Electrical stimulation of the omnipause neuron region (OPN) at saccade onset results in interrupted saccades (IS) —eye movements which pause in midflight, resume after a brief period, and end near the target location. Details on the endpoint accuracy of IS do not exist, except for a brief report by Becker et al. (1981). Their analysis emphasized the accuracy of IS relative to the visual target which remained on during the interrupted period. We instead quantified the metric properties of IS relative to nonstimulated saccades during a target flash paradigm. Our results show that IS tend to be slightly hypermetric relative to the nonstimulated saccades to the same target location. The amount of overshoot is not correlated with target eccentricity. Detailed analyses also indicate that the standard deviations of the endpoint in IS are not significantly larger than those for nonstimulated saccades, although there was a much larger variability produced in eye position during the interruption. Both these latter observations support the notion that saccades are controlled by an internal negative feedback system. Also, the size of the remaining motor error during the interrupted period is one factor influencing when an IS resumes, but the variability in this measure is large particularly for smaller motor errors. Recent results have suggested that the resettable neural integrator involved in the feedback loop may be reset after each saccade through an exponential decay process. To probe the properties of the neural integrator, we varied the duration of interruption between the initial and resumed saccades and sought a systematic overshoot in the final eye position with increasing interruption period and variable initial saccade size. Our results showed the neural integrator does not decay during the pause period of interrupted saccades.

The effects of dark adaptation on the response properties of ganglion cells have been documented extensively in the cat retina. To pinpoint the different retinal mechanisms that underlie these effects, we studied the response characteristics of cat horizontal (H) cells during prolonged dark adaptation. H-cell responses were recorded intracellularly in the optically intact, in vivo eye. To disentangle rod and cone contributions, sensitivity changes during dark adaptation were tracked with white light and with monochromatic lights that favored either rod or cone excitation. Stable, long-lasting recordings allowed us to measure changes of sensitivity for adaptation periods up to 45 min. Thresholds for white light and 503-nm monochromatic light decreased steadily and in parallel. The maximum increase of sensitivity, after extinguishing a photopic adaptation light, was 1.8 log units only, reached after about 35 min. Sensitivity for 581-nm lights also increased steadily, but at a shallower slope. The steady increase of sensitivity was concomitant with a linear shift in resting membrane potential and with an increase in relative rod contribution to the threshold responses. Even though small-amplitude responses were rod dominated after prolonged dark adaptation, sensitivity to rod signals remained relatively low, compared to sensitivity of cone responses or to the absolute sensitivity of ganglion cells. This suggests that the cone-H-cell pathway plays no role in the dark-adapted cat retina.

We have employed two paradigms to investigate the mechanisms of contrast gain control in cat striate cortex. In the first paradigm, optimal drifting gratings were presented in three consecutive periods. The contrast was near threshold in the first and third periods and accompanied by iontophoretic pulses of glutamate or glutamate receptor (GluR) agonists. The contrast was set to evoke a higher firing rate in the second period. Although both visual and iontophoretic conditions were identical in the first and third periods, responses to glutamate, N-methyl-D-aspartic acid (NMDA), and (1S, 3R)-1-Aminocyclopentane-1, 3-dicarboxylic acid (ACPD) were reduced following the adapting interval. (S)-α-Amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA) responses were not reduced. Administration of ionotropic GluR antagonists did not affect adaptation to the high-contrast grating. The metabotropic GluR antagonist (±)-α-Methyl-4-carboxyphenylglycine (MCPG), which acts at presynaptic glutamate autoreceptors, decreased the degree of adaptation exhibited by striate cells. In a second paradigm, contrast response functions (CRFs) were obtained at various adapting contrasts and least-squares fits to a hyperbolic ratio equation generated for each adapting level. Similar to previous reports, DL-2-amino-5-phosphonovaleric acid (APV) reduced the slope of the CRF and increased the responsiveness of the cells but did not affect the semisaturation constant, σ, or the exponent of the CRF, n. Only MCPG significantly altered the distribution of σ and n for 19 cells. The effect on α suggests that this drug can interfere with the cell's ability to shift its operating point to match the adapting contrast. These results suggest the involvement of a presynaptic mechanism for contrast adaptation. The decrease in neuronal responsiveness immediately following the high-contrast period may reflect an additional, postsynaptic effect in which there is a decrease in the NMDA-mediated component of the visual response.

