The retina of Xenopus laevis has previously been shown, using autoradiographic methods, to develop in the normal animal by the annular addition of cells at the ciliary margin.

The development of the retina in animals with surgically produced ‘compound eyes’ was subsequently studied. In these animals the eye cup was split along the dorsoventral axis and the resulting half-eyes were recombined so as to form animals with a double-nasal eye.

The retina in experimental animals was found to develop as in the normal animal. No labelling of cells with radioactive thymidine was seen along the cut edge of each half-eye; thus in terms of cell division each half of the compound eye remains a half.

We are concerned with the mode of growth of the retina and with the mode of development of the connexion pattern between retinal ganglion cells and the tectum in Xenopus. In a normal Xenopus the connexions that form between ganglion cells in the retina and the tectum lead to the existence, in the adult, of the well- organized retinotectal projection which has previously been described (Gaze, Jacobson & Székely, 1963; Gaze, Keating, Székely & Beazley, 1970).

During normal development in Xenopus the eye connects with the contralateral tectum in such a way that, in the adult, most-nasal retina sends fibres to most-caudal tectum while most-temporal retina connects with most-rostral tectum; central retina projects to central tectum. In normal animals the retina increases in size throughout larval development by the addition of cells at the ciliary margin (Straznicky & Gaze, 1971). This process continues until after metamorphosis. Thus the retinal cells that project, in the adult, to the rostral and caudal poles of the tectum are among the youngest cells in the retina, having differentiated late in development; whereas the retinal cells which project, in the adult, to the central regions of the tectum are among the oldest cells in the retina, having differentiated early in development.

In Xenopus embryos at developmental stage 32 (Nieuwkoop & Faber, 1967) it is possible to make ‘compound eyes’ by, for instance, removing the temporal half of the eye cup and replacing it, in dorso-dorsal orientation, with a nasal half-eye taken from the opposite side of another embryo at the same stage of development (Gaze, Jacobson and Székely, 1963, 1965; Gaze, 1970) so that a double-nasal (NN) compound eye results. Other varieties of compound eye can be constructed in a comparable fashion.

In Xenopus with such surgically produced NN or TT (double-temporal) compound eyes, the projection from the compound eye to its contralateral tectum in the adult animal is abnormal (Gaze, Jacobson & Székely, 1963,1965). If the eye is NN, then both nasal and temporal extremities of the retina project to the caudal pole of the tectum, while cells along the vertical midline of the retina project to the rostral tectal pole. These midline retinal cells in an NN eye thus connect to that part of the tectum which, in a normal animal, receives the projection from the most temporal retinal cells.

The midline retinal cells in an NN eye thus behave, in relation to the connexions they form, as do cells from the temporal pole of a normal eye. A comparable statement can be made about the entire ganglion-cell population of each half of the compound retina, in that its tectal projection resembles that from a whole normal eye in terms of tectal extent and the ordering of the connexion pattern (Gaze, Jacobson & Székely, 1963, 1965).

Two alternative mechanisms have been put forward to account for the connexion pattern formed by compound eyes (Straznicky, Gaze & Keating, 1971). In one, each half-retina comprising the compound eye is deemed to undergo pattern regulation in that the cells at the midline take on the ‘specificity characteristics’ of the cells at the missing pole of the eye; in the other view, regulation of this sort is held not to occur and the connexion pattern actually formed is determined by a matching of the polarity and the extent of the available retina to the available tectum.

The normal developing Xenopus retina is surrounded by a ring of precursor cells at the ciliary margin which mitose during development and thus add to the extent of the neural retina (Straznicky & Gaze, 1971). Since each half of the compound eye projects to the tectum as if it were a whole normal eye, and in view of the possibility that each half of a compound eye is a regulated system in terms of cellular positional information, we thought that it would be worthwhile to find out whether each half of the compound eye also resembles a normal eye in being surrounded by a ring of retinal precursor cells. Cells from the temporal pole of the normal eye take part in the growth of the eye whereas midline cells do not. We have therefore investigated the mitotic history of cells at the cut edge of the compound NN eye, using tritiated thymidine as a marker.

The methods employed have been described in earlier papers (Gaze, Jacobson & Székely, 1963; Straznicky & Gaze, 1971).

