/ . Embryol. exp. Morph. Vol. 50, pp. 253-267, 1979
Printed in Great Britain © Company of Biologists Limited 1979
253
Selection of appropriate medial branch
of the optic tract by fibres of ventral retinal origin
during development and in regeneration:
An autoradiographic study in Xenopus
C. STRAZNICKY, 1 R. M. GAZE 2 AND T. J. HORDER 3
Department of Human Morphology, School of Medicine,
Flinders University of South Australia National Institute
for Medical Research, London, and Department of
Human Anatomy, Oxford
SUMMARY
The formation of the branches of the optic tract has been studied with the use of [3H]proline autoradiography, during development and during regeneration of the optic nerve
in Xenopus with one compound ventral (VV) eye made by the embryonic fusion of two ventral
eye fragments.
The formation of the optic pathway was abnormal in that the lateral branch failed to
develop, suggesting that fibres from a VV retina selectively entered the tectum via the medial
branch during development. Three months after section of the optic nerve of a VV eye,
regeneratedfibreswere present both in the contralateral and ipsilateral tecta. On the ipsilateral
side regenerated fibres entered the tectum via the medial branch only. Retinal fibres entered
the contralateral tectum through both branches in some animals and through the medial
branch only in otheis.
It is concluded that mechanical factors alone are insufficient to explain the phenomenon of
selection of the appropriate medial branch by fibres of ventral retinal origin either during
development or in regeneration. Some form of fibre-substrate interaction seems to be necessary; and this ability of fibres from a VV eye to take the path appropriate for ventral retina
argues strongly that the VV eye is not a regulated system in terms of cell specificities.
INTRODUCTION
An orderly retinotopic projection of the eye to the contralateral tectum has
been described in the frog (Gaze & Jacobson, 1963). In this the temporal
retinal quadrant projects rostrally, the ventral retinal quadrant medially, the
dorsal retinal quadrant laterally and the nasal retinal quadrant caudally on
the tectum (Fig. 1 A). The horizontal meridian of the retina is represented by
1
Author's address: Department of Human Morphology, School of Medicine, Flinders
University of South Australia, Bedford Park, 5042, Australia.
2
Author's address: National Institute for Medical Research, The Ridgeway, Mill Hill,
London, NW7 1AA, U.K.
3
Author's address: Department of Human Anatomy, Oxford, 0X1 3QX, U.K.
Xy
EMB 50
254
C. STRAZNICKY, R. M. GAZE AND T. J. HORDER
a line extending from rostrolateral to caudomedial on the tectum. The retinotopic ordering of optic fibres in the optic tract has also been estabJished (Scalia &
Fite, 1974). The optic tract before entering the tectum opens out into medial and
lateral branches. The medial branch carries fibres from the ventral retinal half to
the rostro-medial half of the tectum, whereas optic fibres from the dorsal retina
reach the caudolateral tectum through the lateral branch of the optic tract.
Earlier observations on compound eyes (made by the embryonic fusion of
two nasal or two temporal halves in Xenopus) have shown that the projection
from each (similar) half spreads out across the whole rostrocaudal extent of the
tectum, instead of being to the corresponding half (Gaze, Jacobson & Szekely,
1963). In animals with compound ventral (VV) eyes (Straznicky, Gaze &
Keating, 1974), the projection from both the host and the transplanted ventral
halves expands from the medial part of the tectum laterally to cover the entire
dorsal surface. Consequently the projection of the horizontal meridian of the
VV eye lies approximately along the lateral margin of the tectum. Although
Holmes' silver stain was used on serial sections of the brains of the operated
animals in these experiments, no major malformation of the optic tract was
noticed and reported in the original paper (Straznicky et ah 191 A). We have
recently re-examined the slides from the original experimental animals and
found that in cases with one VV eye the lateral branch of the contralateral
optic tract, normally carrying fibres of dorsal retinal origin, was missing in all
seven cases which could be surveyed; moreover, in three of these animals the
medial branch of the tract on the affected side was obviously larger than normal. This suggests that preferential pathway selection is operative during
development of the VV eye whereby fibres of dorsal retinal origin reach the
tectum via the medial branch. The whole fibre complement of the optic nerve
of the VV eye seems to have selected the medial branch with associated lack of
development of the lateral branch.