The activities of both protein phosphatases and protein kinases are responsible for the transient changes in the levels of phosphorylation and probably the functions of protein intermediates involved in the biochemical and physiological mechanisms underlying the photoresponse in photoreceptor cells from both vertebrate and invertebrate organisms. Of the known protein serine/threonine phosphatases, various forms of type I (PP 1) and type 2A (PP 2A) protein phosphatases are responsible for dephosphorylating many of the known phosphoproteins including those involved in photoreceptor cell function. In this report, we provide biochemical evidence for both PP 1– and PP 2A-like activities in the visual and nonvisual tissue of the horseshoe crab, Limulus polyphemus, that membrane and soluble forms of both enzymes are present, and that the activities of both enzymes are greater in light- than in dark-adapted lateral eyes. These activities were characterized using glycogen phosphorylase a, a substrate for both PP 1 and PP 2A, and various protein phosphatase inhibitors, including okadaic acid. We also report that okadaic acid, at concentrations required to inhibit PP 1, inhibited physiological functions of photoreceptor cells from the ventral eye, causing a delayed reduction of the resting membrane, and slowing and reducing light responses.

Variations in velocity detection thresholds for full-field visual rotation about various axes are compatible with a simple channel-based system for coding the axis and velocity of the rotation (Harris & Lott, 1995). The present paper looks at the frame of reference for this system. The head-centered, craniotopic reference system and the retinal-based, retinotopic reference systems were separated by using eccentric eye positions. We measured the threshold for detecting full-field visual rotation about a selection of axes in the sagittal plane with the eyes held either 22½ degs up, straight ahead or 22½ degs down in the head. The characteristic features of the variation in detection thresholds did not stay stable in craniotopic coordinates but moved with the eyes and were constant in retinotopic coordinates. This suggests that the coding of head rotation by the visual system is in retinotopic coordinates.

The role of N-methyl-D-aspartate (NMDA) receptors in the mediation of cortical inputs to visual neurones in the superficial layers of the superior colliculus (SSC) has been investigated. Extracellular recording with iontophoresis in the SSC of cortically intact cats has demonstrated that visual responses of most neurones were reduced by iontophoretic application of the NMDA receptor antagonist D-2-amino-5-phosphonopentanoate (APS). Following inactivation of areas 17 and 18 of the visual cortex with topical lignocaine, the visual responses of 11, previously AP5-sensitive, neurones were no longer reduced by APS ejection. The cortical input is generally assumed to influence the directional responses of visual neurones in SSC. However, detailed study of the directional bias showed that the degree of directional tuning in SSC neurones was similar to that of retinal ganglion cells, as previously described by others. Moreover, inactivation of the visual cortex with topical lignocaine did not alter the directional bias of SSC neurones. Likewise, the directional bias of SSC neurones was not reduced by iontophoretic ejection of APS in the SSC. These data imply that NMDA receptors have an important role in mediating the cortical input to the SSC. However, cortical input does not appear to be responsible for conferring directional bias upon SSC neurones' visual responsiveness.

The Drosophila ninaC locus encodes two retinal specific proteins (p132 and p174) both consisting of a protein kinase joined to a myosin head domain and a C terminal with a calmodulin-binding domain. The role of p132 and p174 was studied via whole-cell recording and through measurements of the pupil mechanism, i.e. the pigment migration in the photoreceptor cells, in the ninaC mutants, P[ninaCΔ132] (p132 absent), P[ninaCΔ174] (p174 absent), and ninaCp235 (null mutant). Voltage-clamped flash responses in P[ninaCΔ174] and ninaCp235 showed delayed response termination. In response to steady light, plateau responses in both P[ninaCΔ174] and ninaCp235 were also large. In both cases the defect was significantly more severe in ninaCp235. Responses in P[ninaCΔ132] were apparently normal. P[ninaCΔ174] and ninaCP235 were also characterized by spontaneous quantum bump-like activity in the dark and by larger and longer light-induced quantum bumps. The turn-off of the pupil mechanism in P\ninaCΔ174] and ninaCp235 was also defective, although in this case the rate of return to baseline in both mutants was more or less the same. In all ninaC mutants, the amplitudes of the pupillary pigment migration were distinctly smaller than that in the wild type. The reduction of the amplitude was largest in P[ninaCΔ174]. The light sensitivity of the pupil mechanism of P[ninaCΔ174] compared to that of wild type was reduced by 1.3 log units. Remarkably, the light sensitivity of P[ninaCΔ132] and ninaCP235 was ca. 0.5 log units higher than that of the wild type. The results suggest that the p174 protein is required for normal termination of the transduction cascade. The diverse phenotypes observed may suggest multiple roles for calmodulin distribution for controlling response termination and regulating pigment migration in Drosophila photoreceptors.

