J. Embryol. exp. Morph. 83, 1-14 (1984)
Printed in Great Britain © The Company of Biologists Limited 1984
Fibre order in the normal Xenopus optic tract, near
the chiasma
By J. W. FAWCETT 1 , J. S. H. TAYLOR 2 , R. M. GAZE 2 , P.
GRANT 3 AND E. HIRST 4
1
The Salk Institute, 5-29018, PO Box 85800, San Diego, California 92138,
U.S.A.
2
University of Edinburgh, Department of Zoology, West Mains Road,
Edinburgh, EH93JT, U.K.
3
Department of Biology, University of Oregon, Eugene, Oregon 97403,
U.S.A.
4
National Institute for Medical Research, London NW71AA, U. K
SUMMARY
In juvenile Xenopus retinotopic fibre order in the optic tract near the chiasma was investigated by labelling small groups of opticfibresfrom peripheral retina with HRP. This selective
fibre labelling with HRP was combined with autoradiography following administration of
tritiated thymidine to the eye, so that the HRP-labelled fibres could be located within the
borders of the optic tract.
Fibres arising from the periphery of all four retinal quadrants were superficially located in
the optic tract near the chiasma, with dorsal retinal fibres showing the greatest tendency to
travel deep in the diencephalon. Retinal lesions closer to the optic nerve head labelled fibres
which ran deeper in the optic tract. Near the chiasma, fibres from ventral retina tended to
group rostrally whilefibresfrom dorsal retina tended to group caudally. However, no obvious
localization of fibres arising in temporal or nasal retina was seen in the lower optic tract.
INTRODUCTION
In Xenopus the nerve fibres arising from the retina form an ordered map across
the surface of the tectum. In trying to identify possible mechanisms involved in
the establishment of such maps we are examining the ordering of fibres in this
system.
The first major analysis of the retinotopic ordering of the fibres in the amphibian optic tract was made by Scalia & Fite (1974). They showed that, in Rana
pipiens, fibres from ganglion cells arranged around the circumference of the
retina were distributed across the superficial (or pial) aspect of the optic tract
such that nasodorsal fibres were represented at either margin of the tract and all
parts of the retinal circumference were represented in sequential order in between, with dorsal fibres lying laterally.
It was then suggested (Gaze & Grant, 1978) that the radial order of retinal
2
J. W. FAWCETT AND OTHERS
ganglion cells in Xenopus was represented as radial order of fibres in the tract,
that fibres from central retina, near the optic nerve head, travelled most deeply
through the diencephalon, whereas fibres from peripheral retina travelled most
superficially, close to the pial surface. Because of the mode of growth of the
retina, which adds cells at the ciliary margin, this amounted to saying that the
oldest fibres travel deepest in the tract and the youngest fibres travel most superficially.
This idea of chronotopic ordering arose from observations of the regenerating
system (Gaze & Grant, 1978). In these experiments the optic nerves of midfifties-stage tadpoles, cut close to the brain, regenerated and grew through the
superficial part of the diencephalic tract to reinnervate the optic tectum. This led
to the suggestion that newly growing fibres always travelled in the most superficial parts of the tract. Recently it has occurred to us that an alternative explanation may exist for these findings. From stage 53 in Xenopus a large proportion
of newly growing fibres arise in ventral retina (Jacobson, 1976). In optic nerve
cuts of the type described many of the growing ventral fibres in the nerve would
have been left uncut. The fibre population arising from the cut end of the nerve
would be mixed, comprising initially uncut newly growing ventral fibres and
subsequently a mixture of new and regenerating fibres. If the ventral fibres
maintained their ventral position in the nerve through the chiasma, they would
attain a ventral position in the diencephalic tract, i.e. they would be superficial.
The superficial position of the regenerating fibres could result from other fibres
following these ventral fibres and thus producing an abnormal pattern.
Because of this possibility we thought it necessary to re-investigate the
retinotopic positioning of newly growing fibres in the intact normal optic tract.
We have argued that, if progressively younger retinal ganglion cells do indeed
send their axons along paths progressively more superficial in the growing optic
tract, then peripheral retinal lesions in juveniles should label fibres superficially
placed in the tract, and conversely deeper retinal lesions, closer to the optic nerve
head, should label fibres deeper in the tract.
