Nitric oxide (NO) is a novel neuronal messenger that likely influences retinal function by activating retinal guanylyl cyclase to increase levels of cGMP. In the present study, the localization of neuronal nitric oxide synthase (nNOS, Type I NOS) in the cone-dominant tree shrew retina was studied using NADPH-d histochemistry and nNOS immunocytochemistry. Both NADPH-d and nNOS-immunoreactivity (IR) labeled the inner segments of rods and the myoids of a regular subpopulation of cones, with their corresponding nuclei outlined. The labeled cone myoids were co-localized with a marker for short-wave-sensitive (SWS) cones (S-antigen) and also displayed the regular triangular packing and density (7%) characteristic of SWS cones in tree shrew and other mammalian retinas. These measures confirmed the identity of the labeled cones as SWS cones. Photoreceptor ellipsoids of all cones were strongly labeled by NADPH-d reactivity, but lacked nNOS-IR. Another novel finding in tree shrew retina was that both NADPH-d and nNOS-IR labeled Müller cells, which have not been labeled by nNOS-IR in other mammalian retinas. Consistent with findings in rod-dominant retinas, two types of amacrine cells at the vitreal edge of the inner nuclear layer and a subpopulation of displaced amacrine cells at the scleral edge of the ganglion cell layer were labeled by both NADPH-d and nNOS-IR. Processes of these labeled cells were seen to extend into the inner plexiform layer, where dense punctate label was seen, especially in the central sublamina. These results show that localization of NOS in the cone-dominant tree shrew retina shares some common properties with rod-dominant mammalian retinas, but also shows some species-specific characteristics. The new finding of nNOS localization in tree shrew SWS cones and rods, but not in other cones, raises interesting questions about the roles of NO in the earliest level of visual processing.

To assess the contribution of spiking inner retinal neurons to the multifocal electroretinogram (ERG), recordings were made from four monkeys (Macaca mulatta) before and after intravitreal injections of tetrodotoxin (TTX). TTX blocks all sodium-based action potentials and thus terminates spiking activity of amacrine and ganglion cells. TTX eliminated a large component from the control responses, and this TTX-sensitive component was present as early as 10 ms after the stimulus. Before injection with TTX, the 103 focal ERG responses varied in waveform across the retina. After TTX, the response waveforms were largely independent of retinal position, indicating that it was primarily the TTX-sensitive component of the control response that was dependent upon retinal location. Given that retinal ganglion cells compose a sizable proportion of the retinal elements that produce action potentials, it is likely that part of the TTX-sensitive component is due to the spiking activity of these cells. Further, the systematic change in waveform of the TTX-sensitive component with distance from the optic nerve head suggests that part of the TTX-sensitive component may originate from the activity of the ganglion cell axons. Based on these findings, there is reason to be optimistic that the multifocal technique can be employed to study the effects of glaucoma and other diseases that affect the inner retina.

The zebrafish has recently assumed a central position in the study of vertebrate development. Numerous studies of other fish have shown that their central nervous systems, and especially their visual systems, continue to add new neurons throughout life, which is probably related to their abilities to regenerate axons and whole nervous tissue. Retinal neurogenesis had not been examined in adult zebrafish, and two reports concluded that the optic tectum ceased neurogenesis early in life, so the question arose whether the zebrafish was anomalous in this regard. We labeled embryonic (24- and 48-h postfertilization) and adult zebrafish with the thymidine analog, bromo-deoxyuridine, and, after short and long survivals, examined the retina and brain for labeled cells. They were abundant in both the optic tectum and the retina. Although the rate of retinal growth slows considerably between embryonic and adult stages, the patterns of neurogenesis in both the embryo and the adult are similar to those described in other fish, so these “fish-specific” features of general interest can justifiably be studied in zebrafish.