We investigated the dynamics of neurons in the striate cortex (V1) and the lateral geniculate nucleus (LGN) to study the transformation in temporal-frequency tuning between the LGN and V1. Furthermore, we compared the temporal-frequency tuning of simple with that of complex cells and direction-selective cells with nondirection-selective cells, in order to determine whether there are significant differences in temporal-frequency tuning among distinct functional classes of cells within V1. In addition, we compared the cells in the primary input layers of V1 (4a, 4cα, and 4cβ) with cells in the layers that are predominantly second and higher order (2, 3, 4b, 5, and 6). We measured temporal-frequency responses to drifting sinusoidal gratings. For LGN neurons and simple cells, we used the amplitude and phase of the fundamental response. For complex cells, the elevation of impulse rate (F0) to a drifting grating was the response measure. There is significant low-pass filtering between the LGN and the input layers of V1 accompanied by a small, 3-ms increase in visual delay. There is further low-pass filtering between V1 input layers and the second- and higher-order neurons in V1. This results in an average decrease in high cutoff temporal-frequency between the LGN and V1 output layers of about 20 Hz and an increase in average visual latency of about 12–14 ms. One of the most salient results is the increased diversity of the dynamic properties seen in V1 when compared to the cells of the lateral geniculate, possibly reflecting specialization of function among cells in V1. Simple and complex cells had distributions of temporal-frequency tuning properties that were similar to each other. Direction-selective and nondirection-selective cells had similar preferred and high cutoff temporal frequencies, but direction-selective cells were almost exclusively band-pass while nondirection-selective cells distributed equally between band-pass and low-pass categories. Integration time, a measure of visual delay, was about 10 ms longer for V1 than LGN. In V1 there was a relatively broad distribution of integration times from 40–80 ms for simple cells and 60–100 ms for complex cells while in the LGN the distribution was narrower.

The aim of this project was to investigate the interaction between Y retinal ganglion cells and the cells of the dorsal lateral geniculate nucleus (LGNd) of the cat, with particular reference to center-surround antagonism and intrageniculate inhibition. Responses of cells in the LGNd were studied by stimulating the retina with spots of light of constant contrast but varying size. The peak discharges of nonlagged X (XN) cells were strongly suppressed with increase in spot size but the responses of lagged X (XL) cells and nonlagged Y (YN) cells were inhibited much less strongly. The effect of the Y system on these responses was examined by producing a selective block of conduction in Y fibers in one optic nerve by means of a pressure cuff (Y-blocking). These effects were assessed by measuring the peak discharge rates and by calculation of a “peak suppression index.” Y-blocking had no significant effect on the peak suppression index of XL, cells in either lamina or on YN cells in the normal (not Y-blocked) lamina but had significant effects on the responses of XN cells, causing a decrease in peak suppression index, both for cells in laminae receiving their principal excitatory input from the Y-blocked eye (both lamina A and lamina A1 ) as well as those in lamina A (but not lamina A1 ) receiving their excitatory input from the normal eye. These effects were obtained with relatively large spots of light. Thus Y optic fibers have both intralaminar (monocular) and interlaminar (binocular) inhibitory effects on XN cells. In addition to these suppressive effects, the experiments also show that ipsilaterally projecting Y fibers have facilitatory effects on XN cells in lamina A when small spots of light, about optimal size for the XN cell, are used. These results suggest that the Y system plays a powerful role in shaping the responses of XN cells, possibly enhancing visual acuity.

Adult teleost retinas grow throughout life, in part by the addition of cells from an encircling, proliferative neuroepithelium. In some species, this proliferative growth is asymmetric around the retina. The present study evaluated two hypotheses regarding asymmetric proliferative growth in adult green sunfish retina: (1) the generation of rod photoreceptors in central retina from proliferative rod precursor cells is also asymmetric; and (2) asymmetric proliferative growth patterns are regulated by mechanisms that are organized retinotopically and are independent of body-axis coordinates. In all retinas examined, rod precursor distribution and rod production were asymmetric, and both were in coarse spatial register with the asymmetric pattern of cellular addition at the retinal margin. In adult eyes that were surgically rotated, the asymmetric patterns of proliferative growth were present and appropriate for the retina's prerotation orientation. The results suggest that proliferative growth at both marginal and central adult sunfish retina is asymmetric, and that these asymmetric growth patterns are regulated by a retinotopic mechanism that is independent of body-axis coordinates.