This unexpected observation is of interest for two main reasons. Firstly, it
gives some information on the behaviour of optic fibres during early visual
development and this information may open up further modes of analysis,
leading to better understanding of the directional controls operating on growing fibres. Secondly, the observations say something about the nature of the
compound VV eye. Previous analyses of the projections from compound eyes
have run into difficulties over whether the compound retina is to be considered
a regulated system, in terms of cellular specificities (Gaz;s, 1970). On such a
view, each (similar) half of a compound eye is held to have regulated its specificity-structure (in terms of Sperry's (1943, 1944, 1945) hypothesis of neuronal
specificity) such that each half contains cells bearing specificities appropriate
to a whole eye. In this case the reduplicated nature of the compound eye projection to the tectum would be automatically accounted for. However, if each
half of a compound eye contains a complete range of specificities, and if both
fibre path and tectal projection are determined by these specificities, one would
Pathway selection by optic fibres in Xenopus
255
expect the pathway chosen by the fibres to be normal as well as the tectal
projection. Thus if the pathway is not normal, this argues strongly that the
compound eye is not a regulated structure.
Small unmyelinated fibres in the brain cannot be individually visualized after
silver staining. Since the great majority of the optic fibres in Xenopus are small
and unmyelinated, Holmes' silver stain is inadequate to determine whether both
mylinated and unmyelinated optic fibres from a VV eye select preferentially
the medial branch of the optic tract during development. We have therefore
re-examined this question and the present report is based on a new series of
animals with right VV eyes where 3[H]proline was injected into the operated
eye to reveal the optic pathway during development. It is also of interest to
know whether, in regeneration, optic fibres of ventral retinal origin select their
appropriate pathways to the tectum, or choose alternatives, and we have
investigated this matter also. Some aspects of the problems discussed in this
paper have been briefly reported elsewhere (Cook & Horder, 1977; Gazs, 1978).
MATERIALS AND METHODS
Xenopus laevis was used, bred in the laboratory at a controlled temperature
(24 °C) and under a normal diurnal light cycle. After operation animals were
reared to metamorphosis and beyond in large containers, fed on Heinz beef
and vegetable puree during larval life, on Tubifex worms after metamorphosis
and from 3 months after metamorphosis on minced liver.
Microsurgery
Embryos at stages 32-33 (Nieuwkoop & Faber, 1956) were anaesthetized
with 0-1 % MS222 (Tr.ica.ine Methane Sulphonate, Sandoz) and the right eye
was operated on to form a compound ventral (VV) eye. The operation has been
described in an earlier paper (Straznicky et al. 1974). In short, the dorsal half
of the right eye was replaced by a ventral half taken from the left eye of another
animal of similar age. Thus the normal naso-temporal polarity of the operated
eye was retained. Some of the animals underwent a second operation after
metamorphosis which involved section of the right optic nerve close to the
optic chiasma, 3 months before visual field mapping and immediate sacrifice.'
Visual field mapping
Operated animals were used 6-12 months after metamorphosis (6-8 cm body
length) for electrophysiological mapping of the retinotectal projection from the
VV eyes. The technique used for this procedure has been described in previous
papers (Straznicky et al 1974; Straznicky, 1976). Only animals with the visual
field map characteristic of a VV eye (Straznicky et al. 1974) were used for
further autoradiographic studies.
17-2
256
C. STRAZNICKY, R. M. GAZE AND T. J. HORDER
Autoradiography
Twenty-four hours before the animals were killed or mapped 10 fid 3[H]proline (Amersham, specific activity 53 Ci/m-mole) was injected into the posterior
chamber of the operated eye in order to reveal the extent of the retinotectal
projection morphologically. The head of the animal was later fixed in Bouin's
solution. The brain was processed for paraffin embedding, serially sectioned at
10 /an and prepared for autoradiography according to the techniques of
Rogers (1973). Deparaffinized sections were coated with llford Nuclear Emulsion
K2, and were exposed at 5 °C for 2 weeks before being developed. The sections
were then stained with Harris' haematoxylin. Quantitative grain density analysis
in the optic fibre receiving layer of the tectum was carried out by photometric
measurements in reflected light (Rogers, 1973). 10 /*Ci 3[H]proline was injected
into the right eye of five normal animals of 6-8 cm body length which served
as controls. Serial sections were also taken from W eyes stained with haematoxylin and eosin to check for gross abnormalities of the eyes.
Grain density counts
The autoradiographic grain density was estimated along the mediolateral
extent of layer nine (Lazar & Szekely, 1967) at the maximum mediolateral
tectal width. An oil immersion lens and a photometric eyepiece were used.
