Paris, of the mean

conjunction on March 10, 1320. The year and the place are mentioned by

Vimond himself in a short paragraph following the numerical value of the radix

(f. 1r):

diate post 19 secunda diei que consistunt immediate post () lucis beati

Mathie composite procedendo ab ortu solis usque ad occasum scilicet anno

domini nostri Ihesu Christi 1320 secundum numeracionem annorum roma-

norum qui incipiunt ex inicio diei circoncisionis domini nostri Ihesu Christi

et existentis ad longitudinem civitatis Parisius que distat a medio mundi per

49 g et 30 min ita quod illa civitas est in parte occidentali et etiam distat ab

illo medio per 8 min et 15s diei equalis.

Note that this radix for the mean conjunctions comes immediately after

19 seconds of a day that fall immediately after the () [space for one

word; illegible] (day)light of Saint Matthew, proceeding from sunrise to

early alfonsine astronomy in paris 231

sunset, namely, in the year of our Lord Jesus Christ 1320 according to the

count of Roman years which start from the beginning of the Day of the

Circumcision of our Lord Jesus, for the longitude of the city of Paris which

is distant from the middle of the world by 49 degrees and 30 minutes

because that city is in the western direction and distant from that middle

by 8 minutes and 15 seconds of an equinoctial day.

Madrid, Biblioteca Nacional, ms 4238, f. 66v, has a short text which is very

similar to the paragraph on f. 1r in Paris, Bibliothque nationale de France, ms

lat. 7286c:

Radix opposicionum: dies 14 m 45 2 55

Hec radix est post 19 2 diei que sunt post meridiem diei Mathie anno 1320

secundum romanos. Nota quod annus in meridie diei mathie 25 dies bisexti

erit semper ultimus dies anni. Iste tabule radicum sunt facte Parisius ad

meridiem cuius cenith distat ab equinocciali 49 g 30 m vel 8 m 15 2 diei.

The numerical datum, 0;0,19d (= 0;7,36h), represents the equation of time for

that day. In the Madrid version, the radix of the opposition, 14;45,55d, is half

the length of a mean synodic month which is about 29;31,50d, and this is the

entry for the first opposition in Vimonds calendar (see below). It is clear that,

according to the version of this text in the Paris manuscript, the civil day in

the calendar includes the period of daylight, that is, the time from sunrise

to sunset, in contrast to the astronomical day that goes from noon to the

following noon. Both values given for the longitude of Paris from Arin, called

the middle of the world, are equivalent. Arin, a corruption of Ujjain (a city

in India), was thought to be halfway between the eastern and western limits

of the world (Neugebauer 1962, p. 11, n. 2). The distance from Arin to Toledo

was taken to be 61;30 and, since Paris was generally said to be 0;48h or 12 to

the east of Toledo, its longitude from Arin is 49;30, as in the passage above

(Mills 19431950, p. 49; Kremer and Dobrzycki 1998, p. 194; and F.S. Pedersen

2002, p. 431). Moreover, when a day is taken to be 360, it follows that 49;30

corresponds to 0;8,15d, for 49;30/360 = 0;8,15. The expression in the Madrid

manuscript, Parisius ad meridiem cuius cenith distat ab equinocciali is a

corrupt version of the better reading in the Paris manuscript, for it would imply

that 49;30 is the latitude of Paris, but then its equivalence to 0;8,15d would

become meaningless. We also note that, according to this text, Vimonds tables

were computed for Paris whereas in the early 1320s other Parisian astronomers

who recast the Alfonsine Tables computed them for Toledo, as is the case for the

232 chapter 8

d d d

2 8;15,53 32 6; 6,50 68 27;45,27

3 27; 9,45 36 21;38,37 72 13;45,23

4 15;31,46 40 7;38,33 76 29;17,10

8 1;31,43 44 23;10,19 152 29; 2,29

12 17; 3,29 48 9;10,15 304 28;33, 8

16 3; 3,25 52 24;42, 2 608 27;34,26

20 18;35,12 56 10;41,58 1216 25;47, 1

24 4;35, 8 60 26;13,44 2432 21;43,12

tables with epoch 1321 by John of Murs (see, e.g., Lisbon, Biblioteca de Ajuda,

ms 52-xii-35).

According to our computations based on the Parisian Alfonsine Tables, the

mean conjunction on March 10, 1320 took place in Toledo at 9;10h, civil time

(i.e., counting from midnight) which, with a correction of 0;48h, is 9;58h in Paris

(civil time), that is, 2;2h before noon. Thus, the radix for the tables on f. 1r (as

well as the radices for the planetary tables, as will be seen later) is the time of

the first mean conjunction in March 1320 (March 10, 1320, at 9;58 a.m., Paris,

or March 9, 1320, 21;58h, Paris, counting from noon). Indeed, the sexagesimal

part of the radix (0;54,54d) is exactly the sum of 12h and 9;58h. The integer

part of the radix, as will be explained later in reference to the annual calendar

presented on this same folio, is counted from the epoch of the calendar, almost

14 days before the mean conjunction of March 10, 1320, that is, February 25, 1320

or February 24b, 1320, where 24b represents the second day called February

24 in a leap year (such that the last day of February is always day 28 both in

ordinary and leap years).

The entries in this table give the instant of the first mean conjunction after a

certain number of years. We are given entries for 1, 2, 3, and 4 years; for multiples

of 4 years up to 76 (= 19 4) years; and for 152, 304, 608, 1216, and 2432 years.

The entries represent the excess of days after an integer number, n, of synodic

early alfonsine astronomy in paris 233

months have elapsed (where n = 13 for year 1, , and 30,081 for year 2432).

Madrid, Biblioteca Nacional, ms 4238, f. 66v, reproduces this table except for

the last row for 2432 years, which is missing.

The value for the mean synodic month derived from year 2432 is 29;31,50,7,

44,35d 0;0,0,0,0,4d. Thus, for year 1: 13 29;31,50,7,44,35d 365d = 18;53,52d,

in agreement with the tabulated value. In the Parisian Alfonsine Tables, the

mean synodic month is 29;31,50,7,37,27,8,25d: this value is found, for example,

in Lisbon, Biblioteca de Ajuda, ms 52-xii-35, f. 16v, containing the tables for

epoch 1321 by John of Murs. So Vimonds parameter is very similar to, but not

identical with, the parameter in the Parisian Alfonsine Tables.

This annual calendar begins on the day of Saint Matthew (February 24) and

lists the dates associated with several saints, as well as the dates and times of 25

consecutive mean syzygies. The practice of adding the extra day in a leap year

after Feb. 24 goes back to the Roman calendar as revised by Julius Caesar, when

the additional day followed Feb. 24 and was called bis-sextus ante calendas

martias (the sixth day before the calends of March). In a leap year February

lasted 29 days, but the last day was numbered 28, for the 24th was assigned

to two consecutive days. This is what is intended in Vimonds calendar where

the year begins on that very day. We know of no other calendar in the late 13th

century or early 14th century beginning on Feb. 24; in particular, the calendars

composed by Geoffreoy of Meaux and William of St.-Cloud do not begin on

that day (Chabs and Goldstein 2003a, pp. 245247). It is worth noting that

Vimonds calendar which lists mean syzygies together with saints days is in the

tradition of these two astronomers who were active in Paris shortly before him:

they displayed planetary data in calendars and depended on the Toledan Tables

for their computations. We also note that the feast of St. Matthew is mentioned

in the canons to the Parisian Alfonsine Tables by John of Saxony as the last day

in a leap year (see Poulle 1984, p. 36, line 41). Vimond offers no explanation for

basing his calendar and his tables on syzygies; we can only conjecture that he

was being faithful to some unknown source.

234 chapter 8

[Saints day / No. syzygy] [date] [time since epoch]

d

Romanus February 28 4

Perpetua virgo March 7 11

1 Opposition March 10 14;45,55

Gregorius papa 12 16

early alfonsine astronomy in paris 235

[Saints day / No. syzygy] [date] [time since epoch]

d

Circoncisio domini Ihesu

Christi initium anni [January 1] 311

Iuliana virgo 16 357

Petrus ad cathedram 22 363

25 Opposition February 28 369; 7,57

In Table 2, columns 1 and 2 have no heading, but column 3 has the heading

days, minutes, and seconds. In the manuscript the name of the month is usu-

ally given in col. 2, and occasionally in col. 1 which has about 90 entries such

as: Annunciatio Domini, Dyonisius, Lucas Evangelista, Innocentes, etc. The syzy-

gies are numbered from 1 to 25, and they are transcribed above. The numbers

in column 3 are integers when a saints day is meant and indicate the number

of days that elapsed since the epoch (day 1) of the calendar, that is, February

25, 1320 (Julian) or what we have called February 24b, 1320. Vimond seems

to use here civil days (from midnight to midnight) rather than astronomical

days (from noon to noon), which makes sense in a calendar. When a conjunc-

tion or an opposition is indicated, we would expect the number in column 3

to refer to the accumulated time from the radix (the conjunction on March

236 chapter 8

10, 1320) in multiples of half a mean synodic month, i.e. 14;45,55d but, in fact,

we are given the accumulated time from the mean syzygyan opposition

immediately preceding the radix, which occurred on February 24, 1320. If this

was the authors intention, it is not clear how the user of these tables was to

take account of the radix given at the beginning of them. Moreover, despite

the coherence of the arithmetic in this calendar, something is seriously wrong

with it, for we find the word oppositio next to March 10, when a conjunction took

place, and the word coniunccio next to March 25, when an opposition occurred.

The same pattern is followed throughout the calendar. There is an explana-

tory note on f. 1r concerning the calendar, but we were unable to make sense

of it.

Year 1324 might be considered as an alternative date for the calendar for,

according to computations with the Parisian Alfonsine Tables, a mean oppo-

sition occurred on March 10 (counting from noon) or on March 11 (counting

from midnight). This would conform with the character of the syzygy men-

tioned in the calendar, and the computations associated with this date yield

results that are quite close to (but not exactly the same as) the information

given in the text. Indeed, in our preliminary discussion of these tables, this

near agreement mislead us to think that 1324 was the radix of Vimonds tables

(Chabs and Goldstein 2003a, p. 270). However, as indicated previously, year

1320 is specifically mentioned, and it fits much better with the radix of mean

conjunctions and with the radices for planetary positions displayed on ff. 1v and

4r.

f. 1v Radices for the argument of solar anomaly, the argument of lunar anomaly

(henceforth, solar and lunar anomaly, respectively), the solar apogee, and the

lunar ascending node:

Lunar anomaly 1s 3; 6,14

Solar apogee 2s 29;56,15

Ascending node 10s 13;14,43

Note the use of zodiacal signs of 30, a characteristic of all tables in this set.

A short text below these parameters explains that the radices for the motion

of the solar apogee and the ascending node are counted from the beginning

of Aries on the 9th sphere, indicating that tropical coordinates are used here.

These radices were calculated for March 10, 1320, at the time of the mean con-

junction of the Sun and the Moon. According to the Parisian Alfonsine Tables

the solar apogee for March 10, 1320 is 89;23,50, a value which differs by about

early alfonsine astronomy in paris 237

half a degree from the entry in the text. Both values in turn differ from the solar

apogee for 1320 in the tables for 1322 by John of Lignres (89;24,22) that is found

by adding two values given on f. 9v of this same manuscript: the solar apogee

(81;7,15,39) and the motion of the 8th sphere at that time (8;17,6,48). The

same result, 89;24,22, can also be found in another copy of John of Lignress

tables, Erfurt, ms ca q 362, f. 21ra. For the rest of the radices, recomputation with

the Parisian Alfonsine Tables for the epoch, March 9, 1320, at 21;10h in Toledo

(counting from noon), yields results which are very close to the values in the

text, especially for the Moon:

Lunar anomaly 33; 6,14 33; 6,28

Solar apogee 89;56,15 89;23,50

Ascending node 313;14,43 312;54,39

The solar longitude is the sum of the solar anomaly and the solar apogee:

and again the agreement is very good. Since this is the time of a mean con-

junction, the mean lunar longitude will be equal to the mean solar longitude.

According to the Parisian Alfonsine Tables, the mean lunar longitude at this

epoch was 356;11,3, i.e., it differed from the mean solar longitude by only 0;0,13

(note that the Moon travels this distance in about 20 seconds of time which is

below the accuracy of 1 minute for the time of mean conjunction). Hence the

absence of a radix for lunar mean motion simply reflects the fact for the epoch

of Vimonds tables the mean longitude of the Moon is the same as the mean

longitude of the Sun.

The agreement for the radix of lunar anomaly to the minute is particularly

impressive since the motion in lunar anomaly is about 0;30/h. Lunar anomaly

is not subject to precession and it is independent of solar motion. (We use the

term precession for a constant motion of the eighth sphere, and trepidation for a

238 chapter 8

s () s () s ()

Lunar anomaly 11 5;37, 8 9 15;25,15 8 21; 2,22

Solar apogee 0 0; 1,12 0 0; 2,18 0 0; 3,30

Ascending node 11 9;43,23 10 20;58,27 10 0;40,50

s () s ()

Lunar anomaly 1 5;51,58 4 8;22, 2

Solar apogee 0 0; 9, 7 0 11;32,36

Ascending node 6 25;27,27 4 29;26,44

variable motion of the eighth sphere.) So, even though Vimond and the authors

of the Parisian version of the Alfonsine Tables differ on matters of definition

and made slight changes in mean motions, it is unlikely that either of them

would change the motion in anomaly significantly from what it had been in

their common source.

This table displays the radices for the solar anomaly, the lunar anomaly, the

solar apogee, and the lunar ascending node for 1, 2, 3, and 4 years; for multiples

of 4 years up to 76 years; and then for 152, 304, 608, 1216, and 2432 years, as in

Table 1. A selection of the entries is displayed in Table 3.

The entries for year 1 represent the progress made by the Sun, the Moon,

the solar apogee, and the lunar node in a year of 13 mean syzygies of the same

kind (henceforth lunations) of 29;31,50,7,44,35d. To be sure, the difference

between the solar anomalies for year 2 and year 1 is 349;15,44, meaning that

year 2 contains 12 lunations, for 349;15,44/ (29;31,50 0;59, 8, ) = 12, whereas

the difference between year 3 and year 2 is 378;22,2 (the same value as that

early alfonsine astronomy in paris 239

for year 1), indicating that year 3 contains 13 lunations (378;22,2/ (29;31,50

0;59, 8, ) = 13), as is the case with year 1. With this procedure, we see that 50

is the total number of lunations in the first 4 years, 99 in the first 8 years, ,

7,521 in the first 608 years, and so on. That is, for Vimond 1 year is equivalent

to 13 mean lunations, 2 years is equivalent to 25 mean lunations, etc. Where

possible, we have derived the associated mean motions from the entries for

year 608, because those for years 1216 and 2432 are not completely legible in

the manuscript.

The mean motion in solar anomaly resulting from the entry for year 608

(0s 20;6,6), that is, after 7,521 lunations, and the length of the synodic month

obtained before (29;31,50,7,44,35d), is 0;59,8,8,23,30/d, for

= 0;59,8,8,23,30/d.

This daily mean motion implies a year length of 365;15,42,32d which is side-

real. In the Parisian Alfonsine Tables, however, the length of a sidereal year

is variable, and the fixed length of the tropical year is 365;14,33,9,57, d (=

360/0;59,8,19,37,19,13,56/d).

Similarly, the mean motion in lunar anomaly can be computed from the

entry corresponding to 7,521 lunations (year 608: 4s 8;22,2), for 7521 luna-

tions corresponds to 8060 complete revolutions in anomaly with an excess of

about 120 (computed with approximate values for the appropriate parame-

ters). Hence, with the data in the text, the mean motion in lunar anomaly

is

= 13;3,53,57,27,11/d,

in very good agreement with the corresponding value in the Parisian Alfonsine

Tables (13;3,53,57,30,21/d); the difference only accumulates to 1 in well over

10,000 years.

As for the motion of the solar apogee derived from the entry corresponding

to 7,521 lunations (year 608: 0s 11;32,36), we find 0;0,0,11,13,35/d. By the same

reasoning, the mean motion of the lunar ascending node resulting from the

entry of year 8 in the table (6s 25;27,27) is 0;3,10,18,6,48/d, in contrast to the

value found in the Parisian Alfonsine Tables (0;3,10,38,7,14,49,10/d). In this

case, the entry in the manuscript for 608 years is corrupt.

We note that Vimonds value for the motion of the solar apogee includes

precession as well as its proper motion for, if we add the value for the mean

240 chapter 8

Solar anomaly 0;59, 8, 8,23,30/d 0;59, 8,19,37,19/d

Lunar anomaly 13; 3,53,57,27,11 13; 3,53,57,30,21

Solar apogee 0; 0, 0,11,13,35

Ascending node 0; 3,10,18, 6,48 0; 3,10,38, 7,14

motion in solar anomaly (which is sidereal) to the motion of the solar apogee,

we find 0;59,8,19,37,4/d, in close agreement with the corresponding value

of the mean motion in solar longitude (tropical) in the Parisian Alfonsine

Tables. In the Almagest, the planetary apogees are sidereally fixed whereas

the solar apogee is tropically fixed. In the 9th century astronomers in Bagh-

dad fixed the solar apogee sidereally so that it too was subject to precession

(or trepidation). But in the 11th century Azarquiel realized that the solar

apogee had a proper motion in addition to precession, and fixed its amount

as 1 in 279 Julian years or about 0;0,0,2/d (Chabs and Goldstein 1994, p.

28). In one Andalusian tradition, this proper motion of the solar apogee was

applied to the planetary apogees as well (see Sams and Mills 1998, p. 269;

cf. Mestres 1996, pp. 394395). If we take al-Battns value for precession of

1 in 66 years or about 0;0,0,9/d and add it to the proper motion of the

solar apogee, the result is about 0;0,0,11/d. There is no hint of this proper

motion for either the solar apogee or the planetary apogees in the Parisian

Alfonsine Tables where these apogees are all sidereally fixed and, instead of

precession, the Parisian Alfonsine Tables have tables for trepidation; hence,

there is nothing in those tables with which to compare directly the motion

of the solar apogee in Vimonds tables. We see, then, that the parameters in

Vimonds tables are not identical with those in the Parisian Alfonsine Tables,

and some of these parameters (e.g., the length of the solar year) are defined

differently.

This table displays the radices for the solar anomaly, the lunar anomaly, the

solar apogee, and the lunar ascending node for 25 consecutive syzygies after

the corresponding integer numbers of semi-lunations have elapsed. An excerpt

is shown in Table 4.

early alfonsine astronomy in paris 241

s () s () s ()

Lunar anomaly 6 12;54,30 0 25;49, 1 4 22;42,27

Solar apogee 0 0; 0, 3 0 0; 0, 6 0 0; 1, 9

Ascending node 11 29;13,10 11 28;26,20 11 10;29,13

The entries represent the progress made by the Sun, the Moon, the solar

apogee, and the lunar node in 1, 2, , 25 mean semi-lunations of 29;31,50,7,44,

35d/ 2 = 14;45,55,3,52,17d. The entries in this table agree with those in Table 3,

for in each case the value for 26 consecutive semi-lunations (the sum of the

entries for Syzygy 1 and Sygygy 25 in Table 4) equals the value for 13 lunations

(year 1 in Table 3).

f. 2r Table 5: Sun

(argumentum) at 3-intervals in signs and degrees from 0s 3 to 12s 0; this

is the mean solar anomaly. Column 2 displays the true solar anomaly (motus

completus) in signs, degrees, and minutes. Column 3 (motus gradus) displays

the increment in true anomaly per degree of the argument. Column 4 gives

the solar velocity, in units of minutes and seconds of arc in a minute of a day

(minutum diei), i.e., in a sixtieth of a day. Column 5 displays the time (also called

argumentum), in days, with sexagesimal fractions of a day, that the Sun takes

to complete the arc indicated in column 1.

To obtain an entry in column 5 multiply the corresponding entry in column 1

by the daily mean motion in solar anomaly; the entry for 360 (365;15,42d)

represents the length of the sidereal year, in good agreement with the value

deduced from 99 mean synodic months in Table 3.

As shown in Table 5a, the difference between the argument (col. 1) and the

true anomaly (col. 2) represents the solar equation, with a maximum of 2;10

as in the Parisian Alfonsine Tables. To emphasize the solar equation, we have

added a third column for the differences between entries in columns ii and i,

labeled ii i.

242 chapter 8

argum. motus c. motus g. min. diei argum.

s () s () min. min. d

0 6 0 5;47 57;51 0;57 6; 5,16

3 0 2 27;50 59;59 0;59 91;18,55

3 3 3 0;50 60; 3 0;59 94;21,33

3 6 3 3;50 60;16 0;59 97;24,11

3 9 3 6;51 60;19 0;59 100;26,49

6 0 6 0; 0 62;24 1; 1 182;37,51

6 3 6 3; 7 62;22 1; 1 185;40,29

8 24 8 26;10 60; 3 0;59 267;51,31

8 27 8 29;10 59;59 0;59 270;54, 9

9 0 9 2;10 59;47 0;59 273;56,46

9 3 9 5; 9 59;43 0;59 276;59,24

12 0 12 0; 0 57;51 0;57 365;15,42

i ii ii i

argumentum motus completus

s () s () ()

0 3 0 2;54 0; 6

0 6 0 5;47 0;13

2 27 2 24;51 2; 9

3 0 2 27;50 2;10

early alfonsine astronomy in paris 243

i ii ii i

argumentum motus completus

s () s () ()

3 3 3 0;50 2;10

3 6 3 3;50 2;10

3 9 3 6;51 2; 9

5 27 5 26;53 0; 7

6 0 6 0; 0 0; 0

6 3 6 3; 7 0; 7

8 21 8 23; 9 2; 9

8 24 8 26;10 2;10

8 27 8 29;10 2;10

9 0 9 2;10 2;10

9 3 9 5; 9 2; 9

11 27 11 27; 6 0; 6

12 0 12 0; 0 0; 0

The entries for the solar equation are not explicit in Vimonds table; they can

be graphed as a smooth curve but they do not allow us to decide which specific

table for the solar equation he used. The reason is that Vimonds entries are

only given to minutes in contrast to most other tables in which the maximum

equation is 2;10,0 where entries are given to seconds, and rounding those

values produces Vimonds entries.

This table has the same format as Table 5. An excerpt is displayed in Table 6.

Column 1 gives the argument (argumentum) at 1-intervals in signs and

degrees from 0s 1 to 6s 0 and its complement in 360 from 6s 0 to 11s

29. For columns 2, 3, and 4, one enters with the mean argument of lunar

anomaly, whereas for column 5 one enters with the argument of lunar latitude.

Column 2 displays the lunar equation of center (motus completus) in degrees

and minutes with a maximum of 4;56 as in the Parisian Alfonsine Tables.

244 chapter 8

argumentum motus c. motus min. min. diei latitud.

s () s () () sec. min. ()

0 1 11 29 0; 5 5 12; 9 0; 5,13

0 2 11 28 0;10 5 12; 9 0;10,27

3 0 9 0 4;55 0 13; 5 4; 0, 0*

3 1 8 29 4;55 0 13; 6 4;59,58

3 2 8 28 4;56 0 13; 8 4;59,50

3 3 8 27 4;56 0 13; 9 4;59,35

3 4 8 26 4;56 0 13; 9 4;59,15

3 5 8 25 4;56 0 13;11 4;58,51

3 6 8 24 4;56 0 13;13 4;58,21

3 7 8 23 4;56 0 13;14 4;57,45

3 8 8 22 4;55 0 13;15 4;57, 4

5 29 6 1 0; 6 6 14;25 0; 5,13

6 0 6 0 0; 0 6 14;25 0; 0, 0

divided by 60 (for purposes of interpolation). Column 4 gives the lunar velocity,

in minutes and seconds, in a minute of a day (minutum diei). The minimum

corresponds to 0;30,23/h and the maximum to 0;36,3/h: for a comparison with

other tables for lunar velocity, see Goldstein 1996. Column 5 displays the lunar

latitude, with a maximum of 5;0,0 as in the Parisian Alfonsine Tables and the

Almagest. It is surprising that the expression motus completus is used here for

the lunar equation of center, whereas in Table 5 it was used for the true solar

anomaly; clearly, it has a range of meanings and cannot be translated by a single

expression.

early alfonsine astronomy in paris 245

() () () () ()

2 3; 2 15 2;47 13 2;34 11 2;23

3 4;33 23 4;10 19 3;51 16 3;35

4 6; 4 31 5;33 25 5; 8 22 4;46

5 7;35 38 6;57 32 6;25 28 5;57

6 9; 6 45 8;20 38 7;42 33 7; 9

7 10;36 53 9;43 45 8;58 38 8;20

* In the ms, gradus velocitatis appears above this column but it refers to the headings

of the other columns, labeled: 11, 12, 13, 14.

There are two subtables for computing the time from mean to true syzygy:

see Tables 7(1) and 7(2). The first subtable is a double-argument table where,

on analogy with the other subtable, the vertical argument seems to be the

elongation between the Sun and the Moon (at 1-intervals from 1 to 7) and the

horizontal argument, the velocity in elongation (i.e., the difference between the

lunar and the solar velocities) in degrees per minute of a day (only four values

for the velocity in elongation are given: 11, 12, 13, and 14).

An entry, e, in this subtable was derived by means of the formula (expressed

in modern notation)

where is the true elongation at mean conjunction (or the result after sub-

tracting 180 at mean opposition), and the velocity in elongation, vm(t) vs(t),

is the difference between the daily velocities of the Moon and the Sun at the

time of mean syzygy. We cannot give a satisfactory explanation for the factor

16;40 (= 100/6) or for the headings of the columns indicating that the entries

are in degrees and minutes (rather than in units of time). Between these four

columns, one finds the differences, in minutes (but labeled seconds), between

two consecutive entries in the same row, to facilitate interpolation.

The second subtable is also a double-argument table giving the time in days

as a function of the elongation (at intervals of 0;1 from 0;1 to 1, or 60 minutes)

246 chapter 8

min. sec. min. sec. min. sec. min.

2 10;55 55 10; 0 46 9;14 40 8;34

60 5;27 27 5; 0 24 4;37 20 4;17

* In the ms, gradus velocitatis appears above this column but it refers to the headings

of the other columns, labeled: 11, 12, 13, 14.

and the velocity in elongation in degrees per minute of a day (again, only 4

values for the velocity in elongation are given: 11, 12, 13, and 14). Between these

four columns, one finds the differences, in minutes of a day, between successive

entries in the same row, to facilitate interpolation. Some selected rows of this

subtable are displayed in Table 7(2).

The entries in this subtable were computed by means of the formula (ex-

pressed in modern notation)

t = / [vm(t) vs(t)]

where t is the time interval between mean and true syzygy, is the true elon-

gation, and the velocity in elongation, vm(t) vs(t), is the difference between

the daily velocities of the Moon and the Sun at the time of mean syzygy. This

approach to the problem of finding true syzygy was followed by a number of

medieval astronomers and differs from that presented by Ptolemy in Almagest

vi.4 (Chabs and Goldstein 1997, pp. 9396; cf. Kremer 2003, pp. 305329).

Madrid, Biblioteca Nacional, ms 4238, f. 67r, reproduces both subtables ex-

cept that the last row of Table 7(2) corresponds to the argument of 9 min.

early alfonsine astronomy in paris 247

table 8 Correction of the lunar position for each day between syzygies ( ff. 3v4r)

motus c. min motus c. min motus c. min

s () s () s () s ()

0 24 0 10; 7 12; 1 0 22; 8 12; 6 6 6;41 14;26

1 6 0 9;24 12;11 0 21;35 12;19 6 7;38 14;10

2 0 0 21; 4 12;53

6 0 0 14;45 14;45 0 29;30 14;41 6 4;27 11;51

6 12 0 15;47 14;35 1 0;22 14;26 6 3;30 11;55

7 18 1 1;38 12;52

8 12 0 18;11 13; 7

9 6 5 29;30 13;28

11 18 0 12;48 11;48 0 24;36 11;42 6 3;27 14;46

12 0 0 11;51 11;49 0 23;40 11;47 6 4;29 14;45

ff. 3v4r Table 8: Correction of the Lunar Position for Each Day

between Syzygies

This double-argument table displays two columns for each day, from day 1 to

day 14. The days in the horizontal argument refer to the time from conjunction

to opposition. The vertical argument is given at intervals of 12, from 0s 12

to 12s 0. The heading calls it elongatio lune ab auge epicicli and it represents

the mean lunar anomaly at mean syzygy. For each day, the first column gives

the increment in lunar longitude, here called motus completus, in signs and

degrees, to be added to the mean lunar longitude at the preceding mean syzygy,

whereas the second column displays one sixtieth of the differences between

248 chapter 8

successive entries in the same row, here called motus ad minutum diei, and

given in arc-minutes. The entries in the second column thus represent the true

lunar velocity in a minute of a day for that particular day.

