Note to the Reader
This is a revised and annotated version of a paper first published in The
Journal of Neuroscience. The quality of figures 8 and 12 (two darkfield
micrographs) are improved. Two new figures have been incorporated into this
revision. Annotation and additions are delimited by brackets [...].
Enlarging images. Thumbnail versions of all figures are embedded in the
paper. Full-size images—under 300K—will download into a separate window when
you select the thumbnails. Drag this second window to the side of the text
window. In some cases, high-resolution images—usually under 600K—that match
the quality of the original micrographs can be downloaded by selecting the
text at the bottom the corresponding figure legend. These image files can be
viewed with Adobe Photoshop, NIH Image, or equivalent.
Revised HTML edition <http://www.nervenet.org/papers/SC1982.html> copyright
© 1998 by Robert W. Williams
Prenatal Development of Retinocollicular Projections in the Cat: An
Anterograde Tracer Transport Study
Robert W. Williams and Leo M.
Department of Psychology and Physiology Graduate Group, University of
California, Davis, California 95616
The Journal of Neuroscience 2:604–622 (1982)
Materials and Methods
The retinal projection to the superior colliculus of the cat was
examined at four gestational ages, embryonic days 38, 46, 56, and 61,
using the anterograde transport of horseradish peroxidase and tritiated
leucine. Gestation in the cat averages 65 days. By the 38th day of
gestation (E38) the entire superior colliculus, both ipsilateral and
contralateral to an injection, was labeled intensely. The peroxidase label
had an extremely coarse-grained texture. As late as E46, the ipsilateral
retinal projection extended to the presumed caudal border of the
colliculus, while by E56, as in the mature cat, the ipsilateral terminal
field was largely excluded from the rostral and caudal tectal poles. As in
younger fetal material, at E56 the contralateral terminal field extended
across the entire tectal mantle. By E61 clear gaps of label were observed
in the contralateral stratum griseum superficiale. At this age patches or
bands of label were accentuated on the ipsilateral side. In double-labeled
tissue, it could be demonstrated that the ipsilateral patches corresponded
quite closely to the gaps in the contralaterally-derived label. Thus, a
few days before birth, the retinal projection to the superior colliculus
has a pattern which resembles that of the adult.
During the last decade, the advent of axoplasmic transport tracing
methods has provided a wealth of data on developing retinofugal
projections in a variety of vertebrates ranging from the frog (Currie and
Cowan, 1975) to the monkey (Rakic, 1977a). Recent work has shown that the
intitial distribution of retinal ganglion cell axons in fetal and neonatal
mammals is more diffuse and extensive than can be demonstrated in the
adult (Rakic, 1977a; Cavalcante and Rocha-Miranda, 1978; Frost et al.,
1979; Land and Lund, 1979; Sanderson et al., 1982).
Remarkably little is known about the in utero development of
the feline visual system. With the exception of the work of Cragg (1975),
Anker (1977), and Shatz and her colleagues (Shatz and DiBeradino, 1980;
Kilot and Shatz, 1981), the genesis of retinal projections has not been
examined. The predominant position of the cat in research involving the
postnatal maturation of visual function warrants a considerably more
complete description of the formation of this species’ visual system.
Our work has been directed at establishing the normal sequence of
events leading to the formation of the mature retinocollicular projection.
The tectal anlage does not undergo substantial rotation during development
as does, for example, the dorsal lateral geniculate nucleus (DLGn) (Rakic,
1977b; Kalil, 1978; Shatz and DiBerardino, 1980). This simplifies the
analyis and reconstruction of the distribution of axonal label and makes a
comparision between animals of different gestational ages relatively
straightforward. In addition, an excellent background for the
developmental work is provided by the substantial number of studies which
have examined retinocollicular connections in the adult cat using a
variety of anatomical techniques (Laties and Sprague, 1966; Sterling,
1973; Kelly and Gilbert, 1975; Graybiel, 1975, 1976; Harting and Guillery,
1976; Magalhães-Castro et al., 1976; Wässle and Illing, 1980; Bowling and
Michael, 1980; Behan, 1981; Mize, 1981).
We have found that in the cat, as in the monkey (Rakic, 1977a), the
retinal projections to the superior colliculus of fetal animals are more
widespread than those demonstrated after parturition. The mature pattern
of termination appears to be sculpted from a profuse prenatal projection
field by the selective retraction and partial segregation of crossed and
uncrossed retinal efferents. The details of this reformation of retinal
input to the superior colluculs are the main emphasis of this paper.
Abstracts summarizing some of our finding have been published recently
(Williams and Chalupa, 1980, 1981).
Materials and Methods
Timed pregnancies were obtained by exposing an estrous female to a
potent male cat for 24 hr. To ensure that the female was receptive, a
number of matings were observed. The end of the 24-hr exposure period
marked the beginning of embryonic day 1 (E1). This method of timing
conception is only accurate to within 1 day. Gestation in the domestic cat
has a duration ranging from 62 to 67 days.
Anesthesia was induced in the pregnant cat with a 4% halothane vapor
in oxygen. Following intubation and during the entire surgical procedure,
the halothane concentration was maintained at 1.0 to 1.5%. A midline
abdominal incision was made under asceptic conditions and the uterine
horns were withdrawn in turn. Each fetal head has located by palpation and
a small longitudinal incision of the uterus was made to expose the head.
Membranes and skin overlying the eye were parted. Each eye then received
an intravitreal injection of either horseradish peroxidase (HRP) or [3H]-leucine.
HRP was diluted in distilled water to a concentration of 30%. The
peroxidase injections ranged from 5 µl for the youngest fetuses up to 20
µl for perinatal animals. Injections of [3H]-leucine
(specific activiy, 57 Ci/mmol) ranged from 20 µCi in 0.4 µl to 100 µCi in
2.0 µl, again varying with age. Due to the large volume fo the peroxidase
injections, leakage around the site of puncture was common. For this
reason, we are unable to specify the exact quantitiy of peroxidase delived
into the vitreous. The amount spilled was frequently sufficient to label
substantial numbers of neurons innervating extrinsic ocular muscle. The
injections were made using a Hamilton microliter syringe fitted with a 26
gauge needle or with a fine-tipped glass pipette equipped with a plunger.
To avoid penetrating the retina, the insertion site was made far anterior,
roughly at the site of the ciliary body.
Following completion of the injection, the uterine and abdominal
incisions were closed with absorbable suture, and anesthesia was
discontinued. The female was usually alert within 20 min after surgery.
Approximately 24 hr later the cat was reanesthetized, and the fetuses were
delivered by cesarean section. Fetuses were perfused immediately by
transcardial infusion of physiological saline, followed by a
phosphate-buffered 2.5% glutaraldehyde, 1.0% paraformaldehyde solution,
and finally 10% sucrose in phosphate buffer. The mother’s incisions were
closed and anesthesia was terminated.
