The retinotectal projections from double-nasal (NN), double-temporal (TT) and double-ventral (VV) compound eyes in Xenopus were studied autoradiographically and electrophysiologically during development. Early TT projections were confined to rostrolateral tectum and spread with advancing age to cover most of the tectum by shortly after metamorphosis. Early VV projections showed a decreased density of label on lateral tectum. Early NN projections appeared to extend across the entire rostrocaudal length of the available tectum at all stages of development, but showed a decrease in label density on rostral tectum. The results are discussed in relation to various hypotheses about the formation of retinotectal connexions.

‘Compound eyes’ may be formed in Xenopus embryos by surgical operation to fuse two eye fragments in the same orbit. Such an operation is usually performed at stage 32 (Nieuwkoop & Faber, 1967); that is, at tail-bud stage, after the eye anlage has become spatially determined in terms of the orientation of the map it will eventually form, but before any optic fibres have yet passed from the eye to the tectum. If the temporal half of an eye anlage is removed and replaced by a nasal half, taken from an eye of opposite laterality, the result is a double-nasal (NN) compound eye. In a comparable fashion doubletemporal (TT) and double-ventral (VV) eyes can be made.

If an animal with an NN, TT or VV eye is allowed to develop until after metamorphosis and the retinotectal projection from the compound eye is then examined electrophysiologically, it is found, as expected, that each such eye gives a reduplicated projection to the tectum. The orientation of the tectal projection from each half of the compound eye is appropriate to the embryonic origin of that half-retina. However each (similar) half of the compound eye spreads its connexions across most of the tectum, instead of confining its projection to the appropriate half of the tectum, as would such a half-retina in a normal animal (Gaze, Jacobson & Székely, 1963; Straznicky, Gaze & Keating, 1974). The orderly spreading of the projection from a half-retina (comprising half of a compound eye) raises interesting questions about the developmental relationships between the eye and the tectum.

In a normal animal the entire retina maps, in a continuous fashion, over the entire tectal surface. The development of a normal retinotectal projection, as well as its restoration when the optic nerve is cut and allowed to regenerate, is usually ascribed to the existence of specific chemical affinities between recognition markers on retinal and tectal cells (Sperry, 1943, 1944, 1945, 1951, 1963, 1965). The spreading of the projection from each half-retina of a compound eye presents problems for such a mechanism. Difficulties arise because there are several possible interpretations of the compound-eye results.

One possibility would be for the compound-eye operation to be followed by regulation of the specificity-structure, or distribution of markers in the eye, such that each half of the compound eye acquires a full set of cell markers. This could account for the spreading of the projection. We have recently furnished strong evidence, however, that this is not the case. Each component half of a compound eye, when allowed to regenerate fibres to the ipsilateral tectum, behaves as if it possesses only the markers characteristic of that half-eye (Gaze & Straznicky, 1979, 1980; Straznicky, Gaze & Horder, 1979).

Another possible explanation for the spreading of the compound eye projection, is that only that area of tectum develops which is innervated by optic fibres (Sperry, 1965). Thus fibres from an NN eye may permit only caudal tectum to develop, since this is where nasal fibres normally go; fibres from a TT eye may permit only rostral tectum to develop, and fibres from a VV eye may permit only medial tectum to develop. If this were to happen, we might expect the resulting tectum to be approximately normal in size, even though derived from only half the original tectal structure, since the optic input comprises fibres from two (similar) half-retinae, rather than just one. We could call this the ‘overgrown half-tectum’ hypothesis.

A further possibility is that recognition by optic axons of localized tectal markers may not be involved in the initial establishment of the visual projection. It could be that some other mechanism is responsible for the first setting up of the retinotectal map (Hope, Hammond & Gaze, 1976; Willshaw & von der Malsburg, 1979) and the tectal markers (which we know to exist in later life, since they are involved in nerve regeneration, and which could either be cellular or related to fibre debris) are then placed on the tectum by the optic nerve fibres themselves.

Any adequate theory that seeks to account for the formation of topographically ordered retinotectal connexions must accommodate the orderly connectivity found in the projections from compound eyes. More information might be provided about the elaboration of ordered maps from such eyes by observing the development sequence through which the projection passes on its way to the adult state. We have previously shown, in normal animals, that because of the different patterns of growth of the retina and tectum (Straznicky & Gaze, 1971, 1972), the normal development of the retinotectal projection requires a continuous rearrangement of functional connexions as the system grows (Gaze, Chung & Keating, 1972; Gaze, Keating & Chung, 1974; Gaze, Keating, Ostberg & Chung, 1979). Thus the parts of the tectum with which the various regions of the retina connect are not necessarily the same in the young larva as in the post-metamorphic animal. Throughout development the general orientation of the retinotectal map remains the same but the specific retinotectal connexions, which underlie the map, change. We felt, therefore, that it would be useful to observe the way in which the projections from compound eyes responded to this developmental situation.