As mentioned above, John of Spira composed canons to some of Vimonds

tables. In particular, the canon in Vatican, Biblioteca Apostolica, ms Ottob. lat.

1826, ff. 152v153r, describes the use of Table 8, here entitled Tabula veri loci

lune ad dies datos post mediam coniunccionem vel opposicionem solis et lune.

The canon ends with an explicit reference to John Vimond working in Paris:

Explicit canon tabule sequentis que est una tabularum quas composuit

Magister Johannes Vimondi. Iste autem canon est undecimus canonum

quos composuit magister Johannes de Spira supra tabulas predicti magistri

Johannis Parisius.

On ff. 153v155v we find a copy of Table 8, but in this case the entries in the

second column (the true lunar velocity in a minute of a day) are given to one

sexagesimal place.

We know of only a few similar tables for the same purpose, but the entries in

them differ from those given by Vimond. Erfurt, ms ca 2 388, is a 15th-century

manuscript which, according to Poulle (1973), contains one of the rare copies

of John of Lignres Tabule magne. On ff. 30r32v, there is an expanded version

of Table 8, with the same structure and the same columns. In this case, the

horizontal argument runs from day 1 to day 15 and the column for velocity

gives entries in minutes and seconds per hour which result from the entries

in Vimonds Table 8 by multiplying them by 2;30 (= 60/24) for conversion from

arc to time. Another example is furnished by Levi ben Gerson (d. 1344) who

compiled a double-argument table, based on his own model, for finding the

lunar position between syzygies as a function of the number of days since

syzygy from 1 to 14 and the mean lunar anomaly at 10-intervals from 0 to 350

(Goldstein 1974, pp. 148149, 246254). Yet another such table is found in an

anonymous zij in Hebrew for year 1400: this double-argument table shares the

same structure, but the anomaly is given at intervals of the daily increment

in mean lunar anomaly from day 0 to day 27 (cf. Goldstein 2003, p. 166). The

zij of Judah ben Verga (ca. 1470) also includes a table with the same structure

(Goldstein 2001, pp. 247, 269270).

early alfonsine astronomy in paris 249

In a small table, the text gives the following values for the radices of the

planets:

Venus 3s 3;46,55

Mars 3s 15; 9,42

Jupiter 6s 4; 8, 5

Saturn 9s 14; 0,46

When recomputed for the instant of the mean conjunction on March 10, 1320,

these radices that depend on the mean longitudes or mean arguments of

anomaly (henceforth, simply anomaly) confirm the use of this date as epoch.

In the case of the superior planets the radix can be represented by the following

formula:

Rx(planet) = 0 a(Sun)

where 0 is the mean longitude of the planet at epoch, and a(Sun) is the apogee

of the Sun at that time. According to the Parisian version of the Alfonsine

Tables, the mean motions for the superior planets on that day, in Toledo at 9;10

a.m. (= 9;58 a.m. in Paris), counting from midnight, are:

Saturn 13;57, 1

Jupiter 274; 4,20

Mars 195; 5,57

If we subtract the value of the solar apogee for this epoch (89;56,15) given by

Vimond (f. 2r), we obtain:

Jupiter 184; 8, 5

Mars 105; 9,42

in perfect agreement with the radices given in the text. Note that using the

standard Alfonsine value for the solar apogee at that time (89;23,50) yields no

agreement, confirming the authors preference for his value, 89;56,15. The rea-

son for subtracting the solar apogee is that for Vimond the planetary apogees

partake in the motion of the solar apogee.

250 chapter 8

For Venus and Mercury, Vimonds radices can be obtained by adding the

planets anomaly and the solar longitude and subtracting from the sum the

value for the solar apogee at epoch. For Venus we compute according to the

Parisian Alfonsine Tables at Vimonds epoch:

= 24754;52,55d 0;36,59,27,23,59,31/d + 45;45,55,19 + 356;10,50 89;56,15

= 93;46,58

where t, the time from epoch Alfonso to epoch Vimond, is 24754;52,55d; v, the

mean motion in anomaly for Venus, is 0;36,59,27,23,59,31/d; 0(Venus), the

radix for Venuss mean anomaly at Alfonsos time, is 45;45,55,19; 0(Sun), the

mean longitude of the Sun at Vimonds epoch, is 356;10,50; and a(Sun), the

solar apogee at Vimonds epoch, is 89;56,15. This result, 93;46,58, differs from

the radix in Vimonds text by only 0;0,3.

For Mercury, we compute according to the Parisian Alfonsine Tables at

Vimonds epoch, as for Venus, where 0(Sun) a(Sun) = 266;14,35:

14,35 = 346;6,15

whereas Vimonds text has 11s 16;6,10 (= 346;6,10), in excellent agreement with

our recomputation.

A short text below these radices tells us that we should add two quantities,

the radix for the planet and the solar apogee. For Vimond the solar apogee

and each of the planetary apogees share the same motion; hence the differ-

ence between them is always the same. In particular, since Venuss apogee is

always the same as that of the Sun, nothing is given for Venus. The text then

displays values for each planet of the distance of its apogee from the solar

apogee:

Saturn 5s 12 = 162

Jupiter 2s 22 = 82

Mars 1s 14 = 44

Venus

Mercury 3s 29 = 119

These values agree closely with those of Ibn Isq (early 13th century) [Mestres

1996, p. 395]. They are used as shifts in subsequent tables for the planets,

and can be derived from the radices used in the Parisian Alfonsine Tables by

early alfonsine astronomy in paris 251

s () s () s ()

Venus 8 15; 2,47 3 12;46,53 0 3;48,53

Mars 6 1;10, 7 0 26;51,41 3 1;58,32

Jupiter 1 1;53,32 2 1;19,53 8 2;52,21

Saturn 0 12;50,22 0 24;41,28 3 7;46,36

subtracting the solar apogee for the time of Alfonso from the radix of the apogee

for each planet (see, e.g., the editio princeps of the Alfonsine Tables printed

by Ratdolt (1483), c8d1; note that the signs used there are physical signs of

60):

Jupiter 2,42;48,38,41 1,20;37,0 = 82;11,38

Mars 2, 4;23,51,41 1,20;37,0 = 43;46,51

Venus 1,20;37, 0 1,20;37,0 = 0

Mercury 3,19;51,11,41 1,20;37,0 = 119;14,11

These results, when rounded to the nearest degree, are in perfect agreement

with Vimonds data. Therefore, the conclusion is that Vimond started with the

same planetary apogees as those in the Parisian Alfonsine Tables.

This table displays the radices for the five planets for 1, 2, 3, and 4 years;

for multiples of 4 years up to 76 years; and then for 152, 304, 608, 1216, and

2432 years, as in Table 3. As was the case for the radices for the Sun and the

Moon, 1 year is equivalent to 13 mean lunations, 2 years is equivalent to 25 mean

lunations, , 8 years is equivalent to 99 mean lunations, etc.

The mean daily motion in longitude resulting from the entries for year 8

(computed in the same way that was used for finding the mean motions in

Table 3) are shown below under the heading Vimond. If we add the daily

motion of the apogees (0;0,0,11,13,35/d), as we did in the case of the Sun, we

252 chapter 8

obtain the entries displayed in the second column, in good agreement with the

values for the mean motions in longitude in the Parisian Alfonsine Tables (see

Ratdolt 1483).

Vimond of the apogee Parisian Alf. t.

Jupiter 0; 4,59, 4, 1,19 0; 4,59,15,14,54 0; 4,59,15,27, 7

Mars 0;31,26,27,26,34 0;31,26,38,40, 9 0;31,26,38,40, 5

Venus 1;36, 7,35,47,21 1;36, 7,47, 0,56 1;36, 7,47, 1,19

Merc. 4; 5,32,16, 5,55 4; 5,32,27,19,30 4; 5,32,27,20, 0

It is most unusual for the mean motions of Venus and Mercury to be the sum of

their mean motions in anomaly and the solar mean motion, but there can be no

doubt that this is what Vimond did, as is confirmed by the note on f. 4vb. In fact,

we know of no other medieval astronomer writing in Latin who presented the

mean motions of the inferior planets in this way. For purposes of comparison,

the entries for Venus and Mercury under Parisian Alfonsine Tables are the

sum of the mean motions in anomaly and the solar mean motion: for Venus

0;36,59,27,24,0/d and 0;59,8,19,37,19/d, and for Mercury 3;6,24,7,42,41/d and

0;59,8,19,37,19/d. Note that in Ptolemys models the solar mean motion is also

the mean argument of center for Venus and Mercury.

This table displays the radices for the five planets for 25 consecutive syzygies.

The entries in this table are based on the same motions as those embedded in

the previous table. As was the case for the monthly radices in Table 4, for each

planet the entries for Syzygy 1 and Syzygy 25 add up to the entry corresponding

to Year 1 in the previous table (except for 1 for Mercury, Mars, and Jupiter).

For Venus and Mercury the mean motions extracted from Table 9 give exact

agreement, confirming the interpretation given above. Thus, in the cases of

Venus and Mercury one has obtained the sum of the solar anomaly and their

mean anomalies, respectively, at any syzygy (see Figure 8.1). This quantity is not

the argument in the table of equations (see Tables 12 and 15, below), and it is

early alfonsine astronomy in paris 253

s () s () s ()

Venus 0 23;39,20 1 17;18,41 7 21;23,27

Mars 0 7;44,14 0 15;28,28 6 13;25,52

Jupiter 0 1;13,36 0 2;27,12 1 0;29,57*

Saturn 0 0;29,38 0 0;59,16 0 12;20,44

of Vimonds tables for the mean

motion for Venus

not clear that there is any advantage to this method as against computing the

mean anomaly directly.

f. 6v), and 23 (Saturn, f. 7r): Equation of Center and First Station

Tables 11 and 12 are to be used together to compute the true longitude of a planet

from its mean longitude. In most zijes in the Ptolemaic tradition, there is only

one such table for each planet, but Vimond has separated those functions that

depend on the mean argument of center from those that depend on the mean

anomaly and put them in different tables. A similar idea is already found in

the zij of Ibn Isq, described in Mestres 1996. Ibn Isqs parameters for the

maximum equations of center for Mars and Mercury are those of al-Battn,

254 chapter 8

but for Saturn, Jupiter, and Venus they are not; rather, they are 5;48 for Saturn,

5;41 for Jupiter, and 1;51 for Venus. The tables for planetary equations () are

divided into two groups: the first group contains the tables for the equation of

centre and the interpolation function. () The second group (two tables for

each planet) contains the tables for the equations of anomaly at apogee and

perigee and for the middle position (Mestres 1999, p. 234). So, the arrangement

of Vimonds tables bears a similarity to an Andalusian/Maghribi tradition that

is not otherwise attested in Latin.

However, it is not uncommon to find later sets of tables associated with

the Parisian Alfonsine Tables where the planetary equations are split into two

tables for each planet: see, e.g., Erfurt, ms ca q 362, ff. 28r36r, where the entries

are displayed at intervals of 1 and the radices are given for Paris (1320) as well

as for London and Brugge (1366).

Besides offering two tables for the equations of each planet, Vimonds tables

give additional information arranged in a presentation which is certainly pecu-

liar, as explained below.

s () s () min min min s ()

0 12 0 14;57 60;30 60 59 4 24;30

0 18 0 21; 0 60;20 59 58 4 24;32

0 24 0 27; 2 59;40 59 57 4 24;35

1 0 1 3; 0 59;30 59 54 4 24;38

1 6 1 8;57 59; 0 58 51 4 24;44

1 12 1 14;51 58;50 58 48 4 24;50

1 18 1 20;44 58;20 57 44 4 24;56

1 24 1 26;34 58;10 57 40 4 25; 7

2 0 2 2;23 58;10 57 35 4 25;20

2 6 2 8;12 57;50 57 29 4 25;37

2 12 2 13;59 57;40 57 24 4 25;53

2 18 2 19;45 57;30 57 19 4 26; 9

2 24 2 25;30 57;30 57 14 4 26;24

3 0 3 1;15 57;30 57 10 4 26;38

3 6 3 7; 0 57;30 57 6 4 26;50

3 12 3 12;45 57;30 57 4 4 27; 2

3 18 3 18;30 57;30 57 2 4 27; 9

early alfonsine astronomy in paris 255

s () s () min min min s ()

4 0 3 29;59 57;10 56 0 4 27;14

4 6 4 5;40 57;30 57 1 4 27;12

4 12 4 11;25 57;30 57 3 4 27; 7

4 18 4 17;10 57;30 57 5 4 26;59

4 24 4 22;55 57;30 57 7 4 26;46

5 0 4 28;40 57;30 57 11 4 26;34

5 6 5 4;25 57;30 57 16 4 26;19

5 12 5 10:10 57;40 57 21 4 26; 4

5 18 5 15:56 58; 0 57 26 4 25;48

5 24 5 21;44 58;10 57 31 4 25;30

6 0 5 27;33 58;10 57 37 4 25;16

6 6 6 3;22 58;30 58 41 4 25; 3

6 12 6 9;13 59; 0 58 45 4 24;54

6 18 6 15; 7 59;10 58 49 4 24;48

6 24 6 21; 2 59;30 59 52 4 24;42

7 0 6 26;59 59;50 59 55 4 24;37

7 6 7 2;58 60;30 60 57 4 24;34

7 12 7 9; 1 60;40 60 59 4 24;31

7 18 7 15; 5 61; 0 60 60 4 24;30

7 24 7 21;11 61;40 61 60 4 24;29

8 0 7 27;21 61;50 61 60 4 24;29

8 6 8 3;32 62; 0 61 59 4 24;29

8 12 8 9;44 62;10 61 59 4 24;30

8 18 8 15;57 62;20 61 58 4 24;32

8 24 8 21;11 62;30 62 57 4 24;34

9 0 8 28;26 62;40 62 56 4 24;36

9 6 9 4;44 63; 0 62 55 4 24;38

9 12 9 11; 2 63;10 62 54 4 24;39

9 18 9 17;21 63;10 62 54 4 24;40

9 24 9 23;44 63;20 62 53 4 24;41

10 0 10 0; 3 63;10 62 53 4 24;42

10 6 10 6;22 63;10 62 53 4 24;41

10 12 10 12;41 63;10 62 54 4 24;40

10 18 10 19; 0 63; 0 62 54 4 24;39

10 24 10 25;18 62;50 62 55 4 24;37

256 chapter 8

s () s () min min min s ()

11 6 11 7;51 62;40 62 57 4 24;33

11 12 11 14; 7 62;10 61 58 4 24;31

11 18 11 20;20 62; 0 61 59 4 24;30

11 24 11 26;32 61;50 61 60 4 24;29

12 0 12 2;43 61;20 60 60 4 24;29

s () s () min min min s ()

0 12 0 11;34 57;50 57 0 5 15;54

3 0 2 27;50 60; 0 59 31 5 17;11

3 6 3 4;50 * 60;20 59 33 5 17;17

3 12 3 9;52 60;30 60 36 5 17;23

8 24 8 26;10 60; 0 59 33 5 17;17

9 0 9 2;10 59;50 59 31 5 17;11

9 6 9 8; 9 59;30 59 27 5 17; 2

12 0 12 0; 0 57;50 57 0 5 15;50

The table for the equation of center of each of the five planets has six columns.

Column 1 gives the argument (argumentum) at 6-intervals in signs and degrees

from 0s 6 to 12s 0. Column 2 displays the entry in col. 1 corrected for the equa-

tion of center (motus completus), in signs, degrees, and minutes. The author

early alfonsine astronomy in paris 257

s () s () min min min s ()

0 12 0 17;36 50; 0 26 4 5 8;21

1 18 1 17;16 49;10 25 0 5 7;31

7 18 7 18;54 73;10 38 59 5 19;13

12 0 12 7;24 51;10 26 8 5 9; 6

follows here the same pattern as that for the true solar anomaly (see Table 5).

Column 3 (motus gradus) gives the increment of the true argument per degree

of the argument, in minutes and seconds. Most entries in this column are gen-

erated by dividing by 6 the differences between two successive entries in col. 2

and thus were probably intended for interpolation in col. 2. Column 4 (motus

diei) displays the velocity in minutes of arc per day, and the range of values

for each planet is the same as in the column labeled motus centri or motus

puncti (that only depends on the argument of center) in the table for planetary

velocities associated with the Toledan Tables and the Castilian Alfonsine Tables

(Chabs and Goldstein 2003a, pp. 170182); for the other component of the plan-

etary velocity, see Tables 12, 15, 18, 21, and 23, col. 4, below. So, the entries in this

column are only one component of the planets velocity. Column 5 is intended

to provide minutes of interpolation and is headed diametri (perhaps to distin-

guish these linear minutes from minutes of an hour, minutes of a day, and

minutes of a degree). Finally, column 6 lists the first station in signs, degrees,

and minutes.

But for a shift of the entries, the equations of center for Mercury, Mars, and

Saturn that can be derived from cols. 1 and 2 are basically the same (with

258 chapter 8

s () s () min min min s ()

0 12 0 17;31 57;20 5 20 4 5; 9

2 12 2 12;59 44;10 4 0 4 4; 6

2 18 2 18;23 44; 0 4 0 4 4; 5

2 24 2 23;48 44; 0 4 0 4 4; 5

3 0 2 29;12 44;10 4 0 4 4; 6

8 18 8 17;33 66;30 6 60 4 7;11

8 24 8 14;14 66;20 6 60 4 7;11

9 0 9 0;52 66;20 6 60 4 7;10

11 24 11 29;56 59;10 5 29 4 5;39

12 0 12 0;51 * 58;40 5 26 4 5;29

the Toledan Tables (Toomer 1968, pp. 6068; F.S. Pedersen 2002, pp. 1259

1308).

The maximum value for Mercury (3;2) occurs at about 0s 24 and 7s 6, that

of Mars (11;24) at 4s 18 and 10s 12, and that of Saturn (6;31) at 2s 12 and 8s

12. However, for the other two planets the entries differ systematically from

those in the above-mentioned zijes: for Venus the maximum value is 2;10 at

3s 0 and 3s 6, and 8s 24 and 9s 0; and for Jupiter the maximum value is

5;57 at 5s 24 and 11s 18. The entries for Mercury, Mars, Jupiter, and Saturn

are shifted by about 119, 44, 82, and 162, respectively, in relation to those

in the zij of al-Battn and the Toledan Tables. No such shift appears in the

table for Venus. As mentioned above, these shifts result from the difference

between the apogee of each of the planets and that of the Sun. Because of

these shifts, for the superior planets one enters these tables in col. 1 directly

early alfonsine astronomy in paris 259

s () s () min min min s ()

0 12 0 15;24 66; 0 2 55 3 25;19

5 12 5 12; 0 53;30 2 0 3 22;44

5 18 5 17;21 53;40 2 0 3 22;45

11 12 11 18;43 67;10 2 60 3 25;28

11 24 11 25;25 67; 0 2 59 3 25;27

12 0 12 2; 7 66;30 2 58 3 25;25

with their mean motions for a given syzygy (the radix plus the motion in

years and semi-lunations); for Venus and Mercury one enters with the solar

anomaly for a given syzygy. Clearly, Vimond intended to make this table more

user-friendly than the standard version of the table for the equation of cen-

ter.

Vimond has a double motion of the solar apogee: precession and proper

motion. The planetary apogees are fixed with respect to the solar apogee (i.e.,

they are subject to both precession and the proper motion of the solar apogee).

If we add the solar apogee (about 90) to the values for the shifts listed above,

we find that the planetary apogees are 209 for Mercury, 90 for Venus, 135 for

Mars, 172 for Jupiter, and 252 for Saturn. In the Toledan Tables, the apogees of

the Sun and of Venus are both 77;50 (Toomer 1968, p. 45), that is, about 12 less

than 90. Adding this difference to the planetary apogees in the Toledan Tables

rounded to degrees, we find the following:

260 chapter 8

Apogees

v (from the shifts) v (from the radices) tt + 12

Venus 90 90 90

Mars 135 134 134

Jupiter 172 172 176

Saturn 252 252 252

Mercury 3; 2 3; 2 3; 2 3; 2

Venus 1;59 1;59 2;10 2;10

Mars 11;24 11;24 11;24 11;24

Jupiter 5;15 5;15 5;57 5;57

Saturn 6;31 6;31 6;31 6;31

The agreement of Vimonds data with the apogees in the Toledan Tables

shows that Vimond has included the motion of the solar apogee in the motions

of the planetary apogees, thus following a theory for which there was no

previous evidence outside al-Andalus and the Maghrib (Sams and Mills 1998,

pp. 268270). We know of no other set of planetary equation tables arranged in

this way. See also Table 27 (equation of access and recess), below, for yet another

shift in Vimonds tables.

The maximum values for the equation of center in Vimonds planetary

tables are the same as in the editio princeps of the Alfonsine Tables (see Table

11a).

Despite their agreement for the values of the maximum equations, the

structure of Vimonds tables is very different from that of the Parisian Alfonsine

Tables and would seem to be independent of it. Moreover, it is significant that

the maximum equation of center for Jupiter in both cases is 5;57, for this value

is not known in any text prior to the Parisian Alfonsine Tables, indicating a

strong connection between the tables of Vimond and the work of his Parisian

contemporaries. The origin or derivation of this parameter for Jupiter is not

described in any extant text, and it is likely that this value was simply taken

early alfonsine astronomy in paris 261

from an earlier work: the most reasonable candidate is the Alfonsine Tables as

they existed in Castile.

For all planets, except Mercury, an entry, c, in column 5 can be computed,

but for shifts, from the modern formula

c = 60 (1 cos )/2,

where is the mean argument of center. The same approach is found in Levis

lunar theory (Goldstein 1974, table 35, col. ii: see p. 54).

The entries for Mercury in col. 5 do not follow the same pattern as that for the

rest of the planets. The entries can be recomputed, approximately, according to

the following formula:

where d is the maximum distance of the center of the epicycle from the

observer, d is the minimum distance, and r() is the distance as a function of

the mean argument of center, .

A similar formula for interpolation was already used by abash in the 9th

century (as-Saleh 1970, pp. 137138). In Ptolemys model for Mercury d is

69 for argument 0, d is 55;34 for an argument close to 120, and r(180)

is 57 (O. Pedersen 1974, pp. 313324). Hence, formula [1] can be replaced

by

is likely that Vimond (or his source) used approximations (if this, indeed, was

the formula he had in mind). We computed the distances from the observer

to the center of Mercurys epicycle according to formulas in modern terms

given by O. Pedersen (1974, p. 320, equations 10.34 and 10.35), and then used

them in equation [2], above. A comparison of our results for c5() with the

entries in Vimonds table is displayed in Table 11b. Col. ii has the values for

c5() that depend on the distances computed according to the formulas given

by O. Pedersen and equ. [2], above; col. iii has the arguments in Vimonds

table (with the shift); and col. iv has the entries in Vimonds Table 11, col. 5.

Although the agreement is not exact between col. ii and col. iv, the trend is

clear. Vimonds value for 180 (col. iii), 37, has the poorest agreement, but this

entry should probably be corrected to 36, judging from the surrounding

values.

262 chapter 8

figure 8.2 Vimonds equation of center, col. 5, for Venus, Mars, Jupiter, and Saturn

early alfonsine astronomy in paris 263

Mercury with recomputation

i ii iii iv

c5( ) (Vimond) c5( ): Vimond

0 0; 0 120 0

30 10;50 150 11

54 29;24 174 31

60 34;15 180 37

66 38;53 186 41

90 53;33 210 55

120 60; 0 240 60

150 56;38 270 56

180 53;34 300 53

(Copernicus 1543, ff. 177v178r), his col. 4 (for interpolation) shows the same

trend as Vimonds col. 5. We are convinced that column 5 in Vimonds tables for

the equation of center is intended to be used for interpolation with column 5

in the tables for the equation of anomaly, and this is analogous to Copernicuss

use of his col. 4 (see below). Indeed, Vimonds col. 5 serves much the same

purpose as col. 8 in Ptolemys tables for the planetary equations (Almagest,

xi.11) but, since the definitions for the columns that yield the equation of

anomaly are different, so is the function for interpolation. Moreover, in contrast

to the geometric methods in the Almagest used for computing the coefficients

of interpolation for each of the four planets (Venus, Mars, Jupiter and Saturn),

Vimond has approximated the results that would be derived from the geometry

of the models by introducing a single trigonometric function in those cases.

In Almagest, xi.11 (Toomer 1984, pp. 549553), col. 8 in the planetary equa-

tion tables is intended for interpolation as a function of , the mean argument

of center, and the entries are given to minutes and seconds (for Ptolemys

method of computation and a graph of the entries in his col. 8, see Neugebauer

1975, pp. 184186, 1267). A similar set of values, given only to minutes, is found in

al-Battns zij in the tables for the planetary equations, col. iv (Nallino 1903

1907, 2:110137), and in corresponding tables in the Parisian Alfonsine Tables,

col. 3 (Ratdolt 1483, e7rg5v).

As for the entries for the first station of each planet, they are essentially

the same as in previous tables of the same kind (Almagest, Handy Tables,

264 chapter 8

al-Khwrizm, al-Battn, and the Toledan Tables) with the same shifts that we

noted above.

(Jupiter, f. 7r), and 24 (Saturn, f. 7v): Equation of Anomaly

The tables for the equation of anomaly for each of the five planets have seven

columns. Table 12 displays a selection of values for the equation of anomaly for

Mercury.

Column 1 gives the mean argument of anomaly (argumentum) at 6-intervals

(at 3-intervals for Mars and Venus) from 0s 6 to 6s 0 and its complement

in 360 from 6s 0 to 11s 24. Column 2 displays the correction due to the

argument of anomaly at maximum distance (motus completus) in degrees and

minutes and represents the difference between the equation of anomaly and

the correction for maximum distance (cf. Almagest, xi.11, columns 6 and 5; and

Neugebauer 1975, pp. 183184). The only other text of which we are aware that

treats the equation of anomaly in this way is the zij of Ibn al-Bann (d. 1321)

where this presentation is applied in his tables for Saturn and Jupiter but not

in those for the other planets (see Sams and Mills 1998, pp. 278285). The

extremal values in col. 2 that appear in the text are shown below; they are

followed by the corresponding entries for col. vi and col. v in the zij of al-Battn

(Nallino 19031907, 2:109137):

Vimond al-Battn

Venus 44;49 (= 45;59 1;10) at 4s 15

Mars 36;44 (= 40;58 4;14) at 4s 6

Jupiter 10;34 (= 11; 3 0;29) at 3s 12

Saturn 5;53 (= 6;12 0;19) at 3s 0 and

(= 6;13 0;20) at 3s 6

These corrections agree with those that follow from the Almagest as well as

the zij of al-Battn, the Toledan Tables, and the editio princeps of the Alfonsine

Tables (with minor variants: 40;59 rather than 40;58 for Mars; 6;12 and 0;19

correspond to 3s 0 rather than 3s 1 for Saturn), and this means that Ptolemys

eccentricities underlie them even though, in the case of Venus and Jupiter,

early alfonsine astronomy in paris 265

s () s () () min min min sec sec

0 6 11 24 1;28 15 45 0;18 2 8

0 12 11 18 2;56 15 45 0;33 2 8

0 18 11 12 4;24 14 44 0;48 2 8

0 24 11 6 5;50 14 44 1; 3 2 8

1 0 11 0 7;15 14 42 1;18 2 8

1 6 10 24 8;37 13 42 1;33 2 8

1 12 10 18 9;58 13 40 1;48 2 8

1 18 10 12 11;15 12 39 2; 0 2 8

1 24 10 6 12;30 11 36 2;18 3 9

2 0 10 0 13;39 10 34 2;35 3 9

2 6 9 24 14;44 10 31 2;53 3 9

2 12 9 18 15;44 9 28 3;10 4 8

2 18 9 12 16;38 8 25 4;14 4 8

2 24 9 6 17;25 7 19 4;30 4 8

3 0 9 0 18; 4 5 15 4;45 4 8

3 6 9 24 18;34 3 9 4;57 4 8

3 12 8 18 18;53 1 3 5; 5 4 8

3 18 8 12 19; 1 1 3 5;10 3 6

3 24 8 6 18;56 3 9 5;13 2 4

4 0 8 0 18;39 6 18 5; 6 1 3

4 6 7 24 18; 4 9 27 4;55 1 1

4 12 7 18 17;12 11 35 4;29 0 1

4 18 7 12 16; 4 15 45 4;55 2 6

4 24 7 6 14;36 18 55 4;12 4 13

5 0 7 0 12;49 21 66 4;29 6 18

5 6 6 24 10;42 24 74 3;55 7 22

5 12 6 18 8;18 26 81 3;12 9 29

5 18 6 12 5;42 28 87 2;15 11 34

5 24 6 6 2;53 29 89 1;10 12 36

6 0 6 0 0; 0 29 89 0; 0 12 36

the eccentricities were modified for computing the equation of center (cf.

North 1976, 3:196). Similarly, in the tables of Ibn al-Bann the eccentricities

underlying the equations of anomaly are taken from the Almagest, but his

266 chapter 8

maximum equations of center for Venus and Jupiter are not those of either

Ptolemy or of Vimond (Sams and Mills 1998, p. 276).