Less than 2 hr after fixation of the fetuses, the eyes and brains
were removed, weighed, and photographed. The brains were immersed in a 20%
sucrose solution for 24 hr, blocked in the frontal plane, and cut frozen
at a thickness of 40 µm using a sliding microtome. All sections were saved
in phosphate buffer and alternate sections were mounted for conventional
autoradiographic processing (Hendrickson and Edwards, 1978). Tissue was
cleared in xylene, air-fried in a humid dust-free envrionment, and then
dipped in Kodak NTB-2 emulsion. The sections were stored at 4 °C for 8
weeks and developed in Kodak D-19. The remaining tissue was reacted to
demonstrate the presence of peroxidase enzyme using the chromogen
tetramethyl benzidine (TMB) (Mesulam, 1978). Every 10th section reacted
for peroxidase histochemistry was also counterstained with cresyl violet.
The midbrain of the fetal cat is a conspicuous and relative large
structure which extends 6.0 mm in the rostrocaudal axis by the 5th week of
gestation. (Fig. 1). The tectal cleft, setting off the superior from the
inferior collicular anlagen, is not easily distinguished at this age.
However, if the most caudal extension of retinal ganglion cell axons is
considered to mark this boundary, then as early as E38, the superior
colliculus extends 2.7 ± 0.2 mm. A comparable figure for the mature
superior colliculus is 5.0 ± 0.5 mm. At E38 and E46, no demarcation can be
made between the rostral collicular margin and the pretectum (Fig. 2). By
E56, the pretectal nuclei, especially the nucleus of the optic tract and
the pretectal olivary nucleus, can be differentiated from the more dorsal
fasciculi of retinal efferents [Williams and Chalupa, 1984], and at the
rostral collicular margin, a portion of these efferents becomes capped
dorsally by the superficial gray. At this age, then, the anterior border
of the superior colliclus is well defined.
Figure 1. Drawings based on photographs of the dorsal view
of the fetal and early postnatal cat brain. The mesencephalon, including
the superior and inferior colliculi, is clearly visible as the shaded
portion of the drawings of the fetal brains. This drawing includes an E51
fetus not otherwise covered in the Results. Calibration bar, 1 cm.
Injections of tracer substances were made into the eyes of 36 fetal
cats ranging in age from E30 to E61. The results presented in this paper
are based on the 10 fetuses (see Table 1) in which there was no retinal
damage and in which fixation by perfusion was successful so that axonal
transport to the superior colliculus could be demonstrated for at least
one of the tracers. All photomicrographs and photomontages are identified
both by the age and number of the individual cat, in parentheses, from
which material was obtained.
Table 1: Summary data on cases
|Eye weight (mg)
|Vol HRP (µl)
Table 2: Quantitative Analysis of Labeling
*Adult data from Graybiel 1975
As early as E30, a prominent optic chiasm is present (Anker, 1977; R.W.
Williams, unpublished observation, [1983 thesis]), and it is therefore
probable that optic axons have penetrated the roof of the mesencephalon at
least a few days before the end of the first month of gestation. Shatz and
DiBerardino (1980) have demonstrated a bilateral projection to the cat
diencephalon as early as E32, and at this age Shatz (personal
communication) has noted the presence of a weak contralateral
Figure 2. The retinal projection to the superior
colliculus of fetus E38(5) demonstrated by the anterograde transport of
HRP. The rostral pole of the superior colliculus is shown in the lowest
photograph of the montage, and the caudal pole is shown in the
uppermost photograph. Theleft side of the montage is contralateral to
the eye which received the peroxidase injection.l Note how extensive the
uncrossed projection to the superior colliculus is at this gestational
age, especially in comparison with that in older fetuses (see Figs. 7 and
9). Adjacent sections are spaced approximately 150 µm apart. Calibration
bar, 1 mm.
E38. In our material, a bilateral projection to the mantle of
the superior colliculus anlage could be demonstrated with both tracers as
early as E38. The distribution of peroxidase label in one of the three
successful E38 fetuses is shown in Figure 2. A gradient of the density of
the TMB chromogen through the depth of the tectal plate is evident. This
plate is characterized by a very high cell population density with a
maximum thickness of 300 µm. The intense label ends abruptly at the
ventral margin of the plate. Medial and lateral portions of the
contralateral colliculus were labeled more heavily that was the central
band. This erosion of label toward the middle was even more marked on the
ipsilateral side (Fig. 2, right side). In spite of these gradients in the
density of label, the entire rostrocaudal extent of the contralateral
tectal plate was labeled heavily.
most remarkable feature of peroxidase labeling at E38 was its extremely
coarse-grained texture (Fig. 3A). In coronal sections through the
midbrain, the label was made up entirely of large clumps of polymerized
TMB chromogen. This coarse label may be due to a combination of the
following factors: (1) the absence of fine terminal arbors, (2) the
relatively small number of retinal efferents that have penetrated the
collicular mantle, and (3) the presence of large labeled terminal
varicosities or growth cones. Sections cut through the dorsal lateral
geniculate nucleus (DLGn), in which axons course parallel to the plane of
section, revealed the dilated axons tips characteristics of axonal
extension (Fig. 3B). The ubiquity of what appear to be growth cones in the
DLGn suggests that all portions of this nucleus were being penetrated by
extending axons. As in the DLGn, the coarse collicular label was
omnipresent, and this suggests that at E38 the superior colliculus also
was being overgrown by retinal fibers.
Figure 3. HRP-TMB label at E38. A. Coarse-grained
peroxidase label in the caudal ipsilateral superior colliculus of fetus
E38(5). B, Peroxidase-labeled growth cone in the contralateral DLGn
of the same fetus. Calibration bar, 50 &mciro;m in A; 5 µm in B.
Heavy coarse grained label also filled the medial and lateral portions
of the ipsilateral tectal mantle (Fig. 2, right side). As on the
contralateral side the central band was considerably less intensely
labeled. Along the anteroporterior axis the heavy label extended roughly
2.6 mm from the pretectal border to within a few hundred micrometers of
the caudal tip of the superior colliculus. The peroxidase label in the
most posterior 300 µm was not heavy, but it was nevertheless distinct. The
meager uncrossed projection to the extreme caudal pole of the E38 (5)
fetus may represent the precursor of the monocular crescent seen in mature
cats (Graybiel, 1976).
Autoradiographic material at E38 did not match the quality of the
peroxidse material. The label was only marginally above the background
level at the dorsal margin of the ipsilateral tectal anlage, and although
the contralateral label was moderately intense, it failed to resolve the
fine details of the retinal projection evident in the peroxidase material.
