The quality of the foveal cone mosaic in human and primate retinas is a basic parameter of spatial vision function. The present study uses digital-texture analysis procedures to analyze the crystalline order of inner segment sections containing the rod-free portions of foveal cone mosaics. Definition of the cone cross-sectional centers made possible by adequate preprocessing allows precise mapping of lattice vertices and differentiation of hexagonal positions by procedures for direct neighbor recongnition.

In a further step, the existence of subunits within the hexagonal areas is revealed by the determination of axial orientation. The lattice of the subunits is characterized by similar orientation and high positional correlation of its hexagonal units.

The axial orientation of the areas differs from that of neighboring subunits by angular shifts of 10–15 deg and linear series of nonhexagonal irregularities demarcate the borders. Although larger patches with continuous hexagonal order occur in the surrounding rod-free regions, elevated degrees of disorder (30%) are found within the foveolar center (ca. 300 cones). Analysis of a mosaic showing labeled B cones (Szél et al., 1988) demonstrates that lattice disorder is in part associated with the blue cone subpopulation. The foveal mosaic from a glaucomatuous eye reveals severe lattice degradation throughout the rod-free zone, presumably due to extensive receptor loss.

The low-frequency superstructure results in local sets of sampling grids (5'–8') with differing orientational bias. Besides a horizontal/vertical difference of mosaic compression (ca. 1:1.15), the present analysis gives no hints for the existence of systematic meridional anisotropies at the receptor mosaic level. The study reveals a discontinuous organization of the foveal mosaic and points to possible sources for the induction and location of lattice disorder.

In the fish retina, retinomotor movement, spinule formation, and alteration of connexon density within gap junctions occur in response to changes in ambient light conditions. All of these morphological parameters can also be influenced by the application of dopamine. This study examines whether the morphological alterations of these structures are correlated with the activity of endogenous dopamine during an entrained 12-h light/12-h dark cycle and after 1-h sort-term adaptation periods.

The two measured parameters of retinomotor movement, cone inner segment length and pigment dispersion, were well-correlated with endogenous cyclic dopamine activity. However, retinomotor movement was initiated already at the end of the entrained dark period, before the onset of light and before the onset of dopamine turnover. Furthermore, a 1-h dark-adaptation period in the middle of the light phase reduced dopamine activity but did not affect retinomotor movement. At the switch from light to dark and after a 1-h light period at midnight retinomotor movement correlated exactly with dopamine turnover and illumination conditions. The formation of spinules was correlated with dopaminergic activity during all phases of the light/dark cycle and during short-term adaptation periods. Spinules were expressed in the light when dopamine activity was high and they were retracted when dopamine activity was reduced during darkness. Connexon density of horizontal cell gap junctions showed a weaker correlation with the endogenous dopamine turnover. In this case, a high activity of endogenous dopamine was paralleled by a high density of connexons.

Our results suggest that endogenous dopamine is involved in the cyclic regulation of the observed morphological alterations and that dopamine is part of the light signal for these mechanisms.

Calbindin-D28K and calretinin are homologous calcium-binding proteins localized in many neurons of the central nervous system. We have compared polyclonal antibodies against calbindin and calretinin and have shown by western blots using purified calbindin and calretinin from rat that (1) anti-calretinin does not recognize calbindin and (2) anti-calbindin presents some cross-reactivity with calretinin.

In this report, we have compared by immunohistochemistry the localization of both calcium-binding proteins in the retina of monkey, pig, sheep, rat, cat, pigeon, and salamander. These results are compared with previous data for chick. There are many differences between species and not within species, but some aspects of the distribution are conserved. All species, except rat and monkey, have some cones which contain calbindin only. Most species also have some bipolar cells containing calbindin only. Calretinin is rarely seen in photoreċeptors or bipolar cells. All species have horizontal cells which contain calretinin or calbindin or both. All species have amacrine cells and ganglion cells containing one or other protein.

