(Received 4 July 1977)

Transcription

1 J. Phyaiol. (1978), 276, pp With 7 text-figure8 Printed in Great Britain EFFECTS OF PICROTOXIN AND STRYCHNINE ON RABBIT RETINAL GANGLION CELLS: CHANGES IN CENTRE SURROUND RECEPTIVE FIELDS BY J. H. CALDWELL* AND N. W. DAW From the Department of Physiology and Biophysics, Washington University School of Medicine, 660 South Euclid Avenue, St Louis, Missouri 63110, U.S.A. (Received 4 July 1977) SUMMARY 1. The effects of picrotoxin and strychnine on the centre surround types ofganglion cell (X, Y, sluggish sustained and sluggish transient with on or off centres, and colour coded) were studied in the rabbit retina. 2. Picrotoxin changed the centre surround balance in favour of the centre for Y cells and sluggish transient cells but not for X cells or sluggish sustained cells. 3. Inhibition by a moving radial grating was abolished by picrotoxin for off centre Y cells, but not for on centre Y cells. 4. Picrotoxin abolished the surround response for six on centre sustained cells. These were hybrid cells with conduction velocities and receptive field properties characteristic of more than one of the X, Y and sluggish categories. The surround was not abolished by picrotoxin for any of the cells which fell in the standard X, Y and sluggish categories. 5. Strychnine did not affect the centre surround balance substantially in any of the cells tested. Strychnine did affect the transients: in general strychnine shortened or abolished them, while picrotoxin made them larger. INTRODUCTION Recent evidence in the cat has suggested that picrotoxin affects the centre surround balance of Y cells, but not X cells in the retina (Kirby & Enroth-Cugell, 1976). The purpose of the present study was to see if a similar effect would be found in the rabbit, and to extend the results to sluggish centre surround cells and colour coded units. Since most bipolar cells have centre surround receptive fields (Werblin & Dowling, 1969; Kaneko, 1970) and picrotoxin and strychnine are both likely to affect the inner plexiform layer rather than the outer plexiform layer in mammals (Caldwell, Daw & Wyatt, 1978), one would not expect either drug to abolish the surround of these receptive fields altogether. Consequently, we decided to look at several other aspects of the receptive field organization, such as the radial grating effect (Werblin, 1972; Cleland, Levick & Sanderson, 1973) and how transient or sustained the response was. As it turned out, despite the expectations, the surround was abolished in a few unusual cases. * Present address: Department of Physiology, University of Colorado School of Medicine, 4200 East Ninth Avenue, Denver, Colorado 80262, U.S.A.

2 300 J. H. CALDWELL AND N. W. DAW METHODS The methods have been described in the two previous papers (Caldwell & Daw, 1978; Caldwell et at. 1978). RESULTS The results were obtained from the same fifty-eight rabbits used in the previous paper (Caldwell et al. 1978). Eight subclasses of centre surround cells were studied for receptive field alterations caused by picrotoxin or strychnine, and fifty-four examples were obtained (Table 1). These centre surround cells have physiological properties similar to those of the cat (Caldwell & Daw, 1978). A ninth centre surround type is colour coded. All cells classified as brisk (X and Y) or sluggish had receptive fields which retained Table 1. Effects of picrotoxin upon centre surround cells Effect on Effect on Number centre surround Cell type tested response response Y on centre 9 Increase or Increase or no change no change Y off centre 15 Transient off increase On centre sluggish transient Off centre sluggish transient X on centre colour coded X off centre On centre sluggish sustained Hybrid on centre sustained units 6 More sustained No change or more sustained Increase 6 Increase Increase or more sustained Effect on centre/ surround balance Altered in favour of centre Altered in favour of centre Altered in favour of centre Altered in favour of centre Spontaneous activity No change or small increase Large increase Slight increase No change 4 No change No change No change Increase 3 No change No change No change Large increase 5 No change No change No change Slight or less increase surround inhibition (1) 6 No change Elimination Increase of surround Inhibition excitation Surround antagonism abolished Other No change in radial grating inhibition Elimination of radial grating inhibition No change in radial grating inhibition Possible decrease in radial grating inhibition

