A postembedding immunogold procedure was used on thin sections of the dorsal lateral geniculate nucleus (LGN) and perigeniculate nucleus (PGN) of the cat to estimate qualitatively and quantitatively, at the electron-microscopic (EM) level, the intensity of glutamate or aspartate immunoreactivities on identifiable synaptic terminals and other profiles of the neuropil. On sections incubated with a glutamate antibody, terminals of retinal and cortical axons in the LGN, and of collaterals of geniculo-cortical axons in the PGN, contain significantly higher density of immunogold particles than GABAergic terminals, glial cells, dendrites, and cytoplasm of geniculate cells. By contrast, in sections incubated with an aspartate antibody, terminals of retino-geniculate, cortico-geniculate, and geniculo-cortical axons did not show a selective enrichment of immunoreactivity, but instead the density of immunogold particles was generally low in the different profiles of the neuropil, with the exception of nucleoli. These results suggest that glutamate, but not aspartate, is a neurotransmitter candidate in the retino-geniculo-cortical pathways.

Direction-selective neurons in cat striate cortex were tested with bar-shaped stimuli, sequentially flashed at spatially displaced positions chosen to elicit maximal direction selectivity. Temporally overlapping flash exposures of prolonged duration (400–1000 ms) were employed at a series of onset asynchronies to explore the nature of temporal tuning of the direction-selective mechanism. In most neurons studied, direction selectivity was found to be supported by a surprisingly broad range of stimulus onset asynchronies, which was greater for longer exposure durations. These findings imply the existence of a sustained input to the direction-selective mechanism, in spite of the relatively transient nature of most cortical neurons' step responses. A model is described to illustrate how different front-end temporal filters can affect the dependence of two-flash direction selectivity on stimulus onset asynchrony. The versions of the model which successfully predict the form of the observed responses are those which combine inputs from sustained and transient filters.

We have recorded steady-state PERGs from five macaque monkeys in response to red-green plaid patterns reversed sinusoidally in contrast. The patterns had either a pure luminance contrast (red-black, green-black, yellow-black), pure red-green color contrast, or a variable amount of luminance and color contrast. By varying the relative luminance of the red-to-total luminance (color ratio) of red-green patterns, a value could be obtained at which the PERG amplitude was either minimum or locally maximum, and the phase was most lagged. This value was very similar to that producing equiluminance in human observers, and was considered to be equiluminance for the monkey. The phase of the PERG to chromatic stimulus was systematically lagged compared with that to luminance stimuli, by an amount corresponding to about 10–20 ms under our experimental conditions. The variation of phase with temporal frequency suggested an apparent latency of about 80 ms for color contrast compared with 63 ms for luminance. These estimates were confirmed with separate measurements of transient PERGs to abrupt contrast reversal. As a function of temporal frequency, the chromatic PERG function was clearly low-pass with a cutoff around 15 Hz, whereas that to luminance was double-peaked and extended to higher temporal frequencies, around 30 Hz. For both luminance and chromatic stimuli, the amplitude of PERGs increases with increasing stimulus contrast. By summing vectorially the luminance and chromatic responses of appropriate contrasts, we were able to predict with accuracy the response as a function of color ratio. In two monkeys, the optic chiasm was sectioned sagittally causing total degeneration of ganglion cells in the nasal retina, without affecting the temporal retina (verified by histology). In these animals, there was a strong response to both luminance and chromatic patterns in the temporal retinae, but none to either type of pattern in the nasal retinae, suggesting that the PERG to both luminance and chromatic stimuli arises from the inner-retinal layers. Electrophysiological studies suggest that the PERG to chromatic stimuli is probably associated with the activity of P-cells. P-cells may also make a major contribution to the PERG of luminance stimuli, although M-cells may also participate. The above findings on normal monkeys all agree with those reported in the accompanying paper for humans (Morrone et al., 1994), so similar conclusions can probably be extended to human PERG.

