Vous êtes sur la page 1sur 45

IMAGE EVALUATION

TEST TARGET (MT-3)

1.0

;siiM iiM

^
I.I

IIIIM
IIIII12

II

2.0

1.8

1.25

CIHM/ICMH
Microfiche
Series.

CIHM/ICIVIH Collection de
microfiches.

s
Canadian
Institute for Historical Microreproductions
Institut

Canadian de microreproductions historiques

1980

Technical and Bibliographic Notes/Notes techniques et bibliographiques

The

has attempted to obtain the best copy available for filming. Features of this copy which may be bibliographically unique, which may alter any of the images in the reproduction, or which may significantly change the usual method of filming, are checked below.
Institute

original

qu'il lui

microfilm^ le meilleur exemplaire a 6t6 possible de se procurer. Les details de cet exemplaire qui sont peut-dtre uniques du point de vue bibliographique, qui peuvent modifier une image reproduite, ou qui peuvent exiger une modification dans la mdthode normale de filmage sont indiquds ci-dessous.
L'Institut a

D D D D
I I

Coloured covers/ Couverture de couleur

I I

Coloured pages/ Pages de couleur

Covers damaged/ Couverture endommag^e

Pages damaged/ Pages endommagdes and/o Pages restored and/or laminated/ Pages restaur^es et/ou pelliculdes
foxed/ Pages discoloured, stained or foxei Pages d^colordes, tachetdes ou piqu6es

Covers restored and/or laminated/ Couverture restaurde et/ou pelliculde

Cover
Le

title

missing/

titre

de couverture manque

Coloured maps/ Cartes gdographiques en couleur


Coloured ink (i.e. other than blue or black)/ Encre de couleur (i.e. autre que bleue ou noire) Coloured plates and/or illustrations/ Planches et/ou illustrations en couleur

Pages detached/ Pages d6tach6es

I I

Showthrough/ Showthrough/ Transparence


print Quality of prir varies/ Qualitd indgale de I'impression

I I

Bound with other material/ Relid avec d'autres documents

Comprend du

material/ Includes supplementary materii matdriel supplementaire

D D

Tight binding may causr shadows or distortion along interior margin/ La reliure serr^e peut causer de I'ombre ou de la distortion le long de la marge int^rieure

Only edition available/ Seule Edition disponible

Blank leaves added during restoration may appear within the text. Whenever possible, these have been omitted from filming/ II se peut que certaines pages blanches ajout6es lors d'une restauratlon apparaissent dans le texte, mais, lorsque cela 6tait possible, ces pages n'ont pas 6t6 filmdes.
Additional comments:/ Commentaires suppl6mentaires:

obscured by errata have been refilmed to ensure the best possible image/ Les pages totalement ou partiellement obscurcies par un feuillet d'errata, une pelure, etc., ont 6t6 filmdes d nouveau de fapon d obtenir la meilleure image possible.

Pages wholly or

partially

slips, tissues, etc.,

This item

is

Ce document
10X

filmed at the reduction ratio checked below/ est filmd au taux de reduction indiqud ci-dessous

14X

18X

22X

The copy filmed here has been reproduced thanks


to the gene/osity of:
Library,

L'exemplaire filmd fut reprodult grSce d g6n6rosit6 de:


BibliotMque,

la

Geological Survey of Canada

Commission Giologique du Canada

The images appearing here are the best quality possible considering the condition and legibility of the original copy and in keeping with the
filming contract specifications.

Les images suivantes ont 6t6 reproduites avec le plus grand soin, compte tenu de la condition et de la netteti de l'exemplaire filmd, et en conformity avec les conditions du contrat de filmage.

Original copies in printed paper cover'; are filmed beginning with the front cover and ending on the last page with a printed or illustrated impression, or the back cover when appropriate. All other original copies are filmed beginning on the first page with a printed or illustrated impression, and ending on the last page with a printed

or illustrated impression.

Les exemplaires originaux dont la couverture en papier est imprim6e sont filmds en commengarit par le premier plat et en terminant soit par la dernidre page qui comporte une empreinte d'impression ou d'illustration, soit par le second plat, selon le cas. Tous les autres exemplaires originaux sont film6s en commengant par la premidre page qui comporte une empreinte d'impression ou d'illustration et en terminant par la dernidre page qui comporte une telle empreinte.

The

last

recorded frame on each microfiche


"),

shall contain the

TINUED

symbol ^- (meaning "CONor the symbol V (meaning "END"),

whichever applies.

Un des symboles suivants apparaitra sur la dernidre image de chaque microfiche, selon le signifie "A SUIVRE". le cas: le symbols signifie "FIN". symbole

Maps, plates, charts, etc., may be filmed at different reduction ratios. Those too large to entirely included in one exposure are filmed
beginning
in

be
to

the upper

left

hand corner,

left

right and top to bottom, as many frames as required. The following diagrams illustrate the

method:

Les cartes, planches, tableaux, etc.. peuvent dtre film4s d des taux de reduction diffirents. Lorsque le document est trop grand pour dtre reproduit en un seul clich6, il est filmd d partir de Tangle sup6rieur gauche, de gauche d droite. et de haut en bas, en prenant le nombre d'images ndcessaire. Les diagrammes suivants
illustrent la

mdthode.

"*
1

SMITHSONIAN CONTRIBUTIONS TO KNOWLEDGE.


281

ON THE

GENERAL INTEGRALS

PLANETARY MOTION
BY

SIMOx\
i-IiorKSSIlH
III-

NEWT 0MB,
.XITKI, STATKS XAVY.

JIATllKMAilCS

ArCKPIKll

K(1R IMMU.ICATIflV.

(I

CT

II 11

KH

S 7

.(

Pin LA nicM'M
CiJi.hlNs,

a;

nUNTKU.

70q

JAYNK 8TUEET.

ADYEHTISEMENT.
The
N".

following ]\[omoir, on the " General Integrals of Planetary Motion," was


II.

submitted to Prof

A. Newton, of Yale College, and Mr.

(J.

^\. Hill, of

Nyuck,

Y., and has received their approval for publication in the " Smithsonian Con-

tributions to

Knowledge."

JOSEPH HENRY,
Secretary Smithsonian Institution.
WASiriNOTON,
^^.

0.,

Deccnilx'i', 1874.

(iii)

"-''

tiillii

Vfi'

'"iTl.J^

IM{K1'

ACK.

Thk
tiou of
III.

|)n'S('iit

memoir may he fumsidcnd


i)ai)ci-.s liy
i)

iis.

in pnrt,

nn rxtciision

aiul jj;mprnliza</<

two foniicv

the aiitiior:

tiic first

Ixuiig T/n'orir i/rsjirr/iir/iii/iniin


in J.i<umUi''n Joiirihul,

l.iiiir i/ii!.s,ni./

i/iir.'i

/'(iHloii. ih'n

P/anc/cs,

piil)lisli('(l

tome

xvi., 1.S71

and

tlu'

second, Siir

mi.

Thro mm:

</r Mrriiu.ii/iir

Crh.slr, pnhlisiied in tiie


is

Coiiipkfi Jiendits,

tome Ixxv.
tlio snhji-rt,

Notwitiistandinj,'

its

extent, the antlior

ronscions,

Ml his

treatment of

of s(-veral gaps, wliicli nniy detract from entire rigor.

iill

some of tiiese are of snch n natnre tliat tlie reader can readily them, while the remaind(-r wonld have led hito long digressions, and tiins caused great delay in the publication of th(> paper. To the fornu-r class belong
lie believes tliat (1) tJK! analogy Ix'tween

the

expressions for

tlu-

rectangular co-ordinates

and
tlie
all

//,

wliieh differ only hi that


is

tlie latter is

comiiosed of jn-oducts of sines, wliile


;

former

composed of similar

])rodncts of cosines

and

(,')

the omission of
tinit

consid(rations of the modifications

growing out of

tlu- fact

in ecpiation (1)
all

one value of

/i

vanishes.

To

the latter class belong the omission of

considera-

tions r(>specting

the convergence of the series encountered, respecting terms of


tlie

long

period,

and respecting the occurrence of relations among

arguments,
sati'llites
tiie

such as that
of Jupiter.

known

to subsist betwi'cu tlu; will

These subjects

mean motions naturally come up for

of three of the

consideration whi'U

process of actually integrating


the most general
integration
is

tlie

differential ecpiations of planetary

motion

in

way

is

midert.dveu.

Xo method

for

the actual execution of this

given at present, partly l)ccause


it,

tlie

papcn-

may be

considered com-

plete without

partly because the author has not succeeded in

working out any

method
to

satisfactory to himself It is true that a large part of the paper is devote<I reviewing the general forms met with in a certain integrating process, but the actual execution of this process, even for a single approximation, may be consid-

ered impracticable on account of the enormous labor involved in it. It is shown, by a bird's eye view, that a certain olyect is, in tliV nature of things, attainable

but a practicable

way

of actually reaching

extremely agreeable
fully

to the

it is yc-t to be point(>(l out. It would be author to learu that abler hands than his were success-

working

to effect the actual solution of this noble

problem

in its

most general

form.
(

V)

CONTENTS.

PAUU

1.

Introduction

2.

Canonical Transformntioii of

tlic

Equations of Motion

.....

3.

Approximation
in
ft

to tlic Kcquirt'il Solution Ijy tlio Viiriutions of

First

Approximate Solution

.......
tlio

Arbitrary Constants

4.

Formation of the Lagraiigian ('uelDcionts (n


Canonical Form

at),

and lloduetiou of

tlio

Equations to a
11

5.

Fundamental Relations between


Functions of
c,, c,,

etc.

G.