The majority of E46 material was derived from one fetus (E46(2)) in which
both the peroxidase and autoradiographic methods were successful. The
distribution of peroxidase label at this age is shown in Figure 4. The
contralateral projection demonstrated by the peroxidase method was
extremely dense and moderately coarse grained. In distinction to the
younger fetuses, the label was distributed with greater uniformity along
the mediolateral axis. Small clusters of the TMB chromogen were
occasionally noted as deep as 400 µm below the dorsal surface. Closer to
the tectal margin the density of chromogen at E46 was too high to
distinguish individual clusters, but it is probable that even at this age
the superior colliculus is the recipient of new axon ingrowth.
Figure 4. The distribution of label within the superior
colliculus fo fetus E46(2). This series of drawings was made using a
Bausch and Lomb projection microscope. Conventions are as in Figure 2. Two
tiers of label are evident on the ipsilateral (right) side. Adjacent
sections are spaced approximately 200 µm apart. Calibration bar, 1 mm.
The contralateral autoradiographic label confirmed the picture obtained
using the peroxidase method. Figure 5A, an autoradiograph through the
anterior contralateral superior colliculus, demonstrates a significant
change in the quality of label with depth. The dorsal 100 µm of the
collicular label is dense and uniform, whereas the ventral 150 µm of label
is mottled. The significance of this variation in the quality of
autoradiographic and peroxidase label with depth is not known. It is
possible that the mottled appearance in the deeper half of the plate is
caused by (1) heavily labeled fasciculi of retinal fibers, (2) aggregates
of growth cones, or (3) the particular types of retinal axons arborizing
or growing in the deeper portions of the retinorecipient laminae. In a
study of the prenatal development of the retionfugal projection in the
rat, Lund and Bunt (1976) found that the initial complement of retinal
fibers coursed tangentially along the surface of the tectum. The majority
of these fibers eventually came to lie under the stratum opticum. At E38
the heaviest label was located just below the pia, whereas by E46 some of
the heaviest and most coarse label was located along the ventral margin of
the tectal mantle. The mottled appearance of label in far rostral sections
through the E46 midbrain thus may represent the formation of the stratum
E38 and E46, there was a significant change in the quality and
distribution of ipsilateral label. The granularity of the TMB chromogen at
E46, although still coarse compared to older fetal and neonatal material,
was finer than that seen 8 days earlier, especially in the more rostral
setions. The granularity of the label was most heavily textured along the
ventromedial portion of the tectal plate in the region corresponding to
the area of highest cell population density in Nissl-counterstrained
sections (Fig. 5B). The more superficial label was more homogeneous than
that seen along the ventral margin of the tectal plate except in the far
caudal collliculus where the sparse superficial label was particlarly
coarse. However, it should be pointed out that the tectal plate was
thinner in more caudal sections so that a comparison based on depth may
not be appropriate. The change in the grain of peroxidase label may be due
to the elaboration of finer axonal arbors in the upper half of the mantle.
Figure 5. A, Autoradiograph showing the pattern of
label in the rostral contralateral superior colliculus of fetus E46(2).
This dark-field-illuminated micrograph shows the change in the quality of
label with depth: the superficial label is relatively homogeneous; the
deeper label is mottled. The deeper label may presage the formation of the
statum opticum. B, Granularity of peroxidase label along the medial
edge of the ipsilateral superior colliclus of fetus E46(2). This coronal
section is from the caudal one-third of the superior colliculus. Medial is
to the left. Calibration bars, 50 µm.
distinct tiers of ipsilateral peroxidase and autoradiographic label were
evident in sections through the middle and caudal half of the superior
colliclus as shown under dark-field illumination in Figure 6. At this
gestational age, cellular lamination was not apparent in Nissl-counterstained
sections, and thus it was not possible to related these tiers of label to
specific collicular layers. The upper tier of more dense and finer grained
label extended as far as 35 µm below the pia mater, whereas the lower tier
of more coarse label was heaviest close to the bottom of the tectal plate
80 to 140 µm beneath the dorsal surface. If is possible that the lower
tier of label is composed of stained axons of passage and growth cones
that may presage the development of the statum opticum, but if this were
the case, it is difficult to understand why the bottom tier would be most
evident in more caudal sections. An alternate possibility is that each of
the two tiers receives a distinct class of retinal input, since it is
known that, in the mature cat, neurons giving rise to the uncrossed
projection form a heterogeneous population (Fukuda and Stone, 1974; Wässle
and Illiing, 1980).
Figure 6. Two relatively distinct tiers of HRP-TMB label
are evident in this coronal section through the ipsilateral superior
colliculus of fetus E46(2). As in Figure 4 the label is lighter than
background. This coronal section was taken 1.9 mm posterior to the rostal
collicular border. Medial is the the left. Calibration bar, 200 µm.
Even as late as E46, uncrossed label covered the entire tectal plate.
This expansive projection was seen in both peroxidase and autoradiographic
material. At E46 there was no indication that the undecussated
retinocollicular projection had begun to retract from either the anterior
or posterior poles of the superior colliculus. The reorganization that
occurs between E38 and E46 suggests either the onset of the elaboration of
finer axonal plexuses or the segregation of particular types of ganglion
Between E46 and E56 there was only one notable change in the overall
distribution of peroxidase and autoradiographic label in the contralateral
superior colliculus. At E56, the central portion of the anterior third of
the contralateral colliculus was only sparesely labeled. This region of
light labeling is seen most clearly in Figure 7. At this gestational age,
the incipient strata of the superior colliculus could be distinguished in
Nissl-counterstained sections, and it appeared that the erosion of label
was restricted to the upper middle portion of the SGS. The set of
autoradiographs through the anterior third of the superior colliculus from
this fetus (E56(2)) also demonstrated less intense label. In the mature
cat, monkey, and ape (Graybiel, 1975; Hubel et al., 1975; Tigges et al.,
1977; Pollack and Hickey, 1979), the fovea centralis representation on the
superior colliculus has been shown by anterograde transport methods to
receive a relatively sparse retinal input. We assume that the lightly
labeled zone noted at E56 represents the central portion of the retina.
Figure 7. A photomontage of 16 sections spaced
approximately 180 µm apart showing peroxidase label in the contralateral
(left) and ipsilateral (right) superior colliculus of fetus E56(2).
Conventions are as in Figure 2. Note the substantial change in the
distribution of ipsilateral label in comparison to E38(5) (Fig. 2). The
ipsilateral label is restricted tothe central region, but in distinction
to the E61 material (Figs. 9 and 10), the label is still diffuse. The
vertical gaps in label seen especially on the contralateral side are
vascular channels. Calibration bar, 1 mm.