Polyclonal (7493 and 7317) and monoclonal (mAb3) antibodies, generated to the α subunit of the voltage-gated sodium channel (αNaCh), were employed to assess the cell types containing αNaCh-like immunoreactivity in the mature cat and monkey retina. Immunoblot analyses of retinal proteins in the cat revealed that the polyclonal and monoclonal antibodies we employed labeled a band in the 260–kDa region which corresponds to the molecular mass of the α subunit of the NaCh. In both the cat and monkey, these antibodies immunolabeled several distinct types of retinal cells. With the polyclonal antibodies immunoreactivity was observed in ganglion cells and their intraretinal axons, in horizontal cells, and unexpectedly, in cones. In addition, in both species, a limited number of heavily labeled profiles, presumed to be bipolar cells, were seen in the inner nuclear layer. In cat and monkey the monoclonal antibody labeled axons in the fiber layer, ganglion cell somata, and a continuous band of immunoreactive cell bodies (presumed bipolar cells) situated in the outer half of the inner nuclear layer. By immunolabeling isolated cells dissociated from the cat retina, it was possible to demonstrate unequivocally that a population of bipolar cells was labeled by the monoclonal and the polyclonal antibodies we employed. The differences in the labeling observed with the monoclonal antibody as compared to the polyclonal antibodies were interpreted as reflecting the presence of different α-subunit subtypes in the mammalian retina. Collectively, our findings suggest that αNaCh-like proteins are expressed by a more diverse population of retinal cells than expected on the basis of previous physiological and immunohistochemical studies.

Visually guided saccadic eye movements to singly presented stationary targets form a bimodal distribution. After superior colliculus lesions, the so called “express saccades” that form the first mode of the distribution are no longer obtained. The aim of this study was to determine what role several other neural systems play in the generation of express and regular saccades, with the latter being those that form the second mode in the bimodal distribution. Lesions were made in the parvocellular and magnocellular portions of the lateral geniculate nucleus to disrupt either the midget system or the parasol system that originates in the retina and areas V4 and MT. The effects of the lesions were examined on the accuracy and latency of saccadic eye movements made to stationary and to moving visual targets. Following magnocellular and MT lesions deficits were observed in smooth pursuit and in the amplitude of saccades made to moving targets. However, none of the lesions produced significant changes in the bimodal distribution of saccadic latencies to stationary targets. The results suggest that express saccades and regular saccades are not selectively mediated by either the midget or the parasol systems or by areas V4 and MT. Neither are the frontal eye fields involved as had previously been shown. We suggest that the superior colliculus plays a central role in producing both express and regular saccades by virtue of highly convergent input from numerous cortical structures.

We have studied the responses of MT neurons to moving gratings, spatially modulated in luminance and chromaticity. Most MT neurons responded briskly and with high contrast sensitivity to targets whose luminance was modulated, with or without added chromatic contrast. When luminance modulation was removed and only chromatic stimulation was used, the responses of all MT neurons were attenuated. Most were completely unresponsive to stimulation with targets whose modulation fell within a “null” plane in color space; these null planes varied from neuron to neuron, but all lay close to the plane of constant photometric luminance. For about a third of the neurons, there was no color direction in which responses were completely abolished; almost all of these neurons had a definite minimum response for chromatic modulation near the isoluminant plane. MT neurons that responded to isoluminant targets did so inconsistently and with poor contrast sensitivity, so that only intensely modulated targets were effective. Whereas the best thresholds of MT neurons for luminance targets are close to behavioral contrast threshold, the thresholds for isoluminant targets lie considerably above behavioral contrast threshold. Therefore, although some MT neurons do give responses to isoluminant targets, they are unlikely to be the source of the chromatic motion signals revealed behaviorally.

Enzymatic treatments that facilitated whole-cell electrophysiological recordings were used on Limulus ventral photoreceptor cells. Ventral optic nerves were treated with either collagenase or collagenase, papain, and trypsin. Either treatment greatly increased the ease of making whole-cell recordings of transmembrane potentials. Light responses obtained from enzyme-treated photoreceptor cells were nearly identical to results obtained without enzyme treatment and compared favorably to in vivo recordings of light responses from the compound lateral eye. Enzyme-treated cells also responded to applied octopamine, as do untreated cells, with an increased phosphorylation of a 122-kD protein. This suggests that the external receptors and internal biochemical machinery required for at least one second-messenger cascade are present after enzyme treatment. The morphological integrity of enzyme-treated photoreceptor cells was examined with light microscopy as well as with scanning and transmission electron microscopy. In general, we found that each enzyme treatment greatly reduced the integrity of the layers of glial cells that surround the photoreceptor cells thereby making these cells easily accessible for whole-cell recordings of transmembrane potentials. The morphology of the rhabdomere was normal after enzymatic degradation of the adjacent glial covering.