Development of

n,

iij,

and Q;

1.
S.

Form

of

Second Approximation

General Tlieoreni

...... ... .......


t..o

Coellieicuts of

tliu

time, b

b,,

etc.,

considered as

10 10

24
20

i).

Summary

of Llesultd

2H

vii )

ON THE

GENEKAL INTEGRALS OF PLANETARY MOTION.


Introilitcdon.

1.

If

wc examine
tlie

wliiit lius

been done
terms of
f-tv
It,

liy

fjfeomefers

towards developinjjf
that
tlie iiio.st

tlie

co

ordinate^ of

pliUiets

in

tlie tlie
i^

time,

we

sliail .see

>,'eiieral

expressions yet found are those


the elements in a periodic form.
of
tlie

development of the secular variations of


well

known

that if

we

negleet (piantities

third order with respect to

th*?
liie

eccentricities

and

inclinations, the intei^ia-

tion of the equations

which

iv(>

secular variations of those elements, and of


llie

the

loii<,'itudes

of the periheh'

iid

of

nodes, h-ads to the conclusion that

tlie

genend expressions of those elenumts


C sin
7T

in

terms of the time are of the form

c COS

71

= S,

Xi COS

(r/,Y

-f- li^)

(1)

sin

^ cos
71

= i\ = Vj Mi cos
(/,,

JUi sill (/// -j- y^)

(/// -|- y,)

beiiif,' tlie

number of

planets-, A', M;,

aiid

//,

being functions of the eccentrici-

,5, and y^ are angles depending also on the positions of the perihelia and nodes at a given epoch. It is to be nmiarkinl that one of the values of hi is zero, the corresponding (piantities J/ and y depending on the position of tlie plane of reference.

ties at a

given epoch and of the mean distances, whih;

The

num(>rical values of these constants for the solar system have been fouiui
'I'he

liy

several geometers.
Jje Verrier

latest

and most complete determiualions are

thos(,'

of

and of Stockwell.'
consider the terms

A\ hen

we

commonly

called

])eriodic, that

is,

those wliich

depiMid on the
tion

mean longitudes

of the

i>lan(>ts,

we

shall find that their determina-

depends on the integration of


'"'^

differentials of the
//

form
/^?o.

,i
tvher(>

('V

+/71' +;ri

+ ^-v +

we put

m' the mass of the disturbing planet.

'

Smithsonian CuuU'ibutioiis to KMowli'dj,^', Xo.

23-2.

\'ol.

XVJII.
j

October, 1874.

G K S K U A L IN T E G U A L
// ii

V VL

A \KTARY M OTIO N
mean
distances of the two

function of the eccentricities, inclinations, and


the

phinets, devekjpable in
/,
/'

powers of the two former quantities.


tlie

mean

ionyitm'

vs

of

phniets.

;t, 7i'

the longitudes of their periiielia.

0, 0'

the longitudes of their nodes.


/,

(,y,

numerical integer

coefficients,
/,' /,

and

in

which

/'

+ +/
/
7i

Tile coefficient

is

/ -f-f of the form


-f-

0.
*

AtA'Jyci^"'' (1

+ A,c'-\it is
/'^'

A.,'"

-f etc.),

while the circular function of whidi


COS
sin
'^''^

a coefficient
+'''"')

may be put

in the

form

+-'"^'

+ +

cos (/7
^i
('''

(7)

c!"

(.^"^

+->"^'

'''

^'''')

+ '0.

As these equations have


developed in powers of
(.l
tlie

different elements arc time, and w( arc tints led to expressions of the form

hitlierto

been integrated the

+ .17 + .l'r + ...,);'!;^^(;7'+//).


sliall
tlie

But

it is

clear, that

we

get more general expressions


time,

if,

instead of using

developnrents in pow(>rs of

wo

substitute

tiie

ments given by equations

(1).

Tlio substitution will

genenU values of the elebe most readily made by

reducing the circular to exponential functions,

i utting in (1) for brevity

n = f "N^^

the equations (1)

may

be put in the form

ell-'

=z=

v.A^A -^

Tn the preceding differential to be integrated the coefficient of


the form
(1 -f

*"^, (/'/'

-f- //) is

of

A,"

A,.'^

+ etc.)

Aeie'J'

0*

,?/"

'^^ (^Vr

+fn' +

m + IM).

If hi the last factor

functions,

its

we sul)stitut(: the preceding ex])oneiitials for the circular product bv ,V^',^V^' in the case of a cosine reduces to half of the sum
(di)> (ciir
(</>)*

(W' +(;,/ (,';,/(x; (^y.

Substituting the values of tliese expressions in t(>rnis of the exponentials just given, developing by the polynomial tlieorem, and then substituting for the expo-

GRNERAL INTEGRALS OF
a scries of terms, eacli of the form
COS
''sin
in eacli of
('^^'

TLA NETA

II

Y MOTION.
sum

nentials their expressions in circular functions,

wc

find tlmt this

reduces to

+ '^^-=+
liavc
'i

+'A.+i,?-'.+y2>.'2+

. .

+./X),

which we

+ '^+

Ji+J.+
The
expressions A^e^
r, /',
(ji

+ '=./+/ +./ = +
/

/.'.

+
.

A^e'^ -\- etc.,

comprising products

and powers of the

sfpiarcs of

and

^,'

by constant

coefficients by the substitutions of the values

(1) reduce tlieniselves to a series of terms of tiie


Ji

form

cos (/,Xi

+ /A+
/,

+ a +
. .

.j\>.\

+J2A+
.

+i>.'),

in wliich

+ /,+

.+;,+^;+.

=0.

operations and by corresponding ones in the case of sines the expressions to be integrated iiually reduce themselves to tlie form
tlicse

By

m'A
in each of

SU)
('"''

cos

+ +
''

''^-^

'""-^-^

+ii^'i

+i?.',),

which the sum of the

integral coefficients of the variable angles van-

ishes, wliile ^1' is a function of the

mean

distances and of

tlie '2n

quantities A; and
so that

Mi.

By
it

integration this expression will remain of the


as a general

same form,

we may

regard

form

for the i)erturbutiun

due

to the

mutual action of two

jJanets, the elements of each being corrected for secular variations.

sider the action of

all

the planets in succession,


longitudes, wliich

angles except their

mean

If we conwe shall introduce no new variable will make mean longitudes in all.

We siiall therefore have, at the utmost, not more than Sii variabk; angles. We may thus conclude inductively that by the ordinary methods of apjjroximation, the co-ordinates

of eacli of

-in planets,

moving around the sun

in nearly cirinfi-

cular orbits,

and subjected

to their
tlie

mutual attractions,

may

be expressed by an

nite series of terms eacli of


/.

form

CCS
,i,;
'

('l>-.

'2?-=

/3?.3)

ii

/;,

being integer

co(>ffici(Mits,

different in each

term

Xi,

?vs

;i2

beinf

each of the form


li-[-h,t
/i,
/.J
.
.

/;,

b(>ing

iiii

arbitrary constants,

and

/;,,

h,

h^Jc,

being functions of Su

other arbitrary constants.

We
series,
tiiat

sliiill
.r

fnrtiier
is

assume that the inclination of the orbit of each jdanet

to tiie

plane of

so small that the co-ordinates

may be

dev(>loped in a convergent
wiiil(> it

arranged according to the powers of this inclination.


tiie

may

be siiown

the general expressions for

nvtangular co-ordinates will be of

tiie

form

^ SI- cos sin y = z = Sv sin


.1

( /,>.,
( /,>.,

A7,'

+ +

/,>., -j/,,X.,

^
-f

,;^,?..,j

4-

/,>.,)

(=i)

(y,>.i

+.y>j

+/,>..)

GKN K
Tlio letter
/, /,

UAL

N T E G U A L S OF P L A N B T A II Y

AI

OT

used to express the sum of an infinite series of similar terms <S' being and y having the signification just expressed, and each system of values of the integers i and J being subjected to the condition

h
It is evident that

+ h + h+

':,

j\+J-2+h+
when
to
>,
i/,

+iu

=
a general one

(3)'

and

z are

expressed in this form, any entire func-

tion of these quantities will reduce itself to the

same form.
(8)
is
:

W(>
tions,

shall

now proceed

show that the form

that

is

to say,

that having an approximate solution of this form,

if

we make

further approxima-

developed in powers of the errors of


(3).

this first solution, every

approximation

can be expressed in the form

We

can

make no

general determination of the limits within which these approxi-

mations will be convergent, we are therefore obliged to assume their convergency.

'2.

Canonical Trdusformatton of

the EQttationn

of Motion.

If
il,

we

i)ut
-|- 1

the potential of the

bodies, that

is,

the

sum

of the products of every


ditt'erential

pair of musses

divided by their nuitual distance, the


;?(/t -\-

equations of

motion will be

1) in

number, each of the form


<1-X:
ill

oil
vx.

If

we

substitute for the co-ordinates themselves their products

by the square

roots of their masses, putting

.\i=W(-.r,; Vj =/;// etc.,


tlie ditt'erential

equations will assinne the canonical form

A\'e

suppose the index


to
//,

to

assume

for

each of the three co-ordinates


-j-

all

values

from
of
tlie

tile valiu;

referring to the sun, and W(> thus have 3(//

1) etinations

form (4) the integration of which will give the co-ordinates


(!(// -f"

in

terms of the
in the follow-

time,

and

') arbitrary constants.

\\ V sliidl

now
;//

(liniinish the
Su])pi)S(>

numlier of variables
that w(> have
f,
ii>

to I)e

determined

ing general manner:


order, l)etw(>en

differential ecpiations of the first

variables

and the time

each being of the form

(It

X
form
;',./)= constant.