Column 3 (motus gradus) gives the increment of the motus completus in

col. 2 per degree of the argument in minutes: in most cases the entry results

from taking the difference between successive entries in col. 2 and dividing

that difference by 6 (or by 3 for Mars and Venus); the purpose of this column

is facilitate interpolation. Column 4 (motus diei) displays the velocity in min-

utes of arc per day, and the range of values for each planet is the same as in

the column labeled motus argumenti (that only depends on the argument of

anomaly) in the table for planetary velocities; see the comments to the Castil-

ian Alfonsine Tables, chapter 27 (Chabs and Goldstein 2003a, pp. 170182). So,

an entry in this column is the second component of the planets velocity and it

complements the first component already displayed in Tables 11, 14, 17, 20, and

23, above. The entries in column 5 (minutum diametri), in minutes and seconds,

actually represent degrees and minutes, and result from adding the correction

for maximum distance to the correction for minimum distance (columns c5

and c7 in Almagest, xi.11). For the extremal values in col. 5 in the text see below;

they are followed by the corresponding entries for col. v and col. vii in the zij

of al-Battn (Nallino 19031907, 2:109137):

Vimond al-Battn

Venus 3;34 (= 1;42 + 1;52) at 5s 12

Mars 13;37 (= 5;34 + 8; 3) at 5s 9

Jupiter 1; 3 (= 0;30 + 0;33) at 3s 24

Saturn 0;46 (= 0;21 + 0;25) at 3s 12

4;56, whereas for 3s 24 in Vimonds table we find

3;12 + 2;1 = 5;13, al-Battns maximum which

occurs at 4s 104s 12.

correction to the planets mean longitude is to be computed, but it seems likely

that one component of this correction is to be computed by adding an entry in

col. 2 to an interpolation factor times an entry in column 5, as is the case with

the tables of Ibn al-Bann for Saturn and Jupiter. The most likely candidate for

this interpolation factor is col. 5 in Table 11, for it depends on the argument of

early alfonsine astronomy in paris 267

(The entries from 5s 18 to 6s 12 are given at 2-intervals, rather than at 3-intervals

as in the rest of the table.)

s () s () () min min min sec sec

0 3 11 27 1;15 25 15 0; 2 0 0

0 6 11 24 2;30 25 15 0; 3 1 0

4 12 7 18 44;44 2 1 2;18 2 1

4 15 7 15 44;49 2 1 2;25 2 1

4 18 7 12 44;44 6 4 2;32 3 2

5 12 6 18 33;25 74 46 3;34 2 1

5 15 6 15 29;43 89 55 3;27 4 2

5 18 6 12 25;25 104 64 3;14 7 4

6 0 6 0 0; 0 144 89 0; 0 22 14

(The entries from 5s 18 to 6s 12 are given at 2-intervals, rather than at 3-intervals

as in the rest of the table.)

s () s () () min min min sec sec

0 3 11 27 1; 8 23 11 0; 8 3 1

0 6 11 24 2;16 23 11 0;17 3 1

4 3 7 27 36;40 1 1 8;53 9 4

4 6 7 24 36;44 0 0 9;19 9 4

4 9 7 21 36;43 3 1 9;46 9 4

5 6 6 24 28;15 46 21 13;30 0 0

5 9 6 21 25;56 53 25 13;37 6 2

5 12 6 18 23;17 62 29 13;19 13 6

268 chapter 8

s () s () () min min min sec sec

5 28 6 2 3; 1 90 42 2;29 74 35

6 0 6 0 0; 0 90 42 0; 0 74 35

s () s () () min min min sec sec

0 6 11 24 0;57 9 8 0; 4 1 1

0 12 11 18 1;52 9 8 0; 8 1 1

3 6 8 24 10;33 0 0 0;59 0 0

3 12 8 18 10;34 1 1 1; 1 0 0

3 18 8 12 10;29 2 2 1; 2 0 0

3 24 8 6 10;15 3 3 1; 3 0 0

4 0 8 0 9;54 5 4 1; 2 0 0

5 24 6 6 1;21 13 12 0; 9 1 1

6 0 6 0 0; 0 13 12 0; 0 1 1

s () s () () min min min sec sec

0 6 11 24 0;34 5 5 0; 3 1 1

0 12 11 18 1; 7 5 5 0; 7 1 1

2 24 9 6 5;46 1 1 0;41 0 0

3 0 9 0 5;53 0 0 0;42 0 0

3 6 8 24 5;53 0 0 0;44 0 0

3 12 8 18 5;51 1 1 0;46 0 0

s () s () () min min min sec sec

5 24 6 6 0;42 7 7 0; 7 1 1

6 0 6 0 0; 0 7 7 0; 0 1 1

superior planets and Venus (not to

scale)

center as it should (see Sams and Mills 1998). Column 6 (motus gradus) seems

to be the increment per degree of argument of the entries in col. 5: in many

cases the entry in col. 6 results from taking the difference between successive

entries in col. 5 and dividing it by 6 (or by 3 for Mars and Venus), and it is for

purposes of interpolation. The entries in col. 6 are given in seconds. The entries

in column 7 (motus diei) are also given in seconds; they are probably associated

with those in the preceding column, for in all cases columns 6 and 7 have their

extremal values for the same arguments, but we have failed to identify their

specific purpose.

Figure 8.4 displays Ptolemys model for the three superior planets and Venus.

o is the observer, d is the center of the deferent circle rac, and e is the equant

point, such that the eccentricity, e = od = de. a is the planets apogee, and =

angle aec, the mean argument of center, is measured from it to the center of the

epicycle about point e. Angle gcp is the mean argument of anomaly, , and the

planet is at point p. Angle hcg is the equation of center and it is also applied to

correct the mean argument of anomaly to yield the true argument of anomaly,

= angle hcp. In the case of the superior planets, cp, the direction from the

270 chapter 8

center of the epicycle to the planet, is always parallel to s, the direction from

the observer to the mean Sun. In the case of Venus, ec is parallel to the direction

from the observer to the mean Sun. The goal is to find the direction from the

observer to the planet, i.e., angle rop is the longitude of the planet, and r is in

the direction to Aries 0.

The mean argument of center, , is angle aec, and the true argument of

anomaly, , is angle hcp. With these arguments, and , we can determine

the equation of anomaly, c(), with Vimonds tables and compare the result

with computations based on the Parisian Alfonsine Tables. According to our

understanding of Vimonds procedure,

where ci refers to the i-th column in the table. Note that c5() is taken from the

table for the equation of center (with the shifts), and c5() is taken from the

table for the equation of anomaly. For instance, for Venus, when = 120 and

= 135

c() = 44;49 + 0;45 2;25

c() = 46;38.

With the same arguments for Venus in the Parisian Alfonsine Tables, we find

c() = c5(135) + c3(120) c6(135)

c() = 45;59 + 0;31 1;15

c() = 46;38

and this is exactly what resulted from Vimonds tables. In the tables for the

planetary equations in Almagest xi.11 and its derivatives in al-Battn and in the

Parisian Alfonsine Tables (among others), the rules for computing the equation

of anomaly require careful attention to algebraic signs. Vimond simplified the

rules for this computation, making his tables more user-friendly. A similar

procedure is described by Copernicus for using his planetary tables in De

revolutionibus, v.23, to compute the equation of anomaly (Copernicus 1543,

ff. 173v179r; cf. Swerdlow and Neugebauer 1984, p. 453).

early alfonsine astronomy in paris 271

f. 7r), and 25 (Saturn, f. 7v): Planetary Latitudes

The tables for the planetary latitudes, both for the superior and the inferior

planets, are in the style of Almagest xiii.5, the zij of al-Battn (Nallino 1903

1907, 2:140141), and some tables associated with the Toledan Tables (Toomer

1968, pp. 7172; F.S. Pedersen 2002, pp. 13221326), as opposed to those in the

Handy Tables and those in the zij of al-Khwrizm.

The table for the planetary latitudes of Mercury has seven columns; the table

for Venus lacks the seventh; and the tables for the superior planets have only

five columns (i.e., cols. 1, 3, 4, 5, and 6).

In all cases column 1 displays the argument (argumentum) at 12-intervals

from 0s 12 to 12s 0. Column 2 is only found in the tables for the inferior planets

and the entries are given in minutes. The heading is radix meridionalis in the

case of Mercury and radix septentrionalis in that for Venus. This column is for

determining the deviation, otherwise called the third component of latitude,

that is, the inclination of the plane of the deferent with respect to the ecliptic.

The entries for deviation can be derived from:

3 = +0;10 c5 for Venus

where c5 is the column for the minutes of proportion in the table for planetary

latitude in Almagest xiii.5 (given there in minutes and seconds). As will be

seen, column 5 for Venus in Table 16, given only to minutes, corresponds to

c5 in Almagest xiii.5. It is noteworthy that column 2 for Mercury is shifted

downwards about 119 whereas there is no shift in the case of Venus. This is

exactly the same feature we noticed in the tables for the equation of center and

the amount of the shift is the same. The column for deviation is certainly not

a common feature in medieval tables (for a survey of the few that have them,

see Goldstein and Chabs 2004), and Vimonds is the earliest set of tables in the

West we know to display such columns.

For the inferior planets, columns 3 and 5 (diametri) give the minutes of

proportion for the inclination and the slant, respectively. We note that columns

3 and 5 for Mercury also exhibit a shift of less than 120, and that no shifts appear

in the case of Venus. We also note that column 5 for Venus lists the rounded

values in the column for the sixtieths found in the corresponding table in the

Almagest xiii.5, the zij of al-Battn, etc.

For the superior planets, columns 3 and 5 give the minutes of proportion for

the northern and southern latitudes, respectively, of the planets. Only half of

272 chapter 8

s () min min min * min min * sec

0 12 13 57 1;44 17 0;12 1

0 24 4 60 1;40 5 0;44 4

1 6 5 60 1;39 7 1; 6 7

1 18 14 57 1;16 19 1;26 9

2 0 23 51 0;59 31 1;44 10

2 12 31 44 0;38 41 2; 0 12

2 24 37 34 0;16 49 2;14 13

3 6 41 23 0;15 55 2;25 14

3 18 44 11 0;48 59 2;29 15

4 0 45 0 1;25 60 2;29 15

4 12 43 13 2; 6 58 2;20 14

4 24 40 25 2;47 54 2; 0 13

5 6 36 36 3;26 48 1;29 9

5 18 29 45 3;54 39 0;48 5

6 0 22 52 4; 5 29 0; 0 0

6 12 13 57 3;54 17 0;48 5

6 24 4 60 3;26 5 1;29 9

7 6 5 60 2;47 7 2; 0 12

7 18 14 52 2; 6 19 2;20 14

8 0 23 51 1;25 31 2;29 15

8 12 31 44 0;48 41 2;29 15

8 24 37 34 0;15 49 2;25 14

9 6 41 23 0;16 55 2;10 13

9 18 44 11 0;38 59 2; 0 12

10 0 45 1 0;59 60 1;44 10

10 12 43 13 1;16 58 1;26 9

10 24 40 25 1;30 54 1; 6 7

11 6 36 36 1;40 48 0;44 4

11 18 29 45 1;44 39 0;12 1

12 0 22 52 1;46 29 0; 0 0

early alfonsine astronomy in paris 273

s () min min min * min min *

0 12 10 12 1; 1 59 0;16

0 24 9 24 0;59 55 0;33

1 6 8 35 0;55 48 0;49

1 18 7 44 0;46 40 1; 5

2 0 5 52 0;35 30 1;20

2 12 3 57 0;29 18 1;35

2 24 1 60 0;18 6 1;50

3 6 1 60 0;10 6 2; 3

3 18 3 52 0;32 18 2;15

4 0 5 44 0;59 30 2;25

4 12 7 35 1;38 40 2;30

4 24 8 24 2;23 48 2;28

5 6 9 12 3;44 55 2;12

5 18 10 0 5;13 59 1;27

6 0 10 12 7;12 60 0; 0

6 12 10 24 5;13 59 1;12

6 24 9 35 3;44 55 2;28

7 6 8 44 2;23 48 2;30

7 18 7 52 1;38 40 2;25

8 0 5 57 0;59 30 2;15

8 12 3 60 0;32 18 2; 3

8 24 1 60 0;10 6 1;50

9 6 1 57 0;19 6 1;35

9 18 3 52 0;29 18 1;20

10 0 5 44 0;35 30 1; 5

10 12 7 44 0;46 40 0; 5 **

10 24 8 35 0;55 48 0;49

11 6 9 24 0;59 55 0;33

11 18 10 12 1; 1 59 0;16

12 0 10 0 1; 3 60 0; 0

** Sic.

274 chapter 8

s () min min * min min *

0 12 50 0; 9 m 0; 4

0 24 55 0;13 m 0; 6

1 6 59 0;16 m 0; 9

1 18 59 0;21 m 0;15

4 12 8 2; 1 m 2;10

4 24 s 2;34 10 2;56

6 0 s 4;21 43 7;30

10 12 s 0;21 2 0;15

10 24 10 0;16 [blank] 0; 9

11 6 22 0;13 m 0; 6

11 18 33 0; 9 m 0; 4

12 0 43 0; 6 m 0; 2

the columns are filled with numbers, the others have capital letters indicat-

ing North [s] and South [m]. Column 3 is shifted about 45 (Mars), about

100 (Jupiter), and about 110 (Saturn) in relation to the corresponding columns

in the Almagest, whereas the shifts for column 5 are increased by 180 in

each case. These shifts are totally consistent with those found for the equa-

tion of center (about 44, 82, and 162 for Mars, Jupiter, and Saturn, respec-

tively). Indeed, subtracting these numbers for each planet, we find 0 (Mars),

about 20 (Jupiter), and +50 (Saturn), in perfect agreement with the dif-

ferences given by Ptolemy in Almagest xiii.6 between the northern limits

on the deferent and the apogees of each superior planet, respectively. Thus,

it is quite clear that the compiler of Vimonds tables, whether Vimond or

not, had a good understanding of this difficult issue as it is presented in the

Almagest.

Columns 4 and 6 display the inclination (declinatio minuti diametri) and

the slant (reflexio minuti diametri) for the inferior planets, and the entries are

given in degrees and minutes, despite the headings, which read minutes and

seconds. For the superior planets, these two columns display the northern and

early alfonsine astronomy in paris 275

s () min min * min min *

0 12 10 1; 8 0 1; 6

0 24 12 1; 9 m 1; 7

3 6 60 1;33 m 1;33

3 18 60 1;39 m 1;39

6 0 12 2; 5 m 2; 8

6 12 0 2; 3 0 2; 6

6 24 s 2; 0 12 2; 3

9 6 s 1;27 60 1;26

9 18 s 1;21 60 1;21

11 18 s 1; 8 34 1; 6

12 0 s 1; 6 12 1; 5

degrees and minutes.

southern limits (both labeled latitudo minuti diametri) and are given in degrees

and minutes.

The extremal values of columns 4 and 6 in the text are shown below:

Venus 7;12 (for 6s 0) 2;30 (for 4s 12 and 7s 6)

Mars 4;21 (for 6s 0) 7;30 (for 6s 0)

Jupiter 2; 5 (for 6s 0) 2; 8 (for 6s 0)

Saturn 3; 2 (for 6s 0) 3; 5 (for 6s 0)

These extremal values in Vimonds tables agree with those in the Toledan

Tables with two exceptions, one of which is a trivial variant for Mercury. But,

as far as we know, the maximum value for the inclination of Venus in Vimonds

table is not attested in any other previous text. It is probably significant that

this value later appeared in the editio princeps of the Alfonsine Tables (1483), as

indicated in Table 13a.

276 chapter 8

s () min min * min min *

0 12 [blank] 2; 5 9 2; 3

0 24 2 2; 7 [blank] 2; 4

1 6 14 2;10 s 2; 7

3 18 60 2;39 s 2;39

5 18 33 3; 1 s 3; 3

6 0 22 3; 2 s 3; 5

6 12 10 3; 1 s 3; 3

6 24 n 2;59 2 3; 0

9 18 n 2;21 60 2;21

11 18 n 2; 5 33 2; 3

12 0 n 2; 3 22 2; 2

Mercury 4; 5 4; 5 4; 5 4; 5 4; 5

2;30 2;30 2;30 2;29 2;30

Venus 6;22 6;22 7;24 7;12 7;12

2;30 2;30 2;30 2;30 2;30

Mars 4;21 4;21 4;21 4;21 4;21

7; 7 7; 7 7;30 7;30 7;30

Jupiter 2; 4 2; 4 2; 5 2; 5 2; 8

2; 8 2; 8 2; 8 2; 8 2; 8

Saturn 3; 2 3; 2 3; 2 3; 2 3; 3

3; 5 3; 5 3; 5 3; 5 3; 5

early alfonsine astronomy in paris 277

s () s ()

argument 0 3;54,46 argument 1 1;16,21

argument 0 7;49,16 argument 2 2;32,27

argument 0 15;38,18 argument 4 5; 4,38

For the inferior planets, between columns 2 and 3 and between columns

5 and 6 we are also given some indications (North and South) to help the

user.

Column 7 appears only in Table 13 (Mercury), and it seems to be outside

the general framework of the table. Its entries are given in seconds and result

from dividing the corresponding entries in column 6 by 10. This probably cor-

responds to the instructions given by Ptolemy in Almagest xiii.6: to compute

the true minutes of proportion for the slant, add 110 when the argument lies

between 90 and 270, or subtract 110 when the argument lies between 0 and

90 or 270 and 360. Whether tabulated or not, these instructions are rarely

found in the medieval Latin literature on the planets (Goldstein and Chabs

2004).

This table displays the radices for the mean motion (motus) and argument

(argumentum) of the fixed stars for intervals of 76, 152, 304, 608, 1216, and

2432 years. Vimond does not give a radix for a specific year but perhaps this

information was in the canons that we have not found. As we shall argue (see

278 chapter 8

Table 27, below), it is likely that the epoch of this table was also 1320 or a date

close to it, that is, the epoch is consistent with our dating of the other radices.

In 76 years the value in the text for the mean motion of the fixed stars is

0;33,32 and in 2432 years it is 17;52,5, corresponding to 0;0,0,4,20,56/d and

0;0,0,4,20,42/d, respectively. These values are equivalent to 48,954 years and

48,999 years, respectively, to complete one revolution, or 1 in about 136 years,

as in the linear term in the standard Alfonsine model for trepidation which

is based on one revolution in exactly 49,000 years. These differences in the

periods depend on the seconds in the entries in Vimonds table and have no

astronomical significance. However, they indicate that Vimond is not using

the standard table for mean motion of the apogees and the fixed stars in the

Parisian Alfonsine Tables (Ratdolt 1483, f. d4v).

In 76 years the value in the text for the mean motion of the argument for

the fixed stars is 3;54,46 and in 2432 years it is 4s 5;4,38, corresponding to

0;0,0,30,26,47/d and 0;0,0,30,24,52/d, respectively. These values are equiva-

lent to 6,992 years and 7,000 years, respectively, to complete one revolution.

The periodic term in the standard Alfonsine model for trepidation is based on

one revolution in exactly 7,000 years, and it corresponds to 0;0,0,30,24,49/d.

These differences have no astronomical significance, but indicate that, once

again, Vimond is not using the standard table for mean motion of access and

recess in the Parisian Alfonsine Tables (Ratdolt 1483, f. d4r).

In fact, an entry for the mean motion of the argument is 7 times the corre-

sponding entry for the mean motion of the linear term.

As in the Parisian Alfonsine Tables, Vimond separates two terms for trepida-

tion: a linear term which corresponds to the difference between the calendar

year of 365;15 days and a fixed tropical year, and a periodic term which corre-

sponds to the difference between a variable sidereal year and the calendar year

of 365;15 days. But in his other tables Vimond has used a fixed sidereal year: we

are unable to account for this inconsistency. To be sure, Vimonds canons may

have explained what he intended.

access and recess (here called motus) is given in degrees and rounded to min-

utes. In Table 27, below, the editors have supplied a minus sign in a few entries,

where appropriate. The table has a maximum of 17;17 for argument 204 and

a minimum of 0;43 for argument 24. These extremal values are 18 apart (=

17;17 + 0;43); hence the amplitude of the sinusoidal curve corresponding to

early alfonsine astronomy in paris 279

() () () ()

0 6 0;19 6 6 16;53

0 12 0;33 6 12 17; 7

0 18 0;41 6 18 17;15

0 24 0;43 6 24 17;17

1 0 0;39 7 0 17;13

1 6 0;29 7 6 17; 3

1 12 0;13 7 12 16;47

1 18 0; 8 7 18 16;26

1 24 0;35 7 14 15;59

2 0 1; 6 8 0 15;28

2 6 1;43 8 6 14;51

2 12 2;24 8 12 14;10

2 18 3; 8 8 18 13;26

2 24 3;56 8 24 12;38

3 0 4;47 9 0 11;47

3 6 5;40 9 6 10;54

3 12 6;35 9 12 9;58

3 18 7;30 9 18 9; 4

3 24 8;26 9 24 8; 8

4 0 9;22 10 0 7;11

4 6 10;18 10 6 6;16

4 12 11;12 10 12 5;22

4 18 12; 4 10 18 4;30

4 24 12;54 11 24 3;40

5 0 13;41 12 0 2;53

5 6 14;24 11 6 2;10

5 12 15; 4 11 12 1;30

5 18 15;39 11 18 0;55

5 24 16;15 11 24 0;19

6 0 16;34 12 0 0; 0

for the equation of access and recess in the Parisian Alfonsine Tables, whose

maximum is 9 for argument 90.

280 chapter 8

Comparison of the entries in both tables shows that the curve representing

Vimonds table is the same as that used by other Parisian astronomers of his

time but shifted in two ways: 247 on the x-axis and 8;17 on the y-axis. In

fact, the entries in Vimonds table can be derived from those in the Parisian

Alfonsine Tables by taking an argument and its corresponding equation in the

latter (where they are given to seconds) and then adding 113 to the argument

and 8;17 to the equation.

Vimonds table begins at a point that in the Parisian Alfonsine Tables cor-

responds to a value of the equation of 8;17 and an argument of about 247.

The value for the equation of access and recess that Vimond thought correct

for his time was 8;17, and he shifted the curve (i.e., the entries in the table)

accordingly; indeed, calculation of the periodic term in trepidation with the

parameters for 1320 in the Parisian Alfonsine Tables yields 8;17 exactly:

radix Incarnation 359;13

Total 67; 6

and

Note that 67;6 + 180 = 247;6 or about 247, and 360 247 = 113 which is

the phase angle of the shift introduced by Vimond.

This table establishes a strong connection between Vimond and the Parisian

Alfonsine Tables, for this theory of trepidation is not found in any previous text.

But again, since the mean motions are different (see Table 26), we see no reason

to assume that Vimond based his theory on the Parisian Alfonsine Tables.

Rather, Vimond may have depended on an Andalusian or Castilian tradition

that was closely related to (but distinct from) the Castilian Alfonsine Tables,

for there is no hint of phase shifts in the Castilian canons.

This table displays the longitude, the latitude, and the magnitude of 225 stars

and nebulae but, in general, their names are omitted. The list is too long to be

related to an astronomical instrument, and the absence of star names makes

us wonder what purpose it was intended to serve. Both coordinates are given

early alfonsine astronomy in paris 281

to minutes. The stars are divided into three groups, in turn divided into sev-

eral subgroups according to the associated planets, a feature which is certainly

not common. Group i has 137 stars that belong to the zodiacal constellations

arranged in 52 subgroups, group ii has 44 stars in northern constellations

(19 subgroups), and group iii has 44 stars in southern constellations (19 sub-

groups); the total number of subgroups is thus 90. We note the balanced repre-

sentation of the stars on both sides of the zodiac.

We have found the same table in an early 14th-century copy, Cambridge,

Gonville and Caius College, ms 141/191, pp. 377382 (for an excerpt, see F.S. Ped-

ersen 2002, pp. 15071508); as well as in Segovia, Catedral, ms 84, ff. 46r51v; and

in Paris, BnF, ms 7482; ff. 61v69v. There are some cases where an entry in one

copy does not agree with the value in, or derived from, Ptolemys treatises in

contrast to the other copy, but there are also examples where entries in both

copies do not agree with those in Ptolemy. On the other hand, in all cases where

there is a blank entry in one copy, it is filled in the other copy.

In the Paris copy only 18 star names are given whereas in the Cambridge

copy this number is reduced to 15. The star names in these copies are generally

not identical, and they are not always ascribed to the same stars. For instance,

the names almalak and almalac are attributed, respectively, to the star in

the 20th subgroup (ms Paris) and to the first star of the 8th subgroup (ms

Cambridge). The star list does not bear a general title in the Paris copy but the

Cambridge copy reads tabula de dispositionibus stellarum fixarum existentibus

ad terminum complementi radicis mediarum coniunctionum solis et lunae quae

alibi signantur. Et primo de dispositionibus illarum stellarum quae sunt prope

viam solis. (Here begins the table on the groups of the fixed stars as they were

at the point of completion [the epoch?] of the radix of the mean conjunctions

of the Sun and the Moon specified elsewhere. First come the groups of those

stars close to the zodiac [lit.: the path of the Sun].)

The first sentence serves as a general title for the table, and the second sen-

tence is a heading for the groups in the zodiacal constellations, corresponding

to the headings in both manuscripts for the groups in the northern and south-

ern constellations. The expression the radix of the mean conjunctions seems

to refer to the radix given on f. 1r, 13;54,54d, which we identified with March

10, 1320. But we do not understand the expression at the end of the comple-

ment.

Madrid, Biblioteca Nacional, ms 4238, ff. 65v66v, reproduces the same star

list except that the signs used here are of 60, contrary to the other manuscripts

containing this list.

We are grateful to Paul Kunitzsch for information on two additional copies

of the same star list: Erfurt, Universittsbibliothek, ms Amplon. 2395, ff. 104v

282 chapter 8

105v; and Munich, Bayerische Staatsbibliothek, ms Clm 26667, ff. 46v47v (cf.

Kunitzsch 1986a, p. 96, n. 10, and p. 98, n. 44). In both manuscripts the list is

anonymous, but in the Erfurt copy a marginal note (in the same hand as the

list) reads: Notandum istas stellarum tabulas fuisse equatas ad annum domini

1338 (f. 105v). As Kunitzsch suggested to us (in a private communication), this

marginal note may have been added by the copyist and not belong to the orig-

inal list; no date appears in the other three manuscripts.

In fact, the list in the Erfurt ms has two extra stars: one is added to the

northern constellations, in subgroup 7 (Bootes), and the other to the southern

constellations, in subgroup 6 (Eridanus). We also note that in the list for the

southern constellations the stars in subgroup 19 (Ara) are located in the Erfurt

ms between subgroups 4 and 5 in the manuscripts in Paris and Cambridge (we

have not seen the manuscript in Munich). Another special feature of the Erfurt

ms is that the subgroups are not numbered; rather, most are given the name of

a star belonging to them or even a generic name. But its main distinguishing

characteristic is that the subgroups have no associated planets, in contrast to

the copies in Paris and Cambridge.

It may be of interest that the 5 manuscripts of which we are aware that

contain this star list are spread all over Europe: 2 in Germany, 1 each in England,

France, and Spain.

The order and the grouping of the stars in this list is peculiar, for they

do not follow the pattern of the catalogue in Ptolemys Almagest that was

generally adopted in medieval star lists and catalogues. Rather, this list is

organized according to Ptolemys Tetrabiblos, a handbook on astrology writ-

ten by Ptolemy after the Almagest. It was translated several times from Arabic

into Latin: in 1138 by Plato of Tivoli, in 1206 anonymously, and in 1256 via

Castilian at the court of Alfonso x by Egidius de Tebaldis (Chabs and Gold-

stein 2003a, p. 232), and was known as the Quadripartitum. In Tetrabiblos

i.9, Ptolemy grouped the stars into three main categories (zodiacal, north-

ern, and southern constellations), following an order differing from that in

the Almagest where the northern constellations precede the zodiacal constel-

lations, and grouped the stars within each category according to their asso-

ciated planets. As an example, we reproduce a passage of Tetrabiblos i.9

corresponding to the stars in the constellation of Aries (Robbins 1940, p.

47):

The stars in the head of Aries, then, have an effect like the power of

Mars and Saturn, mingled; those in the mouth like Mercurys power and

moderately like Saturns; those in the hind foot like that of Mars, and those

in the tail like that of Venus.

early alfonsine astronomy in paris 283

As is readily seen, the order, the subgroups, and the planets associated with

the stars in Aries in Vimonds list perfectly match those in Ptolemys Tetrabiblos.

And this is indeed the case for almost all stars in the 90 subgroups displayed in

Vimonds list.