Neurons of the horizontal cell retinal neural network are subject to modulation by the neurotransmitter nitric oxide (NO). We have examined the effects of NO on glutamate receptor function in isolated horizontal cells from the perch (Perca fluviatilis) using the concentration ramp technique to simultaneously record receptor current and agonist concentration. Dose–response curves for glutamate (0–1 mM) and kainate (0–200 μM) were measured in the presence and absence of 1–2 mM sodium nitroprusside (SNP), 1 mM 8-Br-cGMP, 100 μM cyclothiazide or 200 μM dopamine as modulators. SNP increased the EC50 (i.e. decreased affinity) for glutamate and increased Imax (i.e. increased efficacy), whereas 8-Br-cGMP increased EC50, but not Imax. In the presence of the α-amino-3-hydroxy-5-methyl-4-isoxazole-propionic acid (AMPA) receptor desensitization blocker cyclothiazide, the SNP-induced increase in EC50 persisted, but the increase in Imax was blocked. The increase in EC50, but not the increase in Imax, was also observed when the non-desensitizing agonist kainate (100–200 μM) was applied in the presence of SNP. When 2 mM SNP and 200 μM dopamine were applied together, they increased Imax (740 vs. 2455 pA) and EC50 (422 vs. 682 μM). Our findings indicate that NO modulates horizontal cell glutamate responses by reducing the affinity of receptors for glutamate while simultaneously increasing the maximal current. The shift in affinity is cGMP-mediated and independent of desensitization. The action of NO on horizontal cell glutamate receptors is distinct from, but synergistic with, that of dopamine. Glutamate receptor modulation by NO qualitatively predicts the action of NO on horizontal cell light responses in situ and may alter transmission at visual synapses according to adaptational conditions.

The projection specificity of retinal ganglion cells and the morphology of their terminals were studied in the plethodontid salamander Plethodon jordani. In an in vitro approach, ganglion cells were stained with biocytin and reconstructed by means of light microscopy. Single retinal ganglion cells often have multiple terminal structures in the thalamus, pretectum, and tectum. The projection pattern in the diencephalic neuropils is related to the depth of the terminal arbor within the tectal fiber layer. Terminal arbors in the tectum differ in location, size, and branching pattern. The following types could be distinguished: The most superficial of the optic terminals in layer 1 are relatively small with a diameter of about 100 μm. With the exception of a few varicosities (beads) in the pretectal neuropils, their stem axons have no further collaterals or terminal arbors in the diencephalic neuropils. Intermediate terminals in layer 2 fan out to form a dense plexus with a medio-lateral extent of 180 μm on average. Some terminals in this layer show obvious antenna-like fibers reaching toward the surface of the tectum. The axons of layer 2 projecting neurons have additional collaterals and terminal arbors in the thalamus and pretectum. The deep layer 3 terminals spread out over a diameter of 400 μm on average and their degree of branching is moderate. The axons of layer 3 projecting ganglion cells have dense additional terminal arbors in the thalamus and pretectum. The deepest retinal terminals in the tectum are found within the predominantly efferent fiber layers. This type consists of an unbranched, but beaded axon which runs rostro-caudally with several bends and loops. The stem axon has an additional very dense terminal arborization in the neuropil of the nucleus Bellonci pars medialis and additional sparse collaterals in the pretectal area.

In a visual delayed matching-to-sample task, compared to a control condition, we had previously identified different components of the human EEG that could reflect the rehearsal of an object representation in short-term memory (Tallon-Baudry et al., 1998). These components were induced oscillatory activities in the gamma (24–60 Hz) and beta (15–20 Hz) bands, peaking during the delay at occipital and frontal electrodes, and two negativities in the evoked potentials. Sustained activities (lasting until the end of the delay) are more likely to reflect the continuous rehearsing process in memory than transient (ending before the end of the delay) activities. Nevertheless, since the delay duration we used in our previous experiment was fixed and rather short, it was difficult to discriminate between sustained and transient components. Here we used the same delayed matching-to-sample task, but with variable delay durations. The same oscillatory components in the gamma and beta bands were observed again during the delay. The only components that showed a sustained time course compatible with a memory rehearsing process were the occipital gamma and frontal beta induced activities. These two activities slowly decreased with increasing delay duration, while the performance of the subjects decreased in parallel. No sustained response could be found in the evoked potentials. These results support the hypothesis that objects representations in visual short-term memory consist of oscillating synchronized cell assemblies.