The present results show that, contrary to the electrophysiological evidence from older animals, during tadpole life the projection from each type of compound eye tends to restrict itself to a different part of the available tectum. Thus TT projections are found mainly rostrally on the tectum and VV projections mainly medially, while NN projections extend over the whole available tectum but are best developed caudally.

In the discussion we consider the possible significance of these results for theories of retinotectal connectivity. An abstract of some of this work has been published elsewhere (Straznicky, Gaze & Keating, 1979).

Laboratory-bred Xenopus laevis at various developmental stages were used in this study.

Microsurgery

The right eye of stage-32 to -33 embryos (Nieuwkoop & Faber, 1967) was operated on according to previous descriptions (Gaze & Straznicky, 1980) to obtain an NN eye. In other embryos right TT or W eyes were formed. Animals with right NN, TT or VV eyes were reared separately. Between stage 50 and 8 weeks after metamorphosis the contralateral retinotectal projections from the right operated and left intact eyes were assayed by injecting [3H] proline (3HP) into both eyes. In some animals the visuotectal projection from the operated eye was also mapped electrophysiologically.

Histology

Twenty-four hours before sacrifice 3HP (Radio Chemical Centre, Amersham: specific activity 24 Ci/m-mole) was injected into the posterior chamber of the right compound eye and left intact eye, under anaesthesia with MS222 (Tricaine Methane Sulphonate, Sandoz; 1:1500). One μCi of isotope was given to the animals up to stage 54, 2 μCi from stage 54 to stage 60, 3 μCi from stage 60 to stage 66 (metamorphosis) and 5 μCi in post-metamorphic animals. The head of the animal was fixed in Bouin’s solution, the dissected brain was embedded in paraffin, serially sectioned at 10 μm and mounted on slides. The slides were coated with Ilford K2 emulsion, exposed in light-tight boxes at 4°C for 14 days, developed in Kodak Dektol and counterstained with Harris’s hemotoxylin. Reconstruction of the brains, based on camera-lucida drawings of every fifth section, were made to assess the extent of the projection from the compound and intact eye. The proline distributions were estimated by visual inspection of the sections.

Electrophysiology

Several animals of varying age with right NN, TT or VV eyes were recorded before sacrifice in order to ascertain that the eye operation was successful. Visuotectal mapping procedures were similar to those described previously (Straznicky, Gaze & Keating, 1971; Gaze et al. 1974) thus only a short account is given here. Animals were anaesthetized with MS222. The tectum was exposed, photographed and the animal was set up in a small Perspex globe filled with oxygenated anaesthetic solution (approximately 0·01% MS222) at the centre of an Aimark projection perimeter with the right compound eye centred and the left eye covered.

Visually evoked action potentials were recorded by a Wood’s metal electrode from pre-determined tectal positions and the corresponding receptive fields located. The visuotectal projection from the operated eye was estimated by sampling about 30 tectal recording points.

The results are based on study of 25 animals with one NN eye each, 25 animals with one TT eye each and 18 animals with one VV eye each. The compound eyes were the same size as the normal eyes at the time of autoradiographic analysis, which was performed on all experimental animals. In many cases the tectum contralateral to the compound eye was smaller than that contralateral to the normal eye (Figs 1, 4b,5, 8 and 11) but this was not consistently so. Electrophysiological recording of the visuotectal projection from the compound eye was made in 5 animals with one NN eye, 16 animals with one TT eye and 5 animals with one VV eye, to confirm electrophysiologically that the compound eye construction had been satisfactory. A larger number of TT animals was mapped because of the particular adjustments in retinotectal relationships required by the development of the projection from this kind of eye (see below).

Fig. 1

Reconstruction of the retinotectal projections from the normal and the TT compound eye in animals at various stages of development. Tritiated proline was injected into each eye and later serial transverse sections through the tecta were autoradiographed. The outline of every fifth section was drawn with the aid of a camera lucida and the mediolateral extent of the tectum and of the labelled region, as estimated by visual inspection, was measured with a map measurer. The resulting measurements were marked on graph paper and the overall outlines joined up. Thus each diagram represents a flattened view of the tecta seen from dorsally. The mediolateral extent of each diagram is thus a measured dimension (bar = 1 mm) while the rostrocaudal dimension is arbitrary.

In each case the right eye was compound and projected to the left tectum. The left eye was normal and projected to the right tectum.