The star positions generally agree with those in Gerard of Cremonas version

of Ptolemys star catalogue in the Almagest with an increment in longitude of

17;52 for precession, a value otherwise unattested. If the rate of precession was

taken to be 1 in 66 years, 17;52 would correspond to about 1179 years and,

if we add it to 137ad (the date of the star catalogue in the Almagest), we get

1316ad. But it is not clear that this date had any significance for the author. We

have compared this list to that in the Libro de las estrellas de la ochaua espera

(Madrid, Universidad Complutense, ms 156; see also Rico Sinobas 18631867,

vol. 1, pp. 5145), also known as Libro de las xlviii figuras de la viii spera or even

as Libro de las estrellas fixas. This is an adaptation of the star catalogue for 964

ad by the Persian astronomer al-f (903986) which in turn depended on the

star catalogue in Ptolemys Almagest (see Comes 1990). This work, where the

total precession is 17;8, was compiled in 1256 by Judah ben Moses ha-Cohen,

one of the most distinguished collaborators of Alfonso x. The presentation of

the star data in this Alfonsine text differs substantially from that of a typical

star list although the data themselves are what one would expect, namely,

for each star we are given its name, longitude, latitude, and magnitude. The

associated planets are also given for each star, often adding an indication of

their relative strength, showing that the Alfonsine Libro ultimately relied on

Ptolemys Quadripartitum. However, after comparing the data in the Libro

with those of Vimond, we see no evidence to suggest that the star list found

among Vimonds tables is systematically related to this Alfonsine book. As

Kunitzsch informed us, there is a star list by John of Lignres containing data

for 276 stars, but the longitudes are Alfonsine, i.e., Ptolemys values plus 17;8:

Paris, Bibliothque nationale de France, ms lat. 10264, ff. 36v38v, and Florence,

Biblioteca Nazionale Centrale, ms Conv. soppr. j.4.20, fols. 214v216r. This list

was extracted from the star table that later appeared in the editio princeps of

the Alfonsine Tables (1483), and sheds no additional light on the list included

in Vimonds tables.

Moreover, in the course of examining the star names in the five manuscripts

containing this list, Kunitzsch noticed that the author drew upon a variety of

Latin sources, mainly the translations of the Tetrabiblos but also sources not in

the Tetrabiblos tradition (some of which cannot be identified). Thus, Vimonds

list is dependent on Ptolemy in two ways: the choice of the stars, their order and

grouping, as well as the associated planets, are borrowed from the Quadriparti-

tum; and the numerical data are taken from the Latin version of the Almagest.

284 chapter 8

In sum, we believe that the star list attributed to Vimond in the Paris ms, and

that is anonymous in the Cambridge, Erfurt, Madrid, and Munich mss, derives

from an unknown archetype; we know of no similar star list in Latin in the 14th

century or in the previous Arabic literature with which to compare it.

In Table 28 we present in the first 3 columns a complete transcription of the

Paris copy with translations of the headings and the names of the associated

planets in each case. For the latitudes north is indicated by an abbreviation

of the term septentrionalis, and south by an abbreviation of meridionalis; we

have replaced them with the modern designations + and . Column iv gives the

few star names found in the Paris copy, which were added in interstitial spaces

within the table (some of the star names are partly hidden in the gutter of the

manuscript and cannot be read completely); column v lists the modern star

designation; column vi gives the standard number assigned to each of the 1028

stars in Ptolemys catalogue; column vii offers comparisons and comments,

together with variants in the Cambridge copy; and column viii provides the

identification of the star names.

table 28 Star list ( f. 8r v)

i ii iii iv v vi vii viii

Longitude Latitude Magn. Name Modern Number Comparisons Identification of

(sign, degrees) (degrees) designation (P.-K.) and comments star names

[Zodiacal constellations]

0 24;32 +7;20 3 Ari 362

0 25;32 +8;20 3 Ari 363

early alfonsine astronomy in paris

0 28;52 +7;40 5 Ari 364 C(iii): blank

C(iv): flamai? Unidentified

0 29;22 +6; 0 5 Ari 365 C(iii): blank

C(iv): hercules Unidentified:

see Gem, below

3 [Aries] Mars

1 2;52 5; 5 4 Cet 374 C: +5;15, G: 5;15

1 5;52 1;30 5 Ari 373 C: +1;30

1 7;32 1;20 5 Ari 372 C: +1;20

G: 1;10, +1;10

285

table 28 Star list ( f. 8r v) (cont.)

286

4 [Aries] Venus

1 9;12 +4;50 5 Ari 368

1 11;42 +1;40 4 Ari 369

1 13;12 +2;30 4 Ari 370

1 14;52 +1;50 4 Ari 371

5 [Taurus] Venus, Jupiter C: Moon

1 17;32 9;30 5 30(e) Tau 384

1 21;32 8; 0 3 Tau 385 C: +8; 0

6 [Taurus, The Pleiades] Moon, Mars

1 20; 2 +4;30 5 19 Tau 409

1 20;22 +4;40 5 23 Tau 410

1 20;32 +5; 5 5 27 Tau* 412

1 21;32 +5;20 5 bsc 1188* 411

7 [Taurus] Mars

2 0;32 5;10 1 aldebaran? Tau 393 C(iv): aldebaran G, p. 89 n. 10, etc.

chapter 8

i ii iii iv v vi vii viii

1 26;52 5;45 3 Tau 390 C(iv): almalac If this is a corruption of

Arabic al-malik (the king),

it should designate Leo

(Regulus). See G, p. 101

n. 12.

1 28;42 5;50 3 1 Tau 392 C: +5;50

G: 5;50, 0;50

early alfonsine astronomy in paris

2 3;32 4; 0 4 Tau 399 C: +4; 0

G: 4; 0, +4; 0

9 [Taurus] Mars

2 7;52 3;30 5 106(l) Tau 397

2 8;12 5; 0 5 104(m) Tau 396 C: +5; 0

2 13;32 +5; 0 5 Tau 230/400 G: 3

2 15; 2 2;30 3 Tau 398 G + 17;52: 15;32

10 [Gemini] Mercury, Venus

2 24;22 1;30 4 Gem 437 C: +1;30

2 26; 2 1;15 4 Gem 438 C: +1;15

2 28; 2 3;30 4 Gem 439 C: +3;30

287

table 28 Star list ( f. 8r v) (cont.)

288

3 2;32 10;30 4 Gem 441 C: +10;30

11 [Gemini] Saturn

3 9;32 5;30 3 Gem 435

12 [Gemini] Mars

3 11;12 +9;40 2 ( )annai? Gem 424

13 [Gemini] Mars

3 14;32 [..]6;15 [..] hercules? Gem 425 C(ii): +6;15. C(iii): 2 R, p. 48: Herakles

C(iv): almueredan K1959, p. 127: Vir is

called almuredin

14 [Cancer] Mercury, Mars

3 20;32 +1; 0 5 Cnc 456

3 25; 2 7;30 4 Cnc 457

15 [Cancer] Saturn, Mercury

3 26;12 +11;50 4 Cnc 455

4 4;22 +5;30 4 Cnc 454 G: 5;30

16 [Cancer] Moon, Mars

3 28;12 +0;40 n meollef? gc 2632 449 C: 2. C(iv): mellef? P, f. 15va: meelef

Galaxy m 44

chapter 8

i ii iii iv v vi vii viii

3 28;12 +2;40 4 assinis? Cnc 452 C(iv): asini G

3 29;12 +0;10 4 Cnc 453 G: 0;10

18 [Leo] Saturn, Mars

4 12; 2 +9;30 3 Leo 465

4 12; 2 +12; 0 3 Leo 464 G + 17;52: 12;12

19 [Leo] Saturn, Mars C: Mercury

4 18; 2 +11; 0 3 Leo 466

4 18;32 +4;30 3 Leo 468

early alfonsine astronomy in paris

20 [Leo] Mars, Jupiter

4 20;22 +0;10 1 almalak? Leo 469 G, p. 101 n. 12.

21 [Leo] Venus, Saturn

4 29;12 +13;15 5 60(b) Leo 480 C, G: +12;15

5 2; 2 +13;40 2 Leo 481

5 2;12 +11;30 5 81 Leo* 482

5 4;12 + 9;40 3 Leo 483

5 12;22 +12;50 1 ?? Leo 488 G: +11;50

22 [Leo] Venus, Mercury

5 8;12 +5;50 3 Leo 484

5 8;22 3; 0 5 Leo 487

289

table 28 Star list ( f. 8r v) (cont.)

290

5 9;32 +1;15 4 Leo 485

23 [Virgo] Mercury, Mars

5 14;12 +4;35 5 Vir 497

5 14;52 +5;40 5 Vir 498

5 17; 2 +6; 0 3 Vir 501 G + 17;52: 16;52

5 18; 2 +5;30 5 Vir 500

24 [Virgo] Mercury, Venus

5 26; 7 +1;10 3 Vir 502

6 1; 2 +2;50 3 Vir 503

25 [Virgo] Saturn, Mercury

6 0; 2 +15;10 3 Vir 509

26 [Virgo] Venus, Mercury

6 14;32 2; 0 1 almure? Vir 510 C(iv): alcimech G: ascimech

27 [Virgo] Mercury, Mars

6 24;32 +7;30 4 Vir 518

6 25;12 +2;40 4 Vir 519

6 7;52 +0;30 4 Vir 521 C, G + 17;52: 27;52

7 0;32 +9;50 4 Vir 522

chapter 8

i ii iii iv v vi vii viii

7 5;52 +0;40 2 Lib 529

7 10; 2 +8;30 2 Lib 531

29 [Libra] Saturn, Mercury

7 9;12 [..]1;15 [..] Lib 534 C(ii): +1;15. C(iii): 4

7 11;52 [..]1;40 [..] Lib 533 C(ii): +2;40. C(iii): 4

7 15;22 [..]3;45 [..] Lib 535 C(ii): +3;45. C(iii): 4

7 20;52 [..]4;30 [..] Lib 536 C(ii): +4;30. C(iii): 4

30 [Scorpius] Mars, Saturn

early alfonsine astronomy in paris

7 23;32 5; 0 3 Sco 548 C: +5; 0

7 24;12 +1;20 3 Sco 546

31 [Scorpius] Mars, Jupiter

8 0;29 4; 0 2 Sco 553 G + 17;52: 0;32

32 [Scorpius] Saturn, Venus

8 5;52 15; 0 4 1 + 2 Sco 558 G + 17;52: 6;42

8 11; 2 19;30 3 Sco 561 C: +19;30

8 16; 2 18;50 3 Sco 562 C: +18;50

8 15;52 15;10 3 Sco 564 G + 17;52: 16;52

C(ii): +16;10

8 18;22 16;40 3 1 Sco 563 C: +16;40

291

292

8 15;52 23;30 4 Sco 566 C, G + 17;52: 14;52

C, G: 13;30

8 14;22 13;20 3 Sco 565 C, G + 17;52: 15;22

C(ii): +13;20

34 [Scorpius] Mars, Moon

8 19; 2 13;15 n G Sco* 567 C: 2

+ CGlo 6441

35 [Sagittarius] Saturn, Moon

8 22;22 6;30 3 Sgr 570

9 0;52 3;50 4 Sgr 576 C: +3;50

36 [Sagittarius] Jupiter, Mars

8 24;32 +2; 7 4 Sgr 574

8 26;52 1;30 3 Sgr 573

37 [Sagittarius] Mercury, Jupiter, Sun, Mars C: Moon

9 3; 2 7;45 n 1 + 2 Sgr 577 G: 0;45. C(iv): 2

38 [Sagittarius] Jupiter, Mercury

9 4;12 6;45 3 Sgr 591

9 5;32 2;30 4 Sgr 590 C: +2;30, G: 4;30

9 7;52 2;30 5 Sgr 589 C: +2;30

chapter 8

i ii iii iv v vi vii viii

9 4;52 18; 0 2 Sgr 593

9 5;32 23; 0 2 1 + 2 Sgr 592 C: +23; 0

9 14;32 13; 0 3 Sgr 594 G + 17;52: 24;32

C(ii): +13; 0

40 [Sagittarius] Venus, Saturn

9 16;32 5;50 5 59(b) Sgr 599 C, G + 17;52: 16;22

9 16;32 4;50 5 60(a) Sgr 598 G + 17;52: 15;32

C(ii): +4;50

early alfonsine astronomy in paris

9 17;32 6;30 5 62(c) Sgr 600 C: +6;30

41 [Capricornus] Mars, Venus

9 25;12 +2;20 3 1 + 2 Cap 601

9 25;12 +5; 0 3 Cap 603

9 26;42 +1;30 6 Cap 607

9 26;52 +0;45 6 Cap 605

42 [Capricornus] Mars, Mercury

9 29;32 8;40 4 Cap 612

10 4;32 7;40 4 24(a) Cap 613 C: +7;40

10 8; 2 6;50 4 Cap 614 C: +6;50

10 8;12 6; 0 5 36(b) Cap 615 C: +6; 0

293

table 28 Star list ( f. 8r v) (cont.)

294

10 12;42 +2;10 3 Cap 623 G: +2;10, 2;10

10 14;12 +2; 0 3 Cap 624 G: +2; 0, 2; 0

10 14;42 0;20 4 42(d) Cap 625 G: +0;20

10 15;32 2;50 5 Cap 627 C, G: +2;50

44 [Aquarius] Saturn, Mercury

10 2;32 +8;40 3 Aqr 636

10 4; 2 +8; 0 4 Aqr 635

10 14;22 +8;50 2 Aqr 632

10 24;12 +11;15 4 Aqr 630 G: +11;0. G: 3

45 [Aquarius] Mercury, Saturn

10 19;12 5; 0 4 Aqr 647 G + 17;52: 29;12

10 19;32 7;30 3 Aqr 646 G + 17;52: 29;32

C, G: +7;30

10 22;32 5;40 5 53(f) Aqr 648 C: +5;40

46 [Aquarius] Saturn, Jupiter

11 5;32 1; 0 4 83(h) Aqr 653

11 6;52 7;30 4 1 Aqr 656 C: +7;30, G: 8;30

11 7;52 0;30 4 Aqr 654 C: +7;30

11 8;12 1;40 4 a[n]phora Aqr 655 C: +1;40 See note 1.

chapter 8

i ii iii iv v vi vii viii

11 9;32 +9;15 4 Psc 674

11 12; 2 +7;30 4 Psc 675

11 13;52 +9;20 4 7(b) Psc 676

48 [Pisces] Jupiter, Mercury

11 13;52 +4;30 4 Psc 679

11 17;32 +2;30 4 Psc 680

49 [Pisces] Saturn, Mercury

11 23;52 +6;20 4 Psc 681

early alfonsine astronomy in paris

50 [Pisces] Jupiter, Venus

0 17;12 +15;20 4 Psc 706 G + 17;52: 17;22

0 20; 2 +17; 0 4 Psc 705

51 [Pisces] Saturn, Jupiter

0 13;32 +14;20 4 1 Psc 702

0 14;12 +13; 0 4 2 Psc 703

0 15;32 +12; 0 4 Psc* 704

52 [Pisces] Mars, Mercury

0 20;22 8;30 3 Psc 692

295

table 28 Star list ( f. 8r v) (cont.)

296

[Title:] Then follow the constellations (dispositio) of the other fixed stars in the northern part.

4 5; 2 72;50 2 aliedin UMi 6 C(iv): aliedim See note 2.

4 14; 2 74;50 2 alforcami UMi 7 C(iv): alfoza K1961, p. 58: al-farqadn

(+ UMi)

2 [Ursa Maior] Moon, Venus

5 0; 2 53;30 2 UMa 33 G: 13;30

5 5;52 55;40 2 benezna UMa 34 G: 15;40 K1966, p. 42, no. 23:

benenaz ( UMa)

5 17;42 54; 0 2 UMa 35 G: 14; 0

3 [Draco] Saturn, Mars

5 26;22 84;50 3 Dra 67

5 27;52 88; 0 3 Dra 68 G: 78; 0

4 [Cepheus] Saturn, Jupiter

0 4;32 69; 0 3 Cep 78

0 25;22 71;10 4 Cep 77

11 27;12 72; 0 4 Cep 79

chapter 8

i ii iii iv v vi vii viii

7 28; 2 53;30 4 Her 130 G: 13;30

7 29;42 54;10 3 Her 129 G + 17;52: 21;42

G: 16;10

8 1;52 59;50 3 Her 133 G: 19;50, 59;50

8 3;12 60;20 4 69(e) Her 134

6 [Corona Borealis] Venus, Mercury

6 29;32 46;10 4 CrB 112

7 2;32 44;30 2 alfeca CrB 111 C(iv): alfeca G

early alfonsine astronomy in paris

7 7; 2 44;50 4 CrB 116

7 [Bootes] Mercury

6 9;12 28; 0 3 Boo 107

8 [Lyra] Venus, Mercury

9 5;12 62; 0 1 lilurah Lyr 149 C(iv): lulurach G: allore

9 [Perseus] Saturn, Jupiter

1 17;29 23; 0 2 eiumezuz? Per 202 G + 17;52: 17;32 Unidentified

10 [Perseus] Mars, Mercury

1 22;42 30; 0 2 Per 197

11 [Auriga] Mars, Mercury

2 12;52 22;30 1 alhaioch Aur 222 C(iv): alhaioch G

297

table 28 Star list ( f. 8r v) (cont.)

298

8 12;42 36; 0 3 alhanue Oph 234 C(iv): alhanue K1966, p. 55, no. 33:

alhaue, alhane

13 [Serpens] Saturn, Mars

7 12;12 25;30 3 Ser 271 G: 25;20

7 12;42 36;30 4 Ser 270 C, G: 26;30

7 14;12 24; 0 3 Ser 272

7 16;32 16;30 4 Ser 273

14 [Sagitta] Mars, Venus

9 24;32 39;10 6 Sge 282

9 28; 2 39;20 4 Sge 281

15 [Aquila] Jupiter, Mars

9 21;42 29;10 2 vultur Aql 288 C: 19;10 G

C(iv): vultur

16 [Delphinus] Saturn, Mars

10 6;22 32; 0 3 Del 304

10 8; 2 33;50 3 Del 305

10 9;12 32; 0 3 Del 306

10 11;22 32;10 3 Del 307 G: 33;10

chapter 8

i ii iii iv v vi vii viii

0 15; 2 12;31 3 Peg 316 G + 17;52: 10; 2

G: 12;30. C, G: 2

11 20; 2 31; 0 2 Peg 317

18 [Andromeda] Mars, Venus

0 13;32 15; 7 3 And 345

0 19;42 30; 0 3 And 347

0 19;52 32;30 3 And 348

0 25;42 26;20 3 And 346 G + 17;52: 21;42

early alfonsine astronomy in paris

C(ii): +16;20

1 4;42 23; 0 3 And 349

19 [Triangulum] Mercury

0 28;52 16;30 3 Tri 358

1 3;52 20;40 3 Tri 359

[Title:] Then follow the constellations (dispositio) of the other stars in the southern part.

10 9;42 16;30 4 PsA 1020

299

300

10 16;42 14; 0 4 PsA 1018 G: 14;40

10 18;32 20;20 4 PsA 1012

2 [Cetus] Saturn

0 12;52 20; 0 2 Cet 725

3 [Orion] Mars, Mercury

2 19;52 17; 0 1 Ori 735

4 [Orion] Jupiter, Saturn

2 7;42 31;30 1 Ori 768

2 13;12 24;10 2 Ori 759

2 15;12 24;50 2 Ori 760

2 16; 2 25;40 2 Ori 761

5 [Eridanus] Jupiter

0 18; 2 53;30 1 Eri 805 C(i): 16; 2, G: 13;30

6 [Eridanus] [..] C: Saturn

2 5;12 31;50 4 Eri 772 G + 17;52: 6;12

7 [Lepus] Saturn, Mars

2 12;42 44;20 3 Lep 813

2 13;22 41;30 3 Lep 812

chapter 8

i ii iii iv v vi vii viii

2 13;52 57;40 2 Col 845

2 16;52 59;40 2 Col 844 C: 57;40

9 [Canis Maior] Jupiter, Mars

3 5;32 39;10 1 CMa 818

3 7;32 35; 0 4 CMa 819

10 [Canis Minor] Mercury, Mars

3 13;22 14; 0 4 CMi 847 G + 17;52: 12;52

3 17; 2 16;10 1 CMi 848

early alfonsine astronomy in paris

4 17;52 20;30 2 Hya 905

4 23;52 26;30 4 Hya 906

4 26;32 26; 0 4 1 Hya 907

12 [Crater] Venus, Mercury

5 17;52 18; 0 4 Crt 923

5 17;52 18;30 4 Crt 924 G + 17;52: 24;52

5 20;22 19;30 4 Crt 922

13 [Corvus] Saturn, Mercury

6 2;12 19;40 3 Crv 929

6 6;22 14;50 3 Crv 931 C: 6;12

301

302

3 5; 2 69; 0 1 Car 892 G: 29; 0

15 [Centaurus] Mars, Venus

6 24; 2 25;40 3 Cen 939

7 3;32 22;30 3 Cen 940 C: 2

16 [Centaurus] Venus, Jupiter

6 26;13 41;10 1 Cen 969 G + 17;52: 26;12

C: 26;12

6 27;52 51;10 2 Cru 965

6 29; 2 55;20 2 Cru 968

7 3;12 51;40 2 Cru 966

7 12; 2 45;20 2 Cen 970 C: 45;[..]

17 [Lupus] Venus, Mars

7 13;42 29;10 3 Lup* 973

7 15;52 24;10 3 Lup 972 G: 24;50

18 [Corona Australis] Saturn, Mercury

9 4;22 15;20 4 Cr a 1005 G: 15;10

9 4;42 16; 0 4 Cr a 1004

9 5;12 17;10 4 Cr a 1003 G + 17;52: 4;52

chapter 8

i ii iii iv v vi vii viii

8 8;32 30;20 4 1 Ara 994 G: 5

8 12;52 33;20 4 Ara 996

8 13; 2 34;10 4 Ara 995

early alfonsine astronomy in paris

303

304 chapter 8

Col. i. The number of the zodiacal sign is not repeated in col. vii where variants

are listed; in all cases reported in that column only the degrees and minutes

differed from the entry in the Paris manuscript.

Col. iv: In the manuscript the names of the stars are not presented in a column.

Col. v: The entries in this column have been taken from Toomer 1984.

* indicates that Kunitzsch 1986 and Kunitzsch 1991, pp. 187200, give a

different modern designation.

Col. vi: These numbers are taken from Peters and Knobel 1915 (ultimately from

Baily 1843), and they are also used in Kunitzsch 1986 and 1990.

Col. vii: C refers to Cambridge, Gonville and Caius College, ms 141/191; in certain

cases, it is followed by a column number in Roman numerals. G refers to

Gerard of Cremonas version of Ptolemys star catalogue (Kunitzsch 1990). We

underline entries in Vimonds table for which there is a variant reading. The

entries for longitudes in both copies generally agree with those in G with an

increment of 17;52 for precession; those cases where they differ have been

noted.

(Kunitzsch 1990); K1959 refers to Kunitzsch 1959; K1966 refers to Kunitzsch

1966; P refers to Plato of Tivolis Latin version of the Tetrabiblos (ed. 1493); and

R refers to Robbins 1940.

Note 1. We are informed by Kunitzsch that anphora is not a proper name; rather,

it is a noun used in the description of the stars position: where the water flows

out from the vessel; Erfurt, Universittsbibliothek, Amplon. 2395, f. 105r, in

decursu aque ab anphora; p, f. 16va: In aque vero decursu collocate (without

anphora).

Note 2. As Kunitzsch informed us, aliedim, apparently renders the Arabic al-

jady (the kid), an old Arabic name for UMi (Kunitzsch 1961, p. 62). It is uncer-

tain where the compiler of this list might have found it. In the Tetrabiblos tra-

dition this name never occurs.

early alfonsine astronomy in paris 305

Acknowledgments

We thank Paul Kunitzsch and Beatriz Porres for assistance with the Latin texts

cited in this article, and Fritz S. Pedersen, John D. North, and Julio Sams for

detailed comments on a preliminary version of this paper.

References

Baily, F. 1843. The Catalogues of Ptolemy, Ulugh Beigh, Tycho Brah, Halley, Hevelius.

Memoirs of the Royal Astronomical Society, vol. 13. London.

bsc. See Hoffleit (ed.) 1964.

Boudet, J.-P. 19971999. Le Recueil des plus celebres astrologues de Simon de Phares, 2

vols. Paris.

Casulleras, J. and J. Sams (eds.) 1996. From Baghdad to Barcelona: Studies in the Islamic

Exact Sciences in Honour of Prof. Juan Vernet. Barcelona.

Chabs, J. 2000. Astronoma alfons en Morella a finales del siglo xiv, Cronos: Cuader-

nos Valencianos de Historia de la Medicina y de la Ciencia, 3:381391.

Chabs, J. and B.R. Goldstein 1994. Andalusian Astronomy: al-Zj al-Muqtabis of Ibn

al-Kammd, Archive for History of Exact Sciences, 48:141.

Chabs, J. and B.R. Goldstein 1997. Computational Astronomy: Five centuries of Find-

ing True Syzygy, Journal for the History of Astronomy, 28:93105.

Chabs, J. and B.R. Goldstein 2003a. The Alfonsine Tables of Toledo. Dordrecht.

Chabs, J. and B.R. Goldstein 2003b. John Vimond and the Alfonsine Trepidation

Model, Journal for the History of Astronomy, 34:163170.

Comes, M. 1990. Al-f como fuente del libro de la Ochaua Espera de Alfonso x, in

Comes, Mielgo, and Sams (eds.) 1990, pp. 11113.

Comes, M., H. Mielgo, and J. Sams (eds.) 1990. Ochava espera y astrofsica. Barce-

lona.

Copernicus, N. 1543. De revolutionibus. Nuremberg.

Goldstein, B.R. 1974. The Astronomical Tables of Levi ben Gerson. Transactions of the

Connecticut Academy of Arts and Sciences, 45. New Haven.

Goldstein, B.R. 1996. Lunar Velocity in the Middle Ages: A Comparative Study, in

Casulleras and Sams (eds.) 1996, pp. 181194.

Goldstein, B.R. 2001. The Astronomical Tables of Judah ben Verga, Suhayl, 2:227289.

Goldstein, B.R. 2003. An Anonymous Zij in Hebrew for 1400ad: A Preliminary Report,

Archive for History of the Exact Sciences, 57:151171.

Goldstein, B.R. and J. Chabs 2004. Ptolemy, Bianchini, and Copernicus: Tables for

Planetary Latitudes, Archive for History of Exact Sciences, 58:453473.

306 chapter 8

Hoffleit, D. (ed.) 1964. Catalogue of Bright Stars. Yale University Observatory. New

Haven.

King, D.A. and M.H. Kennedy (eds.) 1983. Studies in the Islamic Exact Sciences. Beirut.

Kremer, R.L. 2003. Wenzel Fabers Table for Finding True Syzygy, Centaurus, 45:305

329.

Kremer, R.L. and J. Dobrzycki 1998. Alfonsine meridians: Tradition versus experience

in astronomical practice c. 1500, Journal for the History of Astronomy, 29:187199.

Kunitzsch, P. 1961. Untersuchungen zur Sternnomenklatur der Araber. Wiesbaden.

Kunitzsch, P. 1966. Typen von Sternverzeichnissen in astronomischen Handschriften des

zehnten bis vierzehnten Jahrhunderts. Wiesbaden.

Kunitzsch, P. 1986a. The star catalogue commonly appended to the Alfonsine Tables,

Journal for the History of Astronomy, 17:8998.

Kunitzsch, P. (ed. and tr.) 1986b. Claudius Ptolemus, Der Sternkatalog des Almagest: Die

arabisch-mittelalterliche Tradition. i: Die arabischen bersetzungen. Wiesbaden.

Kunitzsch, P. (ed.) 1990. Claudius Ptolemus, Der Sternkatalog des Almagest: Die ara-

bisch-mittelalterliche Tradition. ii: Die lateinische bersetzung Gerhards von Cre-

mona. Wiesbaden.

Kunitzsch, P. 1991. Claudius Ptolemus, Der Sternkatalog des Almagest: Die arabisch-

mittelalterliche Tradition. iii: Gesamtkonkordanz der Stern-koordinaten. Wiesba-

den.

Mestres, A. 1996. Maghrib Astronomy in the 13th Century: a Description of Manuscript

Hyderabad Andra Pradesh State Library 298, in Casulleras and Sams (eds.) 1996,

pp. 383443.

Mestres, A. 1999. Materials Andalusins en el Zj dIbn Isq al-Tnis. Doctoral Thesis,

University of Barcelona.

Mills, J.M. 19431950. Estudios sobre Azarquiel. MadridGranada.

Nallino, C.A. 19031907. Al-Battn sive Albatenii Opus Astronomicum, 2 vols. Milan.

Neugebauer, O. 1962. The Astronomical Tables of al-Khwrizm. Copenhagen.

Neugebauer, O. 1975. A History of Ancient Mathematical Astronomy. Berlin.

North, J.D. 1976. Richard of Wallingford, 3 vols. Oxford.