In brief, double-nasal compound eyes were made by removing the temporal half of an eye in a stage 32 (Nieuwkoop & Faber, 1967) Xenopus laevis embryo and replacing it with a nasal half-eye from the opposite side of another stage 32 embryo. At stage 45 or 47, 5 μCi of tritiated thymidine (specific activity 20 Ci/ mmole) were injected into the ventral surface of the larvae in the gut region, and 8 animals were killed 24 h after injection. Tissues were fixed in Carnoy’s fixative and serial sections were cut at 4 μm. Deparaffinized sections were then coated with Ilford Nuclear Research Emulsion G 5 and were exposed for 3–9 weeks before being developed. Sections were then stained with haematoxylin and eosin.

The retinae of the normal eye (Fig. 1 A, B, C) were similar to those reported by Straznicky & Gaze (1971). In summary, at this stage there are about 40 ganglion cells across the retinal equator (as seen in transverse sections) and these are all unlabelled, since they had undergone their final DNA synthesis before administration of the thymidine. Labelled cells are to be found at the ciliary margins, where active incorporation of thymidine was going on at the time the label was given.

Fig. 1.

Autoradiograph of the eyes of a Xenopus larva. [3H]thymidine was given at stage 47 and the tissues were prepared for autoradiography 24 hours later.

(A) Normal eye; the eye is unlabelled except for the ciliary margins, which are shown in higher magnification in B and C. (D) Compound (NN) eye; the eye is unlabelled except for the ciliary margins, which are shown in higher magnification in E and F. The darkly staining cells in the fundal part of the retina are degenerate cells; they are not labelled.

In each eye the plane of section passes through the optic nerve head. The bar in each photograph represents 50 μm.

Fig. 1.

Autoradiograph of the eyes of a Xenopus larva. [3H]thymidine was given at stage 47 and the tissues were prepared for autoradiography 24 hours later.

(A) Normal eye; the eye is unlabelled except for the ciliary margins, which are shown in higher magnification in B and C. (D) Compound (NN) eye; the eye is unlabelled except for the ciliary margins, which are shown in higher magnification in E and F. The darkly staining cells in the fundal part of the retina are degenerate cells; they are not labelled.

In each eye the plane of section passes through the optic nerve head. The bar in each photograph represents 50 μm.

The findings in the retinae of compound eyes were essentially similar to those in the normal eyes. Some disorganization of structure was apparent within the retina, but labelling of cells remained confined to the periphery of the retina and no labelled cells were seen at the position of the cut edges of the two half-eyes (Fig. 1D, E, F).

The results of the present experiments show that the cells at the midline of an NN compound eye, in contradistinction to the cells at the temporal pole of a normal eye, do not take part in the cell division leading to the growth of the eye. Each half of a compound eye is thus not equivalent to a whole eye in terms of its mitotic pattern.

This experimental finding adds support to our previous arguments (Straznicky, Gaze & Keating, 1971) that retinal pattern regulation is unlikely as an explanation for the connexion pattern found in compound eyes. Retinal pattern regulation is a hypothesis which has been invoked to account for the production of a full scale of ‘specificity’ values along half the normal extent of the appropriate retinal axis (compare Fig. 2C and D with Fig. 2 Aand B). In the case of NN eyes, retinal pattern regulation would result in two complete scales of such values, one covering the half-axis of each half-eye.

Fig. 2.

Schematic representation of the growth of the normal eye and the compound (NN) eye in Xenopus. (A) Growth of a normal eye. On the left is shown the small eye of a stage 32 larva. Although such an eye would have several ganglion cells along the naso-temporal axis of the retina, in this figure the eye is divided into only two parts, temporal and nasal, and these are labelled T1 and Nl. Growth of the eye involves adding on successive rings of new cells at the retinal margin, as shown in the diagrams on the right. The highest numbers indicate the most-nasal or most-temporal cells.

(B) Growth of a compound NN eye in which regulation is assumed not to occur. Each half-eye is considered to consist of nasal (initially Nl, Nl) cells only. The addition of new cells occurs in rings as in the normal eye. The eye ends up with all its ganglion cells ‘nasal’ and none ‘temporal’.