3 PICROTOXIN AND STRYCHNINE IN RETINA 301 their centre surround characteristics when either drug was added. Many brisk (X and Y) cells had large increases in spontaneous activity. In contrast, the spontaneous activity of sluggish cells was usually much less than 1/sec and changed very little with either drug. In general, transient responses were differentially altered by the two drugs. Picrotoxin enhanced and prolonged transients, even causing some cells to become sustained. Strychnine, however, shortened or abolished the transients in those cells which were affected. Spot Annulus Diffuse Control. Picrotoxin - L 1 L 200/sec L 1 sec Fig. 1. Centre surround balance shifted by picrotoxin for an on centre Y cell. Responses to a spot and annulus were unaffected. Note the loss of the off response and the increased on response to diffuse light. Centre of receptive field approx. 50 spot diameter 30, annulus inside diameter 60. Picrotoxin effects: on centre cells The characteristic effect of picrotoxin upon on centre Y bells was a shift in the centre surround balance; this was indicated by a stronger on response and weaker off response to diffuse light (Fig. 1). The response to a central spot was increased for half the on centre Y cells and unchanged for the other half; the off response to the annulus was either increased or unchanged. But for almost all units, regardless of the effect upon the spot or annulus response, the transient on became more pronounced in the response to diffuse light. Picrotoxin slightly increased the spontaneous activity of half of the on centre Y cells; that of the rest was unchanged. The transient on response of Y and sluggish transient cells to a central spot is normally inhibited by rotation of a radial grating which illuminates the surround. The reason for using this test is that it is similar to the windmill stimulus used by Werblin (1972) to isolate effects of the inner plexiform layer. By recording intracellularly in mudpuppy retina he was able to show that receptors, horizontal cells, and bipolar cells were equally affected by a stationary and rotating windmill, but amacrine cells were additionally depolarized by the rotation. If the rabbit is similar, it is likely that any effect of rotation upon the ganglion cells is due to amacrine cells. This inhibition of on centre Y cells was unaffected by picrotoxin in the present experiments; this result was very different from that for off centre Y cells (see below). On centre sluggish transient cells were similar to the on centre Y cells with regard to the effect of picrotoxin. The transient responses to the onset of the central spot and the offset of the annulus were both increased by picrotoxin. When the stimulus

4 302 J. H.CALDWELL AND N. W. DAW was diffuse light, the transient on response, which was due to the receptive field centre, increased more than the transient off. Neither the spontaneous activity of these cells nor the inhibition caused by a rotating radial grating was affected by picrotoxin. An effect of picrotoxin with on sluggish transient cells which was not observed with on centre Y cells was a large increase in sustained excitation to a spot in the centre. Some of these on sluggish transient units were altered enough to resemble the on sluggish sustained cells in their response to a central spot (Fig. 2). Both on centre X and on sluggish sustained cells were generally insensitive to picrotoxin. Only one on sluggish sustained cell lost some of the inhibition caused by Spot Annulus Diffuse c LLj_ P Lis _-.---I h 400/sec 1 sec Fig. 2. Conversion of phasic responses to tonic responses by picrotoxin for an on sluggish transient cell. Response to a central spot became extremely sustained. Duration of the off response to the annulus was also increased. Diffuse light response indicates that some centre surround antagonism still existed with picrotoxin. C, control records. P, picrotoxin records. Centre of receptive field approx. 1-50; spot diameter 1.20, annulus inside diameter 3 6. an annulus. The on centre X cells of the rabbit are of two types, colour coded and non-colour coded. Both fit the criteria used for distinguishing X cells. Spectral sensitivity measurements were made for several colour coded cells and they were all blue-on in the centre (sometimes also green-on in the centre) with a green-off surround (Caldwell & Daw, 1978). Only colour coded on centre X cells were tested with picrotoxin. These cells, like the on sluggish sustained cells, were not altered in their receptive field properties. Stimuli and backgrounds of different wave-lengths were used to test the colour components separately (a 420 nm spot on a 530 nm background for the central blue-on response, a 530 nm spot on a 420 nm background for the central green-on response and a 530 nm annulus on a 420 nm background for the surround green-off response). To test the centre surround balance both diffuse white light and diffuse green light (530 nm) were used. None of these responses was altered. The only effect of picrotoxin upon these cells was an increase in spontaneous activity. Hybrid units. Six cells, which were carefully studied, could not be classified as typical X, Y or sluggish cells. These were all on centre with relatively sustained responses to a white spot in the centre. Picrotoxin had dramatic effects upon these cells: it abolished completely the surround for five cells and almost completely for the sixth cell. For the cell in Fig 3 a central spot produced both a transient and