Pedersen, F.S. 2002. The Toledan Tables: A review of the manuscripts and the textual

versions with an edition. Copenhagen.

Pedersen, O. 1974. A Survey of the Almagest. Odense.

Peters, C.H.F. and E.B. Knobel 1915. Ptolemys Catalogue of Stars: A Revision of the

Almagest. Washington.

Poulle, E. 1973. John of Lignres, in The Dictionary of Scientific Biography, 7:122128.

New York.

Poulle, E. 1984. Les tables alphonsines avec les canons de Jean de Saxe. Paris.

Ptolemy. Quadripartitum. See G. Salio (ed.) 1493.

Ratdolt, E. (ed.) 1483. Tabule astronomice illustrissimi Alfontij regis castelle. Venice.

early alfonsine astronomy in paris 307

Rico Sinobas, M. 18631867. Libros del Saber de Astronoma del Rey D. Alfonso x de

Castilla, 5 vols. Madrid.

Robbins, F.E. (ed. and trans.) 1940. Ptolemy: Tetrabiblos. London.

Saby, M.-M. 1987. Les canons de Jean de Lignres sur les tables astro-nomiques de 1321.

Unpublished thesis: Ecole Nationale des Chartes, Paris. A summary appeared as:

Les canons de Jean de Lignres sur les tables astronomiques de 1321, cole Nationale

des Chartes: Positions des thses, pp. 183190.

Saleh, J.A. as- 1970. Solar and Lunar Distances and Apparent Velocities in the Astro-

nomical Tables of abash al-sib, Al-Abhath, 23:129176. Reprinted in King and

Kennedy (eds.) 1983.

Salio, G. (ed.) 1493. Liber quadripartiti Ptholemei; Venice.

Sams, J. and E. Mills 1998. The Computation of Planetary Longitudes in the Zj of Ibn

al-Bann, Arabic Sciences and Philosophy, 8:259286.

Sdillot, J.-J. and L.-A. Sdillot 1834. Trait des instruments astronomiques des Arabes.

Paris. Reprinted Frankfurt a/M (1984).

Swerdlow, N.M. and O. Neugebauer 1984. Mathematical Astronomy in Copenricuss De

revolutionibus. New York and Berlin.

Thorndike, L. and P. Kibre 1963. A catalogue of incipits of mediaeval scientific writings in

Latin. London.

Toomer, G.J. 1968. A Survey of the Toledan Tables, Osiris, 15:5174.

Toomer, G.J. 1984. Ptolemys Almagest. New York and Berlin.

chapter 9

To John D. North (19342008), in memoriam

John of Murs (fl. 13201340), a scholar active in Paris in the first half of the four-

teenth century, was a key figure in the history of astronomy in addition to mak-

ing contributions to music and mathematics. Indeed, his work on astronomy

played a decisive role in the transmission of scientific ideas in the late Middle

Ages.1 He is largely responsible for the introduction of Alfonsine astronomy into

the Parisian milieu, notably a set of astronomical tables originally elaborated

in Toledo by the astronomers in the service of King Alfonso of Castile (d. 1284).

1 L. Gushee, New Sources for the Biography of Johannes de Muris, Journal of the American

Musicological Society, xxii (1969), 326; E. Poulle, John of Murs, in The Dictionary of Scientific

Biography (16 vols, New York, 19701980), vii (1973), 128133; G. Beaujouan, Observations et

calculs astronomiques de Jean de Murs (13211344), in Proceedings of the xivth International

Congress of the History of Science (TokyoKyoto 1974) (Tokyo, 1975), ii, 2730, reprinted in idem,

Par raison des nombres: Lart du calcul et les savoirs scientifiques mdivaux (Aldershot, 1991),

Essay vii; J.D. North, The Alfonsine Tables in England, in Y. Maeyama and W.G. Salzer

(eds), Prismata: Festschrift fr Willy Hartner (Wiesbaden, 1977), 269301; G. lHuiller, Aspects

nouveaux de la biographie de Jean de Murs, Archives dhistoire doctrinale et littraire du

moyen ge, xlvii (1980), 272276; C. Schabel, John of Murs and Firmin of Beauvals Letter

and Treatise on the Calendar Reform for Clement vi, Cahiers de lInstitut du moyen-ge

grec et latin, lxvi (1996), 187215; J. Chabs and B.R. Goldstein, The Alfonsine Tables of Toledo

(Dordrecht and Boston, 2003); M. Lejbowicz, Prsentation de Jean de Murs observateur et

calculateur sagace et laborieux , in C. Grelland (ed.), Mthodes et statut des sciences la

fin du Moyen ge (Villeneuve d Ascq, 2004), 159180; R.L. Kremer, John of Murs, Wenzel

Faber and the Computation of True Syzygy in the Fourteenth and Fifteenth Centuries, in

J.W. Dauben et. al. (eds), Mathematics Celestial and Terrestrial: Festschrift fr Menso Folkerts

zum 65. Geburtstag (Halle [Saale], 2008), 147160.

john of murss tables of 1321 309

Together with John of Lignres and John of Saxony, he recast these tables into

what are now called the Parisian Alfonsine Tables.

Although in 1317 John of Murs was a convinced defender of the Toledan

Tables, as attested in his earliest known work, beginning Auctores kalendarii ,

a few years later he adhered to Alfonsine astronomy and remained faithful to

it thereafter. In particular, in his Expositio intentionis regis Alfonsii circa tabulas

eius (probably composed in 1321) he explained parameters and models already

found in these tables and described many of the features of what he called the

tables of Alfonso, but remained silent on the way he got access to them. In any

case, the Alfonsine material was available to him (and others) in Paris by 1321.2

Reconstructing the transmission of astronomical ideas is a complex task

that is especially difficult for a period when scholars rarely mentioned the

names of their contemporaries or near contemporaries on whose work they

depended, as is the case in the late Middle Ages. In his Expositio John of

Murs mentions only one of his predecessors in Paris, William of Saint-Cloud

(end of the thirteenth century), and he even reproduces parts of Saint-Clouds

Almanach Planetarum of 1292 almost word for word (without marking any of

these passages as quotations).3 Yet some parameters of Alfonsine origin that

John of Murs incorporated into his own tables (such as the maximum value

of the solar equation, 2;10) had already reached Paris, for we find them in the

tables of John Vimond for 1320.4 It seems unlikely that John of Murs did not

know about John Vimond since both astronomers came from the same region,

Normandy, and both worked on planetary tables in Paris at the same time.

John Vimond composed his tables for the use of students at the University

of Paris and was therefore a known participant in the Parisian astronomical

community.

The texts under consideration in this paper have not previously been studied

in detail,5 but we claim that they are of great importance for understanding

the transmission of Alfonsine astronomy from Toledo to Paris. In fact, there

were at least five texts produced in Paris in 1320 and shortly thereafter that

bear on this transmission: (1) John Vimonds tables of 1320, (2) John of Murss

3 See E. Poulle, Jean de Murs et les tables alphonsines, Archives dhistoire doctrinale et littraire

du moyen ge, xlvii (1980), 241271, especially pp. 261265; Chabs and Goldstein, Alfonsine

Tables of Toledo (ref. 1), 263264, 279280.

4 See B.R. Goldstein and J. Chabs, The Maximum Solar Equation in the Alfonsine Tables,

Journal for the History of Astronomy, xxxii (2001), 345348; J. Chabs and B.R. Goldstein, Early

Alfonsine Astronomy in Paris: The Tables of John Vimond (1320), Suhayl, iv (2004), 207294.

5 For a brief account, see North, Alfonsine Tables in England (ref. 1), 284285.

310 chapter 9

Expositio, (3) John of Murss tables of 1321, (4) John of Murss Patefit, and (5)

John of Lignress canons and tables of 1322. We have previously discussed the

evidence in items (1), (2), and (5), and in this paper we focus on items (3) and

(4).6 As we noted in an article published in 1994 (with Jos Luis Mancha), one

lengthy chapter on planetary velocity in the Latin canons to the tables of the

Parisian astronomer, John of Lignres (fl. 13201335), is almost identical with

a chapter in the Castilian canons to the Alfonsine Tables of Toledo, composed

some 50 years earlier.7 As we demonstrate in this paper, John of Murss tables of

1321 and the tables for syzygies in his Patefit are based on the same models and

parameters that underlie the Parisian Alfonsine Tables (although the formats

for these tables are entirely different), indicating that, for matters other than

presentation, John of Murss contribution to Alfonsine astronomy was made all

at once. Some of these parameters are not known from any text or table prior

to those of John Vimond and John of Murs, although others can be discerned

in Castilian material of the late 13th century.8 Hence, what we learn from items

(3) and (4) confirms our results, based on items (1), (2), and (5). In other words,

despite the fact that the Alfonsine Tables of Castile are not extant, the evidence

we present strongly supports the claim that the Parisian material produced in

the 1320s relied on a Castilian tradition associated with King Alfonso x, as John

of Murs himself asserts.9

John of Murs compiled several sets of tables. They have not yet been thor-

oughly examined, except the set called Tabulae permanentes, which is re-

stricted to the computation of the time from mean to true syzygy.10 The tables

of 1321 are his first and most extensive set, and they are entirely devoted to the

planets and the two luminaries, the Sun and the Moon, that is, matters related

6 See, e.g., Chabs and Goldstein, Alfonsine Tables of Toledo (ref. 1), 266284; Chabs and

Goldstein, John Vimond and the Alfonsine Trepidation Model, Journal for the History of

Astronomy, xxxiv (2003), 163170; Chabs and Goldstein, John Vimond (ref. 4).

7 B.R. Goldstein, J. Chabs, and J.L. Mancha, Planetary and Lunar Velocities in the Castilian

Alfonsine Tables, Proceedings of the American Philosophical Society, cxxxviii (1994), 6195.

8 See J. Chabs, Were the Alfonsine Tables of Toledo First Used by Their Authors?, Centau-

rus, xlv (2003), 142150.

9 For the evidence in John of Murss Expositio, see Goldstein and Chabs, Maximum Solar

Equation (ref. 4), 347 n. 2. For contrary views, see E. Poulle, The Alfonsine Tables and

Alfonso x of Castille, Journal for the History of Astronomy, xix (1988), 97113; and idem,

Les astronomes parisiens au xive sicle et l astronomie alphonsine, Histoire littraire de

la France, xliii (2005), 154.

10 B. Porres and J. Chabs, John of Murss Tabulae permanentes for finding true syzygies,

Journal for the History of Astronomy, xxxii (2001), 6372.

john of murss tables of 1321 311

to the daily rotation and trigonometry are not mentioned.11 That John of Murs is

the author of these tables is attested by a note accompanying the text that refers

to them as istas tabulas magistri Jo. de muris (Oxford, Bodleian Library, ms

Can. Misc. 501, f. 106v). A close examination of this material throws light on the

transmission of Alfonsine parameters, although the author does not address

this issue at all.

As far as we know, John of Murss Tables of 1321 are extant in only two

manuscripts: Lisbon, ms Ajuda 52-xii-35 (henceforth ms L); and Oxford, Bod-

leian Library, ms Can. Misc. 501 (henceforth ms O). They are accompanied

by a text (L 65r66v, O 103r105v) consisting of about three pages and begin-

ning Si vera loca planetarum per presentes tabulas volueris invenire a tempore

incarnationis domini dato perfecto deme 1320 After roughly one page we read

Expliciunt canones super revolutiones planetarum. These very short canons

give some indication of the way to use the tables, but the text is too condensed

to be meaningful for anyone without previous familiarity with them, that is

John of Murs was either writing for himself or for a very select audience. This

is consistent with a sentence where he refers to his tables: in hac arte nulli sci-

entifico ignotum est. Several commentsor notesare appended concerning

tables for conjunctions with the Sun, mostly on the periods of anomaly of the

Moon and the planets. At the end of these comments there is another explicit:

Explicit compositio tabularum de certis revolutionibus planetarum.

In this text John of Murs avoided the type of canons which explain at length

how to use a set of tables, such as those written by John of Saxony in 1327 for the

Parisian Alfonsine Tables, filling 30 pages in the editio princeps of the Alfonsine

Tables,12 or the Castilian canons of the Alfonsine Tables of Toledo, completed

no later than 1272 which explain the use of these tables in 54 chapters.13

Unfortunately, as mentioned above, the original tables that corresponded to

these canons are not extant.

A description of the tables follows, based on both manuscripts. We note that

physical signs of 60 are generally, although not systematically, used in both

manuscripts (here incorporated into sexagesimal notation such that a circle

contains 6,0 = 6 physical signs).

11 For a brief account of their contents, see Poulle, Les astronomes parisiens (ref. 9), 2426,

where these tables are dated no earlier than 1325.

12 Tabule astronomice illustrissimi Alfontij regis castelle, edited by E. Ratdolt (Venice, 1483), ff.

a2rb8v.

13 For a transcription of these canons, see Chabs and Goldstein, Alfonsine Tables of Toledo

(ref. 1), 1994.

312 chapter 9

(h) () () ()

1 113 Apr 23 17;29 0,39;54 2,28;31 0,36;43

2 126 May 6 19;41 0,52;34 2,41;10 0,50;35

1 Tables in L and O

L 26rv and O 55rv display a table for the mean conjunctions of the Sun and

Saturn (Table 1). It is called tabula principalis in its heading and lists the dates

of 58 successive conjunctions of the Sun and Saturn from 1320 to 1359 (where

no such conjunction took place in years 21 and 50 after the radix), as well as

the corresponding motus (mean motion), centrum (mean argument of center),

and verus locus (true longitude) of the planet. The title of this table is: Tabula

medie coniunctionis principalis solis et saturni secundum radices per Alfonsum

regem Castele (L adds ultimo) verificatas super Toletum (L: Tollectum) distans

a Parisius in occidente per 48 m. hore (L: 48 hore). It is highly significant that the

title in both manuscripts indicates that the radices, i.e., the values correspond-

ing to the initial time, were verified (verificatas) for Toledo by Alfonso x, king

of Castile. The title also specifies that Toledo is located 48 minutes of an hour

west of Paris (L erroneously reads 48 hours). The same reference to Toledo and

to King Alfonso is also found in all the tables discussed below with radices for

the rest of the celestial bodies, a clear sign that John of Murs took the initial

values from Alfonsine astronomy as computed in Toledo and then converted

these data to the meridian of Paris. And this is indeed the case, as we shall

see.

The first column in Table 1 displays the years from 1320 to 1359; the second

column displays the number of days in that year (counted from the last day

of the previous year) that have elapsed; the third column replaces the number

of days displayed in the second column by the date. Below the table are the

amounts to be added to the first entry in a given column to arrive at the final

entry in that column.

john of murss tables of 1321 313

As derived from the entries, the time between two successive mean conjunc-

tions of the Sun and Saturn, is 1 year, 13 days, and 2;12h (L 65v and O 103v give a

more precise value: 378 days 2;12,13,12h). Thus, in a little over 59 years, 57 mean

conjunctions of the Sun and Saturn occur.

According to our calculations, using a spreadsheet to compute astronomical

positions with the standard Alfonsine Tables,14 a mean conjunction of the Sun

and Saturn took place in Toledo on 10 April 1321 at 15;17h at a mean longitude

of 27;14,25 for both celestial bodies, when the true longitude of Saturn was

22;57,59 and its mean argument of center was 135;51,34, in good agreement

with the values corresponding to the radix. It follows that the epoch 1320

has to be understood as 1320 completed, that is, 1321 current, which is our

usual reckoning.15 It also indicates that these entries were computed using a

solar model with a maximum value of the solar equation of 2;10 and a mean

motion of 0;59,8,19,37,19,13,56/d, as well as a model for the motion of Saturn

with a maximum value of the equation of center of 6;31, a maximum value

of the equation of anomaly of 6;13 and a mean motion of 0;2,0,35,17,40,21/d,

corresponding to the basic parameters of the standard Alfonsine Tables used

in the aforementioned programme for the Sun and Saturn.

14 For recomputations according to the editio princeps of the Alfonsine Tables (1483), we have

used a spreadsheet provided to us by Richard L. Kremer (Dartmouth College, usa), which

was prepared by Lars Gisln (Lund University, Sweden). By the standard Alfonsine Tables

we mean the collection of tables found in the editio princeps, edited by E. Ratdolt (ref. 12)

that, by and large, goes back to a compilation made in Paris in about 1327 with canons by

John of Saxony. Among many others, it includes: tables for the differences between the

eras, tables to transform dates of various eras, a set of radices for various eras, a table for

the movement of the 8th sphere (with a maximum of 9;0), tables of the mean motions

(presented as 60 consecutives multiples of the daily mean motions), equations of the lumi-

naries (with maximum values of 2;10 and 4;56 for the Sun and the Moon), equations of

the planets, etc. There is no modern edition of these tables which would have to be based

on the vast number of extant manuscripts, but we have examined many manuscripts con-

taining them, and none has exactly the same collection as the first edition. However, for

example, Madrid, Biblioteca Nacional, ms 10002, and Vienna, Nationalbibliothek, ms 2288,

share most of the characteristics of the editio princeps although, in contrast to it, they do

not have a star table.

15 There were two conventions for dates counted from the Incarnation in the Middle Ages:

(1) the number of years that have been completed, and (2) the current year which has not

been completed. So for a date as we reckon it in current years, e.g., noon, 15 Feb. 1321, one

might say equivalently that 1320 years have been completed plus 1 month (January) plus

14 days.

314 chapter 9

table 2 Correction for Saturn when not in conjunction with the Sun (excerpt)

() () () () () () () () () ()

3 4 25 4;57 2;56 35 1;46 1;16 29 0;17 5;19 34 4;11 37

54 0 378 2;12 11;24 33 10;18 9; 7 33 10;13 13;44 35 12;34 35

In order to find the true longitude of Saturn when the planet is not in con-

junction with the Sun, another table, called contratabula (Table 2), is needed. It

appears on O 56r58v and on three folios in L, labelled successively 27rv, 26rv,

and 27rv, for the manuscript has two folios numbered 26 and two numbered

27.

In the heading of Table 2, w stands for week, d for day of the week,

and hd for half the difference (in minutes of arc) between two successive

values of the mean argument of center of the planet. In this double argument

table of 34 columns and 30 rows, the first four columns are for the time after

a mean conjunction, that is, the age of Saturn (given as a number of days,

expressed in weeks and days within a week, and hours). The rows are evenly

separated, and the last one corresponds to 54 weeks (378 days or 1 year and

13 days) and 2;12h. As indicated above, this is roughly the time between two

successive mean conjunctions of Saturn and the Sun, and consequently it is

equal to the time between the radix and the first mean conjunction listed in

the table called principalis (see Table 1, above). The headings for the rest of

the columns represent values of the mean argument of center of Saturn, from

0,0 to 5,48 at intervals of 12. The entries give, in degrees and minutes, the

correction to be added, or subtracted, to the mean motion of the planet to

obtain its true longitude.

In Figure 9.1 we represent three rows of this double argument table, showing

the entries (in minutes of arc on the y-axis) for three selected times (12 days,

189 days, and 378 days) after a mean conjunction and for the mean argument

of center from 0 to 360 on the x-axis. As seen at a glance, for a given value of

the age of the planet, the entries are distributed along a sinusoidal curve with

a period of 360.

john of murss tables of 1321 315

figure 9.1 Correction for Saturn as a function of its mean argument of center (the curves, from

lowest to highest, correspond to 12 days, 189 days, and 378 days, respectively)

To illustrate how the contratabula works, let us derive the true longitude of

Saturn exactly one mean conjunction after the radix (text: 36;43; see Table 1,

second row), from the mean values of the preceding conjunction. One begins

by entering Table 2 with the mean argument of center (2,15;51) already given in

the Table 1. After interpolation between the values 2,12 and 2,24 for the mean

argument of center corresponding to a time of 54 weeks and 2;12h, one finds

that 9;29 is the positive correction to be applied. By adding this amount to the

mean motion of the previous conjunction, one finds 36;43 (= 27;14 + 9;29),

which is exactly the entry found in Table 1. The same procedure applies to any

other time after a mean conjunction. Direct recomputation with the standard

Alfonsine Tables confirms this result.

L 28r50r and O 59r80r display tables for the conjunctions with the Sun

(tabula principalis) and corrections (contratabula) for each of the other planets.

In all cases, the titles indicate that the tables were computed for Toledo and

use the radices given by King Alfonso. The times between successive mean

conjunctions with the Sun can be derived from the tables, but the text on

L 65r66v and O 103r105r gives more precise data:

Jupiter 398 days 21;12, 8,24h

Mars 779 days 22;22,34h

Venus 583 days 22;14, 3h

Mercury 115 days 21; 5, 2, 7h

316 chapter 9

Moon (excerpt)

(h) () ()

1 18 2; 3 5, 5;38 4,26;53

2 7 10;51 4,54;55 3,36;41

76 28 11;22 5,16;39 1, 4;49 *

* L: 1,44;49.

These values for the periods of anomaly of the planets are not difficult to

recompute. For the three superior planets, they result from dividing 360 by

the difference between the daily mean motions of the Sun and each planet,

whereas for the inferior planets they are obtained by dividing 360 by the daily

mean motions in anomaly of each planet. All these periods are consistent with

the standard mean motions of the planets in Alfonsine astronomy, as they

appear, for example, in the editio princeps of 1483. As was the case for the

radices of Saturn, the entries corresponding to those of the other planets can

be recomputed from the standard Alfonsine Tables. Therefore, all the values for

the mean motions of the planets that were later used in the Parisian Alfonsine

Tables are already embedded in the Tables of 1321 by John of Murs.

Next we find tables for the conjunctions of the Sun and the Moon (L 50v57v,

O 81r87r). They share the same format as those for the five planets, and

have the same references to King Alfonso and Toledo in the title. The first

table (Table 3), also called tabula principalis, lists the dates of the first mean

conjunction of the Sun and the Moon for each year in 76 consecutive years, as

well as the corresponding motus (mean position) and argumentum lune (mean

lunar anomaly).

Although no date is specified in Table 3, the first row corresponds to the

first conjunction of 1322. Indeed, recomputation with the standard Alfonsine

Tables indicates that a mean conjunction between the Sun and the Moon took

place at Toledo on 18 January 1322 at 2;3h, when the mean longitude of the Sun

and the Moon, here called motus, was 305;38, in perfect agreement with the

tabulated data. The rest of the rows correspond to the first conjunction of the

luminaries (occuring in January) of each year after 1322. This table thus covers

76 years, i.e., the least common multiple of a solar cycle of 4 years and a lunar

john of murss tables of 1321 317

cycle of 19 years. Contrary to the tables for the planets, this table begins in 1322,

and it is possible that the row for the radix (1321, that is, 1320 completed) is

missing in both manuscripts. In that case, one should expect to find 28 January

at 17;14h (time), 5,16;21 (mean longitude), and 5,17;5 (mean lunar argument

of anomaly), according to our computation for Toledo using the standard

Alfonsine Tables. This is confirmed by the table for mean conjunctions given

in the Patefit (see Table 14, below).

L 51r and O 81r display three short tables related to the Moon and they contain

the following data for one lunation:

Mean motion 29; 6,24,12

Mean lunar anomaly 25;49, 6,30

Mean argument of lunar latitude 30;40,13,48

Daily mean motion of the Moon 0;13,10,35, 0/d

Once again, the title refers to King Alfonso and indicates that his radices are

not included in the mean motion: Tabula mediarum coniunctionum solis et

lune infrascriptarum absque radice per Alfonsum. All these data later became

characteristic of the Parisian Alfonsine Tables.

As was the case for the planets, in order to find the true longitude of the

Moon when not in conjunction with the Sun, another table, again called con-

tratabula, is needed. It is displayed on L 51v57r and O 81v87r (Table 4).

In this double argument table the entries are presented in 60 columns and

30 rows. The vertical argument is the time after a mean conjunction, that is,

the age of the Moon (given here in days) and the horizontal argument is the

mean lunar argument of anomaly from 0,6 to 6,0 at intervals of 6. Again, hd

stands for half the differences (in minutes of arc) between two successive values

of the lunar argument for a given time. For example, 0;14 is half the difference,

rounded to minutes, between 0,11;24 and 0,10;57, corresponding to 0,6 and

0,12, respectively, for day 1; hence the entry 14 in the first column labelled

hd in that row. The entries, in degrees and minutes, give the correction to

be added to the mean motion of the Moon to obtain its true longitude, as

can be seen in the following example. Consider the mean conjunction for 18

January 1322, which occurred at 2;3h at Toledo; the mean motion and the mean

lunar argument of anomaly at that time were 5,5;38 and 4,26;53, respectively

(see Table 3). If we wish to know the position of the Moon one or more

days after that time, we enter in Table 4, with the mean lunar argument of

318 chapter 9

table 4 Correction for the Moon when not in conjunction with the Sun (excerpt)

() () () () () () () ()

2 0,23;14 12 0,22;50 0,29;30 0,24; 7 13 0,23;40 13

30 0,32;26 12 0,32; 3 0,38; 6 0,33;14 13 0,32;52 12

* L: 0,20;6.

** L: 11.

anomaly. By interpolation between the values 4,26 and 4,30 we find, for

exactly, say, one day after mean conjunction, 17;56 as the correction to be

applied. Adding this value to the mean motion at conjunction, we find 5,23;34;

recomputation based on the standard Alfonsine Tables yields 5,23;34,41, in

very good agreement with it.

Figure 9.2 shows the behavior of the correction for the Moon (in degrees) as

a function of the mean lunar argument of anomaly for a given time of the syn-

odic month or age (taken here as 20 days after a mean conjunction), whereas

Figure 9.3 displays the correction of the Moon as a function of the age of the

Moon for a given value of the mean lunar argument of anomaly (taken here as

4,0).

As was the case for the planets, the values for the mean motion of the

Moon as well as the maximum lunar equation (4;56) and the maximum solar

equation (2;10), all parameters later used in the Parisian Alfonsine Tables, are

already embedded in John of Murss tables. In his Patefit, John of Murs also

compiled other tables for true syzygies, and they will be examined in section 4,

below.

It is very interesting to contrast John of Murs tables for the mean conjunc-

tions of the luminaries and the position of the Moon (Tables 3 and 4) with

those for the same purpose by John Vimond.16 Vimonds tables are only extant

16 Chabs and Goldstein, John Vimond (ref. 4), 213 and 229, Tables 1 and 8.

john of murss tables of 1321 319

figure 9.2 Correction for the Moon for a given age (20 days)

figure 9.3 Correction for the Moon for a given value of mean lunar argument of anomaly (4,0)

and the same manuscript also contains the canons and tables for 1322 by

John of Lignres. In Vimonds tables the epoch of the first mean conjunction

is 10 March 1320 (computed for Paris), whereas in John of Murss table it is

18 January 1322 (computed for Toledo), indicating that John Vimonds tables

preceded those by John of Murs. Both astronomers use values of the mean

synodic month that are almost identical: the value in John Vimonds tables

is not explicit, but it can be derived from them, namely, 29 days 12;44,3,6h

320 chapter 9

12;44,3,3h. The difference amounts to 3 sixtieths of a second of an hour. It also

happens that both use double argument tables to find the true position of

the Moon between syzygies, and the tables look very similaralthough the

vertical and horizontal arguments are switchedand the entries are given

to the minutes in both cases. However, John of Murss table is larger, for the

argument of anomaly is taken at 6-intervals in contrast to 12-intervals in the

tables of John Vimond, and covers 30 days whereas that of Vimond covers only

14 days; hence, it has more than four times as many entries (60 30 for John of

Murs and 30 14 for John Vimond). We note that in John of Murss tables the

entries for half the differences between the corrections corresponding to two

consecutive values of the argument of anomaly are given to one place (minute

of arc), whereas in John Vimonds tables the entries for interpolation represent

another quantity, the difference between two consecutive corrections for the

same value of the argument of anomaly, and are displayed to two places (second

of arc). That is, we are offered two interpolation schemes that complement

each other: that of John de Murs applies for a fixed day, whereas that of John

Vimond is for a fixed argument of anomaly.

The 420 entries for the corrections common to both tables agree, except for

copyists errors (see Table 5 for a comparison between two selected columns in

both tables).

This does not happen by chance. We are thus faced with two possibilities: the

most likely is that John of Murs used the principle established by John Vimond

and expanded the table, but it could also be that they both depended on a table

in an otherwise unknown prior text.

The approach for the motion of the lunar nodes (L 59v, O 87v) parallels that

for the Sun. In this table we are given the true position of the node for 1320

(4,57;14) and then at four year intervals to 1392, adding a final row for 1393,

thus using the standard period of 93 years used by astronomers in the Alfonsine

tradition.17 At the bottom of this page is a short table for the increments in years

1, 2, and 3 within the four-year calendaric cycle (for year 1 the entry is 5,4;40

in both manuscripts, whereas they should read 5,40;40). Another sub-table

lists the true motion of the node for the days of the year (given by the month,

beginning in January, and the day within the month), at intervals of 6 days.

Note that the entry corresponding to 31 December is correctly given as 5,40;40.

