Behavioural Processes 70 (2005) 122–131

The role of edges in the selection of a jump

target in Mantis religiosa
Karin Hyden, Karl Kral ∗
Institute of Zoology, Karl-Franzens University Graz, Universitaetsplatz 2, A-8010 Graz, Austria

Received 28 February 2005; received in revised form 19 May 2005; accepted 20 May 2005

Abstract

Before jumping to a landing object, praying mantids determine the distance, using information obtained from retinal image
motion resulting from horizontal peering movements.
The present study investigates the peering-jump behaviour of Mantis religiosa larvae with regard to jump targets differing in
shape and size. The experimental animals were presented with square, triangular and round target objects with visual extensions
of 20◦ and 40◦ . The cardboard objects, presented against a uniform white background, were solid black or shaded with a gradation
from white to black.
It was found that larger objects were preferred to smaller ones as jump targets, and that the square and triangle were preferred to
the round disk. When two objects were presented, no preference was exhibited between square and triangular objects. However,
when three objects were presented, the square was preferred.
For targets with a visual angle of 40◦ , the amplitude and velocity of the horizontal peering movements were greater for the
round disk than for the square or triangle. This amplification of the peering movements suggests that weaker motion signals are
generated in the case of curved edges. This may help to account for the preference for the square and triangle as jump targets.
© 2005 Elsevier B.V. All rights reserved.

Keywords: Behaviour; Visual target; Image motion; Shape determination; Range determination; Mantis religiosa

1. Introduction body slowly from side to side in the horizontal plane,

Before jumping to a stationary landing target,
locusts and praying mantids engage in a distinctive
type of behaviour referred to as “peering”. While star-
ing straight ahead, the insect shifts its head and upper
∗ Corresponding author. Tel.: +43 316 380 5614;
fax: +43 316 380 9875.
E-mail address: karl.kral@uni-graz.at (K. Kral).
counter-rotating its head while moving so as to keep
it oriented straight ahead. In the case of a station-
ary target, the lateral shift in the insect’s eye position
shifts the angular position of the image of the target on
the retina. Manipulation experiments, where the tar-
get is moved, indicate that the neural system of the
insect can determine the jump distance from the veloc-
ity of retinal image motion (Wallace, 1959; Sobel,
1990; Poteser and Kral, 1995; for review, see Kral and