In the cat ganglion cell layer, the calretinin antisera define a new, asymmetric, type of cell.

The ocular-following responses elicited by brief unexpected movements of the visual scene were studied in human subjects. Response latencies varied with the type of stimulus and decreased systematically with increasing stimulus speed but, unlike those of monkeys, were not solely determined by the temporal frequency generated by sine-wave stimuli. Minimum latencies (70–75 ms) were considerably shorter than those reported for other visually driven eye movements. The magnitude of the responses to sine-wave stimuli changed markedly with stimulus speed and only slightly with spatial frequency over the ranges used. When normalized with respect to spatial frequency, all responses shared the same dependence on temporal frequency (band-pass characteristics with a peak at 16 Hz), indicating that temporal frequency, rather than speed per se, was the limiting factor over the entire range examined. This suggests that the underlying motion detectors respond to the local changes in luminance associated with the motion of the scene. Movements of the scene in the immediate wake of a saccadic eye movement were on average twice as effective as movements 600 ms later: post-saccadic enhancement. Less enhancement was seen in the wake of saccade-like shifts of the scene, which themselves elicited weak ocular following, something not seen in the wake of real saccades. We suggest that there are central mechanisms that, on the one hand, prevent the ocular-following system from tracking the visual disturbances created by saccades but, on the other, promote tracking of any subsequent disturbance and thereby help to suppress post-saccadic drift. Partitioning the visual scene into central and peripheral regions revealed that motion in the periphery can exert a weak modulatory influence on ocular-following responses resulting from motion at the center. We suggest that this may help the moving observer to stabilize his/her eyes on nearby stationary objects.

The functions of the primate color-opponent and broad-band channels were assessed by examining the visual capacities of rhesus monkeys following selective lesions of parvocellular and magnocellular lateral geniculate nucleus, which respectively relay these two channels to the cortex. Parvocellular lesions impaired color vision, high spatial-frequency form vision, and fine stereopsis. Magnocellular lesions impaired high temporal- frequency flicker and motion perception but produced no deficits in stereopsis. Low spatial-frequency form vision, stereopsis, and brightness perception were unaffected by either lesion. Much as the rods and cones of the retina can be thought of as extending the range of vision in the intensity domain, we propose that the color-opponent channel extends visual capacities in the wavelength and spatial-frequency domains whereas the broad-band channel extends them in the temporal domain.

Spectral-sensitivity curves were derived from the a−, b−, and d−waves of rhesus monkey ERGs after injection of bicuculline, strychnine, or no drug. Without drug injection, the a− and d−wave curves were well-fit by an additive model of weighted photoreceptor absorption spectra, while the b−wave curve requires inhibitory terms to produce an adequate fit. Bicuculline, but not strychnine, reduced the weight assigned to the inhibitory terms in a dose-dependent fashion, to the point that no inhibition was evident. The results suggest that GABAergic synapses are required for the expression of red/green color opponency in primate bipolar cells.

Subtypes of bipolar cells were identified by protein kinase C (PKC)-like immunoreactivity in the goldfish retina. The PKC-like immunoreactivity was visualized by either the avidin/biotin peroxidase method or immunofluorescence method. In frozen cross sections and in wholemounts, the monoclonal antibody against α species of PKC reacted with ON-type bipolar cells, identified by the location of axon terminals in sublamina b of the inner plexiform layer. OFF-type bipolar cells (identified by the location of the axon terminal in sublamina a of the inner plexiform layer) were not immunoreactive. The immunoreactive cells consisted of two subtypes; the rod-dominant ON-type with a large soma and a large bulbous axon terminal, and the cone-dominant ON-type with a small soma and small axon terminal. Antibodies against β and γ species of PKC did not react with any bipolar cells. Of the isolated bipolar cells, enzymatically dissociated from the goldfish retina, 59% were immunoreactive to the monoclonal antibody against α species of PKC. The immunoreactive isolated cells also consisted of two morphological types. Each of them had a morphology typical either to rod-dominant ON-type or to cone-dominant ON-type.