Recent investigations have shown that cortical nerve growth factor (NGF) infusions during the critical period inhibit ocular-dominance plasticity in the binocular portion of the rat visual cortex. The mechanisms underlying the effects of NGF on visual cortical plasticity are still unclear. To investigate whether during normal development intracortical and/or extracortical cells possess uptake/transport mechanisms for the neurotrophin, we injected 125I-NGF into the occipital cortex of rats at different postnatal ages. Within the cortex, only a few labelled cells were observed. These cells were confined to the vicinity of the injection site and their number depended on the animal's age at the time of injection. Labelled cells were absent at postnatal day (PD) 10 but could be detected between PD 14 and PD 18. They then decreased in number over the following period and were not detected in adult animals. Outside the cortex, neurons of the lateral geniculate nucleus (LGN) were not observed to take up and retrogradely transport NGF at any age after birth. In contrast, retrogradely labelled neurons were found in the basal forebrain. Labelled cells were first observed here at PD 14 and then increased in number until reaching the adult pattern. Our results show that intrinsic and extrinsic neurons are labelled following intracortical injections of iodinated NGF. In both neuronal populations, the uptake and transport of NGF is present over a period corresponding to the critical period for visual cortical plasticity. These findings suggest that NGF may play a role, both intra and extracortically, in plasticity phenomena.

It is well established that cortical neurons frequently show different preferred drift directions for random dots and gratings. Dot stimuli often produce two preferred directions which are arranged symmetrically on either side of the preferred directions for gratings. Based on their filter properties in three-dimensional (3-D) Fourier space and on the 3-D power spectra of drifting dot patterns, we estimated the optimal direction to drifting dots for ten neurons in the striate cortex of five adult cats. These estimates frequently gave two optimal directions, one on either side of the optimal direction to gratings. The angle between the two estimated peaks increases with drift speed. Predicted and actual angles were in reasonably good agreement. We conclude, therefore, that the directional selectivity of cortical neurons to drifting random dot patterns can be understood from linear filtering properties. For this reason, the directional tuning to drifting dot patterns seems to reflect the same mechanisms that mediate the responses to sinusoidal gratings and do not require a separate directional mechanism.

The effect of unilateral depletion of retinal dopamine on goldfish visual behavior was studied using a behavioral reflex, the dorsal light reaction (DLR). Retinal dopamine was depleted by intraocular injections of 6–hydroxydopamine (6–OHDA) on two successive days. By 2 weeks postinjection, dopamine interplexiform cells (DA-IPC) were not detected using tyrosine-hydroxylase immunoreactivity (TH-IR). By 6 weeks postinjection, generation of DA-IPC was observed at the marginal zone and by 9 months postinjection, 2–3 rows of DA-IPC were present at the marginal zone. Neurites extended several hundred micrometers toward the central retina. By 2 weeks postinjection, all 6–OHDA lesioned fish tilted 7–15 deg toward the injected eye under uniform overhead illumination. The tilting did not occur under scotopic illumination and reappeared within 1 min of light adaptation. Quantitation of the DLR showed that 6–OHDA lesioned fish behaved as if light were 2.4 log units more intense to the injected eye. Partial recovery was observed by 9 months postinjection, paralleling the reappearance of DA-IPC at the marginal zone. Tilting also was induced by unilateral intraocular injection with Dl and D2 dopamine receptor antagonists, SCH 23390 and S(—)-sulpiride, respectively. Fish did not tilt if they were light adapted at the time of injection. Tilting was observed if the animals were dark-adapted for 3 h and left in the dark for 1 h postinjection. Fish tilted toward the drug-injected eye within 2 min of light adaptation and gradually returned to vertical within 2 h. The tilting response to S(—)-sulpiride was stronger (˜20 deg vs. ˜5 deg) and occurred at lower concentration (1 μM vs. 10 μM) compared to SCH 23390. We conclude that dopamine depletion mimics the dorsal light reaction by increasing the luminosity output of the eye and that dopamine is directly involved in photopic luminosity function.

The effect of light adaptation on the period of photocurrent saturation induced by a bright stimulating flash was examined in rod photoreceptors of the larval-stage tiger salamander (Ambystoma tigrinum). Using suction electrodes, photocurrent responses to brief flashes were recorded from single, isolated rods in the presence and absence of steady background illumination. Background light decreased the saturation period (T) measured at fixed flash intensity (fixed If) and in this respect light-adapted the saturating response. Effects of the background on responses to weak (i.e. subsaturating) and bright flashes were compared with changes in a parameter, where ΔT is the decrease in saturation period, and where TR* is the slope of the line that relates T and ln If in a given state of adaptation. Dark- and light-adapted responses to flash intensities and , respectively, exhibited similar absolute peak photocurrent and falling-phase kinetics when and satisfied the relation, , where Ib is the background intensity. It is argued that ψ approximates the relative PDE*/R* gain of transduction, i.e. the relative peak level of activated cGMP phosphodiesterase (PDE*) produced by a given, small amount of photoactivated visual pigment (R*). Interpreted on this view, the results imply that light adaptation derives largely from a decrease in PDE*/R* gain, rather than from the stimulation of guanylate cyclase activity. The data are consistent with the possibility that modulation of the lifetime of PDE* underlies the background dependence of ψ.