17 E.g. Abraham Zacut: see J. Chabs and B.R. Goldstein, Astronomy in the Iberian Peninsula:

Abraham Zacut and the Transition from Manuscript to Print (Philadelphia, 2000), 60 and

117.

john of murss tables of 1321 321

and 180) of the tables of John Vimond and John de Murs for

the correction of the Moon

Days 0,12 0s12 3,0 6s 0

() (s) () () (s) ()

2 0,22;50 0 22;50 0,29;30 0 29;30

3 0,34;45 1 4;45 0,44; 5 * 1 14;11

4 0,46;48 1 16;48 0,58;38 1 28;38

5 0,59; 3 1 29; 3 1,12;44 2 12;44

6 1,11;46 2 11;46 1,26;20 2 26;20

7 1,24;47 2 24;47 1,39;36 3 9;36

8 1,38; 5 3 8; 5 1,52;36 3 22;36

9 1,51;53 3 21;53 2, 5;11 4 5;11

10 2; 6;12 4 6;12 2,17;22 4 17;22

11 2,20;57 4 20;57 2,29;14 4 29;14

12 2,35;55 5 5;55 2,40;56 5 10;56

13 2,50;51 5 20;51 2,52;39 5 22;39

14 3, 5;37 6 5;37 3, 4;27 6 4;27

The daily motion of the node resulting from the table is 0;3,10,38,11,34/d,

in good agreement with the same parameter in the Parisian Alfonsine Tables

(0;3,10,38,7,14/d), and quite far from other parameters historically used for

the motion of the nodes. As was the case with the previous tables, the title

indicates that the radix was derived from the tables compiled by Alfonso x,

King of Castile, for Toledo. And it is indeed so: our recomputation, using the

standard Alfonsine Tables for noon of the first day of year 1321, indicates that

the true longitude in Toledo was 4,57;12,12, in close agreement with the entry

in the text.

Then come tables for the planets: Saturn (L 60rv, O 88rv), Jupiter (L 60r and

61r, O 88r and 89r),18 Mars (L 61v62r, O 89v90r), Venus (L 62v63r, O 90rv), and

18 The columns for the arguments, the equation of center, and the stations of Saturn and

322 chapter 9

and argument of center of and stations of Saturn

Saturn (excerpt) (excerpt)

Days () () () () () ()

25 0,24 0;51 0,12 5,48 1;17 22;47

38 0,36 1;16 0,18 5,42 1;55 22;49

365 5,48 12;15 2,54 3, 6 0;43 25;28

378 0, 0 12;40 3, 0 3, 0 0; 0 25;30

* L: 0;15.

of the planets with the Sun described above, for each of the five planets we are

given three tables: one for the argument of anomaly and the argument of center,

a second for the equation of center and the stations, and a third for latitude.

Excerpts of the three tables for Saturn follow (Tables 6, 7, and 8).

In Table 6 the first column is for the argument, expressed in days within a

period of anomaly (378 days 2;12,13,12h), as indicated above. The second column

represents the mean motion of the argument of anomaly, in degrees, and the

third column displays the mean motion of the argument of center, in degrees

and minutes.

In Table 7 the first two columns represent the mean argument of center, in

degrees. The third column is for the equation of center of Saturn, in degrees

and minutes. Its maximum, 6;31, agrees with that in the Almagest, the zij of

al-Battn, and the Toledan Tables, as well as the Parisian Alfonsine Tables. The

fourth column represents the first station, in degrees and minutes, but we note

that the zodiacal signs (required for the entries to be meaningful) are not given,

a feature shared by both manuscripts that will be addressed below.

Jupiter are displayed on a single page, whereas the latitudes of these two planets are

presented separately.

john of murss tables of 1321 323

The tables for the planets other than Saturn present the same characteristics.

The maximum values for the equation of center of the other planets are 5;57

(Jupiter), 11;24 (Mars), 2;10 (Venus), and 3;2 (Mercury). This indicates that

John of Murss Tables of 1321 adhere to the tradition represented by the Toledan

Tables for all planets except for Jupiter and Venus, where the latter have 5;15

and 1;59, respectively. However, all 5 parameters used by John of Murs were

already in Paris at the time, for they are found, once again, in the tables of John

Vimond, including the new ones for Jupiter and Venus.19 For the stations of

Saturn and Jupiter there is no column for the signs, in contrast to each of the

other three planets. This is most peculiar in the case of Saturn, for which we

are given entries ranging from 22;45 (for 0,6 = 6) to 25;30 (for 3,0 = 180), as

shown in Table 7. Now, in the Toledan Tables, which use zodiacal signs of 30,

the entries range from 3s 22;44 (for 0s 0) to 3s 25;30 (for 6s 0 = 180), with 3s

22;45 for 0s 6. It would therefore seem that John of Murs, who used physical

signs of 60 in this very same table, took the entries for Saturn from a table,

such as that in the Toledan Tables, using zodiacal signs of 30, and omitted the

signs. In sexagesimal notation the values displayed would range from 1,52;45 to

1,55;30. We mentioned the Toledan Tables as an example only, for other tables

with the same values, using zodiacal signs, were available in Paris at the time:

see, e.g., those of John Vimond which display the same entries, although shifted

differently depending on the planet.20

The tables for planetary latitudes are of great interest because they are

also presented as double argument tables, a feature for which we know of

no precedent. As is often the case, superior and inferior planets are treated

differently. Table 8 reproduces an excerpt of the table for Saturn.

For the superior planets, the entries, in degrees and minutes, are presented

in 7 columns and 31 rows (L 60v61r and 62r; O 88v89r and 90r). The vertical

argument is the argument of center of the planet, shifted +50 in the case of

Saturn, 20 for Jupiter, and with no shift for Mars, in accordance with the

instructions given in Almagest xiii.6. In the heading, hd stands for half the

difference (in minutes of arc) between the entries of two successive columns

for a fixed argument of center. The extremal values appear in the column

corresponding to 180:

Jupiter +2; 5 (North) 2; 8 (South)

Mars +4;21 (North) 6;30 (South)

20 Chabs and Goldstein, John Vimond (ref. 4), 242.

324 chapter 9

() () () () () () ()

316 304 2; 1 3 2; 7 2;55 3 3; 0

40 220 0; 0 0 0; 0 0; 0 0 0; 0

46 214 0;13 0 0;13 0;19 0 0;20

130 130 2; 1 2 2; 5 2;57 4 3; 5

We note that the northern (positive) and southern (negative) limits for

Saturn agree with the tradition (Almagest, al-Battn, Toledan Tables, and also

Vimond) but differ from the values that were to become part of the Parisian

Alfonsine Tables (+3;3 and 3;5, respectively). As for Jupiter, the northern

limit agrees with that in the Toledan Tables and the tables of Vimond, and

departs slightly from that in the Almagest and the zij of al-Battn (+2;4) and

in the Parisian Alfonsine Tables (+2;8), whereas the southern limit agrees with

all the sets of tables mentioned so far. This is also the case for the northern limit

of Mars, but not for its southern limit (Almagest and al-Battn: 7;7; Toledan

Tables, Vimonds tables, and Parisian Alfonsine Tables: 6;30).

For each of the inferior planets, we are given two double-argument tables

for latitude (L 63r and 64r O 91r and 92r): one for the inclination (here called

declinatio) and one for the slant (reflexio). The entries, in degrees and minutes,

are displayed in 7 columns and 16 rows. Tables 9 and 10 display excerpts for the

latitude of Venus.

In all four tables (two for each of the inferior planets), the first four columns

are for the argument, which is shifted +60 in the table for the inclination of

Venus and +90 in the table for the inclination of Mercury. They are small

double argument tables, with 7 columns and 16 rows, where the maximum

value for Venus is 7;22 (cf. Almagest and the zij of al-Battn: 6;22; Toledan

Tables: 7;24; and Vimonds tables and the editio princeps of the Alfonsine

Tables: 7;12) and that for Mercury, 4;5 (in agreement with the values in the

Almagest, the zij of al-Battn, the Toledan Tables, and Vimonds tables, as

well as in the editio princeps of the Alfonsine Tables). Most notable is the

john of murss tables of 1321 325

() () () () () () ()

306 244 114 126 1; 3 3 0;57 3; 2 127 7;16

30 210 30 210 0; 0 0 0; 0 0; 0 0 0; 0

3 latitudo in L), in the tables for the slants of both inferior planets. This is

indeed a very unusual feature in medieval tables, for there are not many which

list values for the deviation. These values were mentioned, but not tabulated,

in Almagest xiii.6, with a limit of 0;10 (north) for Venus and a limit of 0;45

(south) for Mercury. It is significant that the canons to the Castilian Alfonsine

Tables explicitly addressed this particular problem, giving instructions to take

into account the third component of latitude appearing in the tables they

describe.21 To this we can add that John Vimond included the deviation of

Venus and Mercury in his own tables and used the values mentioned in the

Almagest.22 As can be seen in Table 10, John of Murs gave 0;10 (septentrionalis)

for the maximum deviation of Venus, but for that of Mercury he used 0;23

(meridionalis), a value about half that of Ptolemy and all others who worked in

the Ptolemaic tradition. There are no tables for the third component of latitude

in the Parisian Alfonsine Tables.

The last table in both manuscripts is for the days in a year (L 64v, O 102v),

where each day of the year is assigned an ordinal number from 1 to 365.

In addition to the contents of John of Murss Tables of 1321 that appear in

both manuscripts, there are a few other tables closely related to them that are

only found in one of the two manuscripts.

22 Chabs and Goldstein, John Vimond (ref. 4), 257258.

326 chapter 9

() () () () () () () ()

6 354 174 186 0; 0 20 0;41 2;20 70 0; 0 10

90 270 90 270 0; 0 0 0; 0 0; 0 0 0; 0 0

On L 57v58v there is a table for the true positions of the Sun for each day in

a year: see Table 11. The table is also called principalis, and in the title we are

again told that the radices were computed by King Alfonso of Castile for the

city of Toledo. The entries are given in physical signs, degrees, minutes, and

seconds. The date for which the table is valid is not indicated, but it corresponds

to 1321, as shown from the recomputations displayed in Table 11, where the true

positions of a sample of entries are compared with recomputations for 1321.

Note that the values for text and computation do not differ by more than 7

seconds.

Reinforcing our claim that Vimonds tables were composed prior to those by

John of Murs is the fact that Vimond displays a one-year calendar with syzygies

valid for 1320. It would seem odd for either astronomer to compile a one-year

table for the (recent) past. It is far more likely that such a table was produced

at (or near) the beginning of the year in question.

The Sun has also its contratabula (L 59r), of which an excerpt is given in

Table 12, showing the extremal values of the various columns. Note that this

contratabula differs from those reviewed above, among other things for the fact

that it is not a double argument table, although its purpose is substantially the

same.

The argument in Table 12 is the day of the year given at intervals of 6 days,

beginning in January. The second column lists the correction to be subtracted

from the true position of the Sun for 1321, found in the preceding table, in order

to determine its true position for dates corresponding to one, two, or three

years after 1321. For example, let us consider the true longitude of the Sun for

john of murss tables of 1321 327

tc

Text (t) Computation (c) (in seconds)

Feb. 1 5,21;17,33 5,21;17;33 0

Mar. 1 5,48;19,14 * 5,49;19,13 +1

Apr. 1 0,19;46,14 0,19;46,18 4

May 1 0,48;43, 1 0,48;43, 1 0

June 1 1,18;16,45 1,18;16,45 0

July 1 1,46;47,35 1,46;47,36 1

Aug. 1 2,16;24,35 2,16;24,35 0

Sep. 1 2,46;20,24 ** 2,46;26,24 0

Oct. 1 3,16; 2, 2 3,16; 2, 2 0

Nov. 1 3,47;12,50 3,47;12,50 0

Dec. 1 4,17;47,50 4,17;47,57 7

* Instead of 5,49;19,14.

** Instead of 2,46;26,24.

6 January 1322. The procedure is to note the true solar longitude exactly one

year before, 6 January 1321, which is given in Table 11 as 294;53,39. Then we

have to subtract from it the difference for one year displayed in the column

labelled correction for that date in Table 12, and it is 0;14;52; the result is

294;38,47, in good agreement with our recomputation based on the standard

Alfonsine Tables (294;38,27). Note that the column for the correction exhibits

a minimum value of 0;13,48 (31 May30 June) and a maximum value of 0;14;56

(618 December). The approach to finding the position of the Sun in these two

tables is strongly reminiscent of an almanac, in the sense that we are given the

true position of the Sun at noon for all days in a year (tabula principalis) and a

simple procedure to find its true position at noon for each day in the next three

years (contratabula), with no indication of the underlying parameters and no

rule for computing true positions beyond this 4-year period.

The third column gives the hourly velocity of the Sun in minutes and seconds

per hour, from 0;2,23/h to 0;2,34/h, for purposes of interpolation between the

entries in the previous column. These are the same limits as in the Parisian

Alfonsine Tables.23

23 See, e.g., Tabule astronomice illustrissimi Alfontij (ref. 12), ff. g6rg7r.

328 chapter 9

() (/h) (min)

In the fourth column we find the equation of time. There are not many

medieval astronomical tables where the argument is expressed in days of the

year and the entries are given in time, as is the case for this table. In fact, we

only know of two of them, one in Abraham Zacuts ibbur and another in the

Tabule Verificate for Salamanca.24 In Table 12, the extremal values of the entries

are the following:

max: 0;20,58h 6 May

min: 0;12, 0h 1824 July

Max: 0;32,38h 24 Oct

5;14,30, 3;0, and 8;9,30. As far as we know, these values for the equation of

time are unprecedented in the astronomical literature. In particular, this table

differs both in format and in content from the table of Peter of St. Omer for

john of murss tables of 1321 329

and reproduced by John of Lignres, with the following extremal values:25

max: 5;21 Tau 2527

min: 2;49 Leo 5

Max: 7;57 Sco 89

However, all other parameters for the Sun in these tables are those that are

found in the Parisian Alfonsine Tables.

On O 92v98r there are five tables for the conjunctions of the planets with the

Sun that are very similar to those described in section 1. However, the entries are

different, for they correspond, as indicated in the titles, to 1452 perfecto (i.e.,

1453 in the usual reckoning) and were computed for Paris, not Toledo. The titles

add that the radices are those of Alfonso x, king of Castile. For each planet we

are only given the tabula principalis, on the assumption that the corresponding

contratabula to be used is the one already found among John of Murss Tables

of 1321. Note also that the column for the true position has been eliminated and

the accuracy of the entries has been improved. Table 13 displays an excerpt of

the table for the mean conjunctions of the Sun and Saturn, beginning in 1453

(1452 completed).

table 13 Mean conjunctions of the Sun and Saturn for 1453 (excerpt)

of the year and day (h) () ()

1 296 Oct 23 12;21,27 3,41; 0,17 5,28;16,49

2 309 Nov 6 14;33,40 3,53;40,10 5,40;56, 7

25 See Chabs and Goldstein, Alfonsine Tables of Toledo (ref. 1), 186.

330 chapter 9

It is clear that these tables do not belong to John of Murss Tables of 1321 and

are simply an extension of them made more than a century later, indicating

that his table was taken as a model by at least one subsequent user.

The same can be said about the tables on O 98r99v concerning the con-

junctions of the Sun and the Moon beginning in 1453 (1452 completed), which

follow the pattern of Table 3, above.

In addition to the Tables of 1321, John of Murs composed another work con-

taining tables, associated with the canon beginning Patefit ex Ptolomei disci-

plines in libro suo , and traditionally dated 1321.26 The Patefit survives in very

few copies. Three manuscripts have previously been noted: London, British

Library, Royal ms 12.c.xvii; Erfurt, Biblioteca Amploniana, mss 4 360; and

Erfurt, Biblioteca Amploniana, 4 371.27 We can now add to this short list Vati-

can, Biblioteca Apostolica, ms lat. 3116. We have also found other manuscripts

containing extracts of this work, as is the case with the Lisbon manuscript itself

(L 1r22v), described above in relation to the Tables of 1321. Among the tables

associated with the Patefit there are five directly concerning syzygies, two of

which are also double argument tables. For the description of the tables, below,

we have followed the London manuscript (henceforth called B), where zodia-

cal signs are used (except in one case), in contrast to the Tables of 1321 and the

Lisbon manuscript where physical signs are used.

The first is a table for the mean conjunctions and oppositions for a period

of 76 years from 1321 to 1396 (expressed in current years), covering 1880 (=

940 2) successive syzygies. The table has no title and, under the headings

conjunctions and oppositions, four quantities are given in each case: time

in days, hours, and minutes; mean motion of the Sun in zodiacal signs, degrees,

and minutes; mean lunar anomaly in zodiacal signs, degrees, and minutes;

mean argument of lunar latitude in zodiacal signs, degrees, and minutes (see

Table 14, an excerpt of B 155v168r).

The entries in Table 3, above (limited to mean conjunctions), agree with

the corresponding entries in Table 14 of the Patefit, except for the minutes in

a few cases. We also note that in this table, and in general in the Patefit, years

26 According to Poulle, this work was probably composed in the late 1320s: cf. Poulle, Les

astronomes parisiens, (ref. 9), 2627.

27 See, e.g., Kremer, John of Murs, Wenzel Faber (ref. 1), 148.

john of murss tables of 1321 331

Day Time motus arg. lune arg. lat. Day Time motus arg. lune arg. lat.

(h) (s) () (s) () (s) () (h) (s) () (s) () (s) ()

Feb 27 5;58 11 15;28 11 12;54 1 21;20 12 11;36 11 0;54 4 29;59 7 6; 1

Mar 28 18;42 0 14;34 0 8;43 2 22; 1 14 0;20 0 0; 1 5 25;48 8 6;41

Nov 20 0;35 8 7;25 7 5;15 10 27;22 5 6;13 7 22;52 0 22;20 4 12; 2

Dec 19 13;19 9 6;32 8 1; 4 11 28; 2 4 18;57 8 21;58 1 18; 9 5 12;42

Day Time motus arg. lune arg. lat. Day Time motus arg. lune arg. lat.

(h) (s) () (s) () (s) () (h) (s) () (s) () (s) ()

Feb 16 14;47 11 4;44 * 9 22;42 * 1 29;22 * 1 20;25 10 20;11 3 9;47 7 14; 3

Mar 18 3;31 0 3;51 10 18;31 3 0; 3 3 9; 9 11 19;18 4 5;36 8 14;43

* In L these entries are given as 5,34;45 (= 11s 4;45), 4,52;43 (= 9s 22;43), and 1,0;22 (= 2s

0;22), respectively.

Day Time motus arg. lune arg. lat. Day Time motus arg. lune arg. lat.

(h) (s) () (s) () (s) () (h) (s) () (s) () (s) ()

Dec 29 22;48 * 9 17;33 2 19; 5 0 20; ** 15 4;16 9 3; 0 8 6; 9 6 4;55

** Vatican, ms lat. 3116: 0s 20;14. The column for the zodiacal signs corresponding to the lunar

anomaly for oppositions is missing in this manuscript in most cases.

332 chapter 9

refer to the current year. Thus, Table 14 was also computed for the meridian

of Toledo, following the radices set up by King Alfonso, a feature not specified

in the title. The same table is found in Vatican, lat. 3116, ff. 11r23v, and all the

entries displayed above agree with those in B.28 In L the information contained

in this table is split into four different tables: the times of 940 successive mean

conjunctions for the 76-year period, 13211396; the mean motion of the Moon

for 235 successive conjunctions (that is, a lunar cycle of 19 years); the mean

lunar anomaly for 251 successive conjunctions (that is, the cycle according to

which 251 synodic months = 269 returns in lunar anomaly); and the mean

argument of lunar latitude for 223 successive conjunctions (that is, the Saros

cycle of about 18 years): for these cycles see, e.g., Almagest iv.2.29

The second table displays the times of true conjunctions and oppositions for

a period of 76 years (13211396), given in days, hours, and minutes (see Table 15,

an excerpt of B 168r172r). The title, Tabula continens veras coniunctiones et

oppositiones ad Tholetum per 48 m. hore distante a Parisius, indicates that the

entries were computed for Toledo, which is distant by 48 minutes of an hour

from Paris. This table is not in Vat. lat. 3116. In the heading of Table 15, m stands

for month. As expected, the entries in this table were computed with Alfonsine

models and parameters for the Sun and the Moon.

The third table deals with true conjunctions although in the three manu-

scripts we have examined (London, Vatican, and Lisbon) the title refers both to

conjunctions and to oppositions: Tabula vere coniunctionis et opposicionis solis

et lune. Five quantities are given: true longitude of the Moon at mean conjunc-

tion, in physical signs, degrees, and minutes; lunar time correction, in hours

and minutes; hourly lunar velocity, in minutes and seconds; true longitude of

the Sun at mean conjunction, in physical signs, degrees, and minutes; solar time

correction, in hours and minutes (see Table 16, an excerpt of B 172v174r). The

columns dealing with the Moon cover 251 conjunctions, whereas those for the

Sun only cover 235 conjunctions. It is noteworthy that in this table physical

signs of 60 are used, contrary to the other tables in the Patefit. Also in L 9r12v

physical signs are used. The same table is found in Vatican, lat. 3116, ff. 24r25v.

Note that the time correction for the Moon and the time correction for the Sun

are the two terms in which the time from mean to true syzygy, t, is divided.

28 On L 8rv there is another table containing much the same information with the title,

Tabula medie coniunctionis solis et lune in annis ad meridiem Tholeti secundum radices

Alfonsii regis castelle. We are only given entries for the last conjunction of each of the 76

years, but with a higher accuracy, both for time (to seconds of an hour) and for the three

other quantities (to seconds of an arc).

29 G.J. Toomer, Ptolemys Almagest (New York and Berlin, 1984), 174176.

john of murss tables of 1321 333

m. Time Conj. Time Opp. m. Time Conj. Time Opp.

(d) (h) (d) (h) (d) (h) (d) (h)

2 27 7;29 12 20;16 2 16 9;54 2 9;35

3 29 0;36 14 5; 8 3 18 1;36 3 21;26

11 19 17;29 5 5;17 11 9 9;41 24 2; 7

12 19 5; 6 5 1;32 12 9 0; 2 23 21;17

To illustrate this, consider the mean conjunction for 28 January 1321, occurring

at 17;14h after noon (see Table 14). The true conjunction occurs at 13;58h (see

Table 15). Thus, true conjunction precedes mean conjunction, and t = 3;16h.

As readily seem in Table 16, the sum of the lunar correction (6;40h) and the

solar correction (+3;24h) is 3;16h. Computing the time from mean to true

syzygy is an issue that interested many medieval astronomers and they offered

a variety of methods to give a proper answer.30 In a recent paper Kremer

analysed this table and argued convincingly (1) that cols. 2 and 5 are based on

the parameters for the lunar and solar corrections that later appeared in the

Parisian Alfonsine Tables; (2) that col. 4 is based on the lunar velocity table in

the Toledan Tables (which is identical with the corresponding table in the zij

of al-Battn); and (3) that cols. 3 and 6 are based on a computational scheme

similar to the one that underlies the table by Nicholaus de Heybech (fl. 1400)

for finding the time from mean to true syzygy, separating the solar and lunar

components.31

The fourth table has the heading Tabula veri motus lune ad dimidiam luna-

tionem. Argumentum lune ad coniunctionem mediam inventum (see Table 17, an

excerpt of B 177r182v). It is a double argument table where the horizontal scale

is for the argument of lunar anomaly, ranging from 0s 3 to 12s 0 at 3 intervals,

and the vertical scale is the number of days after mean conjunction or oppo-

30 See J. Chabs and B.R. Goldstein, Computational Astronomy: Five Centuries of Finding

True Syzygy, Journal for the History of Astronomy, xxviii (1997), 93105.

31 For additional details, see Kremer, John of Murs, Wenzel Faber (ref. 1), 148155.

334 chapter 9

True lunar Time correction Hourly lunar True solar Time correction

position for the Moon veloc. position for the Moon

() (h) (/h) () (h)

2 5,46;45 2;50 30;24 5,47;35 4;31

3 0,13;49 1;29 30;20 0,16; 8 4;25

236 5,21;22 9;44 32;59 0, 0; 0 0; 0

251 0,37;29 9; 5 31;46 0, 0; 0 0; 0

* We have added a minus sign () to the entries under deme (subtract) in the table.

sition, up to 16 days. The same table is found in Vatican, lat. 3116, ff. 29r34v.

An extract is also found in Brussels, Bibliothque Royale, ms 10861115, f. 29v.

This table is essentially the same as Table 4 but has twice as many columns and

about half as many rows. Another difference is the use of zodiacal signs rather

than physical signs. And yet the most striking difference is that the entries com-

mon to both tables differ slightly in a systematic way. In Table 18 we compare

excerpts of these two tables of John de Murs, one belonging to the Tables of 1321

(Table 4) and other to the Patefit (Table 17).

table 17 True lunar positions for each day between successive syzygies (excerpt)

(s) () () (s) () (s) () (s) () () (s) () ()

2 0 23;24 13 0 23;11 0 21;38 0 23;51 13 0 23;38 14

3 1 5;12 11 1 5; 1 1 5;26 1 5;36 12 1 5;24 12

16 7 4; 7 * 13 7 4;20 7 4;54 7 3;39 14 7 3;53 16

john of murss tables of 1321 335

and 180) of the two tables of John of Murs for the correction of

the Moon

Days 0,12 0s12 3,0 6s 0

() (s) () () (s) ()

2 0,22;50 0 22;47 0,29;30 0 29;33

3 0,34;45 1 4;41 0,44; 5 * 1 14;15

4 0,46;48 1 16;43 0,58;38 1 28;43

5 0,59; 3 1 28;58 1,12;44 2 12;49

6 1,11;46 2 11;41 1,26;20 2 26;25

7 1,24;47 2 24;42 1,39;36 3 9;41

8 1,38; 5 3 8; 0 1,52;36 3 22;41

9 1,51;53 3 21;48 2, 5;11 4 5;16

10 2; 6;12 4 6; 8 2,17;22 4 17;26

11 2,20;57 4 20;54 2,29;14 4 29;17

12 2,35;55 5 5;53 2,40;56 5 10;58

13 2,50;51 5 20;51 2,52;39 5 22;40

14 3, 5;37 6 5;38 3, 4;27 6 4;27

15 3,20;14 6 20;15 3,16;18 6 16;18

16 3,34;38 7 4;44 3,28;14 6 28;11

in the minutes, and the differences range from 0;5 to +0,5. We note, however,

that when computing the true position of the Moon exactly one day after

the mean conjunction of 18 January 1322 at Toledo we found it to be 5,23;34,

using the correction deduced from Table 4 (17;56) and in good agreement with

recomputation, 323;34,41. Had we used Table 17, the correction to be applied

would have been 18;1, and thus the true position of the Moon 5,23;39, in worse

agreement with recomputation.

There is a fifth table whose heading is Tabula invencionis veri loci lune incip-

iendo a coniunctione eius a sole (B 183r188v). Again, it is a double argument

table, for 30 days and the argument of lunar anomaly, and very similar to

Tables 17 and 4, the latter of which is also for 30 days. This table is not in Vat.

lat. 3116.

336 chapter 9

5 Conclusions

John of Murss Tables of 1321 are exclusively concerned with the luminaries and

the planets. Their most significant feature is the organizational principle: the

mean motions of the planets are presented in tables for the mean conjunctions

of each planet with the Sun, and their equations are given in double argument

tables. This approach meant that astronomers could avoid a lot of cumbersome

computations to determine true planetary positions, compared with using

tables previously available in Latin. In setting up his tables that way, John of

Murs took advantage for the planets of the pattern used for the Moon in its

conjunctions with the Sun. We know of no other example of this presentation

for the planets in the Alfonsine corpus.

On the other hand, as far as we know, double argument tables were a novelty

in Europe. North noted that they had been used by Ibn Ynus (tenth century,

Cairo) and al-Baghdd (thirteenth century, Baghdad), but we are not aware of

any double argument tables produced in the Iberian Peninsula prior to 1320.32

In the tables devoted to syzygies in the Patefit, John of Murs split the time

from mean to true syzygy, t, into two separate terms, one for the Sun and

one for the Moon, and was probably the first to introduce this approach to a

complicated problem. The tables he compiled, among them double argument

tables, helped practitioners of astronomy in completing their tasks by reducing

the number of required computations to reach a certain result.

John of Murs compiled his tables with material already in place in Paris

and, in particular, there is evidence that he shared parameters and approaches

with his contemporary, John Vimond. Indeed, we have shown that there is a

stronger relationship between John Vimonds work and John of Murss than

previously thought. As repeatedly indicated in the titles of the tables, John of

Murs used radices for Toledo which he attributed to Alfonso, king of Castile.

