The topographic order of the retinocollicular projection in the rat was examined from birth until maturity. Small, localized deposits of rhodamine-filled latex microspheres were placed into the superior colliculus at different locations. To minimize labeling fibers of passage deposit sites were typically, although not exclusively, placed into the caudal-lateral pole of the colliculus. Examination of the area and density of labeled cells in the retinae of these animals led to the following conclusions: (1) At each age examined, the location of the majority of labeled cells was observed to be in appropriate topographic register with the deposit site in the superior colliculus. (2) Confirming the work of previous investigators, errors in topographic projection were observed. These were present in both the contralateral and ipsilateral retinae and decreased with increasing postnatal age. The mature pattern was present by P10. (3) Quantitatively, the number of retinal ganglion cells terminating nontopographically within the colliculus constituted a relatively minor proportion of the total number of labeled cells in both retinae. It is concluded that the majority of the retinal ganglion cells make topographically appropriate terminations within the superior colliculus during development.

Measurements were made of the optimal velocity for drifting bar-shaped stimuli to excite striate cortex neurons of the cat. These data were compared to the optimal spatial and temporal frequencies of the same neurons, as determined with drifting sine-wave grating stimuli. A systematic relationship was revealed, whereby those neurons preferring higher velocities of bar motion also preferred lower spatial and higher temporal frequencies of gratings. The optimal bar velocity for a given neuron could be quantitatively predicted from the ratio of that neuron's optimal temporal frequency to its optimal spatial frequency.

The laminar structure and cellular distribution of cytochrome-oxidase (CO) reactivity in supragranular puffs of striate cortex was examined in adult macaque monkeys surviving various periods of monocular enucleation, lid suture, and retinal impulse blockage with tetrodotoxin (TTX). Enucleation and TTX produced a rapid and severe loss in the size of the CO reactive region in puffs dominated by the removed or treated eye compared to slower and less marked reductions obtained in deprived puffs of lid-sutured monkeys. In all deprived animals, the cross-sectional areas of deprived puffs decreased most rapidly in the upper layers (2 and 3A). In long-term enucleated (60 wks) and TTX-treated (4 wks) monkeys, puff area was severely reduced in layer 3B, while reactivity in layer 3B appeared partially spared in lid-sutured monkeys. The density of the CO reaction product was significantly and evenly reduced throughout deprived puffs for all of the monkeys examined; however, this decrease was less severe in adult monkeys lid-sutured for 11 wks. Although no evidence for cell loss was obtained, all three forms of visual deprivation led to lower counts of neuronal perikarya with high levels of CO reaction product in both deprived puff and interpuff areas. This effect was less marked in the deprived puffs of monkeys lid-sutured for 2.5 and 3 yrs, suggesting recovery of CO activity in some neurons. Neurons in deprived puffs and interpuffs were generally similar in size to those in nondeprived regions, although CO-reactive cells were significantly smaller in the deprived puffs of monkeys enucleated for 28.5 or 60 wks. These results indicate that the metabolic response of neuronal elements in supragranular striate cortex depends upon the nature of the visual deficit. The partial sparing of CO reactivity in deprived puffs of lid-sutured monkeys may reflect the continued transmission of certain types of visual stimuli through a closed eyelid.

There are several different groups of ganglion cells in the retina of the frog. Although their axons are thought to terminate in different layers of the optic tectum, little is known about the morphology of their terminal arbors or their synaptic targets. The present paper reports the results of a layer-by-layer study of horseradish peroxidase labeled retinal axons in the optic tectum of Rana pipiens. Light and electron microscopy was used to study the axon's laminar distribution, patterns of arborization, and synaptic contacts.

Labeled retinal axons were found in all of the superficial layers of the tectum (A-G). From layer to layer, the retinal axons differed markedly in the diameter of their parent axons (0.2−3.0 μm) and in the morphology and horizontal extent of their terminal arbors.

Five classes of synaptic terminals could be distinguished in the tectum. The retinal terminals belonged to class characterized by round, medium-sized synaptic vesicles. They made synaptic contact with dendrites and other axon terminals in each of the layers. They were always the presynaptic component. The postsynaptic dendrites were often the vertically oriented processes of cells located in the deeper layers. The postsynaptic terminals belonged to a class distinguished by their flat, medium-sized vesicles. These terminals in turn contacted what appeared to be dendrites. In layer eight, the retinal axons were often large, spoon-shaped boutons that ended in apposition with the somata of the layer.