Increment threshold functions of the electroretinogram (ERG) b–wave were obtained from goldfish using an in vivo preparation to study intraretinal mechanisms underlying the increase in perceived brightness induced by depletion of retinal dopamine by 6–hydroxydopamine (6–OHDA). Goldfish received unilateral intraocular injections of 6–OHDA plus pargyline on successive days. Depletion of retinal dopamine was confirmed by the absence of tyrosine-hydroxylase immunoreactivity at 2 to 3 weeks postinjection as compared to sham-injected eyes from the same fish. There was no difference among normal, sham-injected or 6–OHDA-injected eyes with regard to ERG waveform, intensity-response functions or increment threshold functions. Dopamine-depleted eyes showed a Purkinje shift, that is, a transition from rod-to-cone dominated vision with increasing levels of adaptation. We conclude (1) dopamine-depleted eyes are capable of photopic vision; and (2) the ERG b–wave is not diagnostic for luminosity coding at photopic backgrounds. We also predict that (1) dopamine is not required for the transition from scotopic to photopic vision in goldfish; (2) the ERG b–wave in goldfish is influenced by chromatic interactions; (3) horizontal cell spinules, though correlated with photopic mechanisms in the fish retina, are not necessary for the transition from scotopic to photopic vision; and (4) the OFF pathway, not the ON pathway, is involved in the action of dopamine on luminosity coding in the retina.

Vasoactive intestinal polypeptide (VIP) has been shown to potentiate current responses elicited by activation of the GABAA receptor (IGABA) in freshly dissociated ganglion cells of the rat retina. Here we tested the hypothesis that this heteroreceptor cross talk is mediated by an intracellular cascade of events that includes the sequential activation of a stimulatory guanine nucleotide binding (Gs) protein and adenylate cyclase, the subsequent increase in levels of cyclic AMP and, finally, the action of the cyclic AMP-dependent protein kinase (PKA). Intracellular dialysis of freshly dissociated ganglion cells with GTPγs irreversibly potentiated IGABA, while GDPßs either decreased or had no effect on IGABA. Additionally, GDPßs blocked the potentiation of IGABA by VIP. Cholera toxin rendered VIP ineffective in potentiating IGABA, while pertussis toxin had no effect on the VIP-induced potentiation of IGABA. Extracellular application of either forskolin or 8-bromo-cyclic AMP potentiated IGABA, as did the introduction of cyclic AMP directly into the intracellular compartment through the recording pipet. Intracellular application of cyclic AMP-dependent protein kinase (PKA) potentiated IGABA, while a PKA inhibitor blocked the potentiating effect of VIP. These results lead us to conclude that activation of a cyclic AMP-dependent second-messenger system mediates the modulation of GABAA receptor function by VIP in retinal ganglion cells.

Cone photoreceptors in the turtle retina are involved in intricate neuronal interactions with other retinal neurons that modify the responses of the cones to photons absorbed in their outer segments. Therefore, the action spectra of cones strongly depend upon the conditions of measurements. This study describes an attempt to derive the action spectra of turtle cones which are the least distorted by neuronal interactions. To achieve this goal, the photoresponses of cones and horizontal cells were recorded from the turtle retina under different conditions of adaptation using different patterns of the stimulating test flashes. The sensitivity action spectra, derived from small-amplitude (<1 mV) photoresponses, were strongly affected by the recording conditions indicating the contributions of multiple neuronal inputs. Action spectra, constructed from large criterion photoresponses, were less distorted by neuronal interactions and better described the spectral properties of the “isolated” cones. The action spectra of the hyperpolarizing inputs to chromaticity-type horizontal cells were derived by stimulating these cells with mixtures of a saturating red light and a monochromatic light of different wavelength and intensity. The action spectra were constructed from the intensity of the addend component needed to “pull down” the depolarizing response to the red component by a fixed criterion. These spectra, measured in red/green and yellow/blue C-type horizontal cells, are suggested to best represent the “isolated” M-cones and S-cones, respectively.