0376-6357/$ – see front matter © 2005 Elsevier B.V. All rights reserved.
doi:10.1016/j.beproc.2005.05.003
 K. Hyden, K. Kral / Behavioural Processes 70 (2005) 122–131
Poteser, 1997; Poteser et al., 1998; Kral, 1998a, 1998b,
1999, 2003a, 2003b; Kral and Prete, 2004). These find-
ings are confirmed by the linear relationship between
the jump speed of the animal and the target distance
(Sobel, 1990; Poteser, 1998). This implies not only that
an absolute value for the respective object distance is
available as data at the time of the jump, but also that
this can be translated into an appropriate degree of mus-
cle activity.
Behavioural studies in the natural environment show
that a freely moving praying mantis does not jump
blindly, but rather appears to look first for the most
accessible object (Horridge, 1986; Kral and Devetak,
1999). The animal seems to execute peering move-
ments directed toward several objects before deciding
on a target. Clearly, it is important for a mantid to
select an object on which a safe landing is guaran-
teed. Under controlled laboratory conditions it has been
found that in the case of objects at different distances
from the mantid, a stimulus is assessed as a suitable
“jump target” if it is determined to be safely accessible
to the mantid (Walcher and Kral, 1994). The stimu-
lus interpreted as a jump target was always the nearest
object, indicating that the highest peak velocity of reti-
nal image motion of the target arising from peering
movements is the relevant stimulus for a jump (Poteser,
1998). The present study examines whether stimuli dif-
fering in shape and size presented at the same distance
from the experimental animals attract equal numbers
of jumps. In addition, the amplitude and velocity of the
horizontal peering movements are investigated in order
to determine whether there is a behavioural correlation
with the shape and size of the jump targets. The objects
were positioned at a sufficient distance so that the man-
tid would not be able to determine their characteristics
with the aid of other cues, such as binocular cues (Köck
et al., 1993).
2. Materials and methods
2.1. Animals
In the present study, second- and third-stage lar-
vae of the praying mantis Mantis religiosa, with body
lengths ranging from 7 to 10 mm, were examined.
The larvae were held in an incubator with a 12 h
light-darkness cycle (with a brightness of 500–800 lx),
123
at a temperature of 21 ◦ C and a relative humidity
of 50–60%. The animals were housed individually
in numbered plastic cages with a height of 10.5 cm
and a diameter of 4.5 cm. The mantids were sprayed
daily with water and were fed wingless adult fruitflies,
Drosophila melanogaster.
2.2. Arena and behavioural experiments
Between 8:00 and 16:00, individual mantid larvae
were placed on a round island with a diameter of
20 mm. This island platform was positioned in the mid-
dle of a Petri dish with a diameter of 91 mm and a height
of 15 mm. The Petri dish was filled with water to form
a moat surrounding the platform and was placed in the
middle of a round arena made of white cardboard. The
arena had a diameter of 30 cm and was surrounded by a
wall 20 cm high (see Fig. 1). The arena was uniformly
illuminated directly from above. The illumination in
the arena was between 800 and 1000 lx.
Two or three objects differing in shape were pre-
sented simultaneously to each larva as potential jump
targets. In the first part of the investigations targets
made of black cardboard sheets were employed, shaped
as squares, equilateral triangles and round disks. These
Fig. 1. Schematic drawing of the arena used to study the peering-
jump behaviour of mantid larvae (Mantis religiosa). In the middle of
the arena is the round island surrounded by water, where the larva
is placed. Cardboard objects are then presented to the experimental
animal as potential landing targets. On the right-hand side of the
figure the targets used are shown. For more information, see text.
124 K. Hyden, K. Kral / Behavioural Processes 70 (2005) 122–131
particular shapes were chosen so that the objects would
possess vertical, horizontal, curved and oblique edges.
For each shape two sizes of object were used, with
visual angles which measured 20◦ and 40◦ , assuming
that the eyes of the experimental animal were posi-
tioned at the edge of the island. Preliminary studies
showed that a visual extension of 20◦ represents the
threshold visual angle for the jumping behaviour, while
a visual angle of 40◦ corresponds to the central frontal
part of the eye, where the ommatidia and photoreceptor
cells are directed forward (see Köck, 1992). The pre-
liminary studies also showed that black objects were
always preferred to grey ones as jump targets, i.e.,
objects with the greatest contrast to the background
were preferred (K. Kral, unpublished observations). In
the second series of experiments a similar set of objects
was presented, but instead of being black the objects
were shaded with a gradation from white at the top to
black at the bottom (see Fig. 1).
Objects presented simultaneously were equally far
from the edge of the island, in the optimum jumping
range of 15 mm for mantids in the second larval stage
or 17 mm for mantids in the third larval stage. At these
distances, the objects were far enough away that dis-
tance estimation on the basis of binocular cues would
not be possible (Köck et al., 1993; see also Leitinger,
1994). The objects were presented against a uniformly
white, unstructured background. The angle between the
near edges of the objects was 120◦ when two objects
were presented, and 80◦ (for objects with a visual angle
of 40◦ ) to 100◦ (for objects with a visual angle of
20◦ ) when three objects were presented. The vertical
position of each object was chosen in such a way that
the object was aligned with the centre of the mantid’s
frontal visual field, provided that the mantid assumed
an ordinary straight posture, which was usually
the case.
During a trial with an individual animal, the objects
were positioned according to a random sequence. This
eliminated the possibility that the mantid might select a
particular target based on its position. The trial for each
experimental animal lasted 1–2 h. In this time period
an average of two to four target selections with jumps
could be observed. However, a trial was concluded
when the total number of jumps for the particular tar-
get combination reached 100. Following a trial, the
experimental animal was not examined again that day.
All of the targets selected were recorded; however,
jumps were included in the statistical analysis only if
they were preceded by significant object-related peer-
ing movements directed toward each of the potential
landing targets (first toward one object, then another,
etc.) (see Walcher and Kral, 1994). It could thus be
guaranteed that all of the objects were included in the
target selection decision. If a mantid did not show an
interest in the objects by moving to the edge of the
island after an adaptation period of 15–20 min and
beginning object-related peering movements, it was
returned to the incubator and was not tested again
that day.
2.3. Video analysis of peering-jump behaviour
Peering movements occurring immediately prior to
jumping were analysed. They were taped with the aid of
a CCD video camera (SONY, DXC-107 A/107 AP) and
an S-VHS recorder (SONY, SV09620P). The camera
was aimed at the animal from directly above or from the
side. The video recordings were stored in a computer
with the aid of a Cinergy video card and the program
Cinergy TV. The program VirtualDub 1.6.5 was used to
store the film sequences as single frames. In each case,
the two frames were selected that most clearly showed
the starting point and end point of the peering move-
ment. With the aid of the graphic program CorelDraw 8,
the distance between the starting point and end point of
each peering movement was measured in terms of the
displacement normal to vertical straight lines. With its
known diameter, the island in the arena, which was also
recorded on single video frames, served as a reference
for calculating the absolute distance. Once the peer-
ing amplitude was known, the velocity of the peering
movement was determined. The duration of the peering
movement was found by counting the individual frames
of the video recording from the starting point to the end
point. From the amplitude and duration of the peer-
ing movement the velocity could then be determined
(in mm/s).
3. Results
3.1. Black square and round black disk
In the first experiment, mantid larvae (Mantis
religiosa) were presented with a black square and a
 K. Hyden, K. Kral / Behavioural Processes 70 (2005) 122–131 125