The photometric system was calibrated to measure reflected light against the
dark field, from no grains (0) to the blackening of a fully exposed K2 emulsion
with maximum grain density (10). The diameter of the grains varies between 1
and 2 /an and they are packed about 1 /an apart. The thickness of the emulsion
is about 3-4 /an, so 400 /on2 may contain a maximum of 1000 grains. The lower
and upper extremes of the calibration may not represent linear relationships
between the readings and the grain counts, but from 1-9 we can equate one
calibration unit with 100 counts. Grain density was measured in reflected light
over a rectangular area of 20 x 20 /an. By moving the rectangle medio-laterally
along layer 9 in steps of 100 /mi, the grain density profile could be established.
Four measurements were made at each step and on four subsequent sections of
the tectum at its maximum medio-lateral extent. Thus one reading in Fig. 3
and 6 is the average of 16 measurements. Background activity was measured
over brain tissue on the sections.
RESULTS
Out of 30 animals investigated, seven operated animals without and six
with optic nerve section have been included in this report (Table 1). Twelve
animals were discarded because of gross abnormalities in the operated eye,
lack of optic nerve formation or failure of regeneration after optic nerve section.
Few visual responses only
No visual responses obtained
through right eye
No visual responses obtained
through right eye
VV map
Incomplete recording
W map
VV map
VV map
VV map
Only the dorsal field present
early regeneration
VV map partly organized
9 MAM
6 MAM
9 MAM
9 MAM
9 MAM
12 MAM
12 MAM
8 MAM
VV6
VV7
VVONC 1
VVONC 2
VVONC 3
VVONC 4
VVONC 5
WONC 6
Contralateral
tectum
Both medial and lateral
branches are present
Both medial and lateral
branches are present
Lateral branch is missing
Lateral branch is missing
^.
Lateral half of the tectum is
not innervated
Lateral
Lateral branch
branch is
is missing
missing
Lateral branch is missing
^*
Lateral branch is missing
0
CD
oX
os
p
^*
i
1
Lateral branch is missing
Ipsilateral
tectum
[3H]proline autoradiography
Lateral branch is missing
Lateral branch is missing
Lateral branch is missing
Lateral branch is missing
Both medial and lateral
branches are present
Lateral branch is missing
Lateral branch is missing
Both medial and lateral
branches are present
Lateral branch is missing
* MAM, months after metamorphosis.
No visual responses obtaned
through VV eye
No visual responses obtained
through VV eye
Few visual responses only
VVmap
VV map
VVmap
Only the dorsal field present
VVmap
6 MAM*
9 MAM
8 MAM
12 MAM
12 MAM
VVl
VV2
VV3
VV4
VV5
Ipsilateral
tectum
Contralateral
tectum
Animal
Age at
sacrifice
Visual field mapping
Table 1. The visual field and retinal projections to the tectum in operated animals without {VV)
or with optic nerve section (VVONC)
258
C. STRAZNICKY, R. M. GAZE AND T. J. HORDER
A
B
Right visual field (normal)
Right visual field (vv)
Fig. 1. Visual field projection from the right eye to the contralateral (left) tectum in
the normal (A) and in a VV eye animal (B). On the tectal diagrams the arrows point
rostrally. The visual field charts are centred on the optic axis of the eye and extend
out 100° radially. S, superior; N, nasal; I, inferior; T, temporal. For each tectal
position the corresponding field position is indicated by the appropriate number in
the visual field. Any tectal position from which no visual response was recorded is
marked with, an open circle. Broken line on the tecta marks the projection of
the horizontal meridian of the visual field. Note that this line in the VV animal (B)
has shifted close to the lateral margin of the tectum.
In five animals a normal visual field projection through the Y V eye was obtained,
indicating that the operated eye had regulated back to a normal state.
(A) The course of the optic fibres from VV eyes during development
The VV eye projections in five animals were similar to the VV visual field
maps obtained earlier (Straznicky et dl. 191 A). Each half retina of the double
ventral eye projected to the whole dorsal surface of the contralateral tectum,
Pathway selection by optic fibres in Xenopus
259
Fig. 2. Bright field photographs of autoradiographs of the contralateral (left)
tectum in a normal (A-C) and in a VV eye (D-F) animals after injection of [3H]proline into the right eye. A and D correspond to the rostral, B and E to the middle
and C and F to the caudal part of the tectum respectively. Arrow in D indicates the
site of the lateral branch of the optic tract which is devoid of silver grains. The
arrows in E and F mark the decreasing silver grain density towards the lateral
margin and the tectum. Scale is the same in each photograph. Bar in F represents
400 fim.