We examined a white-eyed strain of the norpA mutant (norpA;cn bw) and white (w)norpA+ controls using microspectrophotometry (MSP), electron microscopy (EM), and electroretinography (ERG). These studies revealed that light mediates receptor demise in norpA even though norpA lacks phototransduction. Rhodopsin and the rhabdomere which houses it decrease with increasing age in norpA but not in w with rearing on a 12 h light/12-h dark cycle or in constant light. At higher temperature in norpA;cn bw and w reared in constant light, visual pigment decreases, rhabdomeres diminish, and cells die. Importantly, dark rearing blocked visual pigment loss in norpA;cn bw; the M-potential, an ERG reflection of visual pigment level, corroborated this finding. MSP showed that norpA's visual pigment loss was not due to acute loss of metarhodopsin, rhodopsin's photoproduct. NorpA blocks certain processes expected to be light elicited. The alteration of visual pigment as a function of time of day, present in w controls, is absent in white-eyed norpA, suggesting that light-induced depolarization may be necessary to entrain the rhythm. Microspectrofluorometry using the fluorescent dye, Lucifer yellow, suggested that norpA lacks a light-induced uptake mechanism; using control flies, we determined the stimulus parameters required for uptake in vivo. An attempt to “cure” norpA;cn bw by replacement “therapy” using phospholipase C, missing in norpA's phototransduction cascade, was largely unsuccessful.

The bidirectional axonal transport capabilities of the horseradish peroxidase (HRP) technique facilitated the study of the frontal-eye-field (FEF) input and pretectal output of two regions of extrastriate preoccipital cortex (POC). Following horseradish peroxidase (HRP) gel implants into the middle and dorsal POC in two rhesus monkeys, the middle POC implant demonstrated retrograde frontal cortical labeling largely restricted to the inferior frontal eye field (iFEF) and adjacent inferior prefrontal convexity, whereas the dorsal POC implant showed labeling in the caudal ventral bank of the superior ramus of the arcuate sulcus (sas) and middle-to-dorsal region of the rostral bank of the concavity of the arcuate sulcus (dorsal FEF). Prominent anterogradely labeled efferent preoccipital projections were observed to the ipsilateral pretectal olivary nucleus (PON) and to a lesser extent the anterior pretectal nucleus. Although the middle POC case had heavier projections to the lateral PON, the dorsal case projected more heavily to the medial PON. In addition, both implants demonstrated subcortical connections with the lateral and dorsal inferior pulvinar nuclei, central superior lateral thalamic intralaminar nucleus, caudate nucleus, and middle-to-ventral claustrum. However, while the middle POC implant had efferent projections to the superficial superior colliculus (SC), pregeniculate nucleus (PGN), lateral terminal accessory optic nucleus (LTN), and dorsolateral pontine nucleus (DLPN), resembling those previously reported for the middle temporal (MT) visual area (Maunsell & Van Essen, 1982; Ungerleider et al., 1984), the dorsal implant had projections to the lateral intermediate SC, zona incerta (ZI), PGN, a notably lesser projection to the LTN, and basilar pontine projections to the lateral and lateral dorsal pontine subnuclei (not including the extreme dorsolateral DLPN).

These preliminary results suggest that the preoccipital cortex, which reportedly functions in pupillary constriction, accommodation, and convergence, entertains connections with the PON and other visuomotorrelated structures, and thus could act as an intermediary in the pathway between the iFEF and PON, and provide a possible explanation for pupillary effects that occur with stimulation of the FEF (Jampel, 1960) and within the context of other oculomotor activities. The findings shed light on certain differences in connections of middle vs. dorsal POC with visuomotor-related nuclei, and appear to suggest that the middle region, which receives input from the iFEF, has greater access to the optokinetic (OKN) system by virtue of its projection to the LTN, and to the smooth-pursuit system by virtue of its projection to the DLPN.