To analyze the elemental composition and topology of the extracellular compartments of the compound eye, the eyes of blowflies Calliphora vicina were rapidly frozen and ultrathin cryosections were freeze dried. Three zones of an ommatidium, peripheral cytosol of visual cells, rhabdomeres, and ommatidial cavities were analyzed by X-ray microprobe analysis. The ommatidial cavity was found to contain sodium and potassium in proportion similar to that in the blowfly hemolymph. Potassium-to-sodium ratio in a cytosol was typical for a cytosol. The rhabdomeres displayed an electrolyte content intermediate between the above compartments. Three topologically connected extracellular compartments were characterized by the experiments with tracers, monastral blue and lanthanum: (1) common intercellular space of ommatidia including peripheral clefts between the visual cells, both tracers entered this compartment; (2) the ommatidial cavity, which is not accessible for monastral blue, however, as revealed by our X-ray microanalysis, it was reachable for lanthanum; (3) rhabdomeric loops, which were accessible for lanthanum entering either via the cavity or from the common intercellular clefts. The above characteristics of the ionic content and topology of ommatidial compartments might suggest higher sodium and lower potassium content in the microvilli as compared with the cytosol. The rhabdomeric and “cavital” plasma membranes are assumed to be permeable for these ions so that a voltage of only 25–30 mV, negative inside, is probably formed across them, much lower than the known resting potential −60 mV across the peripheral plasma membrane of a visual cell.

Light-microscopic immunocytochemistry was utilized to localize the different populations of substance P-immunoreactive (SP-IR) neurons in the hamster retina. Based on observation of 2505 SP-IR neurons in transverse sections, 34% were amacrine cells whose pear-shaped or round cell bodies (7–8 μm) were situated in the inner half of the inner nuclear layer (INL) or in the inner plexiform layer (IPL), while 66% of SP-IR somata (6–20 μm) were located in the ganglion cell layer (GCL) which were interpreted to be displaced amacrine cells and retinal ganglion cells (RGCs). At least three types of SP-IR amacrine cells were identified. The SP-IR processes were distributed in strata 1, 3, and 5 with the densest plexus in stratum 5 of the inner plexiform layer. In the wholemounted retina, the SP-IR cells were found to be distributed throughout the entire retina and their mean number was estimated to be 4224 ± 76. Two experiments were performed to clarify whether any of the SP-IR neurons in the GCL were RGCs. The first experiment demonstrated the presence of SP-IR RGCs by retrogradely labeling the RGCs and subsequently staining the SP-IR cells in the retina using immunocytochemistry. The second experiment identified SP-IR central projections of RGCs to the contralateral dorsal lateral geniculate nucleus. This projection disappeared following removal of the contralateral eye. The number of SP-IR RGCs was estimated following optic nerve section. At 2 months after sectioning the optic nerve, the total number of SP-IR neurons in the GCL reduced from 4224 ± 76 to a mean of 1192 ± 139. Assuming that all SP-IR neurons in the GCL which disappeared after nerve section were RGCs, the number of SP-IR RGCs was estimated to be 3032, representing 3–4% of the total RGCs. In summary, findings of the present study provide evidence for the existence of SP-IR RGCs in the hamster retina.

The patterns of glutamate, γ-aminobutyric acid (GABA), and glycine distribution in the zebrafish retina were determined using immunocytochemical localization of antisera at the light-microscope level. The observed GABA immunoreactivity (GABA-IR) patterns were further characterized using antibodies to both isoforms of glutamic acid decarboxylase (GAD65 and GAD67), the synthetic enzyme for GABA. Glutamate-IR was observed in all retinal layers with photoreceptors, bipolar cells, and ganglion cells prominently labeled. Bipolar cells displayed the most intense glutamate-IR and bipolar cell axon terminals were clearly identified as puncta arranged in layers throughout the inner plexiform layer (IPL). These findings suggest the presence of multiple subtypes of presumed OFF- and ON-bipolar cells, including some ON-bipolar cells characterized by a single, large (9 μm × 6 μm) axon terminal. GABA-, GAD-, and glycine-IR were most intense in the inner retina. In general, the observed labeling patterns for GABA, GAD65, and GAD67 were similar. GABA- and GAD-IR were observed in a population of amacrine cells, a few cells in the ganglion cell layer, throughout the IPL, and in horizontal cells. In the IPL, both GABA- and GAD-IR structures were organized into two broad bands. Glycine-IR was observed in amacrine cells, interplexiform cells, and in both plexiform layers. Glycine-positive terminals were identified throughout the IPL, with a prominent band in sublamina 3 corresponding to an immunonegative region observed in sections stained for GAD and GABA. Our results show the distribution of neurons in the zebrafish retina that use glutamate, GABA, or glycine as their neurotransmitter. The observed distribution of neurotransmitters in the inner retina is consistent with previous studies of other vertebrates and suggests that the advantages of zebrafish for developmental studies may be exploited for retinal studies.