Suppose

also that

we have found
,/'(,',, 3'.,

integrals of these equations, each of the

....

Let us assume at pleasure


each of
tlie

)ii

other independent functions of the variables,


'

form
,

f/;,{..,, a-,,,

r,J\

GRXEUAL
so thiit the
tiic tinio
I,

i:\TEGllALS OF
x can be

I>

LA

XETA

1!

Y
/.

MOTION

variablo-s
t\ui

oxpirs.siHl as a function ot

arbitrary constants

and

'

iiik variables
si!^2>
. .

i;ik-

Differentiating the above expression for

c.,

and substituting for

^^' its

value X,

we

sliall

have

(U

vl
.-'s

ex,

-ex,

'"ox,;

By

substituting for the


^1,

in the right liand side of this equation their expres^,

sions ni terms ot

,_.^.,

,i(l

the arbitrarv constants,

we

sliall

problem reduced
variables.

luive the

to tlie

integration of

m~k

equations between that

number of

In

tile

special
//, ;,

ordinates X,

problem now under consid(u-ation, the and tiieir first dcu'ivatives witli respect to

m
tiic

variables are the co-

by winch we

shall seek to reduce the

number

servation of the centre of gravity.


ot X,, r,, etc., so

Wo

time. Tiie integrals of the variables are those of tiie consliall take for c^, etc., lin.-ar functions

chosen
tln'
//

tliat
-\- 1

tlie

form.

Let us take

reduced ecpiations shall maintain the canonical linear functions of tlie co-ordinates .<:
aoa-o

^0

=+ =
fj(

fXnXi

+
+
,

-f

;,^a;

L =oJo+
where wo have put
for

"i'''i

+,i''*'ni

symmetry
w,

= ca

or

Ho,

(6)

being an arbitrary coefficient, while the other coefRrlents urc to bn chosen, so that tlie resnltin;. differential equations shall be of the canonical form Let us represent tiic values of x which we obtain from these equations
c

by

Differentiating any one of the preceding expressions for

^,

and substituting

for

jji

Its

value,

we

ii;jve

<ii'

Wo

''0

"^
"',
f 'J',

"^

'^w,ex'

If
liy

n-placed by their expressions in , ,. ot,,, obtained solving the equations (5), that is, by tlieir values in (7), we shall have
^, etc.,

we suppose

^1
it

^ -mil

Substituting these values iu the preceding equation,

becomes

OEN

li

11

A L

NTKG

11

ALS O

V VL

AN ETARY M O

'I"

N.

i/aioafo _i_au,i
I

WiL-Wia,

."m",-,,

o'li

ji

"^V
In Older
tliat tliis

w?o

"^

m,

'^'m,

,,^_

^^.^.

equiitiou

may roducc
f'aii

to the canonical

form

dtit is

^ oil ~ di
,

'

necessary and sufficient


ojou^o
VI.,

tlnit

the expressions
,

_j_

a>in
1)1,

j2,-2
)llo

aja,
111,,

should vanish whenever


/=/. In otlun- words,
tiie

/ is

it is

to unity wlienevcr necessary and sufficient that the coefficients a sliould


y,

different

from

and should reduce

be so chosen that

(m

+1/^

quantities

]/

m^
(8)

sliould f(U-ni

an ortliogonal system.

The

first line

of coefficients

is

already deter-

mined by the equation


bv the eoiulition

(G), the coefficiint r excepted,

which

is

to

be determined

7)1,,

+
'

:'+
')

+;:=^'
})!

or,

from (6)
W'o

which gives

'l

+....+ c = y m,

C",

putting

for the

sum of

the nuisses of the entire system of bodies.

Having thus

OKX

UAL

]:

11

AL

OP P

[.

AN ETARV
is

-M

OT1OX

Tlio lumihcr of coefficients to be (letevniiiied

now
^'*-

ii(ii -\-

1).

The

total niiiii-

ber of conditions nhicb the system nmst satisfy

is

/"J~-'\butoneoftliese
'^^

beini,'

ahrady

satisfied

by the quantities in
!)y ii{n

tiie first line,

there remain onlv - '^" "^

conditions to be satistied

Z
1) quantities,

we have

therefore

quantities

which
in

Jiiay

be chosen at pleasure.

tiie substitution Avhich we have been oonsidering, and tlu! which the orthogonal system just found may be- formed, have been developed very fully by lladau in a paper in AmiakH ,h V Kmlc Xoniia/v Supcniun;
f>("iieral

The

theory of

various

modes

TomeV. (18(58).' shall, ther(>foro, at i)resent confine ourselves to a brief indication of the special form of the substituticm whicli has been found useful in Celestial Mechanics. AVe first remtirk that if we form the (/* -jequations
1)

We

by giving

in

succession

all

values from

to n,

we

shall

have by the theory of

orthogonal substitutions the

(/< -|-

1) equations

If

we

sui)pose in the first e(piatioiis

we

shall

have from

(.5)

whence,

l)y

substituting tliese valiK's of


,r,

z,

and

//,

in the

second equation, we shall

have

for the expression of

in

terms of

f,

etc. to

replace equation (7)

"-21

\/

111

siHis

IV:

Tlie
as
it

first t(>rm

of this expression

common

to all the values of

.,

represent iiiir,

does, the co-ordinates of

tlu! c<'ntr(!

of gravity of the system.


llie

It

may, there-

fore,

be omitted entiivly, wlieii


it

we

seek only

relative co-onliiuites of the vari-

ous bodies, and, in any case,


motion.

will

disappear from the differential ecpiatious of

The most simple way


equal to zero.
/

of forming

tlie

coefficients

a,j is

to

suppose"^""
/,

of

them

L(>t us first

supimse

a,^

wlienevery>

the

first line, in

which

(I,

being, of course, excepted.


will tlien

The orthogonal system


'

be of the form

Siw

lino Tniii>i;iniuilioii ik's Kiiiiittimi.-

DiircivmifUus

ilu In

DviiaiiiNniL'.

GENKllAL IXTEOllALS OK I'LAXKTAUY iMOTlON.


/'o
\/>lli

l/wio

(10)
<^2()

2I

;
'i/

-^,

l/Wo'

0,,

V"','

""

"'II

"2

Then

a will be determined by the condition

'

?/;

while
tion

all

the other coefficients in the bottom line will be determuied by the condi'

y/ ?,
Takin- the
equation
line

4- ~
/
'

=
_

w<

next the bottom the diagonal coefficient will be detennined by the ^


a?,,

_, -\- af,_,, _,

m_i

'

while the remaining coefficients of the form _,,, will be given hf the equ.itions

m
The
gcMieral values of the coefficients to

which

w.- are

ui tlie

loUuwnig way

thus led

may

be expressed

put
.".

'o

+ =

'i

m,.

by which

will

become

,.

Also, suppose
V,

AVe

Vnj
shall

then have
^,2

.^i

_ ^ "/_!
"i
,

It

is

easy to prove that the coefficients thus formed

f.dfil

the required eoncl.tions.

m.
f

l/vr. y

VI.,

UKNIJUAL l.NTKUUALS OF

1'

L A i\

KT A

11

V MOTION.

Wo

sec that, suDposing

.i

tnil bo.ly,f. is e.iuul to tlie

pkuisure, rehitively to the sun, multiplied by u func^tion of the masses, while c is efjual to tlic co-ordinate of tlie second phmet rehitively to the centre of .'ravity'of the sun and first planet multiplied by another function of tlie masses, and so on 1 lit-se functions t,, when divided by the functions of the masses just alluded to, will (lifter ironi the co-ordinates of the several planets relatively to the sun only by quantities of the order of magnitude of the masses of the planets divided by that ' 01 tiie sun.

co-ordinates of the sini or other cenco-ordinate of the Urst phmet, which may hi. auv one at
ix>i)resi,.iit

to

tlic

r the expressions (3), those quantities will tlu'inselves reduce to expressions of this same form.

and

In what precedes wc have considered only the co-ordiiiat(-s av Of course the other co-ordinates are to be subjected to the same transformation. If we ivpresent by ri and ^ the corresponding functions of // and ^, and if in the expressions for ^
s

we

substitute for x,

v/,

ami

c,

\ 3.

Approximation

to the Rccpdred Solutions hy the Variations of the Arbitrary Constants in a First Approximate Solution.

By

motion

the transformation in question W(have for the determination of the relative ot the n 1 bodies, 3 differential equations, of the canonical form

vi-r

d('-

or:,'

Jt"--c>^,-

^^^^

tuted the second members of (11), the two expressions shall differ only by terms multiplied by small numerieal coeffieients. have to show that when' w^ make a further approximation to quantities o the first order relative to these coefiicients the solution will still admit of being expressed in the form (3). To do this we sliall make the further approximation by the method of the variation of arbitrary constants, remarking, however, that the usual formula, of this method cannot l.^^ appli.>d, because they presuppose that the first approximation is a ri.jorous solution of an approximate dynamical problem, while, in the present cas,>, we are not enti-

Let us now suppose that we have found approximate solutions of these equations 111 the form (:3), the quantities x, //, z being there replaced by , ^, ami r. that IS, solutions which possess the property that, if, on the one hand, eacli expiv.si.m IS twice differentiated, and if, on the other hand, the values (=}) are substi-

We

tled

to

>yords,

we

can be formed, such that we shall find the

assume that our approximate solution (3) possesses this quahtv are not entitled to assume that any function riof the quantities

in oth.r

s
"

,,
'

and

-.in

equations of the form

rigorously and identically satisfied by the approximate expressions, both with respect to the time, and the G.* constants which the solution contains. C'onse-

cannot assume tin, (>xistence of a p,u-turbative function, and must emi)loy other expressious iu place of the derivatives of that function. set out, then, with the three sets of equations, having n in each

qn.-ntly,

we

We

November, 1874,

set

10

OEiNEllAL
I,
)7,

INTKGUALS OF PLANET AH Y MOTIUX.