The purpose of this investigation was to analyze the influence of stimulus orientation on the responses of individual neurons in the monkey's lateral geniculate nucleus (LGN). Our specific goals were to assess the prevalence and the degree of orientation tuning in the monkey LGN and to determine if the preferred stimulus orientations of LGN neurons varied as a function of receptive-field position. The primary motivation for this research was to gain insight into the receptive-field configuration of LGN neurons and consequently into the neural mechanisms which determine the spatial organization of LGN receptive fields in primates.

In both the parvocellular and magnocellular layers, the responses of the majority of individual neurons to sine-wave gratings varied as a function of stimulus orientation. The influence of stimulus orientation was, however, highly dependent on the spatial characteristics of the stimulus; the greatest degree of orientation bias was observed for spatial frequencies higher than the cell's optimal spatial frequency. On a population basis, the degree of orientation bias was similar for all major classes of LGN neurons (e.g. ON vs. OFF center; parvocellular vs. magnocellular) and did not vary systematically with receptive-field eccentricity. At a given receptive-field location, LGN neurons, particularly cells in the parvocellular laminae, tended to prefer either radially oriented stimuli or stimuli oriented more horizontally than their polar axis. Our analyses of the orientation-dependent changes in spatial-frequency response functions, which was based on the Soodak et al., (1987; Soodak, 1986) two-dimensional, difference-of-Gaussian receptive-field model, suggested that the orientation bias in LGN neurons was due to an elongation of the receptive-field center mechanism which in some cases appeared to consist of multiple subunits. Direct comparisons of the orientation-tuning characteristics of LGN cells and their retinal inputs (S potentials) indicated that the orientation bias in the monkey LGN reflects primarily the functional properties of individual retinal ganglion cells. We conclude that orientation sensitivity is a significant property of subcortical neurons in the primate's geniculo-cortical pathway.

The distribution patterns of GABA immunoreactive (+) and immunonegative (−) amacrine cell synapses and profiles in the inner plexiform layer (IPL) were analyzed in three macaque monkey retinas using postembedding electron-microscopic (EM) immunogold cytochemistry. Synapses and profiles were counted at 5% intervals throughout the IPL depth in three EM montages (total area = 6509 μm2), with 0% depth at the inner nuclear layer/IPL border. Nearly 70% of all amacrine synapses were GABA+, and they contacted all major classes of neurons that arborize in the IPL: bipolars (45%), ganglion cells (25%), and GABA+ (20%) and GABA− (10%) amacrines. A major relationship was seen between GABA+ amacrine processes and bipolar terminals: 76% of all amacrine-to-bipolar synapses were GABA+, and 82% of bipolar output dyads contained at least one GABA+ amacrine.

GABA+ amacrine profiles (N = 2455) were concentrated in three wide bands at IPL depths of 0–25%, 40–60%, and 75–100%, corresponding to the dense bands seen with light-microscopic immunocytochemistry. In contrast, GABA+ amacrine synapses (N = 1081) were distributed evenly throughout the IPL depth, rather than being confined to the three dense bands. GABA− amacrine synapses (N = 516) were concentrated at 40% and 60% depths.

Each category of amacrine output synapses had a characteristic pattern of stratification in the IPL. GABA+amacrine-to-bipolar synapses occurred throughout the IPL but were most frequent at 20% and 80% IPL depths, where the dendrites of midget cone bipolars arborize (Polyak, 1941). In contrast, GABA+amacrine-to-ganglion cell synapses were concentrated at 30% and 70% IPL depths, near the dendritic arborizations of parasol ganglion cells (Watanabe & Rodieck, 1989). GABA+ synapses onto bipolars and amacrines were also concentrated at the level of rod bipolar terminals. GABA+ amacrines must play significant but different roles in ON and OFF midget and parasol pathways as well as the rod pathway.