In addition to our re-investigation of the age-dependent distribution of fibres
in the depths of the optic tract, we have investigated the distribution, in the first
part of the optic tract, near the chiasma, of fibres arising from different circumferential positions in the retina. As optic fibres approach the tectum their
retinotopic circumferential ordering can best be seen in wholemount preparations where a localized region of the retina has been labelled with HRP or cobalt
(in Xenopus, Steedman, 1981; Fawcett & Gaze, 1982; in the newt Cynops
pyrrhogaster, Fujisawa, Watanabi, Tani & Ibata, 1981a; in Rana nigromaculata,
Fujisawa et al. 19816; and in Rana pipiens, Reh, Pitts & Constantine-Paton,
1983; Scalia & Arango, 1982). However, retinotopic order further back in the
optic tract, as it leaves the chiasma, is more difficult to assess, for two reasons:
Firstly, the tract here is very narrow and secondly it is embedded in other fibre
systems which make it difficult to distinguish optic fibres from non-optic fibres.
Fibre order in normal Xenopus optic tract
3
fibres. Both these factors are particularly troublesome in Xenopus and we have
therefore had recourse to a double-labelling method where HRP labelling of
selected groups of retinal fibres was combined with autoradiographic demonstration of the extent of the entire tract following intra-occular administration of
tritiated proline.
METHODS
Double labelling
The optimum techniques for HRP staining and for autoradiography are very
different, but the HRP is more demanding, since HRP degrades in animal tissue.
The problem was therefore to make autoradiography work on sections which had
been treated predominantly with the HRP reaction in mind. This proved
relatively easy to do on frozen sections which had been reacted for HRP using
the cobalt-intensified DAB reaction (Adams, 1977), followed by a more or less
standard autoradiographic technique. The cobalt DAB intensification produces
a black reaction product, which is not easy to distinguish from the silver grains
of the autoradiography. However, because of the limited beta emission from the
tritiated proline the silver is restricted to the surface 3 /im of the sections, whereas
the HRP-filled fibres can be seen running through the entire thickness of the
50/im sections. Both the labels are therefore readily distinguishable by focusing
through the sections.
HRP application to selected retinal regions was performed under anaesthesia
using MS222 (Tricaine methane sulphonate, Sandoz). Small peripheral lesions
were made to the retina using fine needles. HRP (Sigma, Type VI) was applied
to the site of the lesion either in recrystallized form or by injection as a 10 %
solution into the vitreous. In some cases HRP partial retinal fills were processed
without proline injection.
Tritiated proline (3H-P; Radiochemical Centre, Amersham, Specific activity
27 Ci/mMol) was concentrated by evaporation under a stream of nitrogen to give
a specific activity of 1/^Ci in 0-1 [A. 24h after HRP application anaesthetized
animals were injected with 0-5 pi\ of proline into the vitreous.
24 h after the proline or 48 h after HRP application the animals were deeply
anaesthetized and perfused with 0-25M-sucrose followed by 2-5% glutaraldehyde in 0-1 M-phosphate buffer with 20 % sucrose at pH7-4. Dissected brains
were immersed in fixative for a further 1 h at 4°C, washed in phosphate buffer
with 20 % sucrose and infiltrated overnight with gelatin albumen. After embedding in gelatin albumen, 50/im parasagital sections were cut using a freezing
microtome and processed for HRP using the cobalt-DAB method of Adams
(1977). For autoradiographic demonstration of proline the sections were washed
and air-dried before being dipped in Ilford K2 emulsion diluted with an equal
volume of 2 % glycerol in water. They were then allowed to air-dry, placed in a
desiccator, exposed for 14 days at 4°C and developed in Kodak D19 developer.
4
J. W. FAWCETT AND OTHERS
Sections were counterstained with neutral red, dehydrated in ethanol, cleared
and mounted.
Lesioned eyes were reconstructed from 10/mi haematoxylin- and eosinstained serial sections of the retina. Every fifth section was drawn using a camera
lucida, digitized, and reconstructed in three dimensions on a PDP 11/23 computer. The distribution of HRP-labelled fibres near the chiasma was recorded as
camera-lucida drawings of the labelled optic tract.
RESULTS
In all cases the tectal projection of the HRP-labelled fibres was correct for the
region of retina labelled; that is, ventral fibres terminated medially, dorsal fibres
laterally, temporal rcstraliy and nasal fibres caudally in the tectum.