The pineal complex of anuran &hibians is a directly photosensory organ, encompassing both an extracranial portion, the frontal organ, and an intracranial portion, the pineal organ proper. The projection neurons of the frontal organ respond differentially according to the wavelengths of the light stimuli. The pineal organ, on the other hand, functions mainly as a luminosity meter. Most of its centrally projecting neurons respond to all increases in ambient illumination with decreases in spontaneous firing of action potentials, although some neural units in the pineal organ may respond according to wavelength. This difference in responses to light stimulation may be reflected in the neural organization of the two parts of the pineal complex. In the present study, we have analyzed the morphology of the projection neurons of the frontal and pineal organs of the frog, Rana esculenta, by backfilling of the neurons with horseradish peroxidase through their cut axons. In the pineal organ, several types of centrally projecting neurons were observed: peripherally situated unipolar and multipolar neurons, the dendrites of which extend into a superficial axon plexus that surrounds the pineal epithelium; smaller unipolar, bipolar, or multipolar neurons situated close to the central pineal tract; and radially oriented bipolar neurons, with short dendritic processes oriented towards the lumen of the pineal organ. This latter type was strongly reminiscent of photoreceptor cells. The centrally projecting neurons of the frontal organ were multipolar, and situated in the ventral part of the organ. One photoreceptor-like bipolar neuron was observed in one frontal organ. The neurons of the frontal organ did not form a superficial plexus of neurites. This difference may relate to the different ratio of chromaticity/luminosity units in the frontal and pineal organs.

The directional selectivity of units within the nucleus of the basal optic root (nBOR) of the accessory optic system (AOS) was studied before and after lesions of the visual telencephalon (visual Wulst) in urethane-anesthetized pigeons. In intact pigeons, most nBOR units preferred upward motion with a temporal component or downward motion with a nasal component. The ipsilateral and bilateral telencephalic lesions generated a dramatic reduction in the number of cells with optimal responses to upward motion. The overall distribution of preferred directions was still bimodal following ipsilateral or bilateral Wulst lesions, with most units showing best responses to a straight temporal or to downward-nasal directions. The contralateral Wulst lesions produced, instead, a marked reduction in downward preferences. The nBOR units which were studied in these cases showed mainly upward-temporal and upward-nasal responses. These data suggest an involvement of the visual Wulst in the determination of the dictional selectivity of nBOR neurons in the pigeon. Specifically, the responses of nBOR units to upward motion appeared to depend on the integrity of the telencephalic descending systems which impinge, in both direct and indirect ways, upon that AOS nucleus. Taken together with data for the mammalian AOS, the present results indicate that nonretinal afferents to AOS nuclei have an important role in the functional organization of that subcortical visual pathway.

Isolated retinal rods of the frog consisting of the outer segment and the ellipsoid were patch-clamped and recorded in the whole-cell mode. The recording pipettes were filled with solutions of different composition in order to alter the cytoplasmic content of sodium, phosphate, and calcium ions, and guanine nucleotides. When a simple medium with potassium as the principal cation was used, the dark voltage slowly approached more negative values. This tendency of spontaneous hyperpolarization was reduced significantly when cGMP or GTP were present in the pipette medium. Sodium ions, on the other hand, clearly increased the speed of hyperpolarization. In the presence of sodium (20 mM), the stabilizing effect of GTP did not occur and that of cGMP was clearly diminished. Phosphate (20 mM) neutralized the sodium effect. High calcium levels (100 μM) did not measurably influence the time course of hyperpolarization. We conclude that the normal cytoplasmic sodium level in rods does not exceed 10 mM and that higher internal sodium concentrations interfere with the sodium–calcium exchange mechanism.

The superficial layers of the frog's optic tectum, Potter's (1969) layers A-G, comprise a complex neuropil made up of many afferent axons, the somata of a few neurons, and many dendrites from the neurons located in the deeper layers. Different types of retinal axons are believed to terminate in different layers (Maturana et al., 1960; Kuljis & Karten, 1988; Sargent et al., 1989), but little is known about the relationships between each type of input and the dendrites of the deep tectal neurons that extend into these superficial layers. The present study used the method of retrograde transport of horseradish peroxidase to study the synaptic contacts on the dendrites of the neurons that give rise to the crossed tecto-bulbar pathway. These cells have apical dendrites that ascend through the superficial retino-recipient layers.