260
C. STRAZNICKY, R. M. GAZE AND T. J. HORDER
1000 n
800 -
600 -
•a
400 -
200 -
0100
500
900
Medio-lateral length of the tectum (/im)
1200
Fig. 3. Grain density counts in normal (N 2) and in VV eye (VV 3) animals along the
maximum mediolateral extent of the contralateral tectum. The background activity
(BG) over the brain tissue is also given. In the VV eye animal there is a steady grain
density decrease laterally. Results giving average + S.E. are based on 16 measurements
each.
that is, most electrode positions on the tectum had two corresponding field
positions, one in the dorsal and one in the ventral half of the visual field,
arranged in rough symmetry about the horizontal meridian (Fig. IB). In
contrast, in normal animals (Fig. 1A) the dorsal field, through the ventral
retina, projects to the rostromedial part of the tectum. In one experimental
animal, only the dorsal visual field (representing the animal's own ventral
retina) projected to the tectum. No noticeable abnormality was detected on the
serial sections of this operated eye; however, it could not be ruled out that the
transplanted retinal half had failed to connect with the brain. In one animal
the visual field mapping was not completed, though a compound ventral field
map was obtained from the rostral half of the tectum.
In normal Xenopus the optic fibres course in the diencephalon towards the
rostrolateral margin of the tectum. At the entrance into the tectum the optic
tract opens out into medial and lateral branches, the positions of which can be
seen at the medial and lateral extremities of the labelled tectum (Fig. 2A-C)
as they course along the medial and lateral margins of the tectum respectively.
The terminal arborization of the optic fibres takes place in the superficial
optic fibre receiving layer (layer 9, according to Lazar & Szekely, 1967) of
the tectum which is well marked by heavy silver grain deposition (Fig. 2B, C).
Pathway selection by optic fibres in Xenopus
261
Although the thickness of layer 9 varies somewhat mediolaterally and rostrocaudally the autoradiographic silver grain density is fairly constant across the
tectum in normal animals (Figs. 3 and 6).
The autoradiograms of serial sections in six of the seven animals with VV
eyes clearly revealed gross abnormalities in the formation of the optic pathway
from the operated eye and in the distribution of W optic fibres in the tectum.
In one animal the pathway appeared normal. A typical example of abnormal
pathway is shown in Fig. 2D-F. Although the diencephalic course of the optic
tract of a VV eye is indistinguishable autoradiographically from normal, at the
level of the rostral tectal pole its separation into medial and lateral branches
cannot be seen. The site of the medial branch is heavily labelled with silver
grains (Fig. 2D) whilst virtually no silver grain deposition is found over the
site of the expected lateral branch. The abnormal distribution of silver grains
indicates that optic fibres entered the tectum via the medial branch only. This
conclusion is supported by the distribution of optic fibres in more caudal
regions of the tectum (Fig. 2E, F). In contrast to the normal situation, the
lateral part of the tectum has a diminishing silver grain density away from the
medial edge (Fig. 2E, F, and Fig. 3). The silver grain density over the lateral
margin of the tectum, which normally carries fibres of dorsal retinal origin,
is so low that it is comparable to the background activity measured over other
parts of the brain. Optic fibres innervating the tectum appear to come from
the medial branch and extend laterally across the tectum as marked by the
decreasing grain density gradient. The autoradiographic observations may be
compared with the results of visual field mapping; both indicate a lateral
spread of the optic fibre projection from VV eyes beyond the region normally
innervated by ventral retina. It is interesting that, although the electrophysiological map extends fairly uniformly across the dorsal surface of the tectum,
the grain distribution is much heavier medially than laterally. The most likely
explanation for this is that the medial grain counts come from terminals and
fibres of passage, whereas more laterally on the tectum the number of fibres
diminishes until eventually there are only arborizations. This difference is
presumably associated with an absolute decrease in the density of innervation
as one passes from medial to lateral on the tectum. The electrophysiological
map, reflecting as it does the position of arborizations, and not their number,
would thus be expected to give a more uniform coverage of dorsal tectum than
the grain counts.