During early mammalian development, inputs from the two retinas intermix within the lateral geniculate nucleus (LGN), then segregate during the first postnatal week into layers that receive input from a single retina. Functionally, the LGN also changes markedly during the first postnatal month; early geniculate responses to retinal input are mainly excitatory, then inhibitory circuits mature within the LGN. These remarkable changes in form and function of the retinogeniculate pathway occur at a time when patterned visual activity is not present, but retinal ganglion cells already manifest spontaneous action potential activity. To examine the role of early retinal activity in these critical developmental processes, we placed the slow release polymer Elvax embedded with tetrodotoxin (TTX) into the vitreous chamber of one or both eyes of neonatal ferrets. Animals receiving monocular injection of TTX had the other eye treated with Elvax containing control citrate buffer. Intraocular injection of horseradish peroxidase was made at the end of the period of TTX treatment to reveal the retinal terminals in the LGN. Chronic monocular or binocular blockade of retinal activity during the first postnatal week did not prevent eye-specific segregation, although it made the boundaries between layers less distinct. Retinal terminals ended preferentially in the appropriate layer, but a large number of terminals were also present in the inappropriate layer. Further segregation was achieved during the second postnatal week of activity blockade, when most retinal terminals ended preferentially in the appropriate geniculate layer and sharper layer boundaries were present. However, a small but significant number of terminals still extended into the inappropriate layer. Together, these findings indicate that monocular as well as binocular blockade of retinal activity resulted in some anomalous retinogeniculate projections and delayed eye-specific patterning, but segregation was largely intact at the end of the second postnatal week. We also report here that intraocular tetrodotoxin had a marked effect on the maturation of intrinsic geniculate circuits prior to eye opening. Whole-cell patch-clamp recordings in the LGN slice preparation revealed that activity blockade prevented the maturation of the slow, but not the fast, hyperpolarizing potential of LGN neurons during the first postnatal month and up to P38, the oldest age studied. In conclusion, these results indicate that spontaneous retinal activity modulates the time course of binocular segregation but does not alone account for the segregation of retinogeniculate terminals. However, early retinal activity plays an important role in developing the intrinsic circuitry of the LGN.

Simultaneous extracellular ERG and intracellular recordings from horizontal and ON-bipolar cells were obtained from the dark-adapted retina of the dogfish. The light intensity–peak response relation (IR) and time course of on-bipolar cell responses closely resembled that of the ERG b-wave, but only at low light intensities [<10 rhodopsin molecules bleached per rod (Rh*)]. Block of on-bipolar cell responses with 50 μM 2-amino-4-phosphonobutyrate (APB) abolished the b-wave and unmasked a vitreal-negative wave. Subtraction from the control ERG resulted in the isolation of a vitreal-positive ERG with an IR which matched that of on-bipolar cells over the full range of light intensities. The D.C. component of the ERG arises as a result of sustained depolarization of on-bipolar cells in response to long (>0.5 s) dim light stimuli, or following bright light flashes. The IR of horizontal cells and the vitreal-negative wave unmasked by APB could be matched by scaling at low light intensities (<5 Rh*). However, horizontal cell responses saturated at about 30 Rh*, while the vitreal-negative wave continued to increase in amplitude. The time course of horizontal cell membrane current with dim flashes could be matched to the rising phase of the vitreal-negative wave, assuming that the delay in generating the voltage response in horizontal cells is due to their long (100 ms) membrane time constant. Blocking post-photoreceptor activity resulted in a much smaller vitreal-negative wave than that unmasked by APB alone. We conclude that the b-wave arises from on-bipolar cell depolarization, while the leading edge of the a-wave is a composite of the change in extracellular voltage drop across the rod layer and a component (proximal PIII) reflecting a decrease in extracellular K+ as horizontal cell synaptic channels close with light.

Previous evidence concerning the physiological cell classes in the medial interlaminar nucleus (MIN) has been conflicting. We reexamined the MIN using standard functional tests to distinguish X-, Y- and W-cells. Discharge patterns to flashing spots also were used to identify some cells as lagged or nonlagged, as previously done for the geniculate A-layers. Also, each cell's response timing (latency and absolute phase) was measured from discharges to a spot undergoing sinusoidal luminance modulation. Of 71 MIN cells, 48% were Y, 27% were W, 8% were X, and 17% were unclassifiable. Lagged and nonlagged discharge profiles were observed in each cell group, with 28% of all cells being lagged. Lagged cells displayed a response suppression and long latency to discharge following spot onset, and a slow decay in firing at spot offset that was often preceded by a transient discharge. These profiles were indistinguishable from those of lagged cells in the A-layers. MIN cells also were heterogeneous in response timing, displaying a range of latency and absolute phase values similar to that in the A-layers. We extended these analyses to 27 cells in the geniculate C-layers. In layer C, 35% of cells were Y, 10% were X, 25% were W, and 30% were unclassifiable. About 11% had lagged profiles, and were X-cells or unclassifiable cells. Layers C1 and C2 contained only W-cells and no lagged profiles. The range of timings in the C-layers was somewhat narrower than in the MIN. Overall, these results show that the MIN contains a greater variety of functional cell classes than heretofore appreciated. Further, it appears that mechanisms which create different timing delays in the A-layers also exist in the MIN and layer C. These timings may contribute to direction selectivity in extrastriate cortex.