Fig. 2. Graphs indicating the extent to which mantid larvae distinguish between a black square and a round black disk with the same (A, B) and
different (C, D) visual angles, when selecting a jump target. For significance levels, see text.

round black disk, with the same or different angular
extensions, as potential jump targets. In preliminary
studies, it was found that with a visual angle of 10◦ ,
the mantids generally exhibited no peering-jump
behaviour, regardless of the shape of the object. In
the experiments, peering-jump behaviour occurred
with objects having a visual angle of at least 20◦ .
In Fig. 2A it can be seen that the mantids did not
make a distinction between the two objects when both
had visual angles of 20◦ (n = 100). When both had
visual angles of 40◦ , the experimental animals jumped
significantly more often to the square than to the disk
(see Fig. 2B; unpaired t-test: P < 0.05; n = 100). If one
of the targets had a visual angle of 20◦ and the other
had a visual angle of 40◦ , the larger object was the
jump target significantly more often (see Fig. 2C and
D; chi-square test: P < 0.001; n = 100).
3.2. Black square and black triangle
In the next experiment, the mantid larvae were pre-
sented with a black square and a black triangle, with
the same or different angular extensions. Preliminary
studies showed that the peering behaviour was the same
regardless of whether the edges of the triangle had a
slope of 45◦ or 60◦ (K. Kral, unpublished observations).
An equilateral triangle with angles of 60◦ was used in
the experiments, so that the visual expansion would be
similar to that of the other target objects. The length of
each side of the triangle was equal to the length of each
126 K. Hyden, K. Kral / Behavioural Processes 70 (2005) 122–131

Fig. 3. Graphs indicating the extent to which mantid larvae distinguish between a black square and a black triangle with the same (A, B) and
different (C, D) visual angles, when selecting a jump target. For significance levels, see text.