(B) The course of optic fibres from a VV eye after regeneration
The visual projections from the VV eyes were recorded in six animals 3
months after optic nerve section. In four animals the visual projection was
almost fully restored and, apart from small irregularities, they were similar to
W maps without optic nerve cut (Fig. 4). In one animal visual responses were
obtained from the dorsal field only and field points were not retinotopically
262
C. STRAZNICKY, R. M. GAZE AND T. U. HORDER
200/im
Contralateral
Ipsilateral
Right visual field (vv)
Fig. 4. Field projection from right eye (VV) to both contia- and ipsilateral optic
tecta after regeneration of the optic nerve in animal VVONC 3. The contralateral
projection represents a fairly good restoration of the' normal' VV map. Even though
optic fibres were present over the medial part of the ipsilateral tectum (Fig. 5D-F)
only a few responses were obtainable from it.
organized. In another a reduplicated map was obtained but the stimulus points
were partly organized only in the dorso-ventral direction in the visual field.
Although in each animal the ipsilateral tectum was also investigated, only two
cases yielded a few visual responses through the operated eye (Fig. 4). This
observation is in agreement with recent findings that an autoradiographically
demonstrable aberrant ipsilateral retinotectal projection may exist in the absence
of electrophysiological responses; the electrophysiological projection, for
reasons that are unclear, may appear much more slowly than the contralateral
projection (Gaze & Keating, 1970; Udin, 1977; Glastonbury & Straznicky,
1978).
Pathway selection by optic fibres in Xenopus
Ipsilateral
Contralateral
Ipsilateral
•
263
Contralteral
#
A
Fig. 5. Bright field photographs of autoradiographs of the optic tecta in VV eye animals with optic nerve cut. Photographs A, B and C were taken from animal VVONC
1, and D, E and F from animal VVONC 3. In VVONC 1 (A-C) both medial and
lateral branches of the contralateral optic tract are filled with silver grains. In
contrast, on the ipsilateral side, only the medial branch is heavily filled with silver
grains. In animal VVONC 3 (D-F) the lateral branch on both the contralateral
and ipsilateral side is missing. The extent of the optic fibre projection on the
ipsilateral side is marked by an arrow with an open circle. The scale is the same
in each photograph. Bar in F represents 500/tm.
264
C. STRAZNICKY, R. M. GAZE AND T. J. HORDER
1000 i
VVONC3
800-
600-
400 -
200 -
0
J
300
700
Medio-lateral length of the tectum (Aim)
1100
Fig. 6. Grain density counts in animal VVONC 3 along the maximum medio-lateral
extent of the tectum. For comparison the measurements on a normal animal (N 1)
and the background activity (BG) are also given. Notice that on the ipsilateral side
there is an abrupt decrease of silver grain density in the lateral third of the tectum.
Results giving average ±S.E. are based on 16 measurements each. To avoid overcrowding the S.E. is not given for N 1.
Autoradiography, after regeneration of the cut optic nerve, showed that
both the contralateral and ipsilateral tecta had become innervated in all cases
(Fig. 5). The regenerated optic fibres occupied layer 9 in the tectum. In three
animals silver grains on the contralateral side were present only over the medial
branch of the tract. In the other three cases the sites of both medial and lateral
branches were filled with silver grains. Examples of each type of results are
given in Fig. 5 A-F, obtained on animals VVONC 1 and W O N C 3. Thus during
regeneration optic fibres of ventral origin may take an alternative entry route
into the tectum. In contrast to the rather variable findings on the contralateral
side, in all six animals regenerated optic fibres were more or less confined to
the medial branch on the ipsilateral side (Fig. 5 A-F). It is worth mentioning
that the diencephalic course of regenerated fibres was abnormal both contralaterally and ipsilaterally, in that they formed a narrow band following the
lateral margin of the diencephalon instead of being spread out in a wide fascicle
between the thalamic gray and the lateral margin as in the normal contralateral
pathway. This abnormal diencephalic course of optic fibres is a consistent
finding when optic fibres regenerate in young Xenopus (Gaze & Grant, 1978).