The distribution of fibres from the whole retina
To provide a topographic framework against which to assess the distribution
of fibres from selected regions of the retina, the extent of the whole tract as seen
after labelling the entire retina with 3 H-P, or the entire nerve with HRP, is shown
in Fig. 1. It may be seen that the optic tract, as it crosses the midline, extends
dorsoventrally throughout virtually the entire extent of the cell-free zone from
the pial surface to the ventricle. Rostrocaudally, the tract occupies a narrow band
25-100 /im thick at the rostral margin of the cell-free zone. To assess the circumferential retinotopic ordering peripheral lesions filling fibres in the widest superficial part of the tract were used. The nature of the double-labelling, from which
the various camera-lucida drawings of the tract were made, is shown in Fig. 2.
The distribution of fibres from the various quadrants of the retina across the
rostrocaudal extent of the tract close to the chiasma, is summarized in Table 1.
In each case the position of the fibres is assessed in the first six (50 fim) sections
beyond the chiasma.
Fibres from ventral retina
In three cases the fibres were located at the pial surface of the diencephalon.
Of these, one case showed the fibres to be confined to the rostral part of the tract
throughout its passage through the chiasma and lower tract (Fig. 3). In the other
two cases by 50-100 /im beyond the chiasma the fibres were clearly distributed
rostrally in the tract. In all cases as the fibres passed up the side of the
diencephalon the rostral localization became obvious and the fibres entered the
medial brachium to innervate medial tectum.
Fibres from temporal retina
In four successful preparations the fibres from peripheral temporal retina ran
close to the pial surface of the diencephalon (Fig. 4). No rostrocaudal localization of the temporal retinal fibres was seen near the chiasma. As the fibres came
Fibre order in normal Xenopus optic tract
5
near to the tectum they gathered in the central region of the rostrocaudal extent
of the tract and entered the tectum directly from the front, without passing
through either brachium.
Fig. 1. (A) Transverse section through the diencephalon in the region of the chiasma
in a juvenile Xenopus. Optic fibres (ot) are seen coming in from the left. Some pass
dorsally towards the ventricle (v), some cross the midline ventrally and pass up the
optic tract near the edge of the diencephalon. The vertical line indicates the
approximate plane of the sagittal sections, from other preparations, shown in Fig.
1(B) and 1(C). Holmes' silver method. Dorsal upwards.
(B) Sagittal section through the midline in a normal juvenile Xenopus. The whole
optic nerve had been labelled with HRP and the fibres (ot) may be seen to be
distributed throughout the entire dorsoventral extent of the cell-free zone (cfz)
extending from the ventricle (v) to the ventral pial surface, t: tectum;/: forebrain.
(C) Sagittal section through the midline in a normal juvenile Xenopus. The entire
eye had been labelled with 3H-P and the opticfibres(ot) may be seen to be distributed
throughout the entire dorsoventral extent of the cell-free zone (cfz). t: tectum; v:
ventricle;/: forebrain. For (B) and (C), the orientation is as given. D, dorsal; V,
ventral; R, rostral; C, caudal. Bar = 1 mm.
J. W. FAWCETT AND OTHERS
***
- -t
B
1
Fibre order in normal Xenopus optic tract
Table 1. Localization offibres from each retinal quadrant across the rostrocaudal
dimension of the optic tract near the chiasma
Grouping of labelled fibres in each section
Section
Retinal Origin
Experiment No.
1
Ventral
1)
2)
3)
Dorsal
1)
2)
3)
4)
5)
R
C
C
C
2
0
R
R
Nasal
Temporal
1)
2)
3)
4)
R
R
R
R
C
2
C
R/-
C
0
-
C/-
1)
2)
3)
4)
R
R
R
C
C/2
R
R
R
C
C/2
-
—
C O
—
R
—
0
—
0
c
—
—
R = rostral grouping; C = caudal grouping; — = no grouping detected; 0 = not determinable, section inadequate; 2 = two groups.
Fibres from nasal retina
There were four successful peripheral retinal lesions in which the fibres were
superficial in the tract, although these fibres showed an increased tendency to
spread deeper in the tract than temporal or ventral fibres (Fig. 5). From the
chiasma no rostrocaudal localization of fibres was seen. Fibres were spread right
across the tract up to the tectum where they entered both the lateral and medial
brachia to terminate in caudal tectum.
Fig. 2. Photomicrographs of a double-labelled optic tract in which fibres from the
entire retina are labelled with pH] proline while fibres from dorsal retina are also
labelled with HRP.