In each diagram rostral is indicated by the arrowhead. The left tectum, receiving the projection from the compound eye, is to the right in each case.

It may be seen that the TT projection was initially restricted to the rostrolateral tectum and that complete tectal coverage has not been reached even by 8 WAM.

Fig. 1

Reconstruction of the retinotectal projections from the normal and the TT compound eye in animals at various stages of development. Tritiated proline was injected into each eye and later serial transverse sections through the tecta were autoradiographed. The outline of every fifth section was drawn with the aid of a camera lucida and the mediolateral extent of the tectum and of the labelled region, as estimated by visual inspection, was measured with a map measurer. The resulting measurements were marked on graph paper and the overall outlines joined up. Thus each diagram represents a flattened view of the tecta seen from dorsally. The mediolateral extent of each diagram is thus a measured dimension (bar = 1 mm) while the rostrocaudal dimension is arbitrary.

In each case the right eye was compound and projected to the left tectum. The left eye was normal and projected to the right tectum.

In each diagram rostral is indicated by the arrowhead. The left tectum, receiving the projection from the compound eye, is to the right in each case.

It may be seen that the TT projection was initially restricted to the rostrolateral tectum and that complete tectal coverage has not been reached even by 8 WAM.

(a) TT projections

In 22 of the 25 animals with one TT eye, the autoradiographic projections from the eyes were studied at different developmental stages between stage 50 and 8 weeks after metamorphosis (Table 1). The projection from the normal eye to its contralateral tectum occupied the greater part of the available tectum, with a small deficit at the caudomedial margin during larval stages 50 to 59, by which time complete coverage of the tectum was reached. At all developmental stages the contralateral projection from the TT eye occupied less tectum than did that from the control eye. At the earliest developmental stages examined, the projection from the TT eye was restricted to a relatively small area of rostrolateral tectum. With development the projection spread caudomedially across greater areas of the tectal surface but even by 8 weeks after metamorphosis there was still no projection from a TT eye to the most caudomedial tectal sector (Fig. 1). Representative dark-field autoradiographs from an 8-week post-metamorphic animal with one TT eye are shown in Fig. 2. It may be seen that, compared to the projection from the normal eye, there is a relative deficit in the TT projection to caudomedial tectum.

Table 1
graphic
graphic
Fig. 2

Dark-field autoradiographs of transverse sections through rostral (top), mid-tectal (middle) and caudal tectum (bottom) in an animal with a TT eye. The animal was TT24, 8 weeks after metamorphosis. Bar = 400μm. The left tectum, contralateral to the TT eye, is to the right. It may be seen that the TT projection is dense rostrally and falls off medially and caudally.

Fig. 2

Dark-field autoradiographs of transverse sections through rostral (top), mid-tectal (middle) and caudal tectum (bottom) in an animal with a TT eye. The animal was TT24, 8 weeks after metamorphosis. Bar = 400μm. The left tectum, contralateral to the TT eye, is to the right. It may be seen that the TT projection is dense rostrally and falls off medially and caudally.

Electrophysiological mapping of the projection through the TT eye was carried out in 16 animals. Fifteen yielded projections characteristic of TT eyes, while in the remaining animal only the host temporal retina had projected to the tectum. A map obtained from a stage-59 animal, together with the histological reconstruction of the autoradiographic projection, is shown in Figure 3. The visuotectal projection shows the whole visual field of the TT eye projecting, with the reduplication and orientation appropriate to a TT eye, to rostrolateral tectum. The autoradiographic reconstruction (Fig. 3 b) confirms that the TT projection is limited to rostrolateral tectum. The map and autoradiography from a TT animal at 8 weeks after metamorphosis are illustrated in Fig. 4. Both the map and the autoradiography demonstrate that by this stage considerably larger tectal areas are covered than at stage 59, but the autoradiography indicates that there is still a significant deficit in caudo-medial tectum.

Fig. 3

(a) Visuotectal map from an animal with a TT eye (TT6, stage 59). The map shows the reduplication of field positions characteristic of TT eyes. In this and each of the other visuotectal maps shown the upper diagram is of the dorsal surface of the left tectum, with the heavy black arrow projecting rostrally. Numbered recording positions and dots on the tectum correspond to positions in the chart of the visual field below. This chart extends out from the centre of the field for 100° and the animal is to be thought of as sitting behind the chart with his right eye looking out at the observer through the centre of the chart. N, nasal; T, temporal; D, dorsal; V, ventral. The large open arrows are to indicate orientation of the field map in relation to the tectum. (b) Reconstruction of the retinotectal projections in this animal. Bar = 1 mm (mediolateral). The left tectum (right in the Figure), which receives the projection from the TT eye, shows label restricted to the rostrolateral region.