Since, at the usual point of the division of the optic tract, fibres from the VV
Pathway selection by optic fibres in Xenopus
265
retina have chosen the medial branch we must assume that some instructional
cues are operative here. Apart from one case the autoradiograms showed that
the optic fibre projections both on the contralateral and on the ipsilateral sides
extend rostrocaudally and mediolaterally across the whole tectum, with a
decreasing density from medial to lateral. Unexpectedly, it was found that in
one animal only the medial half of the ipsilateral tectum had been innervated,
with a sharp cut-off at mid-tectum. The autoradiograms clearly demonstrated
in this case an abrupt decrease of silver grain density approximately along the
midline of the tectum, marking the extent of ipsilateral retinal projection
(Fig. 5D-F). The measurements on this brain have corroborated this observation in that the silver grain density in the lateral half of the tectum is near
to the level of background activity, suggesting very poor if any innervation
from the operated eye (Fig. 6). In contrast, the contralateral tectum showed a
steady decrease of silver grain density in the lateral direction. In Fig. 5E the
ipsilateral tectum (left in photograph) shows a vertical banding pattern comparable to those previously described by Levine & Jacobson (1975) in goldfish
where both eyes had been induced to innervate one tectum.
DISCUSSION
The main finding of this study is that during development fibres from a VV
retina select preferentially the medial branch of the tract for entry into the
tectum. The embryonic eye operation involved the fusion of two ventral halves
along the horizontal midline of the eye. Any departure from the intention of the
operation might have incorporated portions of the dorsal eye blastema in the
VV eye, thus including fibres appropriate to the lateral branch of the tract.
This may have happened in the one reported case where, even though the eye
was shown to be VV by the nature of its tectal map, the presence of a lateral
branch was demonstrated.
To account for the establishment both of the pathway that the optic fibres
take from the eye to the tectum, and of their systematic termination in the
tectum, Attardi & Sperry (1963) suggested that fibres encounter a number of
choice-points. On the basis of guidance cues provided by the micro-environment
of the brain in which the retinal fibres progress to their targets, fibres select
their appropriate routes and avoid alternatives. Probable choice-points include
the optic chiasma, the bifurcation of the optic tract and fibre bundles leading
from the tract towards the particular zones of terminal arborization of the
retinal fibres in the tectum. It has been reported that VV eyes usually have two
ventral fissures and occasionally two nerve heads (Straznicky et al. 1974). The
normal W projection is such that central retina projects laterally while the
dorsal and ventral extremities of the retina project medially. In a sense, therefore, the retinal projection is folded about the horizontal meridian. If the
arrangement of fibres in the optic pathway also shows this dorsoventral folding
266
C. STRAZNICKY, R. M. GAZE AND T. J. HORDER
and if the folded fibre projection occupies a normal extent of the diencephalon,
we might expect fibres from central retina, the region of the fold, to occupy
the lateral branch of the optic tract, provided that mechanisms other than chemoselection are at work. The autoradiographic evidence is against this idea, and
suggests that some form of selective interaction between the fibres and their
substrate is involved in the choice of branches. It would also follow from this,
as argued in the introduction, that the YV eye is not a 'regulated' system,
where each half of the eye contains a complete distribution of cellular affinity
markers. If it were, we would expect each half retina to send fibres throughout
both branches of the tract.
The picture in regeneration is less clear. Where regenerating fibres from a VV
retina pass up the .ipsilateral side of the brain they follow the medial branch
for entry into the tectum. On the other hand, on the contralateral side, where
the original formation of the branches is abnormal, and where the fibre pathway
has degenerated, regenerating fibres from the same eye may use both medial
and lateral routes. We do not known why regenerating ventral fibres sometimes
enter the contralateral tectum via both branches, nor do we know what proportion of fibres behave in this fashion.
One attractive possibility, to account for the differences between ipsilateral
and contralateral pathway selection, is that fibres may like to follow, and grow
alongside, similar fibres. In this case the fibres growing ipsilaterally will meet
up with, and accompany, ventral retinal fibres from the other (normal) eye.
Fibres from the VV eye passing ipsilaterally will therefore be expected to choose
the medial branch of the optic tract, along with the normal fibres. On the
contralateral side, however, there are no fibres to follow in either path, since
the lateral branch did not develop and the fibres that were originally in the
medial branch have degenerated following nerve section.
If fibre-following of this sort is important, the results suggest that there are
differences between first development and regeneration. In first development
it is the rule for VV fibres to pass via the medial branch of the tract but in
regeneration this seems not to be so. The difference between the two situations
could be less dramatic than the appearance, since much could depend on the
initial path chosen by the very earliest fibres, and the environment through
which fibres are growing is quite different in the embryo and in the adult.
These results show that, in half the experimental cases, pathway selection
comparable to that during development occurred in regeneration. In the other
half pathway selection was not obvious. Electrophysiological evidence from the
frog (Udin, 1978) and the goldfish (Horder, 1974) shows that, with normal
eyes, some regenerating optic fibres may enter the tectum by the wrong route.