Ibotenic-acid lesions of the magnocelluar portion of the macaque lateral geniculate nucleus were used to examine the role of the M-cell pathway in spatio-temporal contrast sensitivity. A lesion was place in layer 1 of the lateral geniculate of each of two monkeys. Physiological mapping in one animal demonstrated that the visual-field locus of the lesion was on the horizontal meridian, approximately 6 deg in the temporal field. Visual thresholds were tested monocularly in the contralateral eye, and fixation locus was monitored with a scleral search coil to control the retinal location of the test target.

Three threshold measures were clearly disrupted by the magnocellular lesions. Contrast sensitivity for a 1 cycle/deg grating that drifted at 10 Hz was reduced from about twofold greater than, to about the same as, that for 10-Hz counterphase modulated gratings. Sensitivity for a very low spatial frequency (Gaussian blob), 10-Hz flickering stimulus was reduced so severely that no threshold could be measured. In addition, flicker resolution was greatly reduced at lower modulation depths (0.22), but not at higher depths (1.0). Two of the measured thresholds were unaffected by the lesions. Contrast sensitivity for 2 cycle/deg stationary gratings remained intact, and little or no effect on sensitivity was found for 1 cycle/deg, 10-Hz counterphase modulated gratings.

Together, these results suggest that the magnocellular pathway makes little contribution to visual sensitivity at low to moderate temporal frequencies. On the other hand, some contribution to detection sensitivity is evident at lower spatial and high temporal frequencies, especially for drifting stimuli. It appears that a major role of the magnocellular pathway may be to provide input to cortical mechanisms sensitive to rapid visual motion.

Like many diurnal raptors, the American kestrel or sparrow hawk, Falco sparverius, possesses two foveae in each eye. In this study, we used fundus photography and reversed ophthalmoscopy to plot the projection of these foveae onto the visual field, together with other retinal landmarks such as the pecten and ora terminali. From such data, it was determined that the central monocular fovea (CMF) and temporal binocular fovea (TBF) were separated by 36 deg and that the kestrel has 58 deg of binocular overlap when the eyes are appropriately converged. Single-cell and multi-unit recordings were used to systematically explore the tectal surface and map receptive fields (RF) onto a hemispheric screen. A retinotopic map of the tectal surface was produced from such data and revealed an expanded representation for each fovea on the tectum and a systematic increase in RF size from fovea to periphery. The functional significance of this organization is discussed.

Recordings were taken from infragranular layers of area 17 of anesthetized monkeys with an array of 30 microelectrodes matching about one hypercolumn. From intracortical spike-train correlations, the novel neuronal property “delay scale position” related to retino-cortical delays, was derived. Relationships were established to the degree of spike isolation and to classical response properties. Direction selectivity, spike rate, spike-isolation quality, delay scale, and color selectivity could be linked to an underlying factor upon which the latter variables depend in a fixed way. Neurons with similar factors were characterized by non-delayed correlations. The link was more strict in layer VI than in layer V, and it was related to the parvo/magnocellular subdivision of the visual system.

Accommodation is more accurate with polychromatic stimuli than with narrowband or monochromatic stimuli. The aim of this paper is to develop a computational model for how the visual system uses the extra information in polychromatic stimuli to increase the accuracy of accommodation responses. The proposed model is developed within the context of both trichromacy and also the organization of spatial and chromatic processing within the visual cortex.

The refractive error present in the retinal image can be estimated by comparing image quality with and without small additional changes in refractive state. In polychromatic light, the chromatic aberration of the eye results in differences in ocular refractive power for light of different wavelengths. As a result, the refractive state of the eye can be estimated by comparing image quality in the three types of cone photoreceptor. The ability of cortical neurons to perform such comparisons on image quality with a crude form of spatial-frequency analysis is examined theoretically. It is found that spatially band-pass chromatically opponent neurons (that may correspond to double opponent neurons) can perform such calculations and that chromatic cues to accommodation are extracted most effectively by neurons responding to spatial frequencies of between 2 and 8 cycles/deg.