Despite a careful examination and reconstruction of the contralateral
collicular label, no sign of the representation of the contralateral optic
disc, which is evident in the adult cat as an unlabeled patch (Graybiel,
1975), could be found. Since every section was saved and reacted in fetus
E56(2) and every other section in fetus E56(1), it is improbable that we
simply neglected such a gap. The exclusion of contralateral label from the
area corresponding to the optic disc had not progressed noticeably at this
The most pronounced change between E46 and E56 was apparent in the pattern
of ipsilateral label. As shown in Figure 7, and in greater detail in
Figure 8, the caudal ipsilateral superior colliculus was devoid of
peroxidase label in fetus E56(2). This region was also devoid of
autoradiographic label. In addition, the most rostral segment was only
very sparsely labeled. The quality of the ipsilateral peroxidase label at
E56 differed from that seen in younger fetal material. The label was not
distributed uniformly within the superior colliculus, but it was
nevertheless of an even and fine texture. A distinction similar to that
made at E46 between fine superficial and coarse deep label was not
possible. The most superficial label, restricted to the upper margin of
the stratum griseum superficiale (SGS), was very intense (Figs. 7 and 8).
The more ventral label was diffuse and could not be assigned to any
particular sublamina of the SGS. A diffuse band of label, running
rostrocaudally along the crown of the ipsilateral superior colliculus, is
clearly shown in Figure 8. The focus of this label is over the
intermediate lamina of the SGS. Graybiel (1976) has made high resolution
reconstructions of the ipsilateral retinocollicular projection in mature
cats and she describes numerous bands or columns of label oriented along
the anteroposterior axis. It is probable that this banding is incipient at
Figure 8. Montage of the ipsilateral retinal projection to
the superior colliculus of fetus E56(2). The dark-field effect was
obtained uisng two Polaroid type HN32 linear polarizers crossed for
maximum extinction. Adjacent sections are spaced approximately 120 µm
apart. Banding of the ipsilateral retinocollicular projection may be
incipient at this gestational age. Calibration bar, 1 mm.
A few days before birth, the crossed retinal projection to the superior
colliculus has a distribution pattern similar to that of the mature cat.
E61 was the earliest age at which the stratum zonale was devoid of
retinally derived label (Figs. 9 and 12). In the mature cat, the statum
zonale consists largely of astrocytic processes, intrinsic and cortical
axons, and small neurons (Langer, 1975). It contains few, if any, retinal
terminals at E61, in the neonate (R.R. Mize and P. Sterling, research in
progress), or in the adult (Sterling, 1973).
Figure 9. The retinal projection to the superior
colliculus of fetus E61(4) demonstrated by the orthograde transport of
HTP. Conventions are as in Figures 2 and 7. On the contralateral (left)
side, areas of less intense label are found within the middle of the
statum griseum sperficiale. Adjacent sections are spaced approximately 240
µm apart except between the upper two photographs of the montage between
which the spacing is approximately 500 µm. Calibration bar, 1 mm.
Another feature which distinguished the E61 fetal projection from that
of younger fetuses was the unequivocal presence of two tiers of retinal
label in the anterior one-half of the contralateral superior colliculus
(Fig. 9). The upper tier was composed of a virtually unbroken sheet of
very dense label 60 to 80 µm thick. This dense sheet of label directly
abutted the stratum zonale but was clearly restricted to the superficial
portion of the SGS (SGS-1) of Kanaseki and Sprague, 1974). This tectal
layer is the major recipient of retinal input in the mature cat (Sterling,
1973; Mize, 1981). Of interest in this regard is the recent observation
that focal peroxidase injections made into the upper SGS result in the
retrograde labeling of small ganglion cells of the gamma class (Itoh et
al., 1980). The deep tier of retinal label was 100 to 150 µm thick and
extended down to the stratum opticum. This tier was largely resticted to
the deep lamina of the superficial gray, SGS-3. The more sparsely labeled
stratum sandwiched between these tiers corresponded to SGS-2, which at E61
received the bulk of the uncrossed retinal input. The contralateral
projection to SGS-2 took the form of poorly defined bands of label
extending from SGS-1 to SGS-3, with center-to-center separation ranging
from 150 to 250 µm. Defining the borders of diffuse bands precisely was
difficult in both autoradigraphic and peroxidase material, but their
breadth ranged from 60 to 150 µm.
region of particular sparse label in SGS-2 which lay between bands
presumably received a heavy ipsilateral retinal input. Since alternate
sections were processed to demonstrate the uncrossed peroxidase or the
crossed autoradiographic label, the location of ipsilateral patches or
bands could be related directly to the distribution of contralateral label
(Fig. 10). Discrete patches of ipsilaterally derived label were evident
along the mediolateral axis and appeared to be centered within SGS-2.
Clearly, the notion that crossed and uncrossed terminal arbors are at
least partially segregated is borne out by the double anterograde labeling
technique as has been suggested previously by Graybiel (1975) and Harting
and Guillery (1976). Graybiel and Nauta also have depicted the pattern of
ipsilateral banding in the cat several days before birth (results cited in
Figure 10. The partial segregation of crossed and
uncrossed retinal fibers to the superior colliculus of fetus E61(5).
Alternate sections treated for autoradiography (3H) or for the presence of
peroxidase (HRP) were drawn under dark-field illumination. Peroxidase
sections show the crossed retinal projection to the left, whereas the
autoradiographic sections show the uncrossed projection to the left. Upper
drawings are at more caudal levels of the superior colliculus. Calibration
bar, 1 mm.
the adult cat, the segregation of retinal fibers is most pronounced at the
representation of the optic disc. In only one of three E61 fetuses could
we demonstrate the gap of label at the contralateral representation of the
optic disc (Fig. 11A). This small gap in the peroxidase label was
restricted to SGS-1 and the upper part of SGS-2. In an adjacent
autoradiographic section, a patch of ipsilaterally derived label was found
that corresponded to, and presumably filled, the incipient optic disc gap.
However, not until the day of birth was the contralateral representation
of the optic disc clear of label. The gap extended through the full depth
of the superficial gray. (Fig. 11B).
Figure 11. The optic disc representation gap seen in
HRP-TMB-reacted tissue. A, Partial clearing of label is evident in
the stratum zonale and the upper portion of stratum griseum superficiale
of fetus E61(3). B, In the neonate, P0(1), the gap extends through
the stratum zonale and the entire stratum griseum superficiale.
Calibration bar, 50 µm for A and B.