The somata of the cells that give rise to the crossed tecto-bulbar pathway are located in the superficial half of layer 6, preferentially clustered along the caudal, lateral, and rostral margins of the tectum. The somata of these cells range from 8−30 ¼m in diameter. Their axons are large (2−4 ¼m in diameter) myelinated fibers that arise from either their somata or proximal dendrites. Their axons travel within the deep medullary layer to leave the tectum at the lateral margin. Their dendritic arbors extend obliquely through the superficial layers to reach layer B where they turn and extend within the layer for up to 0.5 mm. The somata of these cells receive only a scant synaptic input. In contrast, their dendrites receive input in every layer, but the nature of this input varies from layer to layer. Synaptic terminals that resemble retinal ganglion cell boutons contact the labeled dendrites in layers B, F, and G. This indicates that the dendrites may receive monosynaptic input from several types of retinal ganglion cells. Terminals with small, flattened vesicles also contact the dendrites of these cells in each layer. In layer F and below, the terminals with flattened vesicles constitute 15% of the contacts; above layer F they constitute only 5−8% of the contacts. Terminals with medium-sized, flattened vesicles also contact the dendrites of these cells in every layer and constitute a large proportion of their input (33−95%). The latter terminals resemble those that are often postsynaptic to retinal terminals.

The photosensory pineal complex of anurans comprises an extracranial part, the frontal organ, and an intracranial part, the pineal organ proper. Although the pineal organ functions mainly as a luminosity detector, the frontal organ monitor the relative proportions of short and intermediate/long wavelengths in the ambient illumination. The major pathway of information processing in the pineal and frontal organs is the photoreceptor to ganglion cell synapse. It is not known whether interneurons form part of the neural circuitry. In the present study, we demonstrate GABA-immunoreactive (GABA-IR) neurons in the pineal and frontal organs of the frog, Rana esculenta. No GABA-IR axons were observed in the pineal nerve between the frontal and pineal organs, or in the pineal tract that connects the pineal complex with the brain. The GABA-IR neurons differed in morphology from centrally projecting neurons visualized by retrograde labeling with horseradish peroxidase. Thus, we suggest that the GABA-IR neurons in the pineal and frontal organs represent local interneurons.

Axons of central origin, immunoreactive with a sensitive antiserum against the tetrapeptide Phe-Met-Phe-Arg-NH2 (FMRFamide), were observed in the intracranial portion of the photosensory pineal organ. The immunoreactive axons enter the caudal pole of the pineal organ via the posterior commissure. The largest density of axons was observed in the caudal part, while fewer axons were detected in the rostral portion. The uneven distribution of the FMRFamide-immunoreactive axons may be related to the distribution of different types of intrapineal neurons. FMRFamide-immunoreactive varicose axons were observed in the extracranial frontal organ. A central innervation of the pineal organ, previously known exclusively from amniotes, is probably not per se linked with the evolutionary transition of the pineal organ from a directly photosensory organ to a neuroendocrine organ. It could rather represent a centrifugal input to a sensory system which has been retained when the directly sensory functions have changed, during phylogency, to neuroendocrine functions.

When a cuff-shaped electrode is placed on the optic nerve of the cat, X and Y axons, by virtue of their different diameters, exhibit different thresholds to electrical stimulation. Large-diameter Y axons have low thresholds, while smaller-diameter X axons have high thresholds. There is very little overlap between the two populations. Given this segregation, the strength of stimulation of the optic nerve required to evoke synaptic potentials in cortical neurons becomes a reliable indicator of the type of visual input a cortical neuron receives. Potentials with thresholds below the thresholds of X axons must be mediated by Y cells of the retina and LGN. Potentials with thresholds above the Y axons of the optic nerve must be mediated by X cells. From previous experiments, one would expect ot find &le input via both types of axon to are 17 of the visual cortex. This was not the case. Of 58 neurons distributed throughout the layers of area 17 from which intracellular records were taken, in only four could substantial Y excitation be detected. Three of these four were located near the border with area 18. All four received large X inputs as well. The 24 neurons studied in area 18 all received large Y inputs but no detectable X input.