Neurons in the mammalian visual cortex have been found to respond to second-order features which are not defined by changes in luminance over the retina (Albright, 1992; Zhou & Baker, 1993, 1994, 1996; Mareschal & Baker, 1998a,b). The detection of these stimuli is most often accounted for by a separate nonlinear processing stream, acting in parallel to the linear stream in the visual system. Here we examine the two-dimensional spatial properties of these nonlinear neurons in area 18 using envelope stimuli, which consist of a high spatial-frequency carrier whose contrast is modulated by a low spatial-frequency envelope. These stimuli would fail to elicit a response in a conventional linear neuron because they are designed to contain no spatial-frequency components overlapping the neuron's luminance defined passband. We measured neurons' responses to these stimuli as a function of both the relative spatial frequencies and relative orientations of the carrier and envelope. Neurons' responses to envelope stimuli were narrowband to the carrier spatial frequency, with optimal values ranging from 8- to 30-fold higher than the envelope spatial frequencies. Neurons' responses to the envelope stimuli were strongly dependent on the orientation of the envelope and less so on the orientation of the carrier. Although the selectivity to the carrier orientation was broader, neurons' responses were clearly tuned, suggesting that the source of nonlinear input is cortical. There was no fixed relationship between the optimal carrier and envelope spatial frequencies or orientations, such that nonlinear neurons responding to these stimuli could perhaps respond to a variety of stimuli defined by changes in scale or orientation.

Single-unit recording and micropressure ejection techniques were used to test the effects of norepinephrine (NE) on the responses of neurons in the superficial layers (the stratum griseum superficiale and stratum opticum) of the hamster's superior colliculus (SC). Application of NE suppressed visually evoked responses by ≥30% in 75% of 40 neurons tested and produced ≥30% augmentation of responses in only 5%. The decrement in response strength was mimicked by application of the α2 adrenoceptor agonist, p-aminoclonidine, the nonspecific β agonist, isoproterenol, and the β1 agonist, dobutamine. These agents had similar effects on responses evoked by electrical stimulation of the optic chiasm and visual cortex. The α1 agonist, methoxamine, augmented the light-evoked responses of 53% of 49 SC cells by ≥30%, but had little effect on responses evoked by electrical stimulation of optic chiasm or visual cortex. The effects of adrenergic agonists upon the glutamate-evoked responses of SC cells that were synaptically “isolated” by concurrent application of Mg2+ were similar to those obtained during visual stimulation. Analysis of effects of NE on visually evoked and background activity indicated that application of this amine did not significantly enhance signal-to-noise ratios for most superficial layer SC neurons, and signal-to-noise ratios were in some cases reduced. These results indicate that NE acts primarily through α2 and β1 receptors to suppress the visual responses of SC neurons. Activation of either of these receptors reduces the responses of SC neurons to either of their two major visual inputs as well as to direct stimulation by glutamate, and it would thus appear that these effects are primarily postsynaptic.