As may be seen in Figure 12, in the rostral half of the ipsilateral
superior colliculus at E61, the discrete banded projection is overlain by
a superficial tier of dense label apposed to the ventral margin of the
stratum zonale. This sheet of superficial label actually os composed of a
set of numerous thin plates of label. Farther posterior the deep tier of
banded or patchy label merges dorsally with the superficial tier. The
superficial label shifts gradually toward the extreme lateral margin of
the SGS before vanishing completely.
Figure 12. Polarized light photomicrograph of the
uncrossed retinal projection to fetus E61(4). Adjacent sections are spaced
approximately 160 µ apart. Patches or bands of label are now clearly
evident within the stratum griseum superficiale. Conventions are as in
Fig. 2. Calibration bar, 1 mm.
The ipsilateral retinal projection described by Graybiel (1976) and
shown in her Figures 1b and 2a and b, resembles that seen in fetus E61(4).
She also noted the presence of a fractured superficial sheet of label in
addition to the more ventral bands in SGS-2. Thus a few days before birth
the retinal projection to the superior colliculus matches that described
by Graybiel (1975, 1976) and Harting and Guillery (1976) in the adult cat.
In particular, at E61, as in the adult cat, the retinocollicular
projection has the following attributes: (i) clear patches of label are
seen in the ipsilateral superior colliculus which, in some instances,
correspond to gaps in label derived from the contralateral retina; (ii)
the stratum zonale is devoid of label; and (iii) the rostral and caudal
margins of the ipsilateral superior colliculus are unlabeled. It is, of
course, possible that the pattern of the retinocollicular projection
undergoes further changes; however, any modificiations in the distribution
of retinotectal fibers after birth are certainly less dramatic that those
that occur during the latter half of gestation.
This study describes the formation of retinal projections to the
superior colliculus of the cat. We have found that by the 38th day of
gestation each retina’s projections to the superior colliculus are
partially segregated. Several days before birth the retinal projection to
the superior colliculus has a pattern similar to that of mature cats. The
retinal innervation of the ipsilateral superior colliculus appears to
involve at least three steps: (i) a generous axonal outgrowth covering the
complete tectal sheet, (ii) a retraction along the anteroposterior axis
such that the projection is excluded from both the rostral and caudal
collicular poles, and (iii) a condensation of the uncrossed terminal
arbors into bands restricted largely to SGS-2.
A number of difficulites could mar the interpretation of the results.
These problems are common to studies employing the orthograde transport of
tracer substances but may be of greater significance in this developmental
study. The uptake of tracer compounds from the vitreous could be
influenced by at least the following variables: uniformity of tracer
distribution in the vitreous, effective vitreal concentration of the
tracer, and uptake competence of retinal ganglion cells as a function of
age, cell class, or retinal position. The first of these
variables–intravitreal distribution and concentration of tracer–obviously
could affect the pattern of terminal label. If it is assumed that all
portions of the developing retina project to the contralateral superior
colliculus, then the pattern of label on the contralateral side provides
an indication of the uniformity of tracer distribution within the eye.
Since, at all ages, heavy label was distributed throughout the full
rostrocaudal extent of the contralateral superior colliculus and since
both eyes received injections of tracer substances that yielded
complementary patterns of labels (see Fig. 10), it is unlikely that the
absence of label seen in rostral and caudal regions of the ipsilateral
superior colliculus at progressively older ages is due to such a
The second factor–uptake and transport as a function of cell birth
date (gestational age at time of final mitosis)–remains problematic. Data
on the time of the last mitosis of matrix [progenitor] cells destined to
differentiate at retinal ganglion cells are not yet available for any
species of carnivore. The results of Sidman (1961) and LaVail and Rakic
(research in progress, cited in Rakic, 1977a) suggest a rough
central-to-peripheral gradient in the time of generation of neurons
destined to occupy the ganglion cell layer of the mammalian retina. Since
we have shown that peroxidase label in the superior colliculus at E38 and
E46 extends from the rostral to the caudal poles, it could be argues that
even as early as E38 axons originating from peripheral temporal ganglion
cells have extended into the most caudal portion of the superior
colliculus. This argument is weakened by the underlying assumption that
fetal topography is similar to that of adults. An alternate possibility is
that early arriving axons originating from the central retina overgrow the
entire tectal mantle and only later become compressed into their adult
terminal fields. The retrograde labeling of ganglion cells from the fetal
superior colliculus could resolve this problem.
Retinal ganglion cells have been categorized using morphological and
physiological criteria into at least three classes (Boycott and Wässle,
1974; Stone and Fukuda, 1974; Cleland and Levick, 1974a, b). There is no
indication in the mature cat that these classes can be distinguished by
the avidity with which they incorporate or transport substances along
their axons. We (unpublished observations) have found that structures
which, in the mature cat, receive predominantly X- and/or Y-type retinal
input (i.e., the main laminae of the DLGn and the medial interlaminar
nucleus (Dreher and Sefton, 1979)) appear to be labeled as well at E38 as
those structures receiving predominantly W-type input (i.e., the ventral
lateral geniculate nuclueus (Spear et al., 1977). It is, therefore,
doubtful that the developmental sequence demonstrated in this paper is
secondary to altered patterns of uptake and transport of tracers by the
three main classes of ganglion cells.
Finally, the retinal position of the ganglion cell soma could
influence the rate of tracer uptake. Neurons within portions of the retina
with a relatively dispersed cell population may accumulate greater
quantities of tracer due to low mobility of the tracer either in the
vitreous humor or within the interstices of the optic fiber and ganglion
cell layers. Stone et al., (1981) have shown that the distribution of
young neurons within the incipient ganglion cell layer is remarkably
uniform at E47. However, by E57 they found an 18:1 central-to-peripheral
gradient, and therefore the possibility of impaired orthograde transport
from the central retina cannot be discounted. In the adult primate, for
example, the central retinal projection to the superior colliculus is not
seen clearly using orthograde transport of tritiated compounds (Hubel et
al., 1975; Tigges et al., 1977) or degeneration methods (Wilson and Toyne,
1970). Since ganglion cell somas are smaller and more densely packed
around the foveal pit, with the ganglion cell layer being up to six cells
deep (Hendrickson and Kupfer, 1976), the availability of tracer per cell
may be consideraly less at central than at peripheral retinal loci. The
recent demonstration by Cowey and Perry (1980) of parafoveal efferents
projecting to the anterior portion of the macaque monkey superior
colliculus after HRP backfills emphasizes the value of this technique in
confirming and extending results obtained with orthograde methods.