No wonder, for John of Murs had a thorough knowledge of what he called

the tables of Alfonso that he had described in his Expositio intentionis regis

Alfonsii circa tabulas eius. Indeed, in his Tables of 1321 John of Murs used

parameters that occur in that text. As Lejbowicz put it regarding John of Murs,

lappropriation des hritages fournit aux novateurs l appui ncessaire leur

travail.33 Moreover, it is also clear that all basic parameters for the planets

and the two luminaries (for mean motions and equations) that some years

32 North, Alfonsine Tables in England (ref. 1), 279 and 293; Chabs and Goldstein, Finding

True Syzygy (ref. 30), 93 and 104 n. 3.

33 Lejbowicz, Prsentation de Jean de Murs (ref. 1), 175.

john of murss tables of 1321 337

later formed the core of the Parisian Alfonsine Tables are already embedded

in his Tables of 1321. Hence credit should properly be given to John of Murs for

these innovations that played a decisive role in the transmission of Alfonsine

astronomy.

Acknowledgements

a draft of this paper.

chapter 10

The Diffusion of an Iberian Astronomical

Tradition in the Late Middle Ages*

Isaac Ibn al-adib (or al-Adab) first appears in the literature, when he was in

Castile in the 1370s, as a student of Judah ben Asher ii (then resident in Burgos),

the great-grandson of Asher ben Yeiel of Cologne (d. c. 1328) who became

chief rabbi of Toledo in 1305. Judah ben Asher ii (d. 1391) composed a set of

astronomical tables that are poorly preserved in a unique copy;1 he was killed

during the anti-Jewish riots that took place all over Spain beginning in 1391 and,

as a result of these riots, many Jews left Spain around that time. Ibn al-adib

was a member of a prominent Jewish family in Castile and arrived in Sicily no

later than 13962 when the island was ruled by Joan i (d. 1396), King of Aragon

and eldest son of Pere iii of the house of Barcelona. Both Pere and Joan were

keen on astronomy and had Jewish scholars at the royal court.3 Ibn al-adibs

main astronomical work was a set of tables in Hebrew for conjunctions and

oppositions of the Sun and the Moon, called Ora selulah (the paved way:

cf. Prov. 15:19). This text is mentioned by several later astronomers, notably

by Abraham Zacut in chapter 5 of his Great composition (ha-ibbur ha-gadol),

composed in 1478.4 Ibn al-adib died in Sicily around 1426.

Flavius Mithridates was a name assumed by William Raymond of Moncada,

a shadowy figure who was active in Italy in the late 15th century.5 Mithridates

1 B.R. Goldstein, Abraham Zacut and the Medieval Hebrew Astronomical Tradition, Journal

for the History of Astronomy, xxix (1998), 177186, p. 179.

2 M. Steinschneider, Mathematik bei den Juden, 2nd edn (Hildesheim, 1964), 168; B.R. Goldstein,

Descriptions of Astronomical Instruments in Hebrew, in Essays in Honor of E.S. Kennedy, ed.

by D.A. King and G. Saliba, Annals of the New York Academy of Science, d (1987), 105141, p. 128.

3 A. Rubi i Lluch, Documents per lhistria de la cultura mig-eval (Barcelona, 19081921); J.M.

Mills, Las Tablas Astronmicas del Rey Don Pedro el Ceremonioso (MadridBarcelona, 1962);

J. Chabs, with the collaboration of A. Roca and X. Rodrguez, Lastronomia de Jacob ben David

Bonjorn (Barcelona, 1992).

4 See, e.g., Oxford, Bodleian Library, ms Opp. Add. 8 42, f. 17b. Cf. F. Cantera Burgos, El judo

salmantino Abraham Zacut, Revista de la Academia de Ciencias Exactas, Fsico-Qumicas y

Naturales de Madrid, xxvii (1931), 63398, espec. pp. 113, 171.

5 See K. Lippincott and D. Pingree, Ibn al-tim on the Talismans of the Lunar Mansions,

isaac ibn al-adib and flavius mithridates 339

reports that he was the son of a Sicilian rabbi, Nissim Ab l-Faraj (Rabi nissim

abu ilfaragh: Vatican, ms Urb. lat. 1384, f. 3a).6 It has been established that he

converted to Christianity in the 1460s and that before his conversion his name

was Samuel ben Nissim.7 His new name was presumably chosen in honour

of his first patron, William Raymond of Moncada, Count of Adern (that is,

Adrano, in Sicily). A document dated 1473 indicates that Guillelmo Raymundo

de Moncata was a student at the University of Naples, and some evidence shows

that by 1477 Mithridates had reached Rome where he attracted the attention of

Federico of Montefeltro, Duke of Urbino (14221482), among others.8 Federico

played a major role in political and military affairs in Italy and he was also a

leading patron of the arts during the Renaissance. Mithridates is best known as

an adviser to the great humanist, Giovanni Pico della Mirandola (d. 1494), for

whom he translated many kabbalistic texts from Hebrew into Latin around the

year 1486.9 We will focus on Mithridatess interest in astronomy, rather than

kabbalah, and on his astronomical tables, uniquely preserved in Vatican, ms

Urb. lat. 1384. Samuels father was a student of Isaac Ibn al-adib10 and, as we

shall see, there is a strong connection between the astronomical tables of Ibn

al-adib and those of Mithridates.

The main purpose of Ibn al-adibs tables is to compute the time and posi-

tion of true syzygy, and the circumstances for solar and lunar eclipses.11 Some of

Journal of the Warburg and Courtauld Institutes, l (1987), 5781, pp. 5859. Although Wil-

liam Raymond of Moncada did not take Flavius Mithridates as his nom de plume until the

1480s, we generally refer to him by the name by which he is best known.

6 Cf. R. Starrabba, Ricerche storiche su Guglielmo Raimondo Moncada, ebreo convertito

siciliano del xv, Archivio Storico Siciliano, iii (1878), 1591, p. 87.

7 A. Scandaliato, Le radici familiari culturali di Guglielmo Raimondo Moncada, ebreo

convertito del rinascimento, nella Sicilia del. sec. xv, in Una Manna Buona per Mantova.

Man Tov le-Man Tovah. Studi in onore di Vittore Colorni, ed. by M. Perani (Florence, 2004),

203240; E. Engel, A Palaeographical Analysis of Mithridates Hebrew Autographs, in

Guglielmo Raimondo Moncada alias Flavio Mitradate: Un ebreo converso siciliano. Atti del

Convegno Internazionale, Caltabellotta (Agrigento), 2324 ottobre 2004, ed. by M. Perani

(Palermo, 2008), 201223.

8 Starrabba Guglielmo Raimondo Moncada (ref. 6), 39, 41, and 4748; M. Steinschneider,

Die hebraeischen bersetzungen des Mittelalters (Berlin, 1893), 986987.

9 G. Busi, et al., The Great Parchment: Flavius Mithridates Latin Translation, the Hebrew

Text, and an English Version (Turin, 2004) pp. 1617; C. Wirszubski, Pico della Mirandolas

Encounter with Jewish Mysticism (Cambridge (Mass.) and London, 1989).

10 Munich, Staatsbibliothek, ms Heb. 246, f. 83a; see M. Steinschneider, Die hebraeischen

Handschriften der K. Hof- und Staatsbibliothek in Muenchen (Munich, 1895), 120.

11 For other approaches to this problem, see J. Chabs and B.R. Goldstein, Computational

340 chapter 10

compiled in the Iberian peninsula and used there in the fourteenth century.

Indeed, there was a rich astronomical tradition in Hebrew in the fourteenth

and fifteenth centuries both in the Iberian peninsula and in southern France,

and a great many sets of tables in Hebrew were composed at that time, each

with its own special characteristics, by Levi ben Gerson, Jacob ben David Bon-

jorn of Perpignan (known as ha-Poel), Immanuel Bonfils of Tarascon, Abraham

Zacut, and Judah ben Verga, among others.12 It is quite unusual to find astro-

nomical tables in Latin compiled in the fifteenth century that do not depend

on either the Toledan Tables or the Parisian Alfonsine Tables. In fact, most

sets of tables in Latin that are independent of these two families derive from

the astronomical tradition in Hebrew; this applies to Mithridatess set, for it

depends primarily on the zij of Ibn al-adib. To be sure, the various tables com-

posed in this period all depend on the Ptolemaic tradition as it was elaborated

in al-Andalus (i.e., Muslim Spain) and then diffused to the Jewish and Christian

communities in the Iberian peninsula, later spreading to other parts of Europe.

Ibn al-adibs tables are preserved in about 20 Hebrew manuscripts of which

we have inspected the following: Vatican, ms Heb. 379; Paris, Bibliothque

nationale de France, ms Heb. 1086; Munich, Staatsbibliothek, ms Heb. 343. We

know of no Latin version of these tables, but they were preserved in Greek:

Venice, ms Marc. gr. 326 (ff. 135139); and Mount Athos, Vatopedi, ms 188

(ff. 113116v). According to Tihon and Mercier,13 Matthew Camariotes (d. 1490/1)

Astronomy: Five centuries of Finding True Syzygy, Journal for the History of Astronomy,

xxviii (1997), 93105. On Ibn al-Kammd (early 12th century), see J. Chabs and B.R. Gold-

stein, Andalusian Astronomy: al-Zj al-Muqtabis of Ibn al-Kammd, Archive for History of

Exact Sciences, xxxxviii (1994), 141; and on Ibn al-Raqqm (d. 1315), see, E.S. Kennedy, The

Astronomical Tables of Ibn al-Raqqm a Scientist of Granada, Zeitschrift fr Geschichte der

Arabisch-Islamischen Wissenschaften, xi (1997), 3572.

12 See, e.g., Immanuel ben Jacob Bonfils, Sefer shesh kenafayim [Hebrew edition of The Six

Wings, bound with Isaac ben Solomon, Sefer or ha-levanah] (Zhitomir, 1872); Mills, Las

Tablas Astronmicas (ref. 3); J. Chabs, The Astronomical Tables of Jacob ben David

Bonjorn, Archive for History of Exact Sciences, xlii (1991), 279314; Chabs, Lastronomia

(ref. 3); B.R. Goldstein, Levi ben Gersons Astronomical Tables (New Haven, 1974); idem,

Astronomy in the Medieval Spanish Jewish Community, in Between Demonstration and

Imagination: Essays in the History of Science and Philosophy Presented to John D. North,

ed. by L. Nauta and A. Vanderjagt (Leiden, 1999), 225241; J. Chabs and B.R. Goldstein,

Astronomy in the Iberian Peninsula: Abraham Zacut and the Transition from Manuscript to

Print (Philadelphia, 2000); B.R. Goldstein, The Astronomical Tables of Judah ben Verga,

Suhayl, iv (2001), 227289.

13 A. Tihon and R. Mercier, Georges Gmiste Plthon: Manuel dastronomie (Louvain-la-

Neuve, 1998), 12.

isaac ibn al-adib and flavius mithridates 341

does not include the tables.

Vatican, Biblioteca Apostolica, ms Urb. lat. 1384, is a manuscript of 88 folios,

probably copied in 14801481, containing three items: (1) De imaginibus coe-

lestibus, a translation into Latin of a work in Arabic attributed to Ibn al-tim;

(2) canons and tables on eclipses; and (3) a translation from Arabic into Latin

of the Quran.14 In item 2, the canons appear on ff. 30r43v, and its title includes

a short dedication to Federico of Montefeltro, Duke of Urbino, together with

the name of the author, given here as Guillelmus Raymundus de Moncata,

i.e., William Raymond of Moncada. There are similar dedications for the two

other texts that begin on ff. 1r and 65r. On f. 31r the epoch of the tables is given

explicitly as Sunday, January 8, 1475. We do not think this date has any spe-

cial significance other than the fact that it is the date of a mean conjunction

of the Sun and the Moon, and that the civil day January 7 (having 12 hours

in common with the astronomical day January 8 that begins at noon on Jan-

uary 7 in the civil calendar) is Saint Raymonds day in the Christian calen-

dar.

Several Greek and Muslim authorities are mentioned in the canons: Aristo-

tle, Ptolemy, and Ibn al-tim on f. 39r; Ibn al-tim on f. 42v; and Ibn Sina

and Ptolemy on f. 43v. On f. 42v we are told that Ibn al-tim observed a solar

eclipse in al-Andalus on July 19, 939.15 Of particular interest is that three other

astronomers are cited on f. 32v: al-Battn, Ibn al-Raqqm, and Ibn al-Kammd:

() apud Il Bactani et Ibn il raccam nec non et Ibn il chimadi. These are the same

three astronomers mentioned by Ibn al-adib in the introduction to his Ora

selulah.16 We are not aware of any other Latin text of the fourteenth or fifteenth

centuries in which Ibn al-Raqqms name appears.

Throughout the canons there are a few references to tables, and there is

even a complete table for the possibility of an eclipse (f. 35v), where the limits

for the distances from a lunar node are 15 for lunar eclipses and 10;16 for

solar eclipses (cf. Vatican, ms Heb. 379, f. 4a). According to Ptolemy the limit for

the possibility of a lunar eclipse corresponds to distances from a lunar node of

12;12; whereas for solar eclipses, the limits are 17;41 to the north and 8;22

14 For a detailed table of contents and discussion of items 1 and 3, see Lippincott and Pingree,

Ibn al-tim (ref. 5); K. Lippincott, More on Ibn al-tim, Journal of the Warburg and

Courtauld Institutes, li (1988), 188190.

15 Cf. Lippincott and Pingree, Ibn al-tim (ref. 5), 58.

16 Vatican, ms Heb. 379, f. 2a.

342 chapter 10

syzygy the true Sun is 3 from the place of the mean Sun at mean syzygy.18 This

means that a lunar eclipse may be possible when the mean Sun at syzygy is

about 15 from a lunar node. In the tables on ff. 47v and 52v, below, the lunar

eclipse limits are 12 from the lunar node. On f. 31v, lines 69, the text refers

to the movement on the 8th sphere, which is called motus proprie et per se,

whereas the movement on the 9th sphere is called motus naturalis. This is

the same terminology as that used in previous astronomical texts in Castilian

and in Arabic.19 On f. 35r, the difference between the 8th and 9th spheres, that

is the difference between sidereal longitudes and tropical longitudes, is said to

be 12;30 without an explanation of the way this value was determined (see also

Tabula gradus solaris, f. 56v, below).

A commentary on Ibn al-adibs tables by Abraham Gascon (Cairo, mid-

sixteenth century) contains a worked example for the solar eclipse of August

11, 1542 (New York, Jewish Theological Seminary of America [jtsa], ms 2571,

ff. 1a8b), based on Ibn al-adibs zij. In this manuscript the authors name

is consistently given as Ibn al-adib and not Ibn al-Adab as in some other

manuscripts.20 In Gascons commentary (f. 3a) the difference between the 8th

and 9th spheres, which he calls the motion in access (tenucat ha-haqbalah), is

taken to be 12, without any explanation of the way this value was determined.

Ibn al-adib does not address this issue in the canons to his tables except to

say that his tables are arranged for 8th sphere, i.e., his coordinates are sidereal

(Vatican, ms Heb. 379, f. 2b). However, in a commentary on these tables by Ibn

al-adibs son, Jacob, the motion in access is said to be 12 (New York, jtsa,

2571, f. 16a), and this may be the source for Gascons remark.21

The tables of Mithridates are on ff. 44r61v, following the canons. We offer a

brief description of them and their relationship to those of Ibn al-adib, with

special attention to the parameters embedded in them, in order to identify lines

of transmission of the astronomical material. The first four tables, together with

two other tables on f. 61rv, list the solar and lunar equations at syzygy, and it is

worth noting that these tables come before the tables for mean motions (see

17 Almagest vi.5; G.J. Toomer, Ptolemys Almagest (New York, 1984), 286287.

18 O. Pedersen, A Survey of the Almagest (Odense, 1974), 229230.

19 See J. Chabs and B.R. Goldstein, The Alfonsine Tables of Toledo (Dordrecht and Boston,

2003), 217218.

20 On Gascon, see B.R. Goldstein, The Hebrew Astronomical Tradition: New Sources, Isis,

lxxii (1981), 237251.

21 Another copy of Jacobs commentary is extant in London, British, Library, Or. 2806, ff. 20b

39b.

isaac ibn al-adib and flavius mithridates 343

ff. 54r55r and 56v57r), whereas the standard order in medieval sets of tables

is the opposite.

This table displays the correction for the solar position (in hours, minutes, and

seconds) at syzygy as a function of the solar anomaly given for each degree. The

maximum is 3;54,22h at anomaly 3s 1 (see Table a). The entries are the same

as those in the column for the correction for the time in Ibn al-adibs tables

in Vatican, ms Heb. 379, ff. 7a8a (see Table aa), where the same maximum

value is found. The letter s at the top left-hand margin of f. 44r means that the

entries are to be subtracted between arguments 0s 1 and 5s 30, and the letter

a at the bottom left-hand margin indicates that the entries are to be added

between 6s 0 and 11s 29. It is worth noting that the maximum occurs at 91; we

are not aware of any other table for the solar equation with a maximum value

at that argument. After inspection of the surrounding entries, it seems that

this particular entry is an isolated error and that the intended maximum was

3;54,20h at an argument of 3s 2. To determine the underlying maximum solar

equation, we have compared Table a with various tables for the solar equation:

Ibn al-Kammds table where the maximum is 1;52,44;22 al-Battns where the

maximum is 1;59,10;23 and the Parisian Alfonsine Tables where the maximum

is 2;10,0.24 Dividing an entry in each of these tables by the corresponding entry

in Table a leads to the result that the best fit is with the entries in al-Battns

table.

2 0; 7,56 2; 0,16 3;23,15 3;54,20 3;21,55 1;53,13

3 0;11,52 2; 4, 0 3;25,17 3;54,18 3;20, 8 1;49,25

4 0;15,48 2; 7,22 3;27,15 3;54, 8 3;18, 1 1;45,35

23 C.A. Nallino, Al-Battn sive Albatenii Opus Astronomicum (2 vols, Milan, 19031907), ii,

78 ff.

24 Tabule astronomice illustrissimi Alfontij regis castelle ed. by E. Ratdolt (Venice, 1483), e2v

e4r.

344 chapter 10

table aa Table for the solar anomaly (Ibn al-adib, Vatican, ms Heb.

379, f. 7a), the first of 6 similar sub-tables

0[s]

Subtract

Anomaly * the Lunar Anomaly the Time the Position

1 0; 2,47 0; 4, 1h 0; 2,12

2 0; 5,32 0; 7,56 0; 4,20

3 0; 8,11 0;11,52 0; 6,29

4 0;11,12 0;15,48 0; 8,40

5 0;13, 3 0;19,41 0;10,49

11[s]

Add

isaac ibn al-adib and flavius mithridates 345

manuscript reads Lunar Anomaly, but in Paris, ms Heb.

1086, 7a8a, the heading is consistently Solar Anomaly.

This table displays the correction for the lunar position (in hours, minutes, and

seconds) at syzygy as a function of the lunar anomaly given for each degree.

The letters a and s in the margins have the same meaning as in the previous

table, but their placement is interchanged. The maximum is 9;42,6h at anomaly

3s 6 (see Table b); comparison with the surrounding entries indicates that

the entries around the maximum value are mistaken. The entries in this table

are the same as those in the column for the correction for the time in the

corresponding table in Vatican, ms Heb. 379, ff. 8b9b (see Table bb), where the

same maximum value is found. Again, when we divide an entry in each of the

tables of other authors for the lunar equation by the corresponding entry in

Table b we find that the best fit is with the entries in al-Battns table, where

the maximum lunar equation is 5;1,0. But an underlying value of 4;56 cannot

be definitely excluded.

2 0;18,40 4;46,52 8;14,25 9;41,35 8;36,15 4;55,47

3 0;28, 2 4;55, 0 8;18,36 9;41,54 8;31,18 4;46,20

4 0;37,20 5; 3, 6 8;22,46 9;42, 4 8;26, 9 4;36,37

5 0;46,40 5;11,12 8;27,50 9;42, 5 8;20,49 4;26,53

6 0;56, 0 5;19,19 8;32, 2 9;42, 6 8;15,14 4;17, 3

346 chapter 10

3s 1. Probably a mistake for 9;41,0h; the same mistake is found in Vatican, ms Heb. 379.

In Hebrew, using alphabetical notation for numbers, 41 is ma (with no space between

the letters) and 40,1 is m a (with a space between the letters).

table bb Table for the lunar anomaly (Ibn al-adib, Vatican, ms Heb.

379, f. 8b), the first of 6 similar sub-tables

0[s]

Add

Anomaly the Lunar Anomaly the Time the Position

1 0; 5, 5 0; 9,20h 0; 0,23

2 0;10,10 0;18,40 0; 0,46

3 0;15,15 0;28, 2 0; 1, 9

4 0;20,20 0;37,20 0; 1,32

5 0;25,25 0;46,40 0; 1,55

11[s]

Subtract

Thus, the two preceding tables (a and b) were intended for computing a first

approximation of the time between mean and true syzygy and, presumably,

were based on tables for the solar and lunar equations with maxima of 1;59,10

and 5;1,0, respectively, both of which are well represented in the astronomical

literature in the Iberian peninsula. On the other hand, the next two tables (c

and d) are intended for computing a first approximation of the difference in

longitude between mean and true syzygy. These tables treat the effect of each

luminary separately and differ substantially from those compiled for the same

purpose at about the same time.25

isaac ibn al-adib and flavius mithridates 347

This table displays the correction for the solar position (in degrees, minutes,

and seconds) as a function of the solar anomaly given for each degree. The

letters s and a in the margins have the same meaning as in the previous

tables. The maximum is 2;8,51 at anomaly 3s 1 (see Table c). The entries are

the same as those in the column for the correction for the position in the

corresponding table in Vatican, ms Heb. 379, ff. 7a8a (see Table aa), where the

same maximum value is found.

2 0; 4,20 1; 6, 5 1;51,23 2; 8,47 1;49,58 1; 2,11

3 0; 6,29 1; 8, 9 1;52,49 2; 8,45 1;49,39 1; 0, 7

4 0; 8,40 1;10, 0 1;53,53 2; 8,40 1;48,49 0;58, 9

5 0;10,49 1;11,50 1;54,56 2; 8,39 1;47,38 0;55,55

The entries in this table, here given in degrees, are equivalent to those in Table

a, given in hours. If ai and ci are the corresponding entries in Tables a and c,

then the expression

ci = 0;32,56 ai,

where 0;32,56/h, the mean lunar velocity, yields fairly good results. This

seems to mean that an entry in Table c is the distance the Moon travels at its

mean velocity corresponding to the time computed in Table a. We note that

0;32,56/h is an approximation of the lunar mean motion in longitude,

13;10,35/d.

348 chapter 10

signis

This table displays the correction for the lunar position (in degrees, minutes,

and seconds) as a function of the lunar anomaly given for each degree. There

is no indication in the table or the margins concerning the addition or sub-

traction of the entries; we have introduced a and s in accordance with the

presentation in Ibn al-adibs corresponding table. The maximum is 0;23,55

at anomaly 3s 4 (Table d). The entries are the same as those in the col-

umn for the correction for the position in the corresponding table in Vati-

can, ms Heb. 379, ff. 8b9b (see Table bb), where the same maximum value is

found.

2 0; 0,46 0;11,46 0;20,21 0;23,53 0;21,13 0;12, 9

3 0; 1, 9 0;12, 7 0;20,30 0;23,54 0;21, 0 0;11,44

4 0; 1,32 0;12,26 0;20,40 0;23,55 0;20,50 0;11,21

5 0; 1,55 0;12,46 0;20,52 0;23,54 0;20,35 0;10,58

The entries in this table, here given in degrees, are equivalent to those in Table

b, given in hours. If bi and di are the corresponding entries in Tables b and d,

then the expression

di = 0;2,28 bi,

where 0;2,28/h, the mean solar velocity, yields good results. This seems to

mean that an entry in Table d is the distance the Sun travels at its mean velocity

isaac ibn al-adib and flavius mithridates 349

approximation of the mean motion of solar longitude, 0;59,8/d.

Two tables on f. 61rv, with the same format as Tables a, b, c, and d, are described

below.

In this multiplication table we find the numbers 1 to 10 across the top and the

numbers from 1 to 10 and then multiples of 5 from 15 to 60 at the beginning of

each row. However, Vatican, ms Heb. 379, f. 15r, displays a multiplication table

with the numbers 1 to 10 across the top and the numbers 1 to 30 (at intervals of

1) at the beginning of each row. On f. 15v the table continues with 1 to 10 across

the top and 31 to 60 (at intervals of 1) at the beginning of each row. Mithridates

kept the heading for the columns, but reduced the number of rows.

latitudinibus

The entries in this table correspond to the time of half-daylight (in hours and

minutes) for various latitudes (0;0, 8;28, 16;50, 24;0, 30;20, 36;0, 40;30,

45;0, 48;30, 51;30, 59;45, 63;0, 64;45, and 66;25) as a function of the solar

longitude. The argument is given at 10-intervals, and the columns for the last 4

latitudes are not completely filled in. This table is similar to that of Ibn al-adib

(Vatican, ms Heb. 379, f. 10a), but the argument there is given at 5-intervals and

for 9 latitudes that are slightly different from those in Mithridatess table.

The table for lunar eclipses on f. 47v and the two tables for lunar eclipses on f. 48r

that follow agree with the corresponding tables in Wing 4 of the Six Wings

by Immanuel Bonfils of Tarascon (Provence, Southern France), composed in

Hebrew c. 1365.

This table displays the digits of a lunar eclipse (in digits and minutes of a

digit), where 1 digit is a twelfth of the lunar diameter, as a function of the

argument of latitude (from 0 to 12, for each integer degree) and the lunar

350 chapter 10

anomaly (from 0s to 11s, for each integer sign). The maximum value is 21;36

digits, corresponding to an argument of latitude of 0s 0. The entries are the

same as those in the columns for the digits of the diameter in Ibn al-adibs

tables for lunar eclipses: see Vatican, ms Heb. 379, f. 11a. The maximum in this

table agrees with the maximum in Wing 4 in the tables of Immanuel Bonfils.26

This maximum is also found in al-Battns tables (Nallino 19031907, 2:90). In

the Almagest the maximum is called entire following an entry of 21 digits;

however, Neugebauer shows that the parameters in the Almagest lead to a

maximum of 21;36 digits.27

The heading for this table refers to the colour of eclipses. Only the frame and the

arguments are given; no other entries are displayed. The argument runs from

1 to 12 digits, for each integer digit. A table for colours of solar eclipses with

these arguments appears in Vatican, ms Heb. 379, f. 13b.28 The text preceding

the tables mentions a table for the colours of eclipses (f. 38v, line 2).

The heading for this table also refers to the colour of eclipses. As in the previ-

ous table, the frame and the arguments are given without any other entries.

The argument runs from 1 to 24 digits, for each integer digit, and there are

two successive numbers in each of the 12 rows. The heading mentions Ibn

al-Kammd.29 There are two tables for the same purpose on f. 53v. These argu-

ments probably refer to digits of a lunar eclipse.

26 See Bonfils 1872, pp. 3840; cf. P. Solon, The Hexapterygon of Michael Chrysococces. (Ph.D.

dissertation, Brown University, 1968); and idem, The Six Wings of Immanuel Bonfils and

Michael Chrysokokkes, Centaurus, xv (1970), 120, p. 7.

27 Toomer, Almagest (ref. 17), 307; O. Neugebauer, A History of Ancient Mathematical Astron-

omy (Berlin and New York, 1975), 136.

28 B.R. Goldstein, Colors of Eclipses in Medieval Hebrew Astronomical Tables, Aleph, v

(2005), 1134.

29 For colours of eclipses according to Ibn al-Kammd, see Chabs and Goldstein, Ibn

al-Kammd (ref. 11), 1819.

isaac ibn al-adib and flavius mithridates 351

This table displays the time of the half-duration of lunar eclipses (in hours and

minutes) as a function of the argument of latitude (from 0 to 12, for each

integer degree) and the lunar anomaly (from 0s to 11s, for each integer sign).

The maximum value is 2;2h, corresponding to an argument of latitude of 0s 0

and a lunar anomaly of 0s. The entries are the same as those in the columns

for the half-duration in Ibn al-adibs tables in Vatican, ms Heb. 379, f. 11a.

We note that the term defectus is used for eclipse; on f. 42v the term ekleipsis

is written in Greek characters. This maximum for the half-duration of a lunar

eclipse is also found in the tables of Immanuel Bonfils where the time from

first contact to the beginning of totality is given as 1;8h and the half-duration of

totality is given as 0;54h, for 1;8h + 0;54h = 2;2h.30

This table displays the time of the half-duration of totality of lunar eclipses (in

hours and minutes) as a function of the argument of latitude (from 0 to 12,

for each integer degree) and the lunar anomaly (from 0s to 11s, for each integer

sign). The maximum value is 0;54h, corresponding to an argument of latitude

of 0s 0 and a lunar anomaly of 0s. The entries are the same as those in the

columns for the half-duration of totality in Ibn al-adibs tables: see Vatican,

ms Heb. 379, f. 11a. They also agree with the values in Wing 4 of the Six Wings by

Immanuel Bonfils.31

In this table only the frame is given with headings for the columns, but there

are no other entries. It has the same pattern as the previous tables for the digits

of the diameter, the half-duration, and the half-duration of totality.