It is not, however, possible by electrophysiological methods, to determine
how many fibres behave in this fashion. The present experiments, though far
from quantitative, provide morphological evidence that a considerable number
of regenerating fibres can take the wrong pathway.
Pathway selection by optic fibres in Xenopus
267
This work has been supported by a grant by the Australian Research Grants Committee.
The authors gratefully acknowledge the skilled histological assistance of Mrs Teresa Clark
and Mrs June Colville.
REFERENCES
ATTARDI:, D. G. & SPERRY, R. W. (1963). Preferential selection of central pathways by
regenerating optic fibres. Expl Neurol. 7, 46-64.
COOK, J. E. & HORDER, T. J. (1977). The multiple factors determining retinotopic order in the
growth of optic fibres into the optic tectum. Phil. Trans. R. Soc. Lond. B 278, 261-276.
GAZE, R. M. (1970). The Formation of Nerve Connections. London: Academic Press.
GAZE, R. M. (1978). The problem of specificity in the formation of nerve connections. In
Specificity of Embryological Interactions {Receptors and Recognition, Series B, vol. 4)
(ed. D. R. Gairod), pp. 53-93. London: Chapman and Hall.
GAZE, R. M. & GRANT, P. (1978). The diencephalic course of regenerating retinotectal
fibres in Xenopus tadpoles. /. Embryo I. exp. Morph. 44, 201-216.
GAZE, R. M. & JACOBSON, M. (1963). A study of the retinotectal projection during regeneration of the optic nerve in the frog. Proc. R. Soc. B 157, 420-448.
GAZE, R. M. & KEATING, M. J. (1970). Further studies on the restoration of the contralateral
retinotectal projection following regeneration of the optic nerve in the frog. Brain Res. 21,
183-195.
GAZE, R. M., JACOBSON, M. & SZEKELY, G. (1963). The retinotectal projection in Xenopus
with compound eyes: /. Physiol., Lond., 165, 484-499.
GLASTONBURY, J. & STRAZNICKY, K. (1978). Aberrant ipsilateral retinotectal projection
following optic nerve section in Xenopus. Neurosci. Letts. 7, 67-72.
HORDER, T. J. (1974). Changes offibrepathways in the goldfish optic tract following regeneration. Brain Res. 72, 41-52.
LAZAR, G. & SZEKELY, G. (1967). Golgi studies on the optic center of the frog. /. Hirnforsch.
9, 329-344.
LEVINE, R. L., & JACOBSON, M. (1975). Discontinuous mapping of retina onto tectum innervated by both eyes. Brain Res. 98, 172-176.
NIEUWKOOP, P. D. & FABER, J. (1956). Normal Table o/Xenopus laevis (Daudin). Amsterdam:
North Holland Publs.
ROGERS, A. W. (1973). Techniques of Autoradiography, 2nd ed. Amsterdam: Elsevier.
SCALTA, F. & FITE, K. (1974). A retinotopic analysis of the central connections of the optic
nerve in the frog. J. comp. Neurol. 158, 455-478.
SPERRY, R. W. (1943). Visuomotor co-ordination in the newt (Triturus viridescens) after
regeneration of the optic nerve. /. comp. Neurol. 79, 33-55.
SPERRY, R. W. (1944). Optic nerve regeneration with return of vision in anurans. /. Neurophysiol. 7, 57-70.
SPERRY, R. W. (1945). Restoration of vision after crossing of optic nerves and after contralateral transplantation of eye. /. Neurophysiol. 8, 15-28.
STRAZNICKY, K. (1976). Reorganisation of retinotectal projection of compound eyes after
various tectal lesions in Xenopus. J. Embryol. exp. Morph. 35, 41-57.
STRAZNFCKY, K., GAZE, R. M. & KEATTNG, M. J. (1974). The retinotectal projection from a
double ventral compound eye in Xenopus laevis. J. Embryol. exp. Morph. 31, 123-137.
UDIN, S. B. (1977). Rearrangements of the retinotectal projection in Rana pipiens after
unilateral caudal half-tectum ablations. /. comp. Neurol. 173, 561-582.
UDIN, S. B. (1978). Permanent disorganisation of the regenerating optic tract in the frog.
Expl Neurol. 58, 455-470.
{Received 4 October 1978, revised 8 December 1978)