(A) Sagittal section at low magnification, to show the position of the optic tract.
Bar = 1 mm. Inset region is shown at higher magnification below.
(B,C) Two photomicrographs of the ventral part of the optic tract, shown in inset
above. (B) The focus is on the autoradiographic grains on the surface of the section.
(C) The focus is on the HRP-labelled fibres as they pass through the 50 ^m section.
By focusing up to where the fibres meet the autoradiographic label, the positions of
the fibres from dorsal retina may be determined in relation to the overall extent of
the tract as shown by proline. Bar = 100^m.
J. W. FAWCETT AND OTHERS
Fig. 3. The distribution of fibres from ventral peripheral retina in relation to the
distribution of all optic fibres in the tract. The outlines on the left are camera-lucida
drawings of the entire optic tract in the first six parasagittal sections beyond the
chiasma, as shown by autoradiography. On these outlines the HRP-labelled fibres
from ventral retina are shown. On the right, the distribution of the fibres is shown
in greater detail. Dorsal is upwards and rostral is to the right. There is a discernable
tendency for the ventral fibres to group rostrally. Each bar = 200 fim.
Fibres from dorsal retina
Of five preparations two showed fibres only in the superficial part of the tract.
In both cases the majority of fibres were in the caudal region of the tract (Fig.
6), although in one of the cases a second group of fibres remained in a more
rostral position throughout the first part of the tract. In the other three cases
fibres spread deeper into the tract. In one case a clear caudal bias in fibre
distribution was noted, but in the other two cases fibres were distributed across
the tract width. Beyond the lower tract the fibres collected in the caudal region
of the tract and entered the lateral brachium.
Deep and superficial lesions
By reconstructing the lesioned retina from histological sections we can demonstrate the extent of the lesion. This method cannot be used to demonstrate which
fibres have been filled, but it can show the maximal extent of damage to the
retinal ganglion cell layer and, therefore, the largest area from which filled fibres
are likely to arise. In all quadrants of the retina central lesions, damaging central
as well as peripheral fibres, label fibres deep in the tract (Fig. 7). However,
Fibre order in normal Xenopus optic tract
Fig. 4. The distribution of fibres from temporal peripheral retina. Conventions as
in Fig. 3.
Fig. 5. The distribution offibresfrom nasal peripheral retina. Conventions as in Fig. 3.
10
J. W. FAWCETT AND OTHERS
Fig. 6. The distribution of fibres from dorsal peripheral retina. Conventions as
in Fig. 3.
Fig. 7. The distribution of fibres from a deep dorsal retinal lesion. On the left the
extent of the lesion, as determined from serial sections, is shown as black dots on the
eye reconstructed as a stereo pair. On the right is shown the distribution of HRPlabelled fibres in the ventral cell-free zone at the chiasma (see Fig. 1). Dorsal is
upwards and rostral to the right. Bar = 100^m.
Fibre order in normal Xenopus optic tract
11
Fig. 8. The distribution offibresfrom a superficial dorsal retinalfill,for comparison
with Fig. 7. Conventions as in Fig. 7.
peripheral lesions label only superficial tract fibres (Fig. 8). In temporal and
ventral retina, deeper peripheral lesions also label only superficial fibres.
DISCUSSION
The experiments reported here are part of an attempt to define the nature and
extent of the retinotopic order of fibres in the optic pathway of Xenopus. We
discuss here order in the tract, as it starts at the chiasma.
Our present results support the suggestions made by Gaze & Grant (1978) by
showing that fibres from peripheral positions in the retina labelled just after
metamorphosis, travel up the most superficial part of the optic tract, whereas
central lesions label fibres that are deeper in the tract. A comparable age-related
distribution of optic fibres in the tract has been observed in Rana pipiens (Reh
etal. 1983), in the goldfish (Dawnay, 1979; Easter, Russoff & Kish, 1981; Bunt,
1982) and in the rat (Bunt, Lund & Land, 1983). There has even been a suggestion that something similar may occur in the cat (Torrealba et al. 1982) although
here the evidence is indirect and the situation somewhat obscure.