^,

= = =

^7.;,
*'/.(

cos

si/i

+ + sin (y,Xi +p.a +


(/,;i,
i,?.,

sin (;,X,

-f- ijXj

+ +

/,a,)
'.m>.i,)

(l^)

+j,,:k,),

in wliich all

the quantities are supposed to be given in terms of


?.

arbitrary con-

stants

and the time, each

being of the form

/(

being an arbitrary constant, wliich each

other arbitrary constants, whieli

we may

b, k, and is given as a function of represent in the most general way by


/'

3?i

entire

So long as no distinction between a and 6m arbitrary constants by

is

necessary,

we may

represent the

Let us now take the complete second derivatives of (12) with respect
supposing
fulfil

to the time,

all

(in

constants variable.

We
in

shall

suppose the variable constants to

Lagrange's conditions,

now 3

number:^

^r,

'scij

dt~^'

jC, ouj at

~
^ "

'

}t,

Siij

dt

~ "'

'

which

will give

dt

dt

'"^*'-

From
shall

have

the second derivatives, combined with the differential equations (11), 'Sn equations of the form

wc

i=i vuj at

O^i

iO-''

which

it is

required to

satisfy.

The

expression in the right-hand

member

of this

ec^uation corresponds to

ni the usual theory,

when

R is the

perturbative function.

Let us multiply
form
in this

tiiis

equation by

^^^^

and add up the 3 equations which we


t^ all

may

way by
obtain

substituting for

the values of

-',

,;,

and

;'

in succession.

We

may thus

*v
j=i

'

v"" '^^/

'':'*

da^
dt
y;

^d^ _
ddf,

'-?
i~i

o^
r/2

^f,
iia,;

/^i da^Oiij
all

the sign S' indicating that

values of
this

and C

as wc-ll as of
'.'

P.

are to be included.
in

The right-hand member of

equation corresponds to

the usual theory.

Let us now multiply the equations (13), the


third by
^^^^'
I
,

first

by

'}', the second

by

^y',

and the

and add together the 3h equations which may be thus formed by giving
If

ail its

values.

we

subtract their

sum from

the last ecpiation, putting

UKiNKUAL IXTKGRALS OF PLANETARY MOTION.


(==i

\ou,,

(jiij

dUj

da J

we

shall

have

(".",)'';;'+(..)'';;'
the sign

etc_f _
/

'v -^.^i,, (i5)


from
1

uicludiiig, as before, not only all values of


r,

to

>t,

but the cor-

rcsijonding terms in
]})

and

'C,.

giving h

all

values in succession from 1 to

6/t,

we

sliall

have a system of

()/*

differential eiinations, the integration of

which

will give the values of

the Km

quantities
iu

terms of the time.


IJy tile

fundamental assumption with wliich we set out, the expressions


tiie rigiit tlie

for ,

>;,

and

'C,

are sueli that

ties

of which

we

neglect

hand members of tiu'se equations are small quantipowers and products. We may, therefore, after solv-

ing these equations so as to get the derivatives iu the form

integrate by a simple quadrature, sujjposing r/

<?,,

etc., iu tlic

second members to
{a,.,

be constant.
to the first

jNIoreover

we

siiall

require the values of the quantities

<i)

only

degree of approximation, and witliin this limit they nuist lu^cessarily conform to the well-known law of Lagrange of being functions of tlie constants oidy, and not contahiiug the time explicitly. This theorem will materially assist
us in
tlu'ir

formation.

-1.

Formation of the Lagmnrjum


E(/iuition/i to

Coefficients ( a,),

and Reduction of

the

a Canonical Form.
/,

llestoring

th(>

two classes of constants represented by a and

we

shall

have three

classes of the functions souglit, included in the forms

Let us now differentiate the equations (Li) with respect


brevity

to the time, puttiu" for

+ '"A 4+ + JA+JA + +
iA
'.^.
'2^.
y.>.i +,/2>...

+ !Jhn = h =N + +./;a = +;,?,, =


i:A,.,

6'

iV';

we

sliall

then have, omitting the index


',

of

f>,

k, luid iV,

Shk

sin

.V
A''

>7',=
^',

SJ>h cos
Sh'k' cos

n,'-,')

N\

To form
witli

the combination
r,

respect to

and

quantities,

two

series

aj) we must differentiate the equations (1'2) and (15') and substitute the results in (14). Li fbrininT these of terms represented by the sign S of summation are to be
(j.,

a^,

I 1

12

UKN K

II

A L

'I'

Ea

II

T,

1'

PLA N KTA U Y M
be

()

T ION.

intiltipliod tof^ctlu"!', wliicli rciulcrs


tlio (ioiiblc
y;,

it

lu'ccssiiiy to

iiion; cxitlicit in vciJVcsciitiiifif


ii.

Kiiinination

wc
to
ii,

tiius

cncouiitcr.

lliiviiifj;

of ciich of

tiic (luimtitit's x,
/,

iiiid

^'

(listiiij,Miislic(l Ity

writing' tlu; viirious values of tlie


tlie
is

index
all

wiiicli takes all

iutej^er values

from

quantities h,

/.,

and

X siiould
it

bo affeeted with
/<,

this

same index.

Ihit

it

not necessary to write

after iV or

because each

N
as

common to all the 's and >;'s, or to all the i,"s, respectively. Again, we have many values of .Vas there are comltinations of the coefficients / /, etc., which
is
/._,,

enter into

it,

while each .V has


/.

its

corresponding coefficients

/.',

in

number.

"NVo

must, therefore, consider

to be written
fu {'.>
'-i-,

''a

'V.Ji

while h and

N arc
/'

affected with the

same

indices, the first excepted.

In other

words,

we have
('l,
i-l-.

'"a

'.)
'ii..)

^Vl^

'V

'a

= =

''/'l

'"l''-l

+ +

'7'3
''i^-i

+ +
V
'3

+ +

iJhn
'M''-Su<S'

Then, in the sense in which have symbolically

wi>

have hitherto used the sign of summation


,=oc
o

we

',

=oc

i,

= OC
/;,)

= OC
),

= OC

To avoid
((,

the comi)lieation of writing so nnuiy indices


(/,,
!.,,

combination, as

by the symbol

we shall represent any one and any other combination by

Wu

shall also put

=n S'= V =
I 1

s.

This summation includes


nute, as ,
y;,

all

the terms in

all

the

values of any one


all
'.hi

co-ordico-ordi-

nates

is

or s, respectively. not here necessary, as


if,

A
/>

sign for a sumuuition including


to ^

and A' are common

and

y;,

while; the corre-

sponding (pnuitities for

being of a diffen-nt form, must be written separately,

^^'e

have, in fact, distinguished them by an accent.

The
will

co-ordinates and their derivatives which enter into the expressions


i

(d^, Uj)
A''.

then assume the following form, the index

being understood after k and

= '%/\u ra = =
i

eos
sin

X^

<V/.'u
'V,
/.',.

X
A"

(Hi)

Ki

sin

;",

a: (////), cos N',

X. I.:

sin X,.

da,

V-

'^.'^

sni AT, -!-/.

'

cos X.,

(17)

r)//

da/

sinA^;+Z'^^^'"/cos A-, ^s
r)iu

OKXKRAL
d

INTIKJIIALS OF

PLANETARY MOTrON.

13

da

'

(IH)

climiuiiijr , into j in tlio tlnrc equations (17), nnd makinf,' tlic reverse elmiie (IS), HL' liiivo the conipltte expressions necessary to form any term of tlie v\press ion
111

By

We

see at once that this expression will be of the form

'i"

A%,,
I
is

-L, sin (y,


to

,Y) +

,17

+ AT-

Since the expression


ties ot
t, r

known

he independent of

the
';

first

(l(>frr,.,>

of approximation,

A=

and

satisfy the oriyinal differential

/, we must hav(>, to quantiand A' by the condition that equations, and the coefficient Au,v must

vanish, unless

wo have
^u

iV;= constant.
l)eing

The

co(>fficicnts /*

/,,

],,,,

supposed incommensurable,

this

can only

happen when we have


and hcnco

in

(:{)'

'ly

'V

V=

'V, etc.,

N^^N,,
when
sin

{S\

X^

will

itself

vanish.

Hence,

(,,

,/.)

containinjr

term whatever, we must liave


("*, "j)

no constant

0.

(19)
?,

A!,Min, differentiating the equations (Ki), the first three with respect to the last three with respect to Jj, we find

and

"'=

Sf^

{JJ.% cos

N'f,

= _ >%

(i/>/,l

cos N,

= S,
= S,

(JJblc),

sin

N,

iJjh'/i'),

sin iV,.

14
!

OKNKIIAL INTKOHALH OF I'LANKTAllY MOTION.


From
tlicsc

expressions

it

may be shown
J =().

timt

in tlic

same
liiive
is

wiiy that \vv tbund

{,i^,

We

next to consider

tiie

combiimtions of

tiic

form (u

Ij),

for

which tho

expression

I (!u^

dij
t

i.lj

Oil,

^ A<t
]

*^

r/j

Tin; terms wiiieii

do not

eoiituin

as u factor are found to be

- -^'V.
-i
<S"
'S;..v'.

OM), ^^^

0A-),

J;l^^

cos i.y,

- A-.) A",).

oyn

^^;'^

+ Uj'^h
l'>.