30 Bonfils 1872, pp. 3840; cf. Solon, The Six Wings (ref. 26), 7.

31 Ibid.

352 chapter 10

The entries in this table for the longitudinal and latitudinal components of the

adjusted parallax are almost identical with those in Ptolemys Handy Tables

for latitude 16;27 and in al-Battns zij for latitude 16;32, according to the

heading.32

The entries in this table for the longitudinal and latitudinal components of the

adjusted parallax are almost identical with those in Ptolemys Handy Tables for

latitude 23;51 and in al-Battns zij for latitude 24;0, according to the heading.

The entries in this table for the longitudinal and latitudinal components of the

adjusted parallax are almost identical with those in Ptolemys Handy Tables

for latitude 30;22 and in al-Battns zij for latitude 30;40, according to the

heading.33 This table appears in Vatican, Heb. 379, f. 12a, with the heading: Table

for parallax in longitude and latitude for [geographical] latitude 30.

The entries in this table for the longitudinal and latitudinal components of the

adjusted parallax are almost identical with those in Ptolemys Handy Tables

for latitude 36;0 and in al-Battns zij for latitude 36;22, according to the

heading.34

32 See W.D. Stahlman, The Astronomical Tables of Codex Vaticanus Graecus 1291 (Ph.D. disser-

tation, Brown University, 1959; University Microfilms, No. 625761), 268269; and Nallino,

Al-Battn (ref. 23), ii, 95.

33 See Stahlman, Astronomical Tables (ref. 32), 272273; and Nallino, Al-Battn (ref. 23), ii,

97.

34 See Stahlman, Astronomical Tables (ref. 32), 274275; and Nallino, Al-Battn (ref. 23), ii,

98.

isaac ibn al-adib and flavius mithridates 353

The entries in this table for the longitudinal and latitudinal components of

the adjusted parallax are almost identical with those in Ptolemys Handy Tables

for latitude 40;56 and in al-Battns zij for latitude 41;15, according to the

heading.35 This table appears in Vatican, Heb. 379, f. 12b, with the head-

ing: Table for parallax in longitude and latitude for [geographical] latitude

40.

The entries in this table for the longitudinal and latitudinal components of the

adjusted parallax are almost identical with those in Ptolemys Handy Tables

for latitude 45;1 and in al-Battns zij for latitude 45;22, according to the

heading.36

f. 51v

This table has no title and very few entries, but it was certainly intended for

parallax for a latitude where the longest daylight (from sunrise to sunset) is

16;30h.

f. 52r

This table has no title and the entries are given in minutes (see Table e). The

argument reads gradus motus differenctia, here meaning the lunar anomaly,

and the entries represent minutes of proportion. Its purpose is to correct

the parallax when the Moon is not at its mean distance at syzygy. Parallax

varies inversely with the lunar distance from the Earth (and lunar distance is a

function of lunar anomaly) but the parallax tables assume that the Moon is at

its mean distance at syzygy. In Ptolemys lunar model for syzygy, the radius of

35 See Stahlman, Astronomical Tables (ref. 32), 276277; and Nallino, Al-Battn (ref. 23), ii,

99.

36 See Stahlman, Astronomical Tables (ref. 32), 278279; and Nallino, Al-Battn (ref. 23), ii,

100.

354 chapter 10

2 5 4 3 0 3 5 28

4 5 4 3 0 3 5 26

6 5 4 3 0 3 5 24

8 5 4 2 1 3 5 22

10 5 4 2 1 3 5 20

12 5 4 2 1 3 5 18

14 5 4 2 1 4 5 16

16 5 4 2 1 4 5 14

18 5 4 1 1 4 5 12

20 5 4 1 2 4 5 10

22 5 4 1 2 4 5 8

24 5 4 1 2 4 5 6

26 5 3 1 2 4 5 4

28 5 3 1 2 4 5 2

30 5 3 1 2 4 5 0

the epicycle is 5;15.37 Hence, the maximum lunar distance is 65;15 at apogee (0

anomaly), the minimum distance is 54;45 at perigee (180 anomaly), where the

mean distance is 60 (near 90 anomaly). Therefore, the correction for parallax

is negative near apogee and positive near perigee and, in general,

p() = p + e() p,

where p is the value in one of the parallax tables, e() is the entry in Table e for

an anomaly , and p() is the corrected value for parallax as a function of .

This table also appears in Vatican, ms Heb. 379, f. 13b, where the heading is:

Table for correcting parallax. Above the columns for 0s, 1s, and 2s, is the word

for subtract, and above the columns for 3s, 4s, and 5s, is the word for add. The

argument is labelled: degrees of [lunar] anomaly [oq]. Similar tables are found

37 In addition to the value 5;15 for this parameter, there are variants 5;13 and 5;14: see Almagest

iv.6; trans. Toomer, Almagest (ref. 17), 197, 202, and 209.

isaac ibn al-adib and flavius mithridates 355

(to two sexagesimal places) in the Tabule Verificate for Salamanca (c. 1461), and

in shorter versions in Abraham Zacuts ha-ibbur ha-gadol as well as in its Latin

version, the Almanach Perpetuum.38

f. 52rv

This table has no title, and its entries represent the time between noon and

nonagesimal, here called medium celum (midheaven), given in hours and min-

utes, for each zodiacal sign. The argument is the geographical latitude (from

30 to 51, for each integer degree). The entries are the same as those in the

corresponding table in Vatican, ms Heb. 379, f. 11b, although the horizontal and

vertical arguments are interchanged. In Ibn al-adibs table the geographical

latitude varies from 30 to 45, at intervals of one degree.

f. 52v

This table has no title, and the entries (in degrees, minutes, and seconds)

represent the lunar latitude as a function of the argument of lunar latitude from

0 to 12 in steps of 1. The maximum value, 1;2,16, is reached at argument 12.

The entries are the same as those in the first two columns of a table in Vatican,

ms Heb. 379, f. 11a which has the heading: Table for eclipses of the Moon and

its latitude at the time of an eclipse. The argument of 12 represents the limit

for the Moons distance from the lunar node at the time of a lunar eclipse and,

with a maximum lunar latitude of 5, the value for 12 of argument would be

1;2,22, which is close to the value here (cf. the table on f. 47v).

f. 53r

Again, this table has no title. The entries represent the velocity of the Moon rel-

ative to that of the Sun, in minutes and seconds of arc per hour, from 0;27,30/h

to 0;32,30/h, as a function of lunar anomaly, given in degrees for every other

integer degree. The relative velocity increases monotonically by 0;0,4/h for

arguments between 0s 0 and 6s 0, and then decreases monotonically by

356 chapter 10

0;0,4/h for arguments between 6s 0 and 12s 0. The entries are the same as

those in the corresponding table in Vatican, ms Heb. 379, f. 11b.

f. 53r

As was the case in several of the previous tables, this one is untitled. The entries

give the fraction of the solar diameter that is darkened in a solar eclipse (in

digits and minutes), where 1 digit is a twelfth of the solar diameter, as a function

of the true lunar latitude (in minutes and seconds of arc): see Table f. It is

noteworthy that the entries for the latitude are not rounded whereas the digits

of the diameter are such that the successive differences between them are 0;20

digits or 0;15 digits. The entries are the same as those in the column for digits

of the diameter in the corresponding table in Vatican, ms Heb. 379, f. 13a. We

note, however, that in Mithridatess table the argument (true latitude) increases

downward, but in Ibn al-adibs table it decreases.

0 0 11 40 16 19 6 0

1 1 11 20 17 15 5 40

1 12 11 0 18 12 5 20

2 44 10 40 19 8 5 0

3 31 10 20 20 4 4 40

4 48 10 0 21 1 4 20

5 8 9 40 22 57 4 0

6 4 9 20 23 53 3 40

7 56 9 0 24 50 3 20

8 54 8 40 25 47 3 0

9 51 8 20 26 44 2 40

10 46 8 0 27 39 2 20

11 45 7 40 28 35 2 0

12 35 7 20 29 31 1 40

13 30 7 0 30 28 1 20

14 21 6 40 31 23 1 0

15 20 6 20 32 20 0 45

33 17 0 30

34 13 0 15

35 0 0 0

isaac ibn al-adib and flavius mithridates 357

f. 53v

This is a double argument table for the half-duration of a solar eclipse and it

is mentioned in the text (f. 42r). The vertical argument is the corrected lunar

latitude, in degrees, at 2-intervals, and the horizontal argument is the velocity

of the Moon relative to that of the Sun, in minutes and seconds of arc per

hour, from 0;27,30/h to 0;32,30/h. The entries give the half-duration of a solar

eclipse, and they are the same as those in the corresponding table in Vatican,

ms Heb. 379, f. 13a. As was the case with the table on f. 53r, the vertical argument

(corrected lunar latitude) increases downward, whereas in Ibn al-adibs table

it decreases. This table is identical (but for minor variants) to those in Ibn

al-Kammds zij (Madrid, Biblioteca Nacional, ms 10023, f. 54r) and in the Tables

of Barcelona (Table 49).39

f. 53v

There are three frames for small tables, also untitled, but two of them have the

word Colores in the heading, as was the case with the two tables on f. 47v whose

entries are missing.

The following tables yield mean positions and times for conjunctions. In the

canons to Ibn al-adibs tables, chapter 2 (Vatican, ms Heb. 379, f. 2b), we are

told that these motions are computed for the western extremity (not, e.g., for

Toledo) and each day begins at noon of the day preceding it, e.g., Sunday begins

at noon on Saturday and ends at noon on Sunday. In other respects the tables

are arranged according to the Jewish calendar with its 19-year cycle, beginning

with Molad Tohu, the conjunction of Tishri, year 1 Anno mundi, i.e., the era of

creation. According to the Jewish tradition, the conjunction of creation took

place on Monday (day 2) at 5h 204 elaqim (where 1h = 1080 elaqim) counting

from sunset,40 at a location whose distance from the western extremity is

75;43,45 in the view of Ibn al-adib. The epoch of the Jewish calendar, Tishri

39 See Chabs and Goldstein, Ibn al-Kammd (ref. 11), 23; Mills, Las Tablas Astronmicas

(ref. 3), p. 237; and J. Chabs, Astronoma andalus en Catalua: Las Tablas de Barcelona,

in From Baghdad to Barcelona: Studies in the Islamic Exact Sciences in Honour of Prof. Juan

Vernet, edited by J. Casulleras and J. Sams (Barcelona, 1996), 477525, pp. 511512.

40 See, e.g., Maimonides: Sanctification of the New Moon, trans. by S. Gandz, introduction by

J. Obermann, astronomical commentary by O. Neugebauer (New Haven, 1956), 116.

358 chapter 10

1, 1 Anno mundi, is equivalent to Oct. 7, 3761bc (jdn 347998). The radices for

the conjunction of creation are given as the entries for the first month, Tishri.

As we will see, the parameters for the mean motion tables are the same for

Mithridates as they are for Ibn al-adib, but the presentation is different and

the initial values are different. In the case of Mithridates there are places where

radices were to be given but they are filled with names which do not correspond

to any numerical values. As we learn from the worked example in Gascons

commentary, Ibn al-adibs mean motion tables are set up such that the entries

in them refer to the current 19-year cycle, the current year within the cycle, and

the conjunction at the beginning of each month.

the solar anomaly (in signs, degrees, minutes, and seconds) for 13 consecutive

months, 19 consecutive years, and groups of 19-year cycles (for cycles 1, 2, 3, , 10,

20, 30, , 100, , 700), respectively. This table also displays a numerical value, 0s

14;33,10, together with the word FRAEDERICO, referring to Federico, Duke of

Urbino. The same value is found for the solar anomaly in Ibn al-adibs tables in

Vatican, ms Heb. 379, f. 6b, and it represents the motion of the solar anomaly in

half a mean synodic month, i.e., the time between successive syzygies. Indeed,

it is so labelled in Vat. 379, f. 6b, and is intended for making the tables usable

for oppositions.

The entries for two successive 19-year cycles show a constant difference of

359;44,50, and this is also the entry for 1 cycle. This line-by-line difference is the

same as in the column for the solar anomaly in Ibn al-adibs table in Vatican,

ms Heb. 379, f. 6a, but that table is presented differently: the information is only

given for cycles 273 to 280 (11s 20;54,8 to 11s 19;7,58), where cycle 273 refers

to the beginning of cycle 273, i.e., 272 cycles have been completed or 272 19

= 5168 years have passed since the creation according to the Jewish tradition.

For an explanation of the way the entries in this column were computed, see

the comments on Tabula gradus solaris (f. 56v), below. The constant difference

of 359;44,50 means that in one 19-year cycle the Sun progresses 6839;44,50.

Now, the first entry in the table for the time on f. 55r, below, associates 2d

16;33,3,20h with a 19-year cycle. (This is also identical to the value underlying the

column for time in Ibn al-adibs table in Vatican, ms Heb. 379, f. 6a.) This time

corresponds to the excess over an integer number of weeks of the time between

two consecutive 19-year cycles, which is taken here as 991 weeks + 2d 16;33,3,20h.

Thus, the time between two consecutive 19-year cycles is 6939;41,22,38,20d.

isaac ibn al-adib and flavius mithridates 359

Dividing the amount progressed by the Sun by this time we obtain a value for

the solar mean motion of 0;59,8,9,16/d and it is a sidereal value. The closest

value of which we are aware is that of Ibn al-Kammd: 0;59,8,9,21,15/d.

In the sub-table for single years the entry for year 1 is 11s 20;41,15, whereas in

the tables of Ibn al-adib the entry for year 1 is 0;0. In the sub-table for months

the entry for month 1 is 9s 7;14,49,42, whereas for Ibn al-adib it is 3s 15;20,10

which, according to Ibn al-adibs canons, is the radix for the conjunction of

creation. We have no explanation for these differences but is seems likely that

Mithridates has changed the epoch.

This table consists of 3 sub-tables giving the lunar anomaly (in signs, degrees,

minutes, and seconds) for 13 consecutive months, 19 consecutive years, and

groups of 19-year cycles (for cycles 1, 2, 3, , 10, 20, 30, , 100, , 700), respec-

tively, as in the previous table. This table also displays a numerical value, 6s

12;54,30, together with the words DUCI URBINI, referring to Federico, Duke of

Urbino. The same value for the lunar anomaly is found in Ibn al-adibs tables

in Vatican, ms Heb. 379, f. 6b, and it represents the lunar motion in anomaly in

half a mean synodic month.

The entries for two successive 19-year cycles show a constant difference of

306;54,52, as in the column for the lunar anomaly in Ibn al-adibs table in

Vatican, ms Heb. 379, f. 6a, where this information is presented differently. The

entry for cycle 1 is 10s 6;54,52. In Ibn al-adibs tables the entry for 273 refers

to the beginning of the cycle (as we learn from Gascons commentary). Since

the radices for this epoch are given by Ibn al-adib as entries for the month

Tishri, the table for cycles should not have radices embedded in them. Ibn

al-adibs entry for cycle 273 is 320;42,12, and it corresponds to the motion

in anomaly in 272 completed cycles of 19 years. To derive this value from the

constant difference of 306;54,52, we first compute:

Now let us take 83480;42,12 and divide it by 272; the result is 306;54,51,40,

which rounds to 306;54,52. This indicates that Ibn al-adibs table was con-

structed from a basic parameter of 306;54,51,40 rather than 306;54,52.

The constant difference of 306;54,52 means that in one cycle the Sun pro-

gresses (251 360) + 306;54,52 = 90666;54,52. Dividing the Moons progress by

the time between two consecutive 19-year cycles, 6939;41,22,38,20d, we obtain a

360 chapter 10

value for the mean motion in lunar anomaly of 13;3,53,55,56,18/d. This is a clas-

sical parameter, to be compared, for example, with those in Ptolemys Almagest

(13;3,53,56,17,52/d) and the Toledan Tables (13;3,53,56,17,57 /d).

In the sub-table for single years the entry for year 1 is 11s 15;38,0, whereas in

the tables of Ibn al-adib the entry for year 1 is 0;0. In the sub-table for months

the entry for month 1 is 2s 12;53,26,31, whereas for Ibn al-adib it is 8s 21;51,25

which, according to Ibn al-adibs canons, is the radix for the conjunction of

creation. We have no explanation for these differences but it seems likely that

Mithridates has changed the epoch.

This table consists of 3 sub-tables giving the excess over an integer number

of weeks (in days, hours, minutes, seconds, and thirds) for 13 consecutive

months, 19 consecutive years, and groups of 19-year cycles (for cycles 1, 2, 3,

, 10, 20, 30, , 100, , 700), respectively, as in the two previous tables. This

table also displays a numerical value, 7d 18;22,1,50h, together with the word

GUILLELMUS, referring to William Raymond of Moncada. Almost the same

value is found for the time in Ibn al-adibs tables in Vatican, ms Heb. 379, f. 6b,

where we have 0d 18;22,2h. It represents the duration of half a mean synodic

month in days (mod 7) and hours. We note that Mithridatess value has a higher

precision than that in Vatican, ms Heb. 379. We also note that 18;22,1,50h =

0;45,55,4,35d and twice this amount is 1;31,50,9,10d, i.e., the excess of a mean

synodic month over 28 days (= 4 weeks).

The entries for two successive 19-year cycles show a constant difference

of 2d 16;33,3,20h (= 2;41,22,38,20d), which is also the value for the first entry.

This is the same value that can be derived from the column for time in

Ibn al-adibs table in Vatican, ms Heb. 379, f. 6a, where this information

is presented differently and with a lower precision (the entries are rounded

to seconds of time). The entry for cycle 273 is 3d 13;51,7h (= 3;34,37,47,30d)

which implies that the entry for cycle 0 would be 4d 7;26,57h = 4;18,37,22,30d

(for 4;18,37,22,30d + 273 2;41,22,38,20d = 4;18,37,22,30d + 6;16,0,25d (mod 7) =

3;34,37,47,30d (mod 7) = 3d 13;51,7h). We have no explanation for Ibn al-adibs

initial value. The value for the constant different in a 19-year cycle is equivalent

to 2d 16h and 595 elaqim and this parameter is exactly that given, for example,

in Maimonidess Sanctification of the New Moon.41

isaac ibn al-adib and flavius mithridates 361

In the sub-table for single years the entry for year 1 is 1d 21;30,10,0h, whereas

in the tables of Ibn al-adib the entry for year 1 is 0d 0;0h. In the sub-table

for months the entry for month 1 is 2d 1;43,6,51h, whereas for Ibn al-adib it

is 2d 6;8,25h which, according to Ibn al-adibs canons, is the radix for the

conjunction of creation. We have no explanation for these differences but it

seems likely that Mithridates has changed the epoch.

This table displays the day of the (Christian) month in a 19-year cycle when

a conjunction takes place (entries range from 1 to 31). The letter p is written

opposite years 2, 5, 8, 10, 13, 16, and 18 in the 19-year cycle, indicating a leap

year, i.e., a year of 13 lunar months.

f. 56r

There are 4 small tables concerning weekday numbers. In one of them the

words Federico, Duke, Count, William, Raymond, Moncata, and Urbino are

assigned to the first day of the month, and in another, numbers between 1 to 7

are assigned to these 7 words. In the two other tables, a number between 0 and

6 is ascribed to each month in a year beginning in January, whether a common

year or a leap year.

This table consists of 3 sub-tables giving the mean longitude of the two lumi-

naries at mean conjunction (in signs, degrees, minutes, and seconds) for 13 con-

secutive months, 19 consecutive years, and groups of 19-year cycles (for cycles

1, 2, 3, , 10, 20, 30, , 100, , 700), respectively, as in previous tables. This table

also displays three numerical values, 6s 0;0,0, 0s 12;30,0, and 0s 14;33,9,59.

The first indicates the amount to the added to the lunar longitude at opposi-

tion; the second is a parameter already mentioned in the text (f. 35r, lines 710),

to be used to change from sidereal to tropical coordinates; and the third repre-

sents the motion of the Sun in longitude in half a mean synodic month. These

values are presented together with the word RAYMUNDUS, referring to William

Raymond of Moncada. Almost the same value for the solar motion for the posi-

tion occurs in Ibn al-adibs tables in Vatican, ms Heb. 379, f. 6b where, for the

362 chapter 10

third value, we find 0s 14;33,10. We note that, as was the case in the analogous

table on f. 55r, Mithridatess value has a higher precision than that in Vatican,

ms Heb. 379, although elsewhere Mithridates gives it as 0s 14;33,10 (f. 54r: see

above).

The entries for two successive 19-year cycles show a constant difference

of 359;48,50, as in the column for the position in Ibn al-adibs table in

Vatican, ms Heb. 379, f. 6a, where this information is presented differently.

This means that in one cycle the Moon progresses (253 360) + 359;48,50 =

91439;48,50. Dividing the Moons progress by the time between two consec-

utive 19-year cycles, 6939;41,22,38,20d, we obtain a value for the lunar mean

motion in longitude of 13;10,34,52,40,30/d. This parameter may be compared

with the value in Ptolemys Almagest (13;10,34,58,33,30,30/d) and those in

the zij of al-Khwrizm, the zij of Ibn al-Kammd, and the Toledan Tables

(13;10,34,52,46/d).

In the sub-table for single years the entry for year 1 is 11s 20;43,15, whereas in

the tables of Ibn al-adib the entry for year 1 is 0;0. In the sub-table for months

the entry for month 1 is 5s 7;3,48,40, whereas for Ibn al-adib it is 5s 16;41,31

which, according to Ibn al-adibs canons, is the radix for the conjunction of

creation. We have no explanation for these differences but it seems likely that

Mithridates has changed the epoch.

We can also deduce from the text the solar mean motion in longitude: (18

360 + 359;48,50)/6939;41,22,38,20d = 0;59,8,11,20/d. The difference between

the solar mean motion in longitude and the motion of the solar anomaly is the

proper motion of the solar apogee: 0;59,8,11,20/d 0;59,8,9,16/d = 0;0,0,2,4/d

(or 1 in about 286 Julian years). Standard values for the proper motion of

the solar apogee are close to this amount, e.g., for Ibn al-Kammd it is 1 in

about 290 Julian years (Chabs and Goldstein 1994, p. 28). If we subtract the

mean solar anomaly at epoch (the conjunction of creation) according to Ibn

al-adib from his value for the mean solar position at epoch, the result is the

solar apogee at epoch: 166;41,31 105;20,10 = 61;21,21.

To verify that a sidereal solar position derived from Ibn al-adibs tables

is approximately correct, we compare the value based on his tables for the

solar eclipse that took place on Aug. 11, 1542 with the sidereal solar position

according to the Toledan Tables: in both cases the result is about 135 (see

below). Hence, the sum of the entries for the solar position in Ibn al-adibs

tables is acceptable (by the standards of his day). But there is a problem:

we expect the radix to be shown in the table for months opposite Tishri,

as explained in Ibn al-adibs canons, corresponding to the conjunction of

creation. But there is another radix embedded in the table for the mean

motions. First, let us compute the mean solar position for Aug. 11, 1542, the mean

isaac ibn al-adib and flavius mithridates 363

conjunction of Elul [month 12], 5302am, which took place on Ab [month 11]

29, in year 1 of cycle 280 (note that the entry for year 1 in Ibn al-adibs table is

0).

Elul: 4s 6;51,11

sum: 4s 14;35, 9

The entry for Elul can be computed by adding 11 times the increase in longitude

in a mean synodic month (= 29;6,20: Vatican, ms 379, f. 6b) to the radix (the

entry for Tishri):

On analogy with the computation of the mean motion in lunar anomaly (see

Tabula motus diuersi, f. 54v), one expects the entry for the beginning of cycle

280 to be 279 times the line-by-line differences in the column for solar position

in the table for cycles, i.e.,

But the entry for cycle 280 is 7;43,58, i.e., it exceeds the expected value by

59;39,28. Now the sum we computed for Elul, cycle 280, can be considered to

have 4 components:

The components 308;4,30 and 320;9,40 are just the result of multiplying

the mean motion parameter, and so they are fixed. But it is unclear why Ibn

al-adib did not consider as radix 107;2,3 (= 166;41,31 59;39,28), combining

the two other components. The only reason that comes to mind is that the

creation, according to the Jewish tradition, took place at the beginning of the

month Tishri which is associated with the autumnal equinox. Moreover, the

purpose of the 19-year cycle is to keep each month in the same season. A radix

of 107;2,3 entered in the month Tishri would mean that the creation took place

closer to summer solstice than to the autumnal equinox.

The same procedure also works for the mean motion in solar anomaly, that

is, we find that the shift in the entries for cycles is the same amount, 59;39,28.

This is appropriate since the difference between the solar position and the solar

364 chapter 10

anomaly is the longitude of the solar apogee; hence, a shift in one requires a

shift in the other.

This table consists of 3 sub-tables giving the argument of latitude (in signs,

degrees, minutes, and seconds) for 13 consecutive months, 19 consecutive years,

and groups of 19-year cycles (for cycles 1, 2, 3, , 10, 20, 30, , 100, , 700), respec-

tively, as in previous tables. It also displays a numerical value, 0s 15;20,7,1,

together with the words DE MONCATHA, referring to William Raymond of

Moncada. Almost the same value for the argument of lunar latitude occurs in

Ibn al-adibs tables in Vatican, ms Heb. 379, f. 6b, where we find 0s 15;20,7. It

represents the motion of the argument of lunar latitude in half a mean synodic

month. As was the case in previous tables, Mithridatess table exhibits a higher

precision than Vatican, ms Heb. 379.

The entries for two successive 19-year cycles show a constant difference of

7;35,20, as in the column for the argument of latitude in Ibn al-adibs table

in Vatican, ms Heb. 379, f. 6a, where this information is presented differently.

This means that in one cycle the Moon progresses (255 360) + 7;35,20 =

91807;35,20. When we divide this amount by the time between two consec-

utive 19-year cycles, 6939;41,22,38,20d, we obtain a value for the mean motion

of the argument of latitude of 13;13,45,39,47,10/d Subtracting the mean motion

in argument of latitude from the lunar mean motion, we find 0;3,10,47,6,40/d.

Again, this value for the mean motion of the node is close to those of al-

Khwrizm and Ibn al-Kammd, whereas the value for the mean motion of the

argument of latitude is close to Ptolemys 13;13,45,39,48,56,37/d.

In the sub-table for single years the entry for year 1 is 5s 14;26,9, whereas in

the tables of Ibn al-adib the entry for year 1 is 0;0. In the sub-table for months

the entry for month 1 is 4s 29;29,28,46, whereas for Ibn al-adib it is 6s 3;40,27

which, according to Ibn al-adibs canons, is the radix for the conjunction of

creation. We have no explanation for these differences but it seems likely that

Mithridates has changed the epoch.

localities, of which many are in Sicily, ordered from west to east. The coordi-

nates are mostly given in degrees and minutes, e. g., for the city called Agrigen-

isaac ibn al-adib and flavius mithridates 365

tum we are given 38;30 (longitude) and 36;24 (latitude). However, those for

Toledo are 10 (longitude) and 42 (latitude). The prime meridian for geograph-

ical longitude is the one used by Ptolemy (and many other medieval tables such

as the Toledan Tables), passing through the Canary Islands, which he referred

to as the Fortunate Islands. It differs from the so-called meridian of water,

usually considered to be 17;30 to the west of Ptolemys prime meridian, and

used in some Andalusian zijes.42

f. 60v

This table is for transforming time-degrees (for each integer degree from 1 to

180) into hours (from 0;4h to 12;0h).

The following two tables have the same format as those on ff. 44r45v.

This table displays the correction for the lunar anomaly (in degrees, minutes,

and seconds) as a function of the solar anomaly, given for each degree. The

letters s and a in the margins have the same meaning as in Tables a, b, and

c. The maximum is 2;43,35 at anomaly 3s 2 (see Table g). The entries are the

same as those in the column labelled correction for the lunar anomaly in the

corresponding table in Vatican, ms Heb. 379, ff. 7a8a (see Table aa), where

the same maximum value is found. We have not succeeded in accounting for

the entries in this table despite the fact that its entries differ from those in

Tables a and c by factors of proportion, and despite our expectation that the

explanation for Table g should be analogous to that for Table h (see below).

This table displays the correction for the lunar anomaly (in degrees, minutes,

and seconds) as a function of the lunar anomaly, given for each degree. There

is no indication in the table or the margins concerning the addition or sub-

Andalus and North Africa, Journal for the History of Arabic Science, x (19921994), 4151.

366 chapter 10

2 0; 5,32 1;24, 4 2;24,45 2;43,35 2;23,30 1;20, 1

3 0; 8,18 1;26,42 2;25,36 2;43,34 2;21,53 1;17,40

4 0;11,12 1;29, 8 2;26,30 2;43,28 2;20,29 1;13,10

5 0;13,50 1;31,33 2;27,42 2;43,19 2;18,40 1;12,30