5 PICROTOXIN AND STRYCHNINE IN RETINA 303 a sustained response, an annulus produced transient inhibition at on and also a transient off response, and diffuse light produced transient responses at on and off. With picrotoxin the central response was more vigorous and possibly more sustained but had basically the same time course. However, the annulus and diffuse light now produced a sustained on response. There was no longer much evidence of an antagonistic surround and the annulus, during picrotoxin infusion, produced a response similar to the centre response. Spot Annulus Diffuse Control 1, Picrotoxin.a*L L iga 1 00/sec 2 sec Fig. 3. Reversal of the response to an annulus for a hybrid on centre unit. Picrotoxin converted the inhibition and off response to an annulus into a sustained on response. Picrotoxin had little effect upon the response to a spot. Response to diffuse light changed from a transient on-off into a sustained on, indicating little or no surround antagonism. Centre of receptive field approx. 80; spot diameter 2.70, annulus inside diameter 80. Four of these cells fell in the category of hybrid on centre sustained cells described in a previous paper (Caldwell & Daw, 1978). They had receptive field responses like sluggish cells, with a conduction velocity characteristic of X or Y cells, and we described these cells as being like on directionally sensitive cells without the directional sensitivity. Another was like a Y cell, except that the response was extremely sustained. The other was a luminosity detector, in the sense that its spontaneous activity depended on the ambient light level, and had receptive field properties characteristic of an X or Y cell, but a conduction latency characteristic of a sluggish cell. Picrotoxin effects: off centre cells Picrotoxin had many of the same qualitative effects upon off centre cells of the Y and sluggish transient types as it did upon the on centre cells of these types. In response to a diffuse light stimulus the transient off always increased more than the transient on (see Fig. 6), suggesting that either (i) the centre had become stronger relative to the surround or (ii) that the interaction of the time courses of the centre and surround had changed. This is similar to the effect of picrotoxin upon the on centre transient cells. In most cases there was a corresponding increase in the centre response (the transient at the offset of the central spot). The surround was never abolished by picrotoxin, but its responses sometimes became more sustained.

6 304 J. H. CALDWELL AND N. W. DAW The inhibition of off centre Y cells by a rotating radial grating is a specific effect of the surround which was abolished by picrotoxin. The transient off response to a central spot is inhibited by a radial grating (Fig. 4); picrotoxin completely eliminated this inhibition and the size of the transient off response was increased as well. The change of inhibition was difficult to assess for off sluggish transient cells due to the increase in the centre response with picrotoxin. The radial grating inhibition was possibly reduced but was not totally abolished. An additional difference between off Y and off sluggish transient cells is that the spontaneous activity of the off Y cells was always sharply increased from a normal rate of 5-10/sec to 50-60/sec after application of the drug, but for the off sluggish transient cells, as with all other sluggish cells, the spontaneous activity was minimally affected. 1MIJL.UJ ULI..L~u~.1L _JL~uLLLI.I=hJ~MILJ~LLU.Control his Picrotoxin * < * ~~~~~~~~~~~~~~spikes/secl I sec Fig. 4. Inhibition of the transient off response by a rotating radial grating was abolished by picrotoxin. The transient off response to a small spot for this off centre Y cell was normally inhibited by rotation of the grating. When picrotoxin was added, the off transient was approximately the same, whether the grating was stationary or rotating. Picrotoxin also increased the size of the off transient slightly as well as the spontaneous activity. The trace beneath the histograms indicates when the small central spot was turned on and off. The grating was stationary the entire time for the histograms on the left and was continuously rotating for the histograms on the right. The period for counting the spikes was the same for all four records and was chosen to include all of the transient off response for the control record with the stationary grating. The receptive field properties of off centre X cells were unaffected by picrotoxin. The responses to the spot, annulus, and diffuse light appeared to be simply superimposed upon the increased spontaneous activity (Fig. 5). Off centre sluggish sustained cells were not tested. Strychnine effects Fewer cells were studied with strychnine than with picrotoxin for reasons given before (Caldwell et at ). The major effect ofstrychnine was upon the transient char-