The projection from striate and prestriate visual cortex to the pontine nuclei has been studied in the macaque monkey by means of anterograde tracer techniques in order to assess the contribution of anatomically and functionally distinct visual cortical areas to the cortico-ponto-cerebellar loop. No projection to the pons was found from central or paracentral visual-field representations of V1 (striate cortex) or prestriate visual areas V2, and V4. Small patches of terminal labeling occurred after injections of tracer into more peripheral parts of V1, V2 and V3, and into V3A. The terminal fields were located most dorsolaterally in the anterior to middle third of the pons and were quite restricted in their rostro-caudal extent. Injections of V5, however, yielded substantial terminal labeling, stretching longitudinally throughout almost the entire pons. This projection could be demonstrated to arise from parts of V5 receiving input from central visual-field representations of striate cortex, whereas parts of V4 receiving similarly central visual-field input had no detectable projection to the pons. Its distribution may overlap to a large extent with the termination of tecto-pontine fibers and with the termination of fibers from visual areas in the medial bank (area V6 or P0) and lateral bank (area LIP) of the intraparietal sulcus, as well as from frontal eye fields (FEF). It appears that the main information relayed to the cerebellum by the visual corticopontine projection is related to movement in the field of view.

A perfusion system was used to monitor the release of [3H]-GABA from isolated retinas of Xenopus laevis. Measurable release was stimulated by glycine at concentrations as low as 200 μM. Glycine-stimulated release was blocked by strychnine, and was not reduced in “calcium-free” Ringer's solution (0 Ca2+/20 mM Mg2+). Glutamate also stimulated calcium-independent release, using concentrations as low as 100 μM. In contrast, release stimulated by 25 mM potassium was reduced by 80% in calcium-free medium.

In most experiments, agonists were applied in six consecutive 4-mm pulses separated by 10-mm washes with Ringer's solution. Under these conditions, the release stimulated by 0.5 mM glutamate or 25 mM potassium decreased by at least 50% from the first to the second pulse, and then gradually decreased with successive applications. In contrast, the response to 0.5 mM glycine at first increased and then only gradually decreased with successive pulses. These patterns of response to different agonists were similar in calcium-free medium.

Somatostatin (—14 or —28) also stimulated release, and this effect was inhibited by AOAA, an inhibitor of GABA degradation. In the presence of AOAA, somatostatin had little effect, except at high concentrations of somatostatin (5 μM), which increased both basal and glycine-stimulated release. In contrast to somatostatin, glycine-stimulated release was much larger in the presence of AOAA.

Autoradiography was used to investigate which cell types released [3H]-GABA under our conditions. Autoradiograms showed that horizontal cells and a population of apparent “off” bipolar cells were well-labeled by [3H]-GABA high-affinity uptake. In addition, light labeling was seen over numerous amacrine cells. After application of glycine, glutamate, or potassium, there was a decrease in label density over horizontal cells.

An aftereffects paradigm was used to behaviorally measure contrast sensitivity of cats to gratings of three different test spatial frequencies after adaptation to gratings of various spatial frequencies, contrasts, and durations. Post-adaptation reductions in sensitivity occurred even after short periods of adaptation (<7 s) and could be as large as 1.0 log unit under some conditions. The magnitude of the adaptation effect varied monotonically with (1) adaptation grating contrast, (2) duration, and (3) the contrast sensitivity for the test grating. Average half-width (at half-height) of the spatial-frequency tuning curves constructed from the data was 1.4 octaves, and was not dependent upon the level of adaptation or the spatial frequency of the test grating. Post-adaptation psychometric functions of the cats showed reduced slopes and maxima suggesting that, unlike humans, in cats apparent contrast grows more slowly with increases in physical contrast after contrast adaptation. All of the characteristics observed are in excellent agreement with electrophysiologically measured properties of neurons in striate cortex of cats. In addition, there was a remarkable similarity of the cat tuning functions, both in shape and bandpass, to those measured in man with a similar paradigm suggesting that (1) the two visual systems are sufficiently similar to make the cat a useful spatial vision model and (2) there is a common functional plan to all mammalian visual systems despite significant anatomical differences between species.