side of the square and to the diameter of the disk. Man-
tids jumped equally frequently to the square and the
triangle when both objects had the same visual angle
of 20◦ or 40◦ (see Fig. 3A and B; n = 100). If the trian-
gle was larger than the square, the mantids jumped to it
significantly more frequently. However, if the triangle
had a visual angle of 20◦ and the square had a visual
angle of 40◦ , the square was the jump target signifi-
cantly more often (see Fig. 3C and D; chi-square test:
P < 0.001; n = 100).
3.3. Black triangle and round black disk
If both the triangle and the disk had visual angles of
20◦ , there was no significant difference in the jump
frequency (see Fig. 4A; n = 100). However, if both tar-
gets had visual angles of 40◦ , the triangle was the jump
target significantly more often than the disk (unpaired
t-test, P < 0.05; see Fig. 4B; n = 100). In this experi-
ment as well, the larger of the two objects was more
frequently selected as the jump target (see Fig. 4C and
D; chi-square test: P < 0.001; n = 100).
3.4. Black square, black triangle and round black
disk

In the third experiment the mantid larvae were pre- In a further experiment each mantid larva was
sented with a black triangle and a round black disk. simultaneously presented with a black square, a black
 K. Hyden, K. Kral / Behavioural Processes 70 (2005) 122–131 127

Fig. 4. Graphs indicating the extent to which mantid larvae distinguish between a black triangle and a round black disk with the same (A, B)
and different (C, D) visual angles, when selecting a jump target. For significance levels, see text.

triangle and a round black disk, all having the same
visual angle. If all of the objects had visual angles of
20◦ , the square was the jump target more frequently
than the triangle, and significantly more frequently
than the disk (unpaired t-test, P < 0.05); the triangle
was more frequently the jump target than the disk,
but the difference was not significant (see Fig. 5A;
n = 100). Similar results were found for objects with
visual angles of 40◦ (see Fig. 5B; n = 100). Here, the
disk was less frequently the jump target than the square
(P < 0.01) or the triangle (P < 0.05), and there was also
a significant difference between the square and the tri-
angle (P < 0.05). The tendency of the mantids to exhibit
preferences among the three shapes was thus more
pronounced than in the case of the smaller objects.
3.5. Square, triangle and round disk shaded with a
gradation from white to black
In an additional experiment a square, a triangle and
a round disk, all with the same visual angle and shaded
with a gradation from white at the top to black at the bot-
tom, were presented to the mantid larvae. With visual
angles of 20◦ , the square was more often the jump target
than the triangle (unpaired t-test, P < 0.05) or the disk
(P < 0.05). The difference between the triangle and the
disk was not significant (see Fig. 6A; n = 100). With a
visual angle of 40◦ , the square was likewise more often
the jump target than the triangle (P < 0.001) or the disk
(P < 0.001), while the difference between triangle and
disk was not significant (see Fig. 6B; n = 100).
128 K. Hyden, K. Kral / Behavioural Processes 70 (2005) 122–131
Fig. 5. Graphs indicating the extent to which mantid larvae distin-
guish between a black square, a black triangle and a round black disk
with visual angles of 20◦ (A) and 40◦ (B), when selecting a jump
target. For significance levels, see text.
3.6. Peering behaviour
For each size and shape of target, video recordings
of 10 peering sequences were made directly from above
or from the side. Analyses of these recordings of object-
related peering movements preceding jumps showed a
linear side-to-side shift of the head once or twice for
all targets. By means of compensatory counter-rotation
about the yaw axis, the head was constantly directed
Fig. 6. Graphs indicating the extent to which mantid larvae distin-
guish between a square, a triangle and a round disk shaded with a
gradation from white to black, with visual angles of 20◦ (A) and 40◦
(B), when selecting a jump target. For significance levels, see text.
forward in such a way that the line of movement of the
eyes was always normal to the vertical part of the target
edges. It was found that the shape and size of the target
had no influence on the general course of the peering
movements.
Fig. 7 shows the amplitude and velocity of peering
movements executed prior to jumps. A comparison of
the amplitude (mean ± S.D.) of the peering movements
showed no difference for targets of different shapes
with a visual angle of 20◦ (see Fig. 7A). For targets
with a visual angle of 40◦ , there was no significant dif-
ference in the peering amplitudes for the square and the
triangle (see Fig. 7B). For the round disk, however, the
magnitude of the peering amplitude was significantly
 K. Hyden, K. Kral / Behavioural Processes 70 (2005) 122–131 129