In human subjects, the blind spot is perceptually filled-in by color and brightness from the surrounding visual field. The present behavioral study examined the occurrence of color filling-in at the blind spot in monkeys. First, the location of the blind spot was determined using a monocular saccade task. The blind spots were located on the horizontal meridian at approximately 15–17 deg from the fixation point in the temporal visual field. Then, filling-in at the blind spot was tested by determining if the monkey could discriminate between an annulus presented on the blind spot and a homogeneous disk in the normal visual field. In this task, the monkey was required to make a saccade to a homogeneous disk of the same color and size as an annulus presented simultaneously in the opposite field. Both stimuli were large enough to cover the blind spot and the inner circle of the annulus was confined inside the blind spot. All four monkeys tested performed this task correctly in over 80% of the trials. However, when one eye was covered and the annulus was presented on the blind spot of the uncovered eye, performance deteriorated significantly. To confirm that these results reflected filling-in, one monkey was trained to maintain fixation when two identical homogeneous disks appeared in opposite visual fields. When only one eye was uncovered, and the annulus was presented on the blind spot of the uncovered eye, the monkey maintained fixation in most of the trials. These results show that monkeys were unable to distinguish an annulus from a homogeneous disk when the annulus was presented on the blind spot. This indicates that color filling-in occurs at the blind spot in monkeys and opens possibility to physiological experiments to study the neural mechanisms of filling-in.

An important issue in understanding the function of primary visual cortex in the macaque monkey is how the several efferent neuron groups projecting to extrastriate cortex acquire their different response properties. To assist our understanding of this issue, we have compared the anatomical distribution of VI intrinsic relays that carry information derived from magno- (M) and parvocellular (P) divisions of the dorsal lateral geniculate nucleus between thalamic recipient neurons and interareal efferent neuron groups within area VI. We used small, iontophoretic injections of biocytin placed in individual cortical laminae of area VI to trace orthograde and retrograde inter- and intralaminar projections. In either the same or adjacent sections, the tissue was reacted for cytochrome oxidase (CO), which provides important landmarks for different efferent neuron populations located in CO rich blobs and CO poor interblobs in laminae ⅔, as well as defining clear boundaries for the populations of efferent neurons in laminae 4A and 4B. This study shows that the interblobs, but not the blobs, receive direct input from thalamic recipient 4C neurons; the interblobs receive relays from mid 4C neurons (believed to receive convergent M and P inputs), while blobs receive indirect inputs from either M or P (or both) pathways through layers 4B (which receives M relays from layer 4Cα) and 4A (which receives P relays directly from the thalamus as well as from layer 4Cβ). The property of orientation selectivity, most prominent in the interblob regions and in layer 4B, may have a common origin from oriented lateral projections made by mid 4C spiny stellate neurons. While layer 4B efferents may emphasize M characteristics and layer 4A efferents emphasize P characteristics, the dendrites of their constituent pyramidal neurons may provide anatomical access to the other channel since both blob and interblob regions in layers ⅔ have anatomical access to M and P driven relays, despite functional differences in the way these properties may be expressed in the two compartments.

Two cDNA clones for nicotinic acetylcholine receptor (nAChR) subunits sensitive to α-bungarotoxin (α-Bgt) have been isolated, the so-called α-Bgt binding proteins α1 (or α7 nAChR subunit) and α2 (or α8 nAChR subunit). Immunohistochemical experiments have shown that both α7 and α8 subunits, as well as subunits insensitive to α-Bgt (β2 and α3), are present in amacrine and ganglion cells of the chick retina. However, only the α8 subunit was observed in presumptive bipolar cells. The present study investigated in detail the pattern of distribution of the bipolar cells containing the α8 nAChR subunit and its relation to the pattern of distribution of bipolar cells immunoreactive to protein kinase C (PKC). Presumptive α8-and PKC-like immunoreactive (α8-LI and PKC-LI) bipolar cells were observed sending their dendrites to the outer plexiform layers and their axons to the inner plexiform layer. Where as α8-LI bipolar cells corresponded to 40–53% of the whole population of bipolar cells, PKC-LI bipolar cells represented only 6–8% of the same population. The soma sizes of the α8-LI bipolar cells were slightly smaller (mean ± s.d.; 4.9 ± 0.8 μm) than the soma sizes of the PKC-LI bipolar cells (5.4 ± 0.9 μm). Double-labeling experiments indicated that probably all PKC-LI bipolar cells also contain α8-LI. This indicates that two distinct groups of cholinoceptive bipolar cells exist in the chick retina, one that contains PKC-LI, and another one that does not.