We studied excitatory local circuits in the macaque primary visual cortex (V1) to investigate their relationships to the magnocellular (M) and parvocellular (P) streams. Sixty-two intracellularly labeled spiny neurons in layers 2–5 were analyzed. We made detailed observations of the laminar and columnar specificity of axonal arbors and noted correlations with dendritic arbors. We find evidence for considerable mixing of M and P streams by the local circuitry in V1. Such mixing is provided by neurons in the primary geniculate recipient layer 4C, as well as by neurons in both the supragranular and infragranular layers. We were also interested in possible differences in the axonal projections of neurons with different dendritic morphologies. We found that layer 4B spiny stellate and pyramidal neurons have similar axonal arbors. However, we identified two types of layer 5 pyramidal neuron. The majority have a conventional pyramidal dendritic morphology, a dense axonal arbor in layers 2–4B, and do not project to the white matter. Layer 5 projection neurons have an unusual “backbranching” dendritic morphology (apical dendritic branches arc downward rather than upward) and weak or no axonal arborization in layers 2–4B, but have long horizontal axonal projections in layer 5B. We find no strong projection from layer 5 pyramidal neurons to layer 6. In macaque V1 there appears to be no single source of strong local input to layer 6; only a minority of cells in layers 2–5 have axonal branches in layer 6 and these are sparse. Our results suggest that local circuits in V1 mediate interactions between M and P input that are complex and not easily incorporated into a simple framework.

Endogenous opioid systems (i.e. opioid peptides and opioid receptors) modulate developmental events in the neonatal mammalian retina. In the present study, the mRNA encoding preproenkephalin A (PPE), the prohormone for the opioid growth factor (OGF), [Met5]-enkephalin, was studied in the developing and the adult retinas of rats. Northern analysis indicated the presence of a 1.4-kb message in the developing and adult retinas corresponding to rat PPE mRNA. Quantitation showed that PPE message was present on postnatal day 1 at 5% of the adult level, and increased during development until the adult quantity was reached by postnatal day 27. In situ hybridization experiments first detected the presence of PPE mRNA in retinal tissues during late gestation. In late prenatal and neonatal retinas, PPE message was associated with areas of the developing retina containing proliferating neuroblasts and postmitotic cells. Later in development, message appeared to be located primarily within the inner retina, with abundant PPE mRNA associated with putative horizontal cells of the inner nuclear layer (INL). The adult retina showed a similar pattern of PPE gene expression in the cells of the INL. These findings document that the gene expression in the retina for PPE begins in the fetus, continues during retinal development, and coincides with the presence of a PPE mRNA derivative ([Met5]-enkephalin) that regulates DNA synthesis during retinal ontogeny. Our results are also the first to show the presence of PPE message in the adult mammalian retina, suggesting transcription of an opioid gene in the mature visual system.

Using patch-clamp recording in combination with reverse transcriptase-polymerase chain reaction (RT-PCR), we show in individual bipolar cells acutely dissociated from the adult rat retina a correlation between the expression of the GABAA receptor ρ1 subunit mRNA and a bicuculline-resistant, diazepam-insensitive component of the GABA-activated whole-cell current response. This “GABAC-like” response, contributing to approximately 42% of the GABA-activated whole-cell current and displaying variable sensitivity to picrotoxin, was found in bipolar cells but not in any of the ganglion cells examined. Expression profiling of GABAA receptor subunit mRNAs in individual electrophysiologically tested retinal neurons revealed that, while both bipolar cells and ganglion cells may express numerous GABAA receptor subunit isoforms, including that of ρ2, the expression of the ρ1 subunit was strictly limited to bipolar cells. We propose a possible link between the presence of a receptor with GABAC-like pharmacological profile and the expression of the retina-specific ρ1 subunit isoform. The results presented in this study constitute the first direct demonstration of such a correlation at the single-cell level.