7 PICROTOXIN AND STRYCHNINE IN RETINA 305 acter of the centre response. This was true for all cell types which had strong transient responses (Y and sluggish transient cells, with both on and off centres; see Figs. 6 and 7) and was also true for off centre X cells (Fig. 5), which have both transient and sustained components. The transients of the spot response were considerably shortened or were abolished. The transients due to the annulus were perhaps altered but never completely abolished. On centre X cells (colour coded and non-colour coded) and on sluggish sustained cells, whose responses were normally extremely sustained, were unaffected, except for the spontaneous activity increase of on centre X cells. Spot Annulus Diffuse Control Ld Picrotoxin 1I_ Strychnine _ f i 200/sec 1 sec Fig. 5. Effects of picrotoxin and strychnine upon an off centre X. Both drugs had similar effects on spontaneous activity but different effects upon the response to a spot. The transient off response to a spot was large with picrotoxin but almost entirely missing with strychnine. The responses to an annulus and to diffuse light were basically unchanged. The loss of the transient response to a central spot was best seen with off centre Y cells (compare the spot response for normal, picrotoxin, and strychnine conditions in Fig. 6). When the spot was turned off in the control record, there was a transient burst; this was unchanged or slightly increased by picrotoxin and almost abolished by strychnine. The central spot during strychnine still had an effect, however, because the firing was decreased while the spot was on. It is important to emphasize the retention of centre surround organization in the presence of strychnine. Of twenty-two cells tested for centre surround alterations, none had a complete loss of either centre or surround. The transients of the on centre Y cell in Fig. 7, for example, are severely reduced for both centre and surround, but are still present. Other cells, such as the off centre X cell in Fig. 5 and the off centre Y cell in Fig. 6, still show sustained inhibition to a spot even though the transients are gone. There were a few inconsistent instances of a shift in centre surround balance. One out of three off centre X cells lost its inhibition by diffuse light and one out of five on sluggish sustained cells had a reduction in the inhibition caused by an annulus, with a corresponding increase in the on response to diffuse light. Brisk (X and Y) cells, but not sluggish cells, had large increases in spontaneous activity and some of these brisk cells began to fire rhythmic bursts of action potentials. This bursting was regular and could be synchronized by the onset and offset of light. The effect of strychnine upon radial grating inhibition was difficult