Fig. 3

(a) Visuotectal map from an animal with a TT eye (TT6, stage 59). The map shows the reduplication of field positions characteristic of TT eyes. In this and each of the other visuotectal maps shown the upper diagram is of the dorsal surface of the left tectum, with the heavy black arrow projecting rostrally. Numbered recording positions and dots on the tectum correspond to positions in the chart of the visual field below. This chart extends out from the centre of the field for 100° and the animal is to be thought of as sitting behind the chart with his right eye looking out at the observer through the centre of the chart. N, nasal; T, temporal; D, dorsal; V, ventral. The large open arrows are to indicate orientation of the field map in relation to the tectum. (b) Reconstruction of the retinotectal projections in this animal. Bar = 1 mm (mediolateral). The left tectum (right in the Figure), which receives the projection from the TT eye, shows label restricted to the rostrolateral region.

Fig. 4

(a) Visuotectal map from an animal with a TT eye (TT23, 8 WAM). Conventions as in Fig. 3. The map shows more extensive tectal coverage than that of Fig. 3. The field positions show the reduplication characteristic of TT eyes. (b) Reconstruction of the retinotectal projections in this animal. Bar = 1 mm (mediolateral). The left tectum (right in the Figure), which receives the projection from the TT eye, still shows a caudomedial deficit at 8 WAM.

Fig. 4

(a) Visuotectal map from an animal with a TT eye (TT23, 8 WAM). Conventions as in Fig. 3. The map shows more extensive tectal coverage than that of Fig. 3. The field positions show the reduplication characteristic of TT eyes. (b) Reconstruction of the retinotectal projections in this animal. Bar = 1 mm (mediolateral). The left tectum (right in the Figure), which receives the projection from the TT eye, still shows a caudomedial deficit at 8 WAM.

(b) VV projections

Successful autoradiography was obtained in 15 of the 18 animals with one normal and one VV eye, the extent of the projection being examined at various developmental stages between stage 50 of larval life and 2 weeks after metamorphosis (Table 1). In two animals the label from the compound eye was restricted to the central region of the contralateral tectum, an exceptional result. In the other 13 animals, label from the VV eye was concentrated medially on the tectum with a relative deficit in the projection to lateral tectum. In contrast, the label from the normal eye covered its own contralateral tectum completely from stage 58. The projection deficit from the VV eye was more marked in the younger animals but was still present at 2 weeks after metamorphosis (Fig. 5), although in the older animals the lateral tectal deficit showed only as a localized decrease in density of label towards the lateral edge of the contralateral tectum (Fig. 6).

Fig. 5

Reconstructions from the retinotectal projections from the normal and VV compound eye in animals at various stages of development. Conventions are as in Fig. 1 except that interrupted lines on the tecta indicate diminished density of projection. Bar = 1 mm (mediolateral).

Fig. 5

Reconstructions from the retinotectal projections from the normal and VV compound eye in animals at various stages of development. Conventions are as in Fig. 1 except that interrupted lines on the tecta indicate diminished density of projection. Bar = 1 mm (mediolateral).

Fig. 6

Photomicrograph of a transverse section at the mid-tectal level in an animal with a VV eye (VV14, stage 66). Bar = 300μm. The left tectum, receiving the projection from the VV eye, is on the right. It may be seen that this projection is most dense towards the midline and tails off toward the lateral edge of the tectum.

Fig. 6

Photomicrograph of a transverse section at the mid-tectal level in an animal with a VV eye (VV14, stage 66). Bar = 300μm. The left tectum, receiving the projection from the VV eye, is on the right. It may be seen that this projection is most dense towards the midline and tails off toward the lateral edge of the tectum.

Electrophysiological mapping of the visuotectal projection through the VV eye was carried out in five animals. All five maps displayed dorsoventrally reduplicated field positions characteristic of VV eyes. The best VV map obtained (Fig. 7) was from a stage-66 animal and shows the type of field reduplication characteristic of VV eyes (Straznicky et al. 1974). Unfortunately in this case the histology was not available. The other VV maps were compound but poorly organized, for reasons that are not understood.

Fig. 7

Visuotectal map from an animal with a VV eye (VV16, stage 66). Conventions as in Fig. 3. The map shows reduplication of field positions characteristic of VV eyes.

Fig. 7

Visuotectal map from an animal with a VV eye (VV16, stage 66). Conventions as in Fig. 3. The map shows reduplication of field positions characteristic of VV eyes.