We have studied, by immunocytochemistry, the ontogeny of GABA, glycine, glutamate, glutamine, and taurine-containing cells in the rabbit retina. Amacrine cells show GABA immunoreactivity by embryonic day 25 (E25) and throughout postnatal life. By contrast, ganglion cells and horizontal cells are only transiently GABA-immunoreactive (-IR); few appear GABA-IR by the third postnatal week. At maturity, glycine is present in amacrine cells and in some bipolar cells. During development, putative ganglion cells transiently contained glycine between E25 and postnatal day 3 (P3), whereas immunolabelling in presumed amacrine cells and bipolar cells persists after birth. Ganglion cells, bipolar cells, photoreceptors, and some amacrine cells are glutamate-IR in the adult retina. Glutamate immunoreactivity first appears in the somata and processes of cytoblastic cells by E20 and is prominent by E25. Surprisingly, ganglion cells are not strongly glutamate-IR until just before eye-opening, at postnatal day 10 (P10), coincident with the appearance of glutamine in their somata and in Müller glial cells. Bipolar cells are glutamate-IR before they or Müller cells contain high levels of glutamine (at P10). Glutamate immunoreactivity in photoreceptors is progressively restricted to the inner segments by eye-opening. At no stage are presumed horizontal cells glutamate-IR or glutamine-IR, but some amacrine cells show glutamate- and glutamine-IR by P10. Taurine is localized to photoreceptors and Müller glial in the adult retina. Some cytoblasts are taurine-IR at E20; with ensuing development, taurine labelling becomes restricted primarily to Müller cells and photoreceptors; some putative bipolar cells may also be labelled. However, for a few days around birth, cells resembling horizontal cells, also show taurine immunoreactivity. The early appearance and often transient expression of these amino acids in retinal cells suggests that these neuroactive molecules may be involved in the structural and functional development of the retina.

We have recorded steady-state visual evoked potentials (VEPs) from patients with vascular damage to their right brain hemispheres, some suffering from unilateral spatial neglect (n = 9), and some not (n = 7). VEPs were recorded in response to sinusoidal gratings of 0.56 cycle/deg contrast-reversed sinusoidally at temporal frequencies from 4–11 Hz. Stimuli were presented either to the left or to the right visual field, or to both. Confirming previous reports, reliable VEPs were recorded from stimuli in the left contralesional hemifield, of comparable amplitude to those of the ipsilesional hemifield and to those of both hemifields of brain damaged patients without neglect. However, analysis of apparent latency derived from phase data showed that the VEPs from the contralesional hemifield were systematically delayed by 30–40 ms compared with those of the ipsilesional hemifield, and compared with both hemifields of the nonneglect groups. This result suggests changes in neural processing in neglect patients.

The content of taurine and the immunocytochemical localization of taurine and γ-aminobutyric acid (GABA) in the rat retina during postnatal development are described. The rat retina is immature at birth; about two-thirds of the cells are undifferentiated neuroblasts, and the taurine content per retina is approximately one-seventh of the adult value. Shortly after weaning the adult morphology and taurine content are attained. Expression of taurine immunoreactivity (taurine-IR) accompanies differentiation; in some cell types (ganglion and horizontal cells) this expression is transient, while in others (photoreceptors, bipolar, and a subpopulation of amacrine cells) it persists into the adult state. At birth, taurine-IR is localized mainly in cells in the position of ganglion cells, especially in their axons within the nerve fiber layer. This reactivity is soon lost from the somata, and disappears from the axons by 10 days of age. At 2 days of age, taurine-IR appeared additionally in somata of amacrine cells flanking the forerunner of the inner plexiform layer, and in growth cone-like processes of photoreceptors. At day 6, taurine-IR was marked in photoreceptor cell inner and outer segments, and in horizontal cells and their lateral processes. Taurine-IR was lost from horizontal cells and most amacrine cells around day 10, and appeared in bipolar cells, where it remained, with that in photoreceptors, into adulthood. Particularly striking was taurine-IR in large synaptic terminal-like processes close to the ganglion cell layer which were first seen around day 16. GABA immunoreactivity was never seen in photoreceptor or bipolar cells, was expressed transiently in horizontal cells at the same time as taurine-IR, but persisted in a subpopulation of amacrine cells and synaptic lamina in the inner plexiform layer and in some fine glial processes in the adult.