The present study examines the differential effects of three excitotoxins, kainic acid (KA), N-methyl-D-aspartate (NMDA), and α-amino-2,3-amino-2,3-dihydro-5- methyl-3-oxo-4- isoxazolepropanoic acid (AMPA) on neurons within the genglion cell layer (GCL) of the chick retina. Two-day-old chicks were given a single, 5 μl, intravitreal injection of KA, NMDA, or AMPA at a range of doses. Following treatment with 40 nmol KA, there was a 21% loss of neurons in the GCL. At 200 nmol KA, the loss increased to 46%. Exposure to KA eliminated mainly small neurons of soma area 5–15μm2, and medium-sized ganglion cells of soma area 15–25μm2. Large ganglion cells (>25μ,2) remained unaffected. The vast majority of small cells were probably displaced amarcrine cells. At a does of 3000 nmol NMDA, no further loss of cells was evident. Exposure to 200 nmol AMPA resulted in a 30% loss of large and some medium-sized ganglion cells. In a further series of experiments, exposure to excitotoxin was followed by a retinal scratch, which eliminated retinal ganglion cells within the axotomized region. The results indicate that only a small proportion of displaced amacrine cells are destroyed by NMDA and AMPA, whereas virtually all displaced amarine cells are sensitive to KA. The findings of this study indicate the existence of subclasses of ganglion cells with specificity towards different types of excitatory amino acids (EAA).

The present study uses the anterograde tracer, Phaseolus vulgaris-leucoagglutinin (PHA-L), to investigate the detailed morphology of individual axons projecting from area V1 to prestriate area V2. Observations are derived from serial reconstructions of 45 axons. Axons are found to differ both in laminar distribution and in arbor size. The majority (25/45; 56%) terminate in the upper half of layer 4 and the lower part of layer 3. Terminal clusters typically measure about 200 μm in diameter (dimensions are uncorrected for shrinkage), and are either in one, two, or occasionally three patches. Patches are separated by 200−500 μm. Of these 25 axons, four also have minor collaterals to layer 5. Of the remaining 20 axons in our sample, eight have one or two terminal arbors (about 200 μm in diameter) mainly in layer 3; another eight have terminations, organized as a single field (about 350 μm in diameter), within layer 4; and four axons have much larger terminal fields (1.0−1.2 mm × 0.3 mm), in layers 3 nd 4. These morphological differences might constitute a gradient or, alternately, indicate distinct subgroups within the striate efferent population. Large terminal fields are asymmetrical, with their long axis oriented in an anterior-posterior fashion toward the depth of the lunate sulcus. Axons with two terminal arbors have a similar bias. As this arrangement is approximately perpendicular to the border of V1, we suggest that striate axons may be extended preferentially along the length of the stripelike compartments in V2. These compartments are also arrayed perpendicular to the border between areas V1 and V2. Reconstruction of small groups of 2–4 convergent axons demonstrates that axons with different morphology (i.e. large or small terminal fields) can occur within the same projection focus. Terminal arbors belonging to different axons can overlap, but tend not to be superimposed exactly.

X- and Y-mediated input to areas 17 and 18 of the cat visual cortex was studied using current-source-density analysis of field potentials evoked by stimulation of the optic nerves. A cuff-shaped electrode was used for stimulation so that Y axons, by virtue of their larger diameters, would have lower electrical thresholds than X axons. The effect in each cortical area of activating Y axons alone could therefore be determined by low-&litude stimulation of the optic nerves. Current-source densities were calculated by two separate methods. (1) In five experiments, field potentials were measured sequentially at different cortical depths with a single tungsten electrode. Current densities were then calculated by computer. (2) In two experiments, current densities were derived in real time from field potentials recorded simultaneously from three sites with a multi-electrode probe. The calculation was performed by an analog circuit specially designed for this purpose. This method has several advantages over the standard, single-electrode method. At stimulus strengths sufficient to activate the majority of Y axons in the optic nerves, but subthreshold to most X axons, the field potentials evoked in area 17 changed little from layer to layer. When the current-source-density analysis was applied to these potentials, no significant sources or sinks were detectable. Only when the stimulus strength was raised to the point that both X and Y axons were activated by the stimulus were any current sources or sinks detected in area 17. The currents were similar in time course and laminar pattern to those recorded after stimulation of the optic chiasm. In area 18, large sources and sinks were evoked by stimulation of Y axons alone. These currents changed little when the stimulus strength was increased to activate X axons as well. Area 18, therefore, in contrast to area 17, seems to be dominated by Y input and receives little X input. These results support the conclusions of the accompanying paper in which synaptic potentials were recorded intracellularly from cortical neutrons. The intracellular experiments failed to show substantial Y input to area 17. The projections of X and Y axons may therefore be much more highly segregated into areas 17 and 18 than previously thought. Alternatively, the nature of the Y input to area 17 may be very different from that to area 18 in that it cannot be easily detected with intracellular or current-source-density techniques.