(c) NN projections

Autoradiographic analysis of the retinotectal projections was obtained in 25 animals with one normal and one NN eye, at various developmental stages from stage 50 to 6 weeks after metamorphosis (Table 1).

The extent of tectal coverage by the projection from the NN eye, in contrast to that by the projections from TT and VV eyes, was essentially similar to the tectal coverage by the projection from the normal eye (Fig. 8). In 21 of the 25 animals, however, it was noted that the autoradiography density of the projection from the NN eye was reduced in the rostral tectum. In the other four animals the NN projection did not show this decreased density rostrally. In contrast to the majority result from the NN eyes, the projection from the normal eye showed, in all but two cases, no reduced density of label of tectum rostrally. The reduction in the density of the rostral tectal label from NN eyes was seen at all developmental stages studied and is illustrated for a stage-50 projection in Fig. 9, and for a stage-61 projection in Fig. 10.

Fig. 8

(a, b) Reconstructions from retinotectal projections from the normal and the NN compound eye in animals at various stages of development. Conventions are as in Fig. 1. Interrupted lines on the tectum indicate decreased density of label. It may be seen that, in most cases, the NN projection shows decreased density rostrally. Bar = 1 mm (mediolateral).

Fig. 8

(a, b) Reconstructions from retinotectal projections from the normal and the NN compound eye in animals at various stages of development. Conventions are as in Fig. 1. Interrupted lines on the tectum indicate decreased density of label. It may be seen that, in most cases, the NN projection shows decreased density rostrally. Bar = 1 mm (mediolateral).

Fig. 9

Photomicrographs of transverse sections through rostral (upper) and mid-tectal (lower) regions in an animal with an NN eye (NNX, stage 50). Bar = 100 μ m. The left tectum, receiving the projection from the NN eye, is to the right. It may be seen that there is some diminution in the density of label rostrally on this tectum.

Fig. 9

Photomicrographs of transverse sections through rostral (upper) and mid-tectal (lower) regions in an animal with an NN eye (NNX, stage 50). Bar = 100 μ m. The left tectum, receiving the projection from the NN eye, is to the right. It may be seen that there is some diminution in the density of label rostrally on this tectum.

Fig. 10

Photomicrographs of rostral (upper), mid-tectal (middle) and caudal (lower) tectum in an animal with an NN compound eye (NN14, stage 61). Bar = 300 μm. The left tectum, receiving the projection from the compound eye, is to the right. It may be seen that this projection is less dense rostrally than caudally.

Fig. 10

Photomicrographs of rostral (upper), mid-tectal (middle) and caudal (lower) tectum in an animal with an NN compound eye (NN14, stage 61). Bar = 300 μm. The left tectum, receiving the projection from the compound eye, is to the right. It may be seen that this projection is less dense rostrally than caudally.

The visuotectal projection through the NN eye was mapped electrophysiologically in five animals and all five yielded field maps characteristic of NN eyes. One such map, from a stage-62 animal, is shown in Fig. 11 a and the autoradiographic reconstruction in Fig. 11 b.

Fig. 11

(a) Visuotectal map from an animal with an NN eye (NN17, stage 62), showing the reduplication of field positions characteristic of NN projections. Conventions as in Figure 3. (6) Reconstruction of the retinotectal projection in this animal. Both tecta are completely covered by label and there is here no decrease in density rostrally on the left tectum (right in Figure). Bar = 1 mm (mediolateral). Arrowhead points rostrally.

Fig. 11

(a) Visuotectal map from an animal with an NN eye (NN17, stage 62), showing the reduplication of field positions characteristic of NN projections. Conventions as in Figure 3. (6) Reconstruction of the retinotectal projection in this animal. Both tecta are completely covered by label and there is here no decrease in density rostrally on the left tectum (right in Figure). Bar = 1 mm (mediolateral). Arrowhead points rostrally.

These experiments were intended to reveal something of the mechanisms by which ordered retinotectal projections are established. The results show that TT projections initially innervate rostrolateral tectum, VV projections initially innervate medial tectum and NN projections initially innervate the more caudal tectal areas most densely. At first sight this differential distribution of projections from different types of compound eye seems to provide support for the chemoaffinity theory, which suggests that the normal adult projection arises as a result of selective affinities of temporal retina for rostral tectum, ventral retina for medial tectum, nasal retina for caudal tectum and dorsal retina for lateral tectum. In fact, however, such an interpretation makes no allowance for the changes that occur in the retinotectal projection with the passage of time, either during the development of the normal projection or of that from a compound eye. We consider first these changes and then return to examine alternative mechanisms for the differential behaviour of the compound eye projections.