Specific antisera, raised in rabbits, against somatostatin 1-14, somatostatin 1-28, the fragment 1-12 of somatostatin 1-28, and prosomatostatin 20-36 were used for immunohistochemistry and gel filtration of the rat retina.

With all antisera, immunoreactive perikarya could be located in the inner nuclear and ganglion cell layers. In the inner nuclear layer, amacrine cells with processes extending predominantly into the first sublayer of the inner plexiform layer were observed. Some processes extended also to the ganglion cell layer. In addition, somatostatin-immunoreactive interplexiform cells were present in the inner nuclear layer.

In the ganglion cell layer, perikarya were found located in the midperiphery and in the far periphery of the retina. The neurons located in the midperiphery of the retina possessed a round perikaryon from which processes could be followed going into the inner plexiform layer, where they dichotomized in the third and first sublayers. The perikarya in the far periphery of the retina near the ora serrata exhibited an ovoid-shaped cell body from which processes extended horizontally in a bipolar manner in the layer itself.

By use of an [35S]-labeled antisense oligonucleotide probe, in situ hybridization of the rat retina showed the presence of perikarya in the inner nuclear layer and ganglion cell layer containing mRNA encoding for prosomatostatin.

Gel filtration of the retinal extracts followed by radioimmunoassay showed the presence of somatostatin 1-14, the fragment 1-12 of somatostatin 1-28, and prosomatostatin 1-64. However, somatostatin 1-28 was not detected.

The results obtained in this study verify the presence of somatostatin 1-14 in the rat retina located in perikarya and processes in the inner nuclear and ganglion cell layers. The positive in-situ hybridization signals show that the intraneuronal somatostatin immunoreactivity is due to synthesis of the peptide and not uptake in the neurons. The presence of the somatostatin propeptide and fragments of this propeptide, in both intraretinal perikarya and fibers, indicate a posttranslational modification of this neuropeptide in the perikarya and the processes as well.

In order to evaluate the normal eye movements of the turtle, Pseudemys scripta elegans, the positions of each eye were recorded simultaneously using two search-coil contact lenses. Optokinetic nystagmus (OKN) was strikingly unyoked in this animal such that one eye's slow-phase velocity was substantially independent of that of the other eye. On the other hand, the fast-phase motions of both eyes occurred more or less in synchrony.

An eye's slow-phase gain is primarily dependent on the direction and velocity of the stimulus to that eye. Using monocular stimuli, the highest mean gain (0.54 ± 0.047; mean ± standard error of mean) occurred using temporal-to-nasal movement at 2.5 deg/s. The mean OKN gain for nasal-to-temporal movement was only 0.13 ± 0.015 at that velocity. Additionally, using the optimal monocular stimulus (temporal-to-nasal stimulation at 2.5 deg/s) only drove the occluded eye to move nasal-to-temporally at 0.085 deg/s, equivalent to a “gain” of only 0.034 ± 0.011.

The binocular OKN gain during rotational stimuli was higher than monocular gain, especially during nasal-to-temporal movement at high velocities. Also the difference in slow-phase eye velocity between the two eyes was smaller during binocular rotational stimuli. In contrast, when each eye simultaneously viewed its temporal-to-nasal stimulus at an equal velocity, two behaviors were observed. Often, OKN alternated between an animal's left eye and right eye. Occasionally, both eyes moved at equal but opposite velocities.

These behavioral data provide a quantitative baseline to interpret the properties of the retinal slip information in the turtle's accessory optic system. Those properties are similar to the behavior of the turtle in that both are tuned to direction and velocity independently for each eye (Rosenberg & Ariel, 1990).