"^'^
'

} ''OS

(x

havinjT the meanin;,' ^iven on pa^'e

Tlie only non-periodic terms in this expression will be those in these terms reduce to

which

/<

v,

mid

or,

by puttnig

fl

we have

(,,/,)=_f^
These
of
(<xi)ressions are

^2)

now

10 be substituted in the differential ecpu.tions repre-

sented by

is

<i3n-

(l.>), wliich will then divide into two classes aecordiuy as the derivativetaken with respect to / /, or /,, or with respect to a, or Having regard to equation (20) we find those of the first class to be of the

form
(/, a,)

(/,, ,,,)

+
(JJ, a,)

j^^i.

a.

J "' _ -f _
-.

v<

-.i
-f^.
_

ii

If,

in the first

member, we substitute

for the coeflicients their values

that

cm, noticiu" ^ "

= _(,,

/^.),

and

in

the second

member put

for brevity

the dia'erential equation reduces to

da,
or

ill

^ da,

'^
>/t

^"f9,,

-^^''
(/<

'^

= il,

(23)

U K >'

!;

It

I,

N T

!;

(1

UA

I,

I' I,

A N KTA

II

V MO

l'

ON

U
of ,,
.,,

Uy
....
poses,

ffiviiijf

all viiliicH in
viiriiitioiis

Niuccssion from
c,,
c.,,

to -hi,

we

sliiiU liiivc ;)n ('(luiitious to


viiriiitioiis

dctciiiiiiif

tho

of

c,,,,

IVom wliich the

.<, lire
it

to

l)c

ohtiiiucd

by

tlu>

iin

('(illations ('^1).

Hut, for our prcscut pur-

will be

more convenient
n.,,

to consider the c's as the


c,,
r.,,

fundamental elements,
c.,

and

to consider

, to be replaced by

in the orijjinal

ecpiations.

The second
"

class of differential e(iuations (15) will,

by

(IJ)),

be represented by

</>

(ft

rd,
the

(-il
first

ol'

(:((,~

(.I-

iki,~

<:i-(ii,\
sliall

Substitutiuf,' for the coetHcients in

member

their values

(W), we

huvo

iiti

ecpiations represented by

dih

ill

i;i,,/t

^
1,
"2

/.,

"^

,"i

\ ft'

in,

"^

"-"'

}
tliis

Putting
form.
(11,

successively equal to
first

;}</,

we

shall
''-'',

have 3m equations of
the second by
'.'",

Let us multiply the

of these eciuations by

tin-

////

cc,

(i\

hy

',

and so on

to the

'iiith,

and add

all

the products, uoticuig that the theory of

functionul deteruunuuts ijives


(

=3ll "

/,,

=+

or

according as

/ is

or

is

not ctjual toy.

Then, by jtutting

we

shall

have
dt
ill,

dt
t

(24)

dl

.111

dl

These 3 equations, combined with the


substituted in
tlie

'^n

equations (23), will give, by simi)le


(i/t

integration by quadratures, the p(n'turl)ation of the

constants, wliich, being

original equations (12), will gi\(' values of the variables

which

satisfy the original differential equations to


satisfi(Ml

terms one order higher than they were

by (12) originally.

It will be observed that if our functions of tho time

and

(\n

arbitrary constants,

which we have repres(>uted by f / and lio of > 'C, and : could be found such tiiat

i!,'

possessed the property that a function


i

for all values of

10

(JKNKRAL INTEGRALS OF
in (2:3)

I'

LA

>' i:

TA

11

MOTION,

we should have

and (24) by

puttinju: li

= il

Liu,

"^

c/j

T).

Fiiiuhniuntal

liilafloii hctu-ccn

the Ciufficioitx
('i,

of
<tc.

the

tiiiiv, />/>,.,,< Ye. ,

vouxhUicI

ax

J'Hiirliaiix "/'

<'j,

In the

pveccdiiijj; section

we haxc found
in ti'rnis of

ourselves able to express the

first

approxi-

mate vahas of the variables


t f

;J//

pairs of arbitrary constants

u
in wliich tlie

''Jn

'ail

two members of eaeh

])air

are

coiijiii/ate to

each other; or possess


o^

tiie

l)roperty tliat tlie expressions (14) all vanish except

when

and

Uj represent tiie

two members of a conjugate

pair, in wliich case


(/ ',)

we have
C^o)

=+

1.

The
tiiey

distinguisliinsi characteristic of tlu> intc^ifvals

we

liave

been investigating
'>]h

is

that

do not contain the time,


3/t

exc(>it
o.

as multiplied

by the

factors
tis

/(,

wliicii

are

functions of the

constants

This ciuiracteristic will enable


coefficients of
to
It

to

deduce a
c.

fundamental relation between tiie differential tlie first phice, we remark that each c has a />

witli respect to

In

which it stands in a ))eculiar relation, ni that the latter, multiplied by the time, is added to t!ie /, whicii is conjugate Tiie theorem in (pu-stion is this; each l)eing to c to form the corresponding
'/..

/>

supposed
values of

to
/

be niar!<ed witli

tlie

index of

its

corresponding

c,

we

shall

have

for all

and J from

to 'Sn,

in other

words, the expression

will

be an exact

difl'erential.

It is ipiitt? possilile tiiat this

theorem
1

may admit

of being deduced immediately


so,

from
fore

tile

preceding theory, but


t
)

have not succeeded in doing


the reverse; form,
'Sii

and have thereas

been obliged

consider

tiie jirolilem in

^^'e have, in starty;,


i,',

ing, sujiposcHl ourselv(>s to liave comjiletely expressed the

co-ordinates g,

functions of the

Lvi

quantities
,.
II.,

and we have
t\,c.,

just

fj.

shown how to rejiiace the first Wn (pmntities liy the quantities W we add to these the first derivatives of the co-ordinates (1(5)

GENE

UAL

X TE

(J II

ALU

OE

1' 1.

ANETAUY MOTION

wc

shall have

im

variables, roprt'sentcd

by c

>; f,

^\ rU, Tf, expressed as functions

of the

G/t (|aantities

\Ve

Let us now suppose those equations solved with respect shall tlien have 6/t equations of the form
Ci

to these last (luantities.

= ^i\ K =
,
tlie
r..
if,

'I'i.

whence
Tlie

/,

=
i|',

?*/,

(06)

^ and
arc

'P

being functions of
first intef;;rals

etc.

first

and

tliird

of

th(\se rxi)r<'ssions
(ail

tlie

(i/*

of

j;iven equations, or, wliat

we may

the

int(f,'nd

functions, bein<,' tliose functions of the co-ordinates,


etjual to arbitrary constants
JiCt us

and the time, which


ijn

rcuiai'

durnig the entiro niovenu-nt.


arbitrary constants by

now, for generality, once more represent the

"n
and
let

">

',

us consider the (())" quantities of I'oisson formed from the "eneral ex-

pression'

the symbol ^\ including, as in (14), the


ting
tiu' giMieral

'in

values of

r,,

and

in succession.

Put-

expression (14) in

tiie

form

(..,.,)=v;['.'-"-';----':=-"-].

forming by multiplication
V

tiie

jiroduct of

tiiis

expression

l)y

(27), then

puttinj.

=y, and forming

the sunuuation

noticing also that

tlie

expression
^

-,
1

itjiii

is

ecpial to unity

whenever
find

x-

and y represent the same symbol, and

to zero in the

opposite case,

we

an expression which
otlier
(

is

itself ecpial to unity wh(-n u

!,

and which vanishes

in all

uses.
Oj,
((

Now
suppose

and

may

here be any of the


/j

(>/(

arbitrary constants.

],(-t
Cj.

us then

a^ to represent

and

/^ respectively,

and

Uj to

represent

This equa-

tion will then

become
[(, '>]
\-

(',> '>)

Oi,

'.) [/,.,

'..]

f (/ ',)V^, r,]
I/npnuip'
purl of

4-

(<te.

or

'

It will 1)0

observed llml the

iiDtntioiis iiitrddiu'cil liy

iiiid

Poissou rcspi'divet)-, uro hero

reversed, a
;t

pripceeiliiiji; wliicli

wiis not liileiilKiiiid

on

tlie

llie

wiitor

November,

1874.

>

18

GEX.ERAL IXTEGRALS OF PLAXETARY MOTIOxV.


l

accoi-chng as

and

^i

repivsent the
(/, c,)

same

or different indices.

tound that

But we
from
/,

liave already

tlie

.>xpressin

vanish(-s

whenever
'^^''

becom7

'

'^'"'' "''^''''
""

"'" ''^""^"

and reduces "^"'''""' '"" "'" "'"^"^""S thus


/ is

different

[A' f,]

=^

1,
[c,, <,]

(28)
vanish.

wliile all other

combinations

[/,,

cj,

[/,, /.]

and

Let us now return

to the integral equations (2(5),

and

first

form the combination

'llw conditions (28) therefore give


i

and
the

t"''"^^^^^
(29)
first

equation applying whenevery

is

different

from

/,

the second

when they

arc

the same.

Let us next consider the combination [f /.] wliich we know must vanish for all vdue^ot I and j. forming the general expression (27) from the integrals (2G), we

[/,,

y=

[<!;,, ,,;.]

y,^^

,,;

.3

y,,^

,,,_-,

I
f

| _^

^,

j-^^^^ ^^^^

^,

This e,,atioa being identically zero, the coefficient of eacli power of vanish Identically. This gives, in the case of tiie middle term,

Forming these expressions by the general formula

(21),

and putting

we

find
'in

p-

r,7

L.