In normal animals the retina grows throughout larval life and into post-metamorphic stages by the accretion of cells at the ciliary margin (Straznicky & Gaze, 1971; Jacobson, 1976; Gaze et al. 1979). Tectal histogenesis, on the other hand, occurs in a curvilinear fashion from rostrolateral to caudomedial tectum and appears essentially complete by stage 58 of larval life (Straznicky & Gaze, 1972). Analysis of the way in which the normal retinotectal projection develops indicates that while the general orientation of the growing map on the growing tectum is preserved, the individual retinotectal connexions underlying the map undergo continuous changes (Gaze et al. 1972; Gaze et al. 1974; Gaze et al. 1979). Thus to account for the development of the normal retinotectal projection we require a mechanism which will preserve the four-dimensional topology of the system while allowing the individual neural elements to vary their interconnexions as development proceeds.

An even more remarkable shift occurs during maturation of the projections from TT and VVeyes. This point may be illustrated, in relation to TT projections, by considering for example Figs. 3 and 4. The topography of a TT map is such that the vertical meridian of visual field, and hence of the retina, projects to the most caudal part of the innervated tectum. Figures 1, 3 and 4 demonstrate, however, that the particular tectal area which constitutes the caudal margin of the projection changes markedly with development. Since neurogenesis in compound eyes, as in normal eyes, occurs at the retinal ciliary margin (Feldman & Gaze, 1972); Straznicky & Tay, 1977) central retina is the oldest part of the retina. This portion of the retina in TT eyes projects during mid-larval life to rostral tectum but by 8 weeks after metamorphosis the same retinal area is projecting to caudal tectum, while the rostral tectal area previously receiving input from central retina now receives from more peripheral retina. Similarly, newly appearing neurons at the nasal and temporal retinal poles project to the rostral tectal margin which earlier received input from older retinal neurons, while the tectal connexions of these older neurons are displaced caudally on the tectum. The maturation of the TT map thus involves fairly radical re-adjustment in individual synaptic relationships between retinal and tectal elements, but these changes are of the same sense, although quantitatively greater, than those that occur in the maturation of the normal map.

In fact, although we cannot make the point so strongly because our electrophysiological maps from VV eyes were poor, consideration of the information about the development of the VV maps leads to similar conclusions. At the earlier larval stages the projection from the retina, which in the adult will form central retina, is to the medial aspect of the tectum. After metamorphosis this retina projects to much more lateral tectum while newly grown dorsal and ventral retina projects now to medial tectum.

With NN eyes the situation is somewhat different. The topography of the NN projection is such that central (oldest) retina projects to rostral tectum while the later appearing nasal and temporal retinal margins project to caudal tectum, so that the growth patterns of retina and tectum in this particular case are more congruent than normal.

With TT eyes, and probably with VV eyes, therefore, marked shifts in connectivity relations between retinal and tectal elements occur during the maturation of the adult map, while the overall topological relations between developing retina and tectum are preserved. The present results, particularly those from TT projections, appear to dispose of the ‘overgrown half-tectum’ hypothesis mentioned in the introduction. A large part of caudomedial tectum, opposite a TT eye, develops a characteristic tectal structure in the absence of optic innervation.

The differential behaviour of the three types of ‘compound eye’ projection requires comment. TT projections initially show a preference for rostrolateral tectum during development, VV projections initially prefer medial tectum and NN eyes initially innervate most densely the more caudal available tectal area. Initially, fibres from TT eyes do not occupy the entire available tectal surface; and both VV and NN eyes show a relative deficit in input to part of the tectum. These clear selective preferences displayed by the different types of compound eye for particular regions of the then-existing tectum appear to mimic the nature of the mature normal adult projection.

In the developing Xenopus visual system we therefore seem to have two conflicting phenomena. The first is that the re-adjustment of connexions required by the different modes of growth of retina and tectum argues against a fixed system of tectal place-specific markers. The second is that the initial differences between the projections from NN, TT and VV eyes indicate that some distinguishing cues exist at the tectal end of the system. We suggest the following mechanisms to account for the findings.

We assume that an optic fibre from any part of the retina can establish connexion, in principle, with cells in any part of the tectum. The ordering of fibres that exists within the optic tract (Steedman, Stirling & Gaze, 1979; and unpublished) may be sufficient, from the earliest stages of development, to deliver temporal retinal fibres preferentially to rostral tectum, ventral fibres to medial tectum (Straznicky et al. 1979) and nasal fibres to the most caudal regions of the available tectum. To account for the more limited spreading of TT and VV projections, in comparison with NN projections, we propose that newly arriving temporal and ventral fibres reach tectal areas already innervated. To establish connexions in these areas they must disrupt established connexions and displace the resident fibres more caudally (temporal fibres) or laterally (ventral fibres) on the tectum. Newly arriving fibres from an NN eye, on the other hand, are delivered from the tracts to newly generated regions of the tectum and do not have to compete with a resident fibre population to establish their connexions. It seems reasonable, therefore, that the covering of available tectal space by fibres from an NN eye would take place more easily, and hence more rapidly, than that by fibres from a TT or VV eye.