Our results suggest that there is a difference in the depth distribution of fibres
arising from the periphery of each retinal quadrant. Dorsal fibres, and to a lesser
extent, nasal fibres tend to spread deeper in the tract than those from temporal
or ventral retina. This observation is consistent with an age-related fibre
distribution since the concentric mode of retinal growth is not even. Just after
metamorphosis, it has been shown that ventral retina is adding most cells to its
peripheral margin, whilst dorsal retinal growth is markedly slower (Jacobson,
1976; Beach & Jacobson, 1979). Likewise nasal retina is growing less than temporal retina (Tay, Hiscock & Straznicky, 1982). By making equal peripheral
lesions we would expect to fill exclusively newly growing fibres in ventral retina,
12
J. W. FAWCETT AND OTHERS
mostly newly growing fibres in temporal retina, fewer such fibres in nasal
retina and considerably older fibres in dorsal retina. These differences are
clearly reflected in the depth of distribution and dorsal and ventral fibres in the
tract.
Gaze & Grant (1978) suggested that the circumferential retinal order
described for Rana by Scalia & Fite (1974) might occur if the wedge-shaped optic
tract were to be formed by opening the retina at a certain position (ND for Rana)
and folding it around the optic nerve head, rather like a fan closing. The actual
order in the Xenopus tract corresponding to the retinal circumferential dimension has been demonstrated by Steedman (1981) and Fawcett & Gaze (1982).
Similar distributions have been described for the newt (Fujisawa et al. 1981a),
for Rana nigromaculata (Fujisawa et al. 1981&) for Rana pipiens, (Scalia &
Arango, 1982; Reh et al. 1983) for the goldfish (Easter et al. 1981; Bunt, 1982)
and for the Rat (Murabe, Fujisawa, Terubayashi & Ibata, 1983). These studies
show that temporal fibres form a narrow band in the middle of the tract, ventral
fibres pass up its rostral edge and enter medial brachium, dorsal fibres concentrate towards the caudal edge of the tract and enter the lateral brachium,
while nasal fibres run in the entire rostrocaudal extent of the tract and approach
their tectal terminations by passing through both brachia.
As the nerve fibres enter the brain they undergo a radical topographic change
from an approximately cylindrical nerve to a deep, narrow ribbon, then to a
wider wedge as they course dorsally around the diencephalon. Near the chiasma
the tract is very narrow, only 100-150 ^um at the pial surface and 20-50 //m deeper
in the diencephalon. In our double-labelled preparations we can see the
rostrocaudal extent of the tract as defined by the dense deposit of silver grains.
Against this background we have observed the rostrocaudal distribution of fibres
from the periphery of each retinal quadrant. There is a detectable segregation
of fibres from dorsal and ventral retina into the caudal and rostral parts of the
tract respectively. Looking at the diencephalic distribution of temporal and nasal
fibres we would expect to find the former occupying a central position and the
latter spread across the width of the tract. However, our preparations do not
clearly demonstrate such differences. We believe that this may reflect the insufficient resolution of our technique, rather than disorder in the projection as
it passes through the lower tract. The distribution of nasal and temporal fibres
in the optic nerve just before the chiasma may significantly contribute to the
difficulty of resolving any variation in rostrocaudal distribution (Scalia & Arango,
1982; Reh et al. 1983; our unpublished data).
There are several ways in which the orderly retinotectal projection could arise
during development. We can envisage processes which would include one or
more of the following components:
1. 'Passive' fibre following where mechanical constraints of the pathway guide
the first fibres and subsequent fibres follow these pioneers.
2. Target recognition, where retinal cells are labelled, tectal cells are labelled,
Fibre order in normal Xenopus optic tract
13
and the axons of the retinal ganglion cells locate and connect with the
appropriately labelled tectal cells.
3. Pathway recognition, where retinal cells are labelled and so are the tissues
encountered by the fibres as they grow. If the fibres are able to actively select
their pathway then target recognition may become unnecessary.
4. Fibre-fibre recognition, where fibres are able to grow in a particular arrangement which they actively maintain.
There are clearly other processes and series of processes, which could result
in the orderly projections we have described. This study demonstrates that
retinotopic ordering in both the radial and circumferential dimensions exist in
the lower optic tract. Previous studies of Xenopus (Fawcett, 1981; Fawcett &
Gaze, 1982; Steedman, 1981) suggest that this order is present throughout the
system. Such ordering could be a 'passive' process resulting from fibre following
and the mechanical constraints of the developing pathway. However, if such
'passive' mechanisms are responsible, experimental manipulation of the system
should show corresponding disorder. Results of compound eye pathway studies
(Straznicky, Gaze & Horder, 1979; Gaze & Fawcett, 1983) and our own
unpublished data showing orderly preferential pathway selection, suggest that
this is not the case. We are currently investigating whether the relative fibre order
is consistent throughout the pathway. The maintenance of fibre relationships
would be strong evidence that some form of fibre-fibre recognition is occurring
in the system.