Fig. 7. Graphs indicating the amplitude (A, B) and velocity (C, D) of peering movements immediately prior to a jump, for different object sizes
and shapes. Results of peering movements oriented toward targets where jumps did not take place are not shown, because they are very similar.
For significance levels, see text.

greater than in the case of the other objects (unpaired
t-test, P < 0.001). A comparison of the peering ampli-
tudes for targets having the same shape, but with visual
angles of 20◦ and 40◦ , showed a difference only for
squares. The peering amplitude was greater for squares
of 20◦ than for those of 40◦ (t-test, P < 0.01).
The velocity of peering movements (mean ± S.D.)
did not differ significantly in the case of the square,
triangle and round disk, with a visual angle of 20◦ (see
Fig. 7C). However, for objects with a visual angle of
40◦ , the peering velocity was greater for the disk than
for the triangle or square (t-test, P < 0.01) (see Fig. 7D).
A comparison of the peering velocity for targets hav-
ing the same shape, but with visual angles of 20◦ and
40◦ , showed a significant difference in the values for
squares and triangles, but not for the disks. The veloc-
ity was greater in the case of the smaller targets (t-test,
P < 0.01). It should be noted that due to the temporal
resolution of the video recordings (30 frames per sec-
ond), the estimated error for the values could vary from
approximately ±4 to ±7%; however, the trends could
be readily discerned despite this degree of error.
3.7. Jumping behaviour
More than 1600 jumps to targets were recorded, and
an additional 100 jumps were examined by means of
video analysis. Landing was successful in every case,
and took place primarily on the vertical, oblique, or
curved lateral part of the edges. In some cases, landing
130 K. Hyden, K. Kral / Behavioural Processes 70 (2005) 122–131
occurred on horizontal edges, but never in the middle
of an object.
4. Discussion
The behavioural experiments with mantid larvae
(Mantis religiosa) in the present study show that, with
the exception of mantids that do not attempt to leave
the island, each animal begins object-related peering
movements as soon as it faces a stationary object so
that the object is within its visual field. The target edges
appear to be the significant visual factor, rather than
the surface of the target. Peering movements seem to
be directed only toward the edges, and landing occurs
only on the edges, never on the surface of the target.
The target selection behaviour also indicates that the
precise size of the surface is not a determining factor.
For instance, in the case of a visual angle of 20◦ , all
shapes are chosen equally often, although the size of
the surface area varies from 23 to 48%; this is also
the case for square and triangular targets with a visual
angle of 40◦ . The present findings correspond to those
for other insects, where visual orientation behaviour
is found to be determined by the extension of edges
and contrast lines, rather than by the size of the sur-
face area of objects (e.g. Wehner, 1981; Srinivasan et
al., 1990). In freely walking flies, detection of edges
occurs as soon as the edges have a visual extension
larger than approximately 10◦ (Wehner, 1972). In the
peering-jump behaviour of mantids, it has been found
that edges are detected if at least approximately 9
vertically arranged or 11 obliquely arranged forward-
looking ommatidia (<1% of all the ommatidia of
each compound eye), or at least 42–54 photorecep-
tor cells (blue-green receptors; see e.g. Heisenberg and
Buchner, 1977), are stimulated simultaneously in each
of the two compound eyes (see Köck, 1992; Köck
et al., 1993).
Furthermore, it appears that an object is more read-
ily identified as a jump target as the length of its edges
increases. Longer edges result in the stimulation of
more ommatidia and photoreceptor cells. With a visual
angle of 40◦ , the largest possible number of forward-
looking ommatidia (18–22) and photoreceptor cells