(C) Growth of a compound NN eye in which regulation is assumed to occur, followed by mitosis along the cut edge of each half-eye. The addition of new cells follows, for each half-eye, the pattern shown for a normal eye (A). In this case two apposed ‘normal’ eyes with opposite nasotemporal polarity would result.

(D) Growth of a compound NN eye in which regulation is assumed to occur and there is no mitosis at the line of junction between the two half-eyes. In this case, since the cells established originally as N1, T1 in each half-eye remain present throughout the further growth of the eye, a continuous manifestation of retinal regulation will be needed to provide the pattern shown for the adult. The ‘name’ of each ganglion cell must thus change each time a new ring of cells is added at the margin.

Fig. 2.

Schematic representation of the growth of the normal eye and the compound (NN) eye in Xenopus. (A) Growth of a normal eye. On the left is shown the small eye of a stage 32 larva. Although such an eye would have several ganglion cells along the naso-temporal axis of the retina, in this figure the eye is divided into only two parts, temporal and nasal, and these are labelled T1 and Nl. Growth of the eye involves adding on successive rings of new cells at the retinal margin, as shown in the diagrams on the right. The highest numbers indicate the most-nasal or most-temporal cells.

(B) Growth of a compound NN eye in which regulation is assumed not to occur. Each half-eye is considered to consist of nasal (initially Nl, Nl) cells only. The addition of new cells occurs in rings as in the normal eye. The eye ends up with all its ganglion cells ‘nasal’ and none ‘temporal’.

(C) Growth of a compound NN eye in which regulation is assumed to occur, followed by mitosis along the cut edge of each half-eye. The addition of new cells follows, for each half-eye, the pattern shown for a normal eye (A). In this case two apposed ‘normal’ eyes with opposite nasotemporal polarity would result.

(D) Growth of a compound NN eye in which regulation is assumed to occur and there is no mitosis at the line of junction between the two half-eyes. In this case, since the cells established originally as N1, T1 in each half-eye remain present throughout the further growth of the eye, a continuous manifestation of retinal regulation will be needed to provide the pattern shown for the adult. The ‘name’ of each ganglion cell must thus change each time a new ring of cells is added at the margin.

In this case, the absence of mitosis around the cut edge of each half-eye makes it necessary to postulate not only an initial retinal regulation, but also a continual manifestation of retinal regulation throughout the period of growth of the eye; that is, until after metamorphosis. In Fig. 2C initial regulation of the compound eye, followed by the addition of new rings of cells as in two normal juxtaposed eyes, has been shown diagrammatically. This can be contrasted with the picture of continuous regulation which must be hypothesized if the idea of retinal pattern regulation is to be maintained in the light of the experimental finding that mitosis only occurs at the free edge of the compound eye. This process demands a continuous change of ‘specificity label’ of all retinal cells each time that a new ring of cells is added to the retinal margin. We have therefore concluded that it seems more likely that retinal pattern regulation does not occur and that retinotectal connexions are formed by some mechanism which takes into account the polarity and extent of available retina (in the present circumstances, a half-retina) and tectum.

Gaze
,
R. M.
(
1970
).
The Formation of Nerve Connections
.
London
:
Academic Press
.
Gaze
,
R. M.
,
Jacobson
,
M.
&
Székely
,
G.
(
1963
).
The retinotectal projection in Xenopus with compound eyes
.
J. Physiol., Lond
.
165
,
484
499
.
Gaze
,
R. M.
,
Jacobson
,
M.
&
Székely
,
G.
(
1965
).
On the formation of connexions by compound eyes in Xenopus
.
J. Physiol, (fond.)
Y16
,
409
417
.
Gaze
,
R. M.
,
Keating
,
M. J.
,
Székely
,
G.
&
Beazley
,
Lynda
. (
1970
).
Binocular interaction in the formation of specific intertectal neuronal connections
.
Proc. R. Soc. B
175
,
107
— 147.
Nieuwkoop
,
P. D.
&
Faber
,
J.
(
1967
).
Normal Table of Xenopus laevis (Daudiri
.. 2nd edition.
Amsterdam
:
North Holland Publishing Co
.
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.
&
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
.