The present results, considered in conjunction with previous work in this field, lead us to the view that several mechanisms co-operate in the establishment of the retinotectal projection. It seems necessary to postulate that early developmental processes lead to the acquisition by retinal cells (and their fibres) of place-related labels (Gaze, Feldman, Cooke & Chung, 1979; Sharma & Holly-field, 1980). We think that the present evidence argues against the existence of independently developed tectal positional markers. Even should further experiments (for instance, graft translocation on a virgin tectum) furnish evidence in favour of independent tectal positional markers, it would be necessary to postulate that such markers must be overridden by mechanisms such as fibre/ fibre interactions and competition among fibres for available tectal space.

Whilst interactions between labelled retinal fibres themselves, and between the fibres and the tectum, seem sufficient to account for both the ordering and the coverage of the projection, it does seem necessary to postulate in addition a mechanism to orient the map on the tectal surface. This mechanism may involve some relatively non-specific positional information on the tectum; but we have drawn attention here to the possibility that orientation of the map may be a consequence of fibre ordering in the tract.

We know little about the mechanisms leading to fibre ordering in the optic tract; and until recently we did not even know that the tract showed retinotopic order (Scalia & Fite, 1974; Steedman et al. 1979, and unpublished). Attardi & Sperry (1963) showed that regenerating optic fibres in adult gold-fish were able to select the correct branch .of the optic tract and they suggested that pathways were ordered by selective affinities between fibres and their substrates. This aspect of the problem has not received the further attention it deserved until very recently.

In Xenopus, fibres are retinotopically grouped as they enter the optic nerve but this ordering is progressively reduced as the fibres pass along the nerve to the chiasma (Fawcett, 1980). Beyond the chiasma, however, the fibres again show a retinotopic ordering, although the arrangement of the fibres is different from that at the optic nerve head (Steedman, Stirling & Gaze, unpublished). We do not know whether this tract order reflects specific fibre/substrate interactions. An alternative possibility is that some rather low-grade positional information influences the pathways selected by the earliest, pioneering, fibres, and that thereafter order is built up by fibres selectively associating with closely related fibres during growth. Clearly further information is required on the mode of development of the optic pathway itself, to permit greater under-standing of the way this factor interacts with others in the establishment of the retinotectal projection.

We thank Mrs June Colville for expert histological assistance.