The distribution of ganglion cells in the retina of the silver lamprey, Ichthyomyzon unicuspis, was revealed by retrograde labeling from the optic nerve with horseradish peroxidase (HRP) and fluorescent-labeled dextrans in live animals and with the fluorescent dye DiI in aldehyde-fixed tissue. The majority of ganglion cells (74%) termed the “outer ganglion cells,” are multipolar and are located at the vitread boundary of the inner nuclear layer. The remaining ganglion cells (26%), termed the “inner ganglion cells” are bipolar and are distributed in a sublamina within the inner plexiform layer. The dense, dendritic meshwork of the outer ganglion cells is largely restricted to the sclerad half of the inner plexiform layer with some cells possessing dendrites which pass through the inner nuclear layer to terminate within the outer plexiform layer. The dendrites of the inner ganglion cells form a thin, dendritic network apposing the inner limiting membrane. Axons from both populations of ganglion cells originate from dendrites or the soma and form fascicles lying adjacent to the outer ganglion cell somata.

Retinopetal fibers, originating from bilaterally distributed neurons of the tegmental midbrain, were thin and varicose and ran parallel to the ganglion cell axons to terminate either with a varicose enlargement or a few short sidebranches in the sclerad third of the inner plexiform layer. The unusual organization of the lamprey retina and outgroup comparison with hagfish suggests that agnathans share a presumably primitive type of retinal ganglion cell organization compared to that of gnathostomes.

This paper describes experiments on GABA-activated whole-cell membrane currents in bipolar cells freshly isolated from the adult rat retina. The main goal was to determine whether bipolar cell responses to GABA could be resolved in terms of mediation by the GABAA receptor, the GABAB receptor, or both. Bipolar cells were isolated by gentle enzymatic dissociation and identified by their distinct morphology. GABA agonists and antagonists were applied focally by pressure and the resultant currents were recorded under whole-cell voltage clamp. In all bipolar cells tested, GABA (0.1–100 μM) induced a monophasic response associated with a conductance increase (IGABA). The shift in reversal potential for IGABA as a function of pipet [CI] paralleled that predicted based on the Nernst equation for Cl−. IGABA was mimicked by muscimol (5–20 μM) and antagonized by bicuculline (20–100 μM). Baclofen (0.1–1.0 mM) produced no apparent conductance change. “Hot spots” of sensitivity to GABA which might be associated with regions of synaptic contact were not found; both the soma and processes of all bipolar cells were responsive to focally applied GABA. Furthermore, all bipolar cells tested responded to glycine.

In conclusion, we have established the presence of GABAA receptors on rat retinal bipolar cells. Our data suggest further that these cells lack GABAB receptors. Finally, our observation that bipolar cells in the rat retina are relatively homogeneous in terms of their sensitivity to GABA and glycine lead us to postulate that the functional significance of the presence of receptors and their distribution on a neuron may be dictated more by the topography of the presynaptic inputs than by its inherent chemosensitivity.

An antiserum directed against tyrosine hydroxylase (TH), the rate-limiting enzyme in the synthesis of dopamine, was used to study the pigeon retina. Labeled cells were observed in both the inner nuclear layer (INL) and ganglion cell layer (GCL). Two populations of TH-immunoreactive neurons were observed in the INL. Some of these cells were 7−10 μ in diameter and gave rise to processes that arborized in three Layers of the inner plexiform layer (IPL). These cells appeared similar to the dopaminergic amacrine cells described previously (Marc, 1988). Other labeled cells in the INL were 12−20 μ in diameter and were recognizable as a previously described subpopulation of TH-immunoreactive displaced ganglion cells (Britto et al., 1988).

A population of labeled cells was observed in the GCL. Counts of these cells in two retinae revealed 5000 and 7000 cells, respectively. They ranged in size from 8−15 μ in diameter in the central retina and from 8−20 m in diameter in the peripheral retina. The density of labeled cells was highest in the central retina and red field and lowest in the retinal periphery. The difference in cell size and cell density as a function of eccentricity is characteristic of the total population of ganglion cells in the avian retina (Ehrlich, 1981; Hayes, 1982). Some of the TH-positive cells in the GCL could be classified as ganglion cells for two reasons: (1) The axons of many of the TH-positive cells in the GCL were TH-immunoreactive as well and could be followed to the optic nerve head. (2) The injection of rhodamine-labeled microspheres into the nucleus geniculatus lateralis, pars ventralis (GLv), resulted in the retrograde labeling of many of the TH-positive cells in the contralateral retina.