Intracellular recording techniques were used to evaluate the effects of norepinephrine (NE) on the membrane properties of superficial layer (stratum griseum superficiale and stratum opticum) superior colliculus (SC) cells. Of the 207 cells tested, 44.4% (N = 92) were hyperpolarized by ≥3 mV and 8.7% (N = 18) were depolarized by ≥3 mV by application of NE. Hyperpolarization induced by NE was dose dependent (EC50 = 8.1 μM) and was associated with decreased input resistance and outward current which had a reversal potential of −94.0 mV. Depolarization was associated with a very slight rise in input resistance and had a reversal potential of −93.1 mV for the single cell tested. Pharmacologic experiments demonstrated that isoproterenol, dobutamine, and p-aminoclonidine all hyperpolarized SC cells. These results are consistent with the conclusion that NE-induced hyperpolarization of SC cells is mediated by both α2 and β1 adrenoceptors. The α1 adrenoceptor agonists, methoxamine and phenylephrine, depolarized 35% (6 of 17) of the SC cells tested by ≥3 mV. Most of the SC cells tested exhibited responses indicative of expression of more than one adrenoceptor. Application of p-aminoclonidine or dobutamine inhibited transsynaptic responses in SC cells evoked by electrical stimulation of optic tract axons. Inhibition of evoked responses by these agents was usually, but not invariably, associated with a hyperpolarization of the cell membrane and a reduction in depolarizing potentials evoked by application of glutamate. The present in vitro results are consistent with those of the companion in vivo study which suggested that NE-induced response suppression in superficial layer SC neurons was primarily postsynaptic and chiefly mediated by both α2 and β1 adrenoceptors.

Zebrafish (Danio rerio) represents an excellent genetic model for vertebrate visual system studies. Because the opsin proteins are ideal markers of specific photoreceptor cell types, we cloned six different zebrafish opsin cDNAs. Based on pairwise alignments and phylogenetic comparisons between the predicted zebrafish opsin amino acid sequences and other vertebrate opsins, the cDNAs encode rhodopsin, two different green opsins (zfgr1 and zfgr2), a red, a blue, and an ultraviolet opsin. Phylogenetic analysis indicates the zfgr1 protein occupies a well-resolved dendrogram branch separate from the other green opsins examined, while zebrafish ultraviolet opsin is closely related to the human blue- and chicken violet-sensitive proteins. Polyclonal antisera were generated against individual bacterial fusion proteins containing either the red, blue, or ultraviolet amino termini or the rod or green opsin carboxyl termini. Immunolocalization on adult zebrafish frozen sections demonstrates the green and red opsins are each expressed in different members of the double cone cell pair, the blue opsin is detected in long single cones, and the ultraviolet opsin protein is expressed in the short single cones. In 120-h postfertilization wholemounts, green, red, blue, and ultraviolet opsin-positive cells are detected in an orderly arrangement throughout the entire retina. The antibodies' photoreceptor-type specificity indicates they will be useful for characterizing both wild-type and mutant zebrafish retinas.

Responses of striate neurons to line textures were investigated in anesthetized and paralyzed adult cats. Light bars centered over the excitatory receptive field (RF) were presented with different texture surrounds composed of many similar bars. In two test series, responses of 169 neurons to textures with orientation contrast (surrounding bars orthogonal to the center bar) or motion contrast (surrounding bars moving opposite to the center bar) were compared to the responses to the corresponding uniform texture conditions (all lines parallel, coherent motion) and to the center bar alone. In the majority of neurons center bar responses were suppressed by the texture surrounds. Two main effects were found. Some neurons were generally suppressed by either texture surround. Other neurons were less suppressed by texture displaying orientation or motion (i.e. feature) contrast than by the respective uniform texture, so that their responses to orientation or motion contrast appeared to be relatively enhanced (preference for feature contrast). General suppression was obtained in 33% of neurons tested for orientation and in 19% of neurons tested for motion. Preference for orientation or motion contrast was obtained in 22% and 34% of the neurons, respectively, and was also seen in the mean response of the population. One hundred nineteen neurons were studied in both orientation and motion tests. General suppression was correlated across the orientation and motion dimension, but not preference for feature contrast. We also distinguished modulatory effects from end-zones and flanks using butterfly-configured texture patterns. Both regions contributed to the generally suppressive effects. Preference for orientation or motion contrast was not generated from either end-zones or flanks exclusively. Neurons with preference for feature contrast may form the physiological basis of the perceptual saliency of pop-out elements in line textures. If so, pop-out of motion and pop-out of orientation would be encoded in different pools of neurons at the level of striate cortex.

We report here a reexamination of the developmental expression of cone opsins in the zebrafish retina. The red- and blue-sensitive opsins appear at 51 h postfertilization (hpf) whereas ultraviolet (UV) opsin is not seen until after 55 hpf. More cells show red cone opsin expression than blue at 51 and 55 hpf, indicating the sequence of cone opsin expression in zebrafish is first red, then blue, and finally UV. Curiously, morphological development of the cones is in reverse order; UV cones appear quite mature by day 6–7 postfertilization (pf), but morphologically, red cones do not appear adult-like until 15–20 days pf.