(108–132) are stimulated simultaneously in each of
the two compound eyes (see Köck, 1992). The larger
number of photoreceptors involved in peering can be
expected to result in a greater flow of information con-
cerning the retinal image motion.
However, the experiments with two targets show
that with edges greater than a certain length, i.e. in
the case of objects with a visual angle of 40◦ , whether
or not the edges are straight or curved plays a role. It
was found that the square and triangle were selected as
jump targets more frequently than the round disk. It is
noteworthy that the amplitude and velocity of peering
movements oriented toward the round disk are greater
than is the case for the square and triangle. Since the
peering movements are executed in such a way that the
line of movement of the eyes is always precisely hori-
zontal, it may be expected that the retinal image motion
associated with vertical and oblique edges could result
in greater motion contrasts or in data that are more
readily processed than is the case with curved or hori-
zontal edges. It is possible that the peering movements
of mantids are accentuated in the case of curved edges
because such edges generate weaker motion signals
than straight edges of a corresponding length. It has also
been found that peering movements can be adjusted to
distance, and that they are accentuated with increasing
distance (see Collett, 1978; Poteser and Kral, 1995).
The amplification of the peering movements in the case
of disks suggests that curved edges generate weaker
motion signals than is the case with squares or tri-
angles. Mantids may therefore perceive the disks as
being further away, which could account for the pref-
erence for square and triangular jump targets. Thus, it
is possible that the attraction of an object as a jump
target may be determined simply by the strength of the
motion signals, without involving explicit shape deter-
mination. Further work is needed to investigate whether
the system is this simple, or whether the mantid has
independent shape determination mechanisms.
It should be noted that an evolutionary (rather than a
mechanistic) aspect may be associated with the present
findings. In previous studies, it has been found that ver-
tical structures are predominant in the natural visual
surroundings of Mantis religiosa (Kral and Devetak,
1999; see also Coppola et al., 1998). The precisely
horizontal movement of the eyes during peering move-
ments might thus be an adaptation for distance determi-
nation of the vertical edges that predominate in the nat-
ural environment. This might likewise help to account
for the fact that vertical edges are preferred as jump tar-
gets. This preference is apparent in the selection among
 K. Hyden, K. Kral / Behavioural Processes 70 (2005) 122–131
three objects, particularly in the case of the objects
shaded from white to black, which may more closely
approximate natural conditions. It is possible that three
objects positioned equidistantly around the arena may
be easier to compare, because two objects are always
within view; however, in the case of two objects posi-
tioned at an angle of 120◦ from one another, only one
may be seen at a time. One reason why the preference
for the square is more apparent for objects shaded from
white to black than in the case of solid black objects
may be that the shaded objects offer less contrast to the
white background, thus making distance determination
more difficult. Whereas in the case of the solid black
objects, the greater overall size of the square does not
appear to play a role, in the case of the objects shaded
from white to black, the fact that the square has the
largest size and the greatest total edge length may be
significant. It is possible that the greater length of the
contrasting boundaries could facilitate distance deter-
mination in this case.
Acknowledgements
We would like to thank Mary Ansell for correcting
the English text and Gerlinde Pauluzzi for the breeding
of Drosophila. We would also like to express our appre-