_I (,f

By

(2f)) all

the terms of these expressions vanish except that one in the

first

wluch the first coeffacient "f Sle'fir?' 'h^'^'^ reduces

'"' "" " ^''" ^'^'^"'"^ i.-vhich/; to l. llcncc

/,

iu

botho

and

(;j())

now

gives
^/^ __
fihj

(31)

GENERAL INTEGRALS OF PLANETARY MOTION.


.

19

Development of D.,

D.J,

Wc

and (^J.
and
n',

have next

to find the forms of the expressions ilj

which enter

into

the equations

{'2'.i)

and

(24).

In the

first

place

we have
J

n = v;
V{Xi

Wi

111

Xj)-

(/A

%)- +
>r,

(Z,

-Zjf

We
>7,

now

substitute for

a-,

y, and

and ^ respectively. By tliis or 'Ct. Hence, wlion wc substitute periodic series, tiie reduced expressions
0)7, jtC,

expressions (9) as linear functions of c, substitution we sliall introduce no terms of tlie form'
for ,

z their

and

C, their

expressions in infinite

will contain cosines only.

In

the forms
^i
r,i

fact, usiu"

= =

SK'i cos Sl'f sin

iV
^V

<:,=: A7/,sin.V',

we

shall

have from (12) when we put

for brevity

C"-"'0^'+("'~"-^>' + XWij IHj I /


\?H,.
9^.
'

tc.

=*

(jf

2/,

2,

Xj := z=
.'/,
2,'-

.S7",j
'S'Z-,^

cos iV
sin iV;
sin iV.
tlie

=--

(32)

A'/.',.,

Each d<>nominator in
^/
(

will therefore

assume

form
sin
\t>\2 .V'

A7- cos .V)-

(,S'A-

sin A')- -f

(A'A-'

Wlien
(-^u

W(<

form these three scjuares


till' fii'st

+ ^^)

1"

s(iuare

is

in

tii(>

second square.

Hence

we find that every term of th(- form // cos destroyed by a corresponding term // cos ( .V -I- K.) the sum of th(-se two squares will only contain ternis

of the form
;,

cos

{N^

X,).
-\-L

Since in each value

(IT))

of .V
!:

we have

'\

we

sludl

have

in

X^

+ +h+

=h

iV,

Also, since in .V the


thins,' will
liol.l

sum of
th(>

these coefficients

is

zero,

it

follows that the smn(<

tru(>

of

third of the precedinj,' s(iuares.

The denominator

in

question

may

therefore be expressed in the form

in wiiich

each

A'^ is

of the form

where
.

+ '+

(,

4-

+ = 0.
'a

mmmmt^^-

J:J..M:^.^d^y,j0^

20

GENERAL INTEGRALS OP PLANETARY MOTION.

The possibility of developing the reciprocal of this denominator in the usual way depends upon the condition tluit the constant term of *S1- cos N is larger than the sum of the coefficients of all the other terms, a condition which, so far as we
yet
tliis

know,

is

fulfilled

by

all

constant term by

1,-^

the planets and satellites of our system. Representing and the quotient of the dum of all the other terms
SI; cos iV

divided by k^ by A, so that

^( 1

+ A)

the developed expression for fl will be

n=s'';;^(i-iA+i-|A-etc.).

When we

develop the powers of

this equation will reduce itself to the

form

n = Sh cos
,., ^ while in each term
.

(/,;i,

+
^i

L?.,

1,?.3

-(- i3,.x,),

(33)

each X being, as before, of the form

= h + ht, + hn =
il) in (23)
0.

y
To form

+ h + h-\ilj

the seco id part of

and of

tiate the expressions (12) twice with respect to the arbitrary constants wliich enter into them.

and (24) we have to differenthe time, and once with respect to


Putting, as before, for brevity,

^=*l^l + *'2^3+
b

+hn^3n

= iA -\-iA-\^2

we have

+ HnK,

= Sh^Jci cos N
sin

^< = -SIH;
^^,*= or
f

(34)

Sh"k',smN'.
il'j

For the other derivatives which enter into


^1

we have

SI: k,

sinN

^f

Sijki cos

(34)'

.>,'=

SJjl:'iCos N'.

Forming
in 4,
it

tlie

sum

of the products which enter into

11^,

in the

manner represented

becomes
!jl-:)
,

^r

S^S, ^(

{h%,^

sin (N.

+ UjA'i\{f>''l'''<)^{^^n(N',

JV;) N'^)-sin(N',-\.N'^))\.

(35)

OENKllAL IXTEGKALS OF PLANETARY MOTION.


This expression reduces to the form

21

S II cos

N, where in each value of iVwe

have
li

0.

In tliis expression it may be worth while to give the complete value of ^corresponding to any value of N. The value of the latter is comphstely determined by
the indices /
i,,

etc.,

which multiply

X,, ?. etc., in its

expression.

Let then

represent the value of

mii'J.i
in (35) all

for whicli we wish to find the corresponding value of by means of (35). The required term will be found by taking combinations of and /i for which we have
/,)
)

N,-N^= N, N\. .v; = N, or .v; -f iV; = K


I-ct

us represent the combination of indices v in


tliat

iV"

by k L, vtc, and those

in

N\

byy,,y^, etc., so

we have
+y,?.,

iV, =i,?.,

+ +

+y:,X3.

Then,

in order that tlie


tlie

sum

or difference of these angles

and of iV^ may make

.V,

according to

formula; just written,

we nmst have
.)>.2

-X'^

=
= =

(,"1

'l)'<.

(,2

and
or

(,,

i3)?.3,

^V
^v
?-;

('l

i,)>-.

4-

('2

.?;)>...

+ ihn J,J>.,nwe have

For the corresponding


(,".

coefficients of the time h,


(ih

I'

(.A

h)^ + - i^h, + a - Q\
A)^.

+ _ (i.
(,

,;)7>,

i.u,)K.

Affecting k and
coefficient IIj{!,,
i
/.,

with the proper indices, as explained in 4, tlie part of the /,) corresponding to any one value of tlie angle N,, will be
)

^Xjljki{lii,^l_,

A-j(,M, i,/<2 <o,

)7>^-

+ J f^jj^^'U^
;/i,|(*o,

./='

Wm I ^',(./.-'.,.y;.-'i,

)/.',(',;

'oi,
)
!.,,

where the values ofh^ anu /^ are those just given. The complete value of //;(/ will be found by taking the sum of all the terms which we can form by giving
ctc.,ji,j.,,

to

that

is,

y in these expressions, all admissible combinations of values, the complete expression will be given by writing before the first line the

symbols
|Ui=OC
i"i

A2=0C
!,

|Uj=OC
|U3

= OC

= oc

= OC

),

22

GENERAL INTEGRALS OF PLANETARY MOTION.


y,=oc
i,

and before the second one

.2 = -cx:

y,=oc

2
,72= oc
/,,

...
.;3,>

= -oc

V _

Differentiating {US) with respect to

wo have

CM)

By

the substitution of these expressions (23)

now assumes

tlie

form
(;57)

(ft

V/jSm.V,

putting for brevity


h'

ij7i -)- If.,

By
(. -.)

the fundamental Iiypothesis tliat

tlae

adopted expressions for

, .,

and C are

and

(3(,) all

n neglected

le^eHlT r" m that approximation.


To form
the equations (24)
.

the terms wh.eh are not of the order of those neglected in the """'' ''^"'' ^ '^''' ^' '' '' '''"^ -^"- ^' ^'-' ^--'ities '

Un,^ mdex

we

wUh which

differentiate (12) with respect to

.,

wlu-rel)y omit-

.,

^, Z, and , are
cos

always to be considered as

:^:d,

dc
j'
'i

A
OK
aS"

A
,,

+
'

,V^

sin A^

f'C:

, cc ?:

. r,f,

sm

/:f)

X-{-tS/,-

: vc
,,

cosiV

(iny

The sum of

the products of these expressions by


(i=/.)^

(:3-t)

which enter into (24)


in
( a;

is

-\

l'[ S-^,

J^

,0, (^;

- ^) (

^^'>'

ilrA%

_ A^)

+ l(i''A-% ^J- (cos


i
/

at;

- A^;) _ cos

( A^;

+ A^;)
I

(^'^A04^' (sin

(A;- a;)

sin (A-;

+ A^;)

while by differentiating

(;33)

we

find

.;^=A(.^cosA^_/^.^^six).
Taking the difference of these two expressions, the equations (24)
will

(07)"

assume the

\ji

=Sh" cos .V+

Sir sin

.V.

(38)

the qmmtities J, and n. have now to

We

r being formed by a process similar to that

used in forming mtegrate the expressions (37) and (3S), and substitute th^

"

OENK

11

L INT K G K A L
<,

OP

PLA N E

'I'

AUY

AI

OTIO X

23

resulting values of

and

/^

in the expressions (12).


h,

Representing the perturhaeach value of X by

tions of eacli quantity by the sign

we

shall

have

to increase

the quantity
^?.,

,M,

thbi.
(V,,

We
into

here have the time

outside the signs s!n or cos in both

from the integra-

tion of (;38),
b?.i.

and

in

(hlji.

We

must next

find the

sum

of

tlie

terms thus introduced

Dilfereutiuting this expression

we have

We

have now

to

form the sum of the terms


/.

in the seeond

wiiich are multiplied by

iJeginniug with the second,

member of this ecpiation we have, omitting the in-

dex of

I)

dh
dt

ch

dc,

c/> dc.

^^^' ~dc,dt '^ dc.dt '^

Substituting for

'

their values in (31), this equation becomes


(

dh

dh

ch

ih

..

which, after multii)lying by


will be
T^C
/-

t, is to be added to the more convenient, instead of using h and It"


7-''*

last

member of

(38).