The gestation period of the cat averages 65 days. Although the time
of origin of young neurons in the cat visual system has not been well
documented as has been done for the mouse by Sidman (1961) and for the
macaque monkey by Rakic (1974, 1976, 1977b), the report of Hickey and Cox
(1970) has shown that the majority of neurons contributing to the DLGn are
postmitotic at midgestation. Since, in both the monkey (Rakic, 1977b) and
the rodent (Brückner et al., 1976), the period of DLGn neurogenesis
overlaps that of the superior colliculus, it is probable that by E38, the
earliest age examined in this study, the majority of collicular neurons
have been generated. Whether the full complement of cells has migrated
into the tectal plate by E38 in unknown. We have observed very dense
packing of cells along the medial margin of the tectal plate at E38 and
particularly E46, and this may be related to the migrational course of the
neuron- or glioblasts.
The partial retraction during development of an ipsilateral
retinocolliculus projection has been described in the rhesus monkey
(Rakic, 1977a), and rat (Land and Lund, 1979), the hamster (Frost et al.,
1979), and the opossum (Cavalcante and Rocha-Miranda, 1978). The
withdrawal of an uncrossed projection may be a general feature of
mammalian, if not vertebrate, ontogeny. Northcutt and his colleagues have
raised the possibility that bilateral retinofugal projections were present
in ancestral vertebrates, partly on the basis of the distinct ipsilateral
projections in an agnathan, the lamprey (Northcutt and Przybylski, 1973),
in a holostean, the garfish (Northcutt and Butler, 1976), an elasmobrach
(Northcutt and Wathey, 1980), and in the primitive Australian lungfish
(Northcutt, 1980). It would be worthwhile to employ the anterograde
transport of peroxidase enzyme in a developmental study of teleostean and
anuran retinofugal projections to see if short-lived ipsilateral retinal
projections could be demonstrated.
The progressive restriction and subsequent segregation of label
within the developing tectal plate of the fetal cat could be accounted for
by at least three phenomena. First, it is conceivable that axoplasmic
transport rates of retinal ganglion cell could vary during prenatal
development. In the extreme case, select groups of fibers which readily
transport tracers early in development may no longer do so at later
gestational ages. At present, this unsettling possibility cannot be ruled
out. Although axoplasmic transport methods, and in particular the
anterograde transport of horseradish peroxidase, have been shown to be
highly sensitive for demonstrating developing retinofugal projections (So
and Schneider, 1978; Land and Lund, 1979), it may nonetheless be
worthwhile to employ degeneration method, especially in conjunction with
electron microscopy, to examine this possibility.
Figure 13. Schematics of possible causes underlying the
retraction of terminal field label in the superior colliculus. Three
'before-and-after' situations are depicted. A, The telodendria of
the retinocollicular axons condense to form restricted terminal fields.
B, Major axons collaterals are eliminated during development. C,
Retinal ganglion cell number is reduced and axon arbors of dying cells
Second, the redistribution of retinocollicular label may reflect
structural alterations in axonal arbors [Fig. 13A and B]. This could
involve a local and selective regression of portions of axons arbors.
Mason (1980), for instance, has shown considerable remodeling of terminal
segments in the kitten’s lateral geniculate nucleus as late as two weeks
after birth. Furthermore, the formation of ocular dominance slabs in the
striate cortex of the cat has been associated with resorption of
geniculostriate axons arbors (Hubel et al., 1977; LeVay and Stryker,
1979). The erosion of label could also result from the elimination of
major axon collaterals. Such a mechanism has been shown to play a role in
the reduction of the number of cortical neurons with interhemispheric
connections (Innocenti, 1981; Ivy and Killackey, 1981). In the mature cat
some ganglion cells axons bifurcate to innervated the lateral geniculate
and the superior colliculus (O’Leary, 1940; Bowling and Michael, 1980;
Giolli, 1980). It is possible that the incidence of collateralization
within the tract, and perhaps even at the chiasm (cf. Ramón y Cajal,
1911), is considerably higher during early stages of development. This
suggestion is testable with retrograde double-labeling methods (e.g.,
Kuypers et al., 1977; Illing, 1980; Jeffery et al., 1981).
Third, the changes in the labeling pattern may result from the
degeneration of numerous ganglion cells whose axons project to the
colliculus during early development [Fig 13C]. There is now compelling
evidence in the chicken (Rager and Rager, 1976; Hughes and McLoon, 1979;
McLoon and Lund, 1980), rat (Kuwabara and Weidman, 1974; Land and Lund,
1979; Cunningham et al., 1981; Jeffery and Perry, 1981), and hamster
(Frost et al., 1979); Sengelaub and Finlay, 1981) that retinal ganglion
cell death is involved in the restructuring of the developing retinotectal
pathway. [For instance, McLoon and Lund (1980) have shown an ipsilateral
retinal projection to the chick tectum on thetenth day of incubation using
the HRP-TMP mehtod. By the 12th day ofincubation this projection has been
eliminated. As noted by McLoon and Lund, this elimination concurs with the
period of peak ganglion cell death (cf., Hughes and McLoon, 1976). If the
complete retraction of the uncrossed retinotectal projection in the chick
is secondary to selective ganglion cell death then a similar mechanism may
underly the partial retraction now demonstrated in numerous mammalian
There is circumstantial evidence that cell death also plays a role in
the early development of retinofugal connections in the cat. Stone et al.
(1981) have estimated that at E47 there are 4 to 8 times as many cells
within the ganglion cell layer as in the adult (Stone, 1978). The sharp
reduction in label within the ipsilateral superior colliculus between E46
and E56 appears to coincide with the precipitous drop in the number of
neuronal profiles in the cat’s ganglion cell layer.
[Additional support for the cell death hypothesis is also found in
the recent work on the development of the rodent visual system. Naturally
occurring cell death has been described by Kuwabara and Weidman (1974) and
by Cunningham et al. (1980) in electron microscopic surveys of prenatal
and neonatal rat retina. Degenerating cells were frequently found adjacent
to apparently normal ganglion cells. Land and Lund (1979) made focal
injections of peroxidase into the superior colliculus of rat pups,and they
report numerous labeled ganglion cells in both nasal and temporal halves
of the ipsilateral retina. The subsequent disappearance of nasal ganglion
cells with ipsilateral projections can be explained either by the
retraction of uncrossed axon collaberalsgienoff at the chiasm or by
selective ganglion cell death. The latter explanation is strengthened by
both Segelaub and Finlays's report (1980) of substntial ganglion cell
death in the retina of postnatal hamsters and by the report of Hsiao and
Schneider (1981) which shows that only very few ganglion cells have
bilateral central connections, even after early enucleation.]