We have previously documented the exquisite motion sensitivity of neurons in extrastriate area MT by studying the relationship between their responses and the direction and strength of visual motion signals delivered to their receptive fields. These results suggested that MT neurons might provide the signals supporting behavioral choice in visual discrimination tasks. To approach this question from another direction, we have now studied the relationship between the discharge of MT neurons and behavioral choice, independently of the effects of visual stimulation. We found that trial-to-trial variability in neuronal signals was correlated with the choices the monkey made. Therefore, when a directionally selective neuron in area MT fires more vigorously, the monkey is more likely to make a decision in favor of the preferred direction of the cell. The magnitude of the relationship was modest, on average, but was highly significant across a sample of 299 cells from four monkeys. The relationship was present for all stimuli (including those without a net motion signal), and for all but the weakest responses. The relationship was reduced or eliminated when the demands of the task were changed so that the directional signal carried by the cell was less informative. The relationship was evident within 50 ms of response onset, and persisted throughout the stimulus presentation. On average, neurons that were more sensitive to weak motion signals had a stronger relationship to behavior than those that were less sensitive. These observations are consistent with the idea that neuronal signals in MT are used by the monkey to determine the direction of stimulus motion. The modest relationship between behavioral choice and the discharge of any one neuron, and the prevalence of the relationship across the population, make it likely that signals from many neurons are pooled to form the data on which behavioral choices are based.

Agrin is a large extracellular matrix protein that directs the accumulation of acetylcholine receptors at the neuromuscular junction. Recent evidence suggests that agrin may be involved in organizing synapses in the visual system as well. Focussing on the pathway that controls accommodation and pupilloconstriction, this study examined the temporal pattern of agrin expression with reference to the organization of cholinergic synapses between embryonic chick Edinger-Westphal and ciliary ganglion neurons. In situ hybridization with an S35-labeled agrin cRNA probe was used to characterize agrin expression in the Edinger-Westphal nucleus during development. Agrin mRNA was detected in the Edinger-Westphal nucleus at all time points studied, from embryonic day 7 (E7, Hamburger and Hamilton stage 31) through newly hatched chicks. Throughout this period, agrin mRNA expression in Edinger-Westphal neurons was lower than in nearby oculomotor and trochlear neurons, suggesting that cells projecting to neuronal targets may require less agrin than those projecting to muscle. Agrin mRNA expression in the Edinger-Westphal nucleus at E7, E8, E9, and E10 was significantly higher than at E12. The early appearance of agrin mRNA coincides with the period during which acetylcholine receptors are being organized on ciliary ganglion neurons, consistent with the possibility that agrin contributes to neuron-neuron synapse formation in this pathway.

We have used light- and electron-microscopic immunohistochemistry to identify the presence of immunoreactivity to neuropeptide Y (NPY) within Müller cells in the retina of the cane toad, Bufo marinus. Müller cells containing NPY-like immunoreactivity (NPY-LI) were identified at the light-microscopic level by the coexistence with immunoreactivity to glial fibrillary acidic protein (GFAP) and at the ultrastructural level by their characteristic relationship to neuron cell bodies and processes. At the light-microscopic level, those cells which contained both NPY-LI and GFAP-LI usually had small cell bodies in the inner nuclear layer, while those cells which contained only NPY-LI were identified as large and small amacrine cells. The radially oriented primary processes in the inner plexiform layer and the vitreal end feet of GFAP-LI Müller cells also expressed NPY-LI. At the ultrastructural level, thin lamellar processes of Müller cells with NPY-LI enclosed some amacrine cell bodies in the inner nuclear layer and amacrine cell dendrites in the inner plexiform layer. These observations suggest that NPY-LI is localized in Müller cells in addition to two types of amacrine cells previously identified in the Bufo retina. This study provides the first evidence that glial elements in the vertebrate retina express NPY-LI.

In the present study, the early postnatal distribution and subsequent fate of visual callosal neurons were studied in neonatal rat pups. Previous studies had indicated that the adult pattern of visual callosal neurons was sculpted from an initially uniform distribution in the neonatal cortex. To reexamine this issue, we used a sensitive tracer, latex microspheres conjugated either to rhodamine or fluorescein, that was injected into the occipital cortex of one hemisphere in pups on the day of birth (PND 1), PND 6, or PND 12. Examination of the resulting retrograde labeling of cortical neurons in the opposite hemisphere indicates that features of the mature visual callosal pattern are present as early as PND 1. At all stages of postnatal development, the relative density of callosal projection cells varies consistently across the mediolateral extent of primary visual cortex —it is always highest in the region of the 17/18a border and lowest in the body of area 17. These data strongly suggest that, from the outset, visual cortical neurons in the region of the 17/18a border preferentially make connections with the opposite hemisphere. The results of experiments in which callosal neurons were labeled on the day of birth indicate that only those neurons that have migrated to their final cortical destinations have extended callosal axons into the vicinity of the visual cortex in the opposite hemisphere. The initial pattern of callosal neurons resembles a dense, compact version of the mature one, and the present study suggests that much of the remaining change in the appearance of this pathway may be accounted for by the decrease in the overall density of neurons that is due to expansion of the cortical gray matter during postnatal life. Taken together, these results suggest that the development of the visual callosal pathway in the rat may be more similar to that in the monkey than has been reported previously.