But

it

in these expressions, to retain

the expressions ,!?,',,', and

',

at

!?

in their present analytical form,

dty;",

dl'
^",

lleprcsentina' them, ^

for brevity,

by

t",

and

the equations (23) and (24) become


.

dcj

dt

_ ~

eq
eij

= i ^ cJi, ,r, t ^

*^
dcj

"^
'^

fV*
'eij

. o^s

^^'efjl

(4)

dt

dcj

^
.

'

'^

'

c-;^

'

6cj

If in the first of these equations


(34)'

we

substitute for the derivatives their values in

and

(3(>i, it

becomes
'^''

'J

='^''

^'

--

(^'"'^''')

^'"
}

^^

+'

"^' (^"^'J ^'^ ^"'^

+-

(^'".A

/''.)

cos ^"

Substituting in the

first

of the above expressions for


dt

we have
.

V dt'n

Sh

oh

rlh

'v.,+'^,.c,+

+'-,.e:

[^^'"^

34

GENERAL INTEGRALS OF
We
have next, in
tl>e

PLANt

.\

R Y MOTION.
tlie

second of equations (40) to substitute


its

expressions for
I.

the derivatives in (37)' nud (37)", nitaininy only the terms multiplied by
gives by substituting for b
b

This

developed expression
ij)i

-\-LJ)i+

+ K
';.,>

+ ,\^,{:,^ + J^+

+ ,..)}

e^

+
Adding
eacli of

.{.r^',(^^^

+ ^*+
find that the

+i,*.)} .^K.
sum
reduces to a series
ot

this expression to (41),

we

.ms

which has a

factor of the form

6b,_dhj

By

(31) these factors arc

all

zero.

Hence the terms of

(39) multiplied by

<

destroy

each other, and we have

the parenthesis around ^

'

indicating that
;

all

the terms multiplied by the time in

that expression are to be omitted

in other words, that, in taking tin; derivatives of

n,
k'

^,

y;,

and ^ with respect

to o

we

are only to consider the coeffici*


to vary
i,, b,, etc.

iits h,

k,

and

as functions of these quantities,

and are not

7.
j

Form of
is

(he

Second Approximation.
:

last

The rest of our process member of (38) being

now

as follows

By

iutegratii.j,'

(37) and (38), the

omitted,
Scj

we have

= S^'-'cosiV
S^''j sin

(,y.)

=_

N.

The
ties
Cj

co-ordinates ,

y;,

and

in (12) being expressed as functions of the quanti-

and

/j,

we

are to suppose these quantities increased by their perturbations,

that

is,

we

are to lind

^^
or, since

= 24^^ + 24^^'
?.<,

we have

replaced

/,

by

K N E
111 (4;J)

UAL

NTK

(J

UAL

O V VLA X

10

TA UY

MUT

ON

25

wc

liuve

hb,=^'p
oc,

= :\n

,'.cj=,S

(b. X 7': COS


/(';

iV,

>-i

ov.

and, integrating,
^^^<

= ('V,)+J'Vv/^
j
-

3/1

f/'

/(',

(7/
iiu iV,

;-

0- cr, I

which, for brevity,

wo may

represont by
h'/.i

S',7v,

sin iV.

(44)

l)iitting

Tn adding tho ofroct of tho ix-rtuvbiUioiis


/,

hci

to ,

>;,

and

^, wl- iiir to

vary only

tile

fxpressions for h^,


<V ='S.
I

etc.,

being
/

hk cos
'^^-

.V

sin

N{IMx
.V(/,,^>.,

+
-f

ip...

h = '%
I

.sin

N+ k cos

Uv,.,

+ +

/;A>.;,)

4,S>.,)
[

W(

We

= S, I a-'sinXf //cosXcy,,^?.. +^V^X2+

+iJ>.J
I

are to put in these expressions

and the values of

(^?.

in (44).

We

tiius find

'f

+ ^ ^<.
;i,

i:.

^'
(

),

Z^. (/,/.,

^J|

+ /,A, +
;,/.,

+ /,A,x
I

cos

a;

_ ^;)

s\,, I V, (^'
;:J'

),

+ k^ {i,L, +
A-,

.;...+ /,z,),
I

sin

(X + X.)
(

+^
5C

^n,..

:^.

(^^

;;' ).

{lA

-V

u. +

/:,A),.

sin
I

a;

- X,)
+ N.)

=i

'^%|:t.(^'^|),

+ ^<(.^Z,+^A,+

+./.Z.J.|sin (X',
+i<Z,X
I
1,

+h

^^.

S<

(^'j;

^),-l/^J,^ +.^A, +
Since, in iV

sin (A-;,

- N,)

we have 'ii=
"

while in
4

^V,

"

ii

= 0,

November.

1874.

it

26
it

GENKllAL IXTEGllALS OF PLANETAUY MOTION'.


follows tlmt
y;,

all

these terms will be of the

same form with those already contained

in t,

and

(12).
inte<,'ration

In the preceding
b,

we have
lint

tacitly siipposed the coefficient of the time,

never to vanish in any case,

some of the values of A'

will necessarily

he

zero,

and

in this case, instead of

having
.'

JA(//cos jy'=

sin iV,

we must put

fi
The
cient of the vanishing term by
/(",

(It

cos

N=
hi.

Id.

only terms of this form are found in

If,

in (3S),

we

represent the coeffi-

we
-V

shall

have

for the terms in question

= - rj.
is

This adds to
quantity
/(".

'/.

the same expression, and

equivalent to diminishing h by the

We
tiie

make

this chang(> not only in the original


(S^,

terms of

r,,

and

'C,

but also in

terms of
tlie

by quantities of

and (^^', because the change will only second order, which we have rejected throughout.
(V,

affect

them

Making

these changes, the expressions

i
will

+ Ki
be
in

-\-

'Vi

"il

if

+ K,
second order,

now

satisfy the differential equations (11) to quaniities of the


still

As we have made this one approximation without changing tlie form of tlie original integrals, so may we make any number of successive approximations. We may, therefore,
while their Ibrni will
all

respects the sani(! as in (12).

regard the form


e
y;

= =

,S7.aS7.-

cos
sin

(;,?.,
(;,?.i

+ + =

?j?.2
*,?.,

4-

(,'

3= S/i sin

(./,Xi -\-jnX.,

+ +
bA
b^

+ +

/;,?.:,)

l,X,)

+j,n>-jJ,

where each

'/>.

is

of the form
?..

'.-{-

7,

being an arbitrary constant, and

/.',

Jc,

and

being each functions of

3)t

other

arbitrary constants, while


'|

fin<l

./i

+ h+ +.h +
to

'),.

+.hn

= = 0,
1,

in each separate term

under the sign S,

be a genc-ral form in which the relative

co-ordinates of n planets, revolving in nearly circular orbits with a nearly uniform

motion,

may be developed when

the approximations are continued indefinitely.

This may, therefore, be regarded as the general form of the integrals of planetary
motion.
8.

Geuernl Theorem,

If

irr I'xprpxx
tlir

the rcldti'rp lirinp force

of

the rnt'nr xt/strDi


/<,

I'li.

Icriiift

of

the ediionlrol

ehnunlx,

eoejjh'irutu i,f the

time b

b.^,

irlH each be eqiinl to the negative

GKNK
of the
iinj
ilrr! rat! I'll

UAL
d/

N TE

UA LS O V
of
the

PLAN E T A

II

MOT ON
1

27

the coHKiunt tvnii

11 riiKf

force with renpert to

Itx corren/iojiiJoi"

canonical element.

Tluit

is

to say,

it'

we

rcpiL'scnt the foiistiiiit

term

the iiviiiy

force
bIiuU

by

V,

and supiiose

to be expressed in terms of the ciinouicul elements,

we

huvo

h=-:

dc.

From

the expressions (9) for

a-,

and

tlie

rorresponding expressions for

i/

and

z, it

will be seen that the expression for the relative living force is

^\./v,.^' V"'u"'

+ ./nJ'^ V
'

+ ^ivk^^+v<'-^
-\-)-

y
etc.
r'

etc.

etc.

corresponding terms in
t', etc.,

and

l,'.

Here the

coefficients of

are those

which we have sliown

to

form an ortho-

gonal system, and, by the properties of


^i;.(i'^
.Substitutuig for
^', r^,

sncii a system,

the expression reduces to

-^''.

+
sin

s'^)-

and

^"

their periodic expressions

- Sbk
t

N
N
be

Shi- cos

the constant term of the living force

is

fonnd

to

the;

sign S* having the signification given on pnge 12.


tiiat

Compare
b^,

tliis

expression
all

with

of

e,

in {'21).

Multiply each

f,

by

its

corresponding

and add

the

vroducts, remembering that


I)

=-

:1
-If!.,!).,

!J|^

-|- etc.

for ^

and

/;,

aod
'It

We

<'t''. for () =JJt^ -\-,)A V thus find, from the expression for l''just given,

V=

h^c^

h.f.,

/|,c,|

+
c^

6;,e,.

Differentiating this expression with respect to


W'e

and substituting

'

'

for

ccj

re;

have

5fo

f)6,
';!

(46)
of the degree

We
(c
f',,

have now to show that


c,).

h is a

homogeneous tunction

3
['"

in
]

Let us represent such a function of the nth degree by

*
Aiirify

98
Let us
,
>:,

(:

N K

I!

F,

N T

i: (i

I!

I,

()

1. 1,

a N K T A

i:

M oTIO
Sinro
of

N.

r.>i)r<-sn,t
ill-'"

the

li

..l,.n,..Mts ..f tl..

systr... l.y r^.


in
ili,.