[The segregation of retinal fibers in the superior colliculus occurs
at a time of vigorous growth of the collicular volume, and does not appear
to require a regression in either the are or volume of the terminal field
in which crossed or uncrossed retinal projections arborize. In this
context the segregation of fibers may be more appropriately thought of as
a consolidation or condensation of axons arbors rather than in terms of
binocular rivalry for postsynaptic sites. it is of interest to not that
segregation of ocular fibers occurs at different gestational ages in the
various retinorecipient nuclei. For example, in the DLGn of the cat,
segregation has begun at E46, more than aweek in advance of the onset of
segregation in the superior colliculus. The elaborationof ocular dominance
laminae in the DLGn may coincide with, and be related to, the reduction of
optic axons projecting to this nucleus. By removing one eye of fetal cats
in advance of segregation (at approximately E-45) it has been possible to
assess the relationship between retraction, segregation, and retinal
ganglion cells death.]
The extent to which the phenomena discussed above contribute to the
transformation in the pattern of retinocollicular label described in this
paper is unknown. Intuitively is seems reasonable that a combination of
factors is involved. If this is the case it will require considerable
effort to assess the relative importance of each of these mechanisms at
different stages of development.
This research was supported inpart by the National Institutes of Health
under grant 532 GM 07416. We thank Dr. Charles J. Sedgwick, Department of
Veterinary Medicine, University of California, Davis, for his guidance and
assistance in the surgical procedures.
Anker RL (1977) The prenatal development of some of the visual pathways
in the cat. J Comp Neurol 173:185–204.
Behan M (1981) Identification and distribution of retinocollicular
terminal in the cat: An electron microscopic autoradiographic analysis. J
Comp Neurol 199:1–15.
Bowling, D. B., and C. R. Michael (1980) Projection patterns of single
physiologically characterized optic tract fibers in cat. Nature 286:
Boycott BB, Wässle H (1974) The morphological types of ganglion cells
of the domestic cat’s retina. J Physiol (Lond) 240:397–419.
Brückner G, Mareš V, Biesold D (1976) Neurogenesis in the visual system
of the rat. An autoradiographic investigation. J Comp Neurol 166:245–256.
Cavalcante LA, Rocah–Miranda CE (1978) Postnatal development of
retinogeniculate, retipretectal and retinotectalprojections in the
opossum. Brain Res 146:231–248.
Cleland BG, Levick WR (1974a) Brisk and sluggish concentrically
organized ganglion cells in the cat’s retina. J Physiol (Lond)
Cleland BG, Levick WR (1974b) Properties of rarely encountered types of
ganglion cells in the cat’s retina and an overall classificiation. J
Physiol (Lond) 240:457–492.
Cowey A, Perry VH (1980) The projection of the fovea to the superior
colliculus in rhesus monkeys. Neuroscience 5:53–61.
Cragg BG (1975) The development of synapses in the visual system of the
cat. J Comp Neurol 160:147–166.
Cunningham TJ, Mohler IM, Giordano DL (1981) Naturally occurring neuron
death in the ganglion cell layer of theneonatal rat: morphology and
evidence for regional correspondence with neuron death in the superior
colliculus. Dev Brain Res 2:203–215.
Currie J, Cowan WM (1975) The development of the retino–tectal
projection in Rana pipiens. Dev Biol 46:103–119.
Dreher B, Sefton AJ (1979) Properties of neurons in cat’s dorsal
lateral geniculate nucleus: a comparison between medial interlaminar and
laminated parts of the nucleus. J Comp Neurol 183:47–64.
Frost DO, So KF, Schneider GE (1979) Postnatal development of retinal
projections in Syrian hamster: a study using autoradiographic and
anterograde degeneration techniques. Neuroscience 4:1649–1677.
Fukuda Y, Stone J (1974) Retinal distribution and central projections
of Y–, X–, and W–cells of the cat’s retina. J Neurophysiol 37:749–772.
Giolli R (1980) A review of axon collateralization in the mammalian
visual system. Brain Behav Evol 17:364–398.
Graybiel AM (1975) Anatomical organization of retinotectal afferents in
the cat: An autoradiographic study. Brain Res 96:1–23.
Graybiel AM (1976) Evidence for banding of the cat’s ipsilateral
retinotectal connection. Brain Res 114:318–327.
Harting JK, Guillery RW (1976) Organization of retinocollicular
pathways in the cat. J Comp Neurol 166:133–144.
Hendrickson A, Edwards SB (1978) The use of axonal transport for
autoradiographic tracing of pathways in the central nervous system. In
Neuroanatomical Research Techniques, (Robertson RT, ed) pp 242–285.
New York: Academic P.
Hendrickson A, Kupfer C (1976) The histogenesis of the fovea in the
macaque monkey. Invest Ophthalmol 14:746–756.
Hickey TL, Cox NR (1979) Cell birth in the dorsal lateral geniculate
nucleus of the cat: a 3H–thymidine study. Soc Neurosci Abst
Hoffmann KP (1973) Conduction velocity in pathways from retina to
superior colliculus in the cat: A correlation with receptive–field
properties. J Neurophysiol 36:409–424.
Hubel DH, LeVay S, Wiesel TM (1975) Mode of termination of retinotectal
fibers in macaque monkey: An autoradiographic study. Brain Res 96:25–40.
Hubel DH, Wiesel TN, LeVay S (1977) Plasticity of ocular dominance
columns in monkey striate cortex. Philos Trans R Soc Lond (Biol)
Hughes WF, McLoon SC (1979) Ganglion cell death during normal retinal
development in the chick: comparisons with cell death induced by early
target field destruction. Exp Neurol 66:587–601.
Illing RB (1980) Axonal bifurcation of the cat retinal ganglion cells
as demonstrated by retrograde double labelling with fluorescent dyes.
Neurosci Lett 19:125–130.
Innocenti GM (1981) Growth and reshaping of axons in the establishment
of visual callosal connections. Science 212:824–827.
Itoh K, Conley M, Diamond IT (1980) Ganglion cell projections to the
superior colliculus, lateral geniculate, pretectum, and vetnral lateral
geniculate in the cat. Soc Neurosci Abstr 6:347.
Ivy GO, Killackey HP (1981) Theontogeny of distribution of callosal
projection neurons in the rat parietal cortex. J Comp Neurol 195:367–389.
Jeffery G, Perry VH (1981) Evidence for ganglion cell death during
development of the ipsilateral retinal projection in the rat. Dev Brain
Jeffery G, Cowey A, Kuypers HGJM (1981) Bifurcating retinal ganglion
cell axons in the rat, demonstrated by retrograde double labelling. Exp
Brain Res 44:34–40.
Kalil R (1978) Development of the dorsal lateral geniculate nucleus in
the cat. J Comp Neurol 182:265–292.
Kanaseki T, Sprague JM (1974) Anatomical organization of pretectal
nuclei and tectal laminae in the cat. J Comp Neurol 158:319–338.
Kelly JP, Gilbert CD (1975) The projections of different morphological
types of ganglion cells in the cat retina. J Comp Neurol 163:65–80.