REFERENCES
ADAMS, J. C. (1977). Technical considerations on the use of HRP as a neuronal marker.
Neuroscience, 2, 141-145.
BEACH, D. H. & JACOBSON, M. (1979). Patterns of cell proliferation in retina of the clawed frog
during development. /. comp. Neurol. 183, 603-613.
BUNT, S. M. (1982). Retinotopic and temporal organisation of the optic nerve and tracts in the
adult goldfish. J. comp. Neurol. 206, 209-226.
BUNT, S. M., LUND, R. D. & LAND, P. W. (1983). Prenatal development of the optic projection in albino and hooded rats. Devi Brain Res. 6, 149-168.
DAWNAY, N. A. H. (1979). 'Chronotopic' organisation of the goldfish optic pathway.
J. Physiol. (Lond.). 296, 13-14P.
EASTER, S. S., RUSSOFF, A. C. & KISH, P. E. (1981). The growth and organisation of the optic
nerve and tract in the juvenile and adult goldfish. /. Neurosci. 1, 793-811.
FAWCETT, J. W. (1981). How axons grow down the Xenopus optic nerve. /. Embryol. exp.
Morph. 65, 219-233.
FAWCETT, J. W. & GAZE, R. M. (1982). The retinotectal fibre pathways from normal and
compound eyes in Xenopus. J. Embryol. exp. Morph. 72,19-37.
FUJISAWA, H., WATANABI, K., TANI, N. & IBATA, Y. (1981a). Retinotopic analysis of fibre
pathways in amphibians. I. The adult newt Cynops pyrrhogaster. Brain Res. 206, 9-20.
FUJISAWA, H., WATANABE, K., TANI, N. & IBATA, Y. (1981b). Retinotopic analysis of fibre
pathways in amphibians. II. The frog Rana nigromacula. Brain Res. 206, 21-26.
GAZE, R. M. & FAWCETT, J. (1983). Pathways of Xenopus optic fibres regenerating
from normal and compound eyes under various conditions. /. Embryol. exp. Morph. 73,
17-38.
14
J. W. FAWCETT AND OTHERS
R. M. & GRANT, P. (1978). The diencephalic course of regenerating retinotectal fibres
in Xenopus tadpoles. J. Embryol. exp. Morph. 44, 201-216.
JACOBSON, M. (1976). Histogenesis of retina in the clawed frog with implications for the
pattern of development of retinotectal connections. Brain Res. 103, 541-545.
MURABE, Y., FUJISAWA, H., TERUBAYASHI, H. & IBATA, Y. (1983). A whole-mount horseradish peroxidase study of the retinal central projection in normal and monocular rats.
Anat. Embryol. 168, 151-158.
REH, T. A., PITTS, E. & CONSTANTINE-PATON, M. (1983). The organisation of thefibresin the
optic nerve of normal and tectum-less Ranapipiens. J. comp. Neurol. 218, 282-296.
SCALIA, F. & ARANGO, V. (1983). The anti-retinotopic organisation of the frog's optic nerve.
Brain Res. 266, 121-126.
SCALIA, F. & FITE, K. (1974). A retinotopic analysis of the central connections of the optic
nerve in the frog. J. comp. Neurol. 158, 455-478.
STEEDMAN, J. G. (1981). Pattern formation in the visual pathways of Xenopus laevis. Thesis,
London University.
STRAZNICKY, C., GAZE, R. M. & HORDER, T. J. (1979). Selection of appropriate medial branch
of the optic tract byfibresof ventral retinal origin during development and in regeneration:
an autoradiographic study in Xenopus. J. Embryol. exp. Morph. 50, 253-267.
TAY, D., HISCOCK, J. & STRAZNICKY, C. (1982). Temporo-nasal asymmetry in the accretion of
retinal ganglion cells in late larval and postmetamorphic Xenopus. Anat. Embryol. 164,
75-84.
TORREALBA, F., GUILLARY, R. W., EYSEL, U., POLLEY, E. H. & MASON, C. A. (1982). Studies
of retinal representations within the cat's optic tract. J. comp. Neurol. 211, 377-396.
GAZE,
(Accepted 27 April 1984)