.,,,

vtc.

u-,

y, 2, a.ul

s,

Mil liiuiir .o-..i(liiiutc.s, -,.

Imvr

(.xpirssioiis (1(1)

tlit"

latter

tin,.- Nv.. .lifH.rcutiiite th.-se xpicssions with resp.Tt to the miiltiply the eoifficieuts by h, u linear liiucticm of /> I,,, cte. Ik'uce

Kv.iy

time

we

= [""\6'^'j.

'I'lic

form of the potential il shows tlmt

""
:/;i;:'\:':::;':.."'';r,

'"" '"

'""""'" "''-"""'

"-

"- -i-

In order that the diftWeutial e,,uation

;^f

= ^^

n..y he satisfied identically

we

must have

K",
or

6'-"]

['--)],

The

expression (-^1) for

,,

/.

hrinfr

Unrar

in a, is

of the form

Ileue.., wh.M, wt>

express

A,

in terms f r

<

cte.,

we must have

Tiie fundamental property of homo,

neons functions

now

gives

Substituting in (4G),

we

find

t/c,

the tlieorem enunciated. This theoren. camiot b,. .lirectly employe,. , ohtuin the values of /,, for the eason tl.at 1 cannot be .Ictern.ined as a InnCou c.f the canonical constants until the equations ot motion are completely iiitei.'r .te.i.
is

which

'*

9. Snmmnri/ of

UrmUK

The

following

is

a brief

summary of some

preceding nivestigation.

of the results which follow from the


for

We first suppose that we have found expressions .uch as ulent.cally satisfy the .lifferential

,",

,,

and ^ of the

forn, (!>),

equations (11).

We

also conceive the

(J

K X
/.

!;

u A
/*

li

!;

ii

i,

i- 1,

a > k

ii

y mot on
i

29

(luiuifitifs
t.,,

and

as

cxpn ssed

in

terms of

'-in

t'anonical coustauts t

f^, f,

so chosiin

tliat tlio

fxpriNsion

(Cj

f/j.

fc^ i/j
sliall

ffj

ell,

Ninill

rcilncc to unity wlicii


',

kJ,

tinil
is

vanish whenever any ot'er of the


/j.

Cii

(iniintities

r
I,

Tlici'.itn
,
i',,,

/,

/,

substituted tor
si'ries

'I'iien:

by

If,

taking the entire


>:.4.

of

',iit

co-ordinates represented

r^i

s'l

Cfu ^ve

multiply the square of each toetticient

k
is,

l)y

the coefficient of the time in the correspond injj angle /jXi -j- I'a?., etc. (that by the corresponding (piantity -\- ih^ -j- etc., ory,/*, -\-JJ>i -\- etc.), and by the
!',<,

coefficient

!j

or

j)

of any one of the


c,

?.'s,

as

>.j,

which

'/.

is

to lie the

same
the

tiiroughout,

then

all

tiie

constants

except
will

r^,

will

identically disappear from


'-.V^.

sum of
is

all

these products, which


in eipiatiou ('21).
T/i'DiTiii II.

sum

reduce identically to
of the time,

I'his

theorem

expressed

The

'An coefficients
'

6,, tj, etc.,

considered as functions

of

ci,

t'a,

etc., fulfil

the

,,

conditions expressed by
vh,
/'''>

where
all
till'

and

./

may have any

values at pleasure from

1
v.,

to 3n.

They
c,.

arc therefore

partial derivatives of

some one function of


is
c,,

c,,

'J7wuiT)ti 111.

This function
The
sum

the negative of the constant term of the expresc,.,,

sion for the living force in terms of


'f/iioniii

etc., as

shown

in the last section.


c,,,

IV.

of the canonical elements

c,, Cj

is

equal to the

" constant of areas." this constant being cither the


velocities

on the plane of
tlu>

A""!', or,

which

is

the

sum of the canonical areolar same, the sum of the products ob-

tained by multiplying the actual areolar velo( ity of each body around any point,
fixed with reference to

centre of gravity of the system, by the mass of the body.


:

This theorem

is

demonstrated as follows

The sum

2 w,(av/,
1=0
is

.r'i2/i)

known
ill

to be a constant
for
:,,

j)ressi()ii ({))

by t!ie principle of conservation of areas. From the exand the corresponding (expression for y introducing the quantity

Uoi as

((S),

we have
('i.'/i
'

illi)

*-

y.'.j',

i^j'.kJt

=0* ^

wr,-

multiplying by ; and then

summing with

respect to

/,

we have

=0

})li

By

the condition of the orthogonal system (S) the


is

sum

in brackets vanishes

ever/

different from
c'y

/,

and

beconi(>s unity

when

these indices are equal.


is

whenMore-

over in (5)

and

vanish whenever the origin of co-ordinates

fixed relatively

30

GENK

II

AL

XTKG H A L

OF

PLAN E T A R Y M O T

X.

SuLstitutins for

^, ,,

C',

and

,'

expression heconies

their oxpressioas (16), the constant term of this

WW.
we a<ld all integrals we have
But
if

the values of

in (-l), noting that

by the

forui of the general cv.in.i.u

h-\-h

we

j\+j2+.h
find, also,

+h+ +
:/;

+ ^,,,=1
+;:,=
S'l>/^,

0,

and hence

neon,,, V.-Thc constant part of the living force, whicli constant //.n the integral of living forces, nsually expressed
IS

is

itself

eqnal to the

in the forin

represented hy

kV,+6,c,
as already

+&,,<,,),

shown

in

}).

The constant part of

itself is

therefore equal to

V,+ft,r-.+

+h,c,.
the preccdii..- ^l.eorv

The equality of // to the constant part of T may be shown by or U may I. easily deduced directly trom the
Jacohi.
I

theorem of lini.g

Lr.

as Jl.own

\or/rsunf/('ii iihrr

Dumimll,;

p.

>[)

>

that the Lagrangiau coefficients ,./,), the sum of the-canonical auH, lar velwcitu-s and the .l.tterence between ,he potential an.l living force are al constant, give r,se to a nun.bcn- of relations b..twe.-ii the quantities /, th. <U>nvat.ves w,th respec. to c, which I have not jet t;.n,l of any use \n the o e ^ turns ot integration. I theirtore omit to c.ite them, especially as tlu-ir ,.ompl..,e i<mpi(i(. expressions are rather complex.

The conditions

ad

are those in which it vould be ..xpress.ons for co-ordinates of th,. planets, if we wished true for all tim,>. The usual expivssious are suici,.utlv -cct tor a few centun,-s, but fail .ititvlv wh,-n w,- exten.l the tim,l.von,! c-itmn bmits But, in the case of th,- plaiu.fary system, we are ol.li,..., , :,., , h,.m tor th,^ r..asn that formulas dev..i,.ped in multipl,. of the >;5 in,I..p,.n.l..t a guments o that system w,udd b,. uun.anag,.abl,- iu p,.a,.ti,.e. Hut, i ,L ,.,J ot the subsuhary systems, as the Tellurian ami Jovian for instanc,-, the scdar
cce.ss,u>

l^e i^nns
to

whi..h

we have been considering


flu.

develop

these expressions to hoi,

wUiLh UHaiccs the ium.b..r

ul

milly imlqu.n.ient arpiim.nts

t..

;)/,-!.

'.'.^.f"

aKX
vaiiatious of
fails

10 15

AL

N T K (J U A L S O

V V L

AX KTA KY MOTION

31

wen

orbits are so rapid that the apimiximatioii in pow(>rs of tlu< tiino for \n-csvn\ uses, llnicc, tiu- lunar theory, eonsidered as a problem
tlic

of

always treated in a manner analogous to that in which the general theory of planetary motion has been eonsidered in the present paper, the three arjrumeiits introdue(<d by the moon being her mean longitude, and the longitudes of her node and perigee. In the theory of Delaunay the analo.ry in (piestiou
is is

three bodies oidy,

nu)st easily seen.

His A, ^\
/<',

7/,

the constant term of

to wliicii

represent three of our canonical' elements e he constantly approximates, is the constant part,

of so much of the expression for th(> living force as contains /., Cr, and 7/, by differentiating which with respect to the latter (pmutities, lie obtains the expressimis for the motions of the thriM- arguments.
Tlie theory of Jupiter's satellites has been treated

by

^l.

Souillart in such a

niam.-r that the <\)-ordinates

contain, instead of the longitudes of the ju-riioves, the varying angles on which these longitudes de|)end. His analytical theory

may

Aiiiialrfi </r /"JCcuIr Xonmtic Sii/u'rieiirc, \o\. '2, INt!,'). he hopiMl that the gen(<ral view of the subject taken in the presi-nt paper will afford a means of introducing a more rigorous system of integration in such cas(>s. One of the special i)rol)Iems growing out of this geiu'ral theorv will !)(< the determination of the coeffici(-nts of the time, A /<.,, etc., eitlier in terms of the canoniIS

given in the
It

may

cal constants c

c,, etc.,

m- of

th(>

largest of the coefJici,-nts

/.

in

the expressions for

the co-(mlinafes of the several planets.

These coefHcients
/>

are,

approximately, the

mean
re.;;s

distances of the planets.

The

(pianlities

of an e(piation of the

'.]u//i

degree, but the writer

ought, perhaj.s, to api)ear as the h.is not yet succeeded in

forn.'iig

any expression
.ijiiares

only the

fitted to give ris.> to such an cpiation, excei)t one in which of the (luantities in nuestiou appear.

ril|)i.,.,aK|l

11V

TIIK
1

.SMITIISOMAN
N U
I'

Ml T

li

T U) N

A a
l>

1)

(
S 7 4

K ! K

M K

II