Kliot M, Shatz CJ (!981) Siamese cat: prenatal development of the
retinogeniculate pathway. Invest Ophthalmol Vis Sci Suppl 20:175.
Kuwabara T, Weidman TA (1974) Development of the prenatal rat retina.
Invest Ophthalmol 13:725–739.
Kuypers HGJM, Catsman–Berrevoets CE, Padt RE (1977) retorgrade axonal
transport of fluorescent substances in the rat's forebrain. Neurosci Lett
Land PW, Lund RD (1979) Development of the rat's uncrossed retinotectal
pathway and its relation to plasticity studies. Science 205:698–700.
Langer TL (1975) Organization of cat superior colliculus: a Golgi and
degeneration study. Anat Rec 181:404.
Laties AM, Sprague JM (1966) The projection of optic fibers to the
visual centers in the cat. J Comp Neurol 127:35–70.
LeVay S, Stryker MP (1979) The development of ocular dominance columns
in the cat. In Society for Neuroscience Symposia. Vol. 4:
Aspects of Developmental Neurobiology (Ferrendelli JA, ed) pp 83–98.
Bethesda MD: Society for Neurosci.
Lund RD, Bunt AH (1976) Prenatal development of central optic pathways
in albino rats. J Comp Neurol 165:247–264.
Magalhães–Castro HH, Murata LA, Magalhães–Castro B (1976) Cat retinal
ganglion cells projecting to the superior colliculus as shown by the
horseradish peroxidase method. Exp Brain Res 25:541–549.
Mason CA (1980) Postnatal development of retinogeniculate axon arbors
in thekitten. Soc Neurosci Abst 6:660.
McLoon SC, Lund RD (1980) Formation and loss of a normal ipsilateral
retinofugal projection in devloping chick. Soc Neurosci Abst 6:647.
Mesulam MM (1978) Tetramethyl benzidine for horseradish peroxidase
neurohistochemistry: a non–carcinogenic blue reaction–product with
superior sensitivity for visualizing neural afferents and efferents. J
Histochem Cytochem 26:106–117.
Mize RR (1981) Morphological identification of retinal synapses in the
cat superior colliculus using quantitative electron micrope
autoradiography. Invest Ophthalmol Vis Sci Suppl 20:185.
Northcutt RG (1980) Retinal projection s in the Australian lungfish.
Brain Res 185:85–90.
Northcutt RG, Przybylski RJ (1973) Retinal projections in the lamprey
Petromyzon marinus (L). Anat Rec 150:400.
Northcutt RG, Butler AB (1976) Retinofugal pathways in the longnose gar
Lepisosteus ossous (Linnaeus). J Comp Neurol 166:1–16.
Northcutt RG, Wathey JC (1980) Guitarfish possess ipsilateral as well
as contralateral retinofugal projections. Neurosci Lett 20:237–242.
O#146;Leary JL (1940) A structural analysis of thelateral geniculate
nucleus of the cat. J Comp Neurol 73:405–430.
Pollack JG, Hickey TL (1979) The distribution of retino–collicular axon
terminals in rhesus monkey. J Comp Neurol 185:587–602.
Rager G, Rager U (1976) Generation and degeneration of retinal ganglion
cells in the chicken Exp Brain Res 25:551–553.
Rakic P (1974) Neurons in the rhesus monkey visual cortex: systematic
relation between time of origin and eventual disposition. Science
Rakic P (1976) Differences in the time of origin and in the eventual
distribution of neurons in areas 17 and 18 of the visual cortex in rhesus
monkey. Exp Brain Res Suppl 1:244–248.
Rakic P (1977a) Prenatal development of the visual system in rhesus
monkey. Phiols Trans R Soc Lond (Biol) 278:245–260.
Rakic P (1977b) Genesis of the dorsal lateral geniculate nucleus in the
rhesus monkey: site and time of origin, kinetics of proliferation, routes
of migration and pattern of distribution of neurons. J Comp Neurol
Ramón y Cajal S (1911) Histologie du système nerveux de l’homme et
d’est vertébrès. Maloine, Paris. Reprinted (1955) by Consego Superior
de Investigaciones Cientificas, Madrid, Vol 2, p. 373.
[Sanderson KJ, Dixon PG, Pearson LJ (1982) Postnatal development of
retinal projections in the brushtailed possum, Trichosurus vulpecula.
Dev Brain Res 5:161–180.]
Sengelaub DR, Finlay BL (1981) early removal of one eye reduces
normally occurring cell death in the remaining eye. Science 213:573–574.
Shatz CJ, DiBerardino AC (1980) Prenatal development of the
theretinogeniculate pathway in the cat. Soc Neurosci Abst 6:485.
Sidman RL (1961) Histogenesis of mouse retina studied with thymidine–H3.
In The structure of the eye (Smelser GK, ed) pp 487–506. New York:
So KF, Schneider GE (1978) Postnatal development of retinogeniculate
projections in Syrian hamsters: An anterograde HRP study. Soc. Neurosci
Spear PD, Smith DC, Williams LL (197) Visual receptive–field properties
of single neurons in cat’s ventral lateral geniculate nucleus. J
Sterling P (1973) Quantitative mapping with the electron microscope:
retinal terminals in the superior colliculus. Brain Res 54:347–354.
Stone J (1978) The number and distribution of ganglion cells in the
cat’s retina. J Comp Neurol 180:753–772.
Stone J, Fukuda Y (1974) A comparison of W–cells with X– and Y–cells. J
Stone J, Rapaport DH, Williams RW, Chalupa LM (1981) Uniformity of cell
distribution in the ganglion cell layer of prenatal cat retina:
implications for mechanisms of retinal development. Dev Brain Res
Tigges J, Bos J, Tigges M (1977) An autoradiographic investigation of
the subcortical visual system in chimpanzee. J Comp Neurol 172:367–380.
Wässle H, Illing RB (1980) The retinal projection to the superior
colliculus in the cat: a quantitative study with HRP. J Comp Neurol
Williams RW, Chalupa LM (1980) Projections of retinal ganglion cells in
fetal cat. Soc Neurosci Abst 6:492.
Williams RW, Chalupa LM (1981) Prenatal development of retinal
projections to the midbrain of the cat. Invest Ophthalmol Vis Sci Suppl
[Williams RW, Chalupa LM (1984) Development of the retinal pathway to
the pretectum of the cat. Neuroscience 10:1249&3150;1267.]
Wilson ME, Toyne MJ (1970) Retino–tectal and cortico–tectal projections
in Macaca mulatta. Brain Res 24:395–406.
Received July 27, 1981; Revised December 14, 1981; Accepted
December 14, 1981
Since 11 August 98 Updated September 2, 2000