A series of psychophysical tests examining early and later aspects of image-motion processing were conducted in a patient with bilateral lesions involving the posterior visual pathways, affecting the lateral parietal-temporal-occipital cortex and the underlying white matter (as shown by magnetic resonance imaging studies and confirmed by neuro-ophthalmological and neuropsychological examinations). Visual acuity, form discrimination, color, and contrast-sensitivity discrimination were normal whereas spatial localization, line bisection, depth, and binocular stereopsis were severely impaired. Performance on early motion tasks was very poor. These include seeing coherent motion in random noise (Newsome & Paré, 1988), speed discrimination, and seeing two-dimensional form from relative speed of motion. However, on higher-order motion tasks the patient was able to identify actions from the evolving pattern of dots placed at the joints of a human actor (Johansson, 1973) as well as discriminating three-dimensional structure of a cylinder from motion in a dynamic random-dot field. The pattern of these results is at odds with the hypothesis that precise metrical comparison of early motion measurements is necessary for higher-order “structure from motion” tasks.

Using an antibody against serotonin (5-hydroxytryptamine, 5-HT), serotonin-like immunoreactive (serotonin IR) neurons were demonstrated in the retina of adult Bufo marinus. All immunoreactive neurons were identified as amacrine cells (ACs). The dendrites of serotonin-IR ACs branched diffusely and densely throughout all levels of the inner plexiform layer (IPL) of the retina. The great majority of these cell somata were located in the vitread part of the inner nuclear layer (INL) and a few of them (ranging from 9–29 cells) were displaced into the ganglion cell layer (GCL). On the basis of the soma sizes, two populations of serotonin-IR ACs, large (type A) and small (type B), were distinguished. 6-Hydroxydopamine (6-OHDA) injected into the eye abolished immunoreactivity in the recently reported tyrosine hydroxylase (TH)-IR ACs (Zhu & Straznicky, 1990), whereas serotonin-IR ACs remained unaffected.

The number of serotonin-IR cells per retina ranged from 23,750–27,390, with a ratio of 1:1.6 to 1:1.9 between type A and B cells. Both cell types were distributed nonuniformly across the retina. Cell densities were slightly lower in the peripheral (96 cells/mm2) than in the central (164 cells/mm2) retina. Linear regression analysis confirmed the presence of a decreasing density gradient from the retinal center to the retinal margin for both small and large cell types. The analysis of the nearest neighbor distances showed that the retinal distribution of serotonin-IR ACs was orderly.

These results have been taken to indicate that 5-HT-IR cells correspond to a population of serotonincontaining ACs. It is suggested that dopamine and serotonin are contained in two different populations of ACs in the rtina of Bufo marinus.

Many studies on human retinal development have cited the third gestational month as a period when the posterior retina undergoes rapid differentiation and maturation, including a lining up of cone precursors. Ultrastructural data on the posterior retina during the third month are very limited, and totally lacking for the cone monolayer. We have examined two human fetal retinas between ten and 11 gestational weeks. Before the appearance of the cone monolayer, the outer neural retina consists of a homogeneous population of undifferentiated neuroblasts. Mitotic figures are still evident, even posteriorally. There is no outer plexiform layer (OPL). The interface of neural retina to retinal pigment epithelium (RPE) is largely featureless. By 11 weeks, the posterior retina has a thin OPL that separates the many rows of cells in the developing inner nuclear layer from the single tier of macular cone precursors. The RPE monolayer consists of cuboidal cells whose apical surface elaborates ridges of cytoplasm and branched processes that project into the subretinal space. The large, cuboidal cones are linked to each other and Müller cells at the outer limiting membrane. They show definitive signs of the structural polarity typical of vertebrate photoreceptors. Their apical cytoplasm contains many organelles common to the inner segment, while the basal cytoplasm has synaptic ribbons and vesicles, and receives invaginating contacts from processes in the OPL neuropil arising from differentiating second-order neurons. Lateral cone surfaces are mutually underlain by large subsurface cisterns.