ciation of the valuable comments on the manuscript
made by Prof. Tom Collett. This study has been sup-
ported by a research grant from the Austrian Science
Foundation (grant no. P14697-Bio to K.K.).
References
Collett, T.S., 1978. Peering—a locust behaviour pattern for obtaining
motion parallax information. J. Exp. Biol. 76, 237–241.
Coppola, D.M., Purves, H.R., McCoy, A.N., Purves, D., 1998. The
distribution of orientation contours in the real world. Proc. Natl.
Acad. Sci. U.S.A. 95, 4002–4006.
Heisenberg, M., Buchner, E., 1977. The role of retinula cell types in
visual behavior of Drosophila melanogaster. J. Comp. Physiol.
117, 127–162.
Horridge, G.A., 1986. A theory of insect vision: velocity parallax.
Proc. Roy. Soc. Lond. B 229, 13–27.
Köck, A., 1992. Morphologische und optische Indizien für eine
postembryonale Entwicklung des binokularen Mechanismus der
Entfernungsmessung bei der Gottesanbeterin Tenodera sinensis.
Master’s Thesis. University of Graz.
131
Köck, A., Jakobs, A.-K., Kral, K., 1993. Visual prey discrimination
in monocular and binocular praying mantis Tenodera sinensis
during postembryonic development. J. Insect Physiol. 39, 485–
491.
Kral, K., Poteser, M., 1997. Motion parallax as a source of distance
information in locusts and mantids. J. Insect Behav. 10, 145–
163.
Kral, K., 1998a. Side-to-side movements to obtain motion depth
cues: a short review of research on the praying mantis. Behav.
Proc. 43, 71–77.
Kral, K., 1998b. Spatial vision in the course of an insect’s life. Brain
Behav. Evol. 52, 1–6.
Kral, K., 1999. Binocular vision and distance estimation. In: Prete,
F.R., Wells, H., Wells, P.H., Hurd, L.E. (Eds.), The Praying Man-
tids: Research Perspectives. Johns Hopkins University Press,
Baltimore, pp. 114–140.
Kral, K., Devetak, D., 1999. The visual orientation strategies of
Mantis religiosa and Empusa fasciata reflect differences in the
structure of their visual surroundings. J. Insect Behav. 12, 737–
752.
Kral, K., 2003a. Behavioural-analytical studies of the role of head
movements in depth perception in insects, birds and mammals.
Behav. Proc. 64, 1–12.
Kral, K., 2003b. Perspectives on the role of head movements in depth
perception. Behav. Proc. 64, 21–22.
Kral, K., Prete, F.R., 2004. In the mind of a hunter: the visual world of
praying mantis. In: Prete, F.R. (Ed.), Complex Worlds from Sim-
pler Nervous Systems. MIT Press, Cambridge, MA, pp. 75–115.
Leitinger, G., 1994. Frühe postembryonale Entwicklung des Kom-
plexauges und der Lamina ganglionaris der Gottesanbeterin nach
Photodegeneration der akuten Zone mit Sulforhodamin. Master’s
Thesis. University of Graz.
Poteser, M., Kral, K., 1995. Visual distance discrimination in praying
mantis larvae: an index of the use of motion parallax. J. Exp. Biol.
198, 2127–2137.
Poteser, M., 1998. Die Bedeutung der Bewegungsparallaxe für die
räumliche Orientierung von Insekten am Beispiel der Gottesan-
beterin Tenodera sinensis. Doctoral Thesis. University of Graz.
Poteser, M., Pabst, M.-A., Kral, K., 1998. Proprioceptive contribution
to distance estimation by motion parallax in a praying mantid. J.
Exp. Biol. 201, 1483–1491.
Sobel, E.C., 1990. The locust’s use of motion parallax to measure
distance. J. Comp. Physiol. 167, 579–588.
Srinivasan, M.V., Lehrer, M., Horridge, G.A., 1990. Visual figure-
ground discrimination in the honeybee: the role of motion paral-
lax at boundaries. Proc. Roy. Soc. Lond. B 238, 331–350.
Walcher, F., Kral, K., 1994. Visual deprivation and distance estima-
tion in the praying mantis larvae. Physiol. Entomol. 19, 230–240.
Wallace, G.K., 1959. Visual scanning in the desert locust Schisto-
cerca gregaria Forskål. J. Exp. Biol. 36, 512–525.
Wehner, R., 1972. Spontaneous pattern preferences of Drosophila
melanogaster to black areas in various parts of the visual field.
J. Insect Physiol. 18, 1250–1268.
Wehner, R., 1981. Spatial vision in arthropods. In: Autrum, H. (Ed.),
Handbook of Sensory Physiology, vol. VII/6C. Springer Verlag,
Berlin, pp. 287–616.