Attardi
,
D. G.
&
Sperry
,
R. W.
(
1963
).
Preferential selection of central pathways by regenerating optic fibers
.
Expl Neurol
.
7
,
46
64
.
Fawcett
,
J. W.
(
1980
).
Fibre organization in the optic nerve of Xenopus laevis
.
J. Physiol
(in the press).
Feldman
,
J. D.
&
Gaze
,
R. M.
(
1972
).
The growth of the retina in Xenopus laevis: an autoradiographic study. II. Retinal growth in compound eyes
.
J. Embryol. exp. Morph
.
27
,
381
387
.
Gaze
,
R. M.
,
Chung
,
S.-H.
&
Keating
,
M. J.
(
1972
).
Development of the retinotectal projection in Xenopus
.
Nature, Lond
.
236
,
133
135
.
Gaze
,
R. M.
,
Feldman
,
J. D.
,
Cooke
,
J.
&
Chung
,
S.-H.
(
1979
).
The orientation of the visuotectal map in Xenopus: developmental aspects
.
J. Embryol. exp. Morph
.
53
,
39
66
.
Gaze
,
R. M.
,
Jacobson
,
M.
&
Székely
,
G.
(
1963
).
The retinotectal projection in Xenopus with compound eyes
.
J. Physiol
.
165
,
484
499
.
Gaze
,
R. M.
,
Keating
,
M. J.
&
Chung
,
S.-H.
(
1974
).
The evolution of the retinotectal map during development in Xenopus
.
Proc. R. Soc. Lond. B
185
,
301
330
.
Gaze
,
R. M.
,
Keating
,
M. J.
,
Ostberg
,
A.
&
Chung
,
S.-H.
(
1979
).
The relationship between retinal and tectal growth in larval Xenopus: implications for the development of the retinotectal projection
.
J. Embryol. exp. Morph
.
153
,
103
143
.
Gaze
,
R. M.
&
Straznicky
,
C.
(
1979
).
Selective regeneration of optic fibres from a compound eye to the ipsilateral tectum in Xenopus
.
J. Physiol
.
293
,
58
59P
.
Gaze
,
R. M.
&
Straznicky
,
C.
(
1980
).
Regeneration of optic nerve fibres from a compound eye to both tecta in Xenopus: evidence relating to the state of specification of the eye and of the tectum
.
J. Embryol. exp. Morph
, (in the press).
Hope
,
R. A.
,
Hammond
,
B. J.
&
Gaze
,
R. M.
(
1976
).
The arrow model: retinotectal specificity and map formation in the goldfish visual system
.
Proc. R. Soc. Lond. B
194
,
447
466
.
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
.
Nieuwkoop
,
P. D.
&
Faber
,
J.
(
1967
).
Normal Table of Xenopus laevis
.
Amsterdam
:
(Daudin) North Holland
.
Scalia
,
F.
&
Fite
,
K.
(
1974
).
A retinotopic analysis of the central connections of the optic nerve in the frog
.
J. comp. Neurol
.
158
,
455
477
.
Sharma
,
S. C.
&
Hollyfield
,
J. G.
(
1980
).
Specification of retinotectal connexions during development of the toad Xenopus laevis
.
J. Embryol. exp. Morph
.
55
,
77
92
.
Sperry
,
R. W.
(
1943
).
Visuomotor coordination in the newt (Triturus viridescens) after regeneration of the optic nerve
.
J. comp. Neurol
.
79
,
33
55
.
Sperry
,
R. W.
(
1944
).
Optic nerve regeneration with return of vision in anurans
.
J. Neurophysiol
.
7
,
57
70
.
Sperry
,
R. W.
(
1945
).
Restoration of vision after crossing of optic nerves and after contralateral transplantation of eye
.
J. Neurophysiol
.
8
,
15
28
.
Sperry
,
R. W.
(
1951
).
Mechanisms of neural maturation
.
In Handbook of Experimental Psychology
(ed.
S. S.
Stevens
), pp.
236
280
.
New York
:
Wiley
.
Sperry
,
R. W.
(
1963
).
Chemoaffinity in the orderly growth of nerve fibre patterns and connections
.
Proc. natn. Acad. Sci. U.S.A
.
50
,
703
710
.
Sperry
,
R. W.
(
1965
).
Embryogenesis of behavioural nerve nets
.
In Organogenesis
(ed.
R. L. De
Haan
&
H.
Ursprung
).
New York
:
Holt, Rheinhardt and Winston
.
Steedman
,
J. G.
,
Stirling
,
R. V.
&
Gaze
,
R. M.
(
1979
).
The central pathways of optic fibres in Xenopus tadpoles
.
J. Embryol. exp. Morph
.
50
,
199
215
.
Straznicky
,
K.
&
Gaze
,
R. M.
(
1971
).
The growth of the retina in Xenopus laevis’. an autoradiographic study
.
J. Embryol. exp. Morph
.
26
,
67
79
.
Straznicky
,
K.
&
Gaze
,
R. M.
(
1972
).
The development of the tectum in Xenopus laevis: an autoradiographic study
.
J. Embryol. exp. Morph
.
28
,
87
115
.
Straznicky
,
C.
,
Gaze
,
R. M.
&
Horder
,
T. J.
(
1979
).
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
.
J. Embryol. exp. Morph
.
50
,
253267
.
Straznicky
,
K.
,
Gaze
,
R. M.
&
Keating
,
M. J.
(
1971
).
The retinotectal projections after uncrossing the optic chiasma in Xenopus with one compound eye
.
J. Embryol. exp. Morph
.
26
,
523
542
.
Straznicky
,
K.
,
Gaze
,
R. M.
&
Keating
,
M. J.
(
1974
).
The retinotectal projection from a double-ventral compound eye in Xenopus laevis
.
J. Embryol. exp. Morph
.
31
,
123
137
.
Straznicky
,
C.
,
Gaze
,
R. M.
&
Keating
,
M. J.
(
1979
).
Selective optic fibre projection from compound eyes to the tectum during development in Xenopus
.
Neurosci. Abstracts
5
,
181
.
Straznicky
,
K.
&
Tay
,
D.
(
1977
).
Retinal growth in double dorsal and double ventral eyes in Xenopus
.
J. Embryol. exp. Morph
.
40
,
175
185
.
Willshaw
,
D. J.
&
Von Der Malsburg
,
C.
(
1979
).
A marker induction mechanism for the establishment of ordered neural mappings: its application to the retinotectal problem
.
Phil. Trans. R. Soc. B
287
,
203
243
.