Previous studies have shown that directionally selective (DS) retinal ganglion cells cannot only discriminate the direction of a moving object but they can also discriminate the sequence of two flashes of light at neighboring locations in the visual field: that is, the cells elicit a DS response to both real and apparent motion. This study examines whether a DS response can be elicited in DS ganglion cells by simply stimulating two neighboring areas of the retina with high external K+. Extracellular recordings were made from ON-OFF DS ganglion cells in superfused rabbit retinas, and the responses of these cells to focal applications of 100 mM KCl to the vitreal surface of the retina were measured. All cells produced a burst of spikes (typically lasting 50–200 ms) when a short pulse (10–50 ms duration) of KCl was ejected from the tip of a micropipette that was placed within the cell's receptive field. When KCl was ejected successively from the tips of two micropipettes that were aligned along the preferred-null axis of a cell, sequence-dependent responses were observed. The response to the second micropipette was suppressed when mimicking motion in the cell's null direction, whereas an enhancement during apparent motion in the opposite direction frequently occurred. Sequence discrimination in these cells was eliminated by the GABA antagonist picrotoxin and by the Ca2+-channel blocker ω-conotoxin MVIIC, two drugs that are known to abolish directional selectivity in these ganglion cells. The spatiotemporal properties of the K+-evoked sequence-dependent responses are described and compared with previous findings on apparent motion responses of ON-OFF DS ganglion cells.

The distribution of glycinergic synapses in macaque monkey retina was investigated. The monoclonal antibody (mAb2b) against the αl subunit of the glycine receptor produced a punctate immunoreactivity that was localized to synapses. In central retina about 70% of the αl subunit-containing synapses were located in strata 1 and 2 of the inner plexiform layer, about 30% were located in strata 3 and 4, and immunoreactivity was absent in stratum 5. Electron microscopy showed that the majority of the synapses in strata 1 and 2 were on cone bipolar axons. The presynaptic profile always belonged to an amacrine cell. Presynaptic and postsynaptic profiles were further characterized using double-label immunofluorescence with cell-type specific antibodies against calcium-binding proteins. An antiserum against calretinin was used to label A<doubt/>II amacrine cells and an antiserum against recoverin was used to label flat midget bipolar cells. In the outer part of the IPL, 75% of the αl-immunoreactive puncta were colocalized with calretinin-immunoreactive An processes and 61% of the αl-immunoreactive puncta were colocalized with recoverin-positive midget bipolar axons. These results suggest that the αl subunit of the glycine receptor is present at the chemical synapse made by A<doubt/>II amacrine cells with flat midget bipolar cells, thus providing a pathway for rod signals to reach midget ganglion cells.

The main route of information flow through the vertebrate retina is from the photoreceptors towards the ganglion cells whose axons form the optic nerve. Bipolar cells of the frog have been so far reported to contact mostly amacrine cells and the majority of input to ganglion cells comes from the amacrines. In this study, ganglion cells of frogs from two species (Bufo marinus, Xenopus laevis) were filled retrogradely with horseradish peroxidase. After visualization of the tracer, light-microscopic cross sections showed massive labeling of the somata in the ganglion cell layer as well as their dendrites in the inner plexiform layer. In cross sections, bipolar output and ganglion cell input synapses were counted in the electron microscope. Each synapse was assigned to one of the five equal sublayers (SLs) of the inner plexiform layer. In both species, bipolar cells were most often seen to form their characteristic synaptic dyads with two amacrine cells. In some cases, however, the dyads were directed to one amacrine and one ganglion cell dendrite. This type of synapse was unevenly distributed within the inner plexiform layer with the highest occurrence in SL2 both in Bufo and Xenopus. In addition, SL4 contained also a high number of this type of synapse in Xenopus. In both species, we found no or few bipolar to ganglion cell synapses in the marginal sublayers (SLs 1 and 5). In Xenopus, 22% of the bipolar cell output synapses went onto ganglion cells, whereas in Bufo this was only 10%. We conclude that direct bipolar to ganglion cell information transfer exists also in frogs although its occurrence is not as obvious and regular as in mammals. The characteristic distribution of these synapses, however, suggests that specific type of the bipolar and ganglion cells participate in this process. These contacts may play a role in the formation of simple ganglion cell receptive fields.