8 306 J. H. CALDWELL AND N. W. DAW Spot Annulus with adapting spot Diffuse Control "I. - J""'All., Picrotoxin -if I.I. Strychnine. 400/secL 1 sec Fig. 6. Different effects of picrotoxin and strychnine upon an off centre Y cell. The transient component of the off response to a small central spot was enhanced by picrotoxin but was severely reduced by strychnine. The off transient in the response to diffuse light was increased by picrotoxin, indicating a shift in the centre surround balance. The annulus may have caused a more sustained response with strychnine. During the annulus records a small central spot was on continuously in an effort to adapt out the centre. Centre of receptive field approx. 30; spot diameter 2, annulus inside diameter 40. Spot Annulus Diffuse I& j Control. Picrotoxin.. Strychnine AL.LLA. -.' _ L 200/sec 1 sec Fig. 7. Effects of picrotoxin and strychnine upon an on centre Y cell. Strychnine almost entirely eliminated the transient responses for all three stimuli. Picrotoxin had little effect except that it slightly increased the responses to a spot and annulus and it increased the on response to diffuse light. Centre of receptive field approx. 2 7 ; spot diameter 2*2, annulus inside diameter 4.5'. to test because the response had already been shortened or reduced by strychnine itself. Nevertheless, strychnine appeared to have little effect upon the radial grating inhibition of on sluggish transient cells. Strychnine was tested upon only one of the hybrid centre surround cells (see above and Caldwell & Daw, 1978). The cell which was tested was slightly direction selective but was definitely not selective enough to be classified as an on direction selective

9 PICROTOXIN AND STRYCHNINE IN RETINA 307 cell. Strychnine had a remarkable effect. The unit became unambiguously direction selective with a clear null direction. This same cell had its antagonistic surround almost entirely eliminated by picrotoxin, as was true for the other hybrid cells. DISCUSSION The most important conclusion from the present experiments is that the fundamental concentric arrangement of the receptive field was not changed by either picrotoxin or strychnine for any cells which fall into the X/Y/sluggish categories. Changes that did occur were usually in the balance between centre and surround. This is to be expected if ganglion cells have a surround formed by both the outer plexiform layer (via the bipolar cell surround) and the inner plexiform layer (via amacrine cells). The changejin response to diffuse light may be due to a change in the relative strengths of the inputs from centre and surround, or a change in the latencies. A change in spontaneous activity is often regarded as uninteresting since it is not a very specific effect. Nevertheless, it is significant that the only cells whose spontaneous activity was sharply increased were X and Y cells. The spontaneous activity of sluggish cells was affected very little by either drug. A few on centre sustained cells which did not fit into the X/Y/sluggish classification lost most of their surround with the addition of picrotoxin. This indicates a large contribution by amacrine cells to the surround of these ganglion cells. If all bipolar cells are centre surround, one would expect some ganglion cell surround to remain even if the amacrine cell input were completely blocked. The cells whose surrounds were abolished by picrotoxin could have received input from non centre surround bipolar cells. The small number of centre surround ganglion cells which lost their surround agrees with the scarcity of reports of non centre surround bipolar cells (Kaneko, 1970; Matsumoto & Naka, 1972; Nelson, 1973; Miller & Dacheux, 1976). Four of the six cells whose surround was almost or completely eliminated by picrotoxin had hybrid receptive field properties, responded only to slow speeds, were on centre, and had moderately fast conducting axons. In these respects they resembled 'on' direction selective cells without the selectivity to direction. In this regard it is interesting that some 'on' direction selective cells do have an off surround (Caldwell & Daw, 1978). Two of the hybrid cells had a slight asymmetry in their response to movement in different directions but not enough to be classified as direction selective. Strychnine caused one of these cells to become strongly direction selective. This suggests that picrotoxin and strychnine have opposing effects on direction selectivity as well as upon certain transient responses. It has been suggested that transient and complex ganglion cells have significant input from amacrine cells but that sustained ganglion cells have primarily a bipolar cell input (Dowling, 1968). This was supported by the observation that all amacrine cells recorded in the mudpuppy have transient responses (Werblin & Dowling, 1969). However, sustained amacrine cells have been found in goldfish (Kaneko, 1973) and catfish (Naka, Marmarelis & Chan, 1975; Chan & Naka, 1976). Thus it is possible that even sustained ganglion cells could receive a substantial contribution from amacrine cells. The hybrid cells whose surrounds were abolished by picrotoxin may have been of this type.

10 308 J. H. CALDWELL AND N. W. DAW Rotation of a radial grating inhibits many transient ganglion cells. Picrotoxin abolished this inhibition for off centre Y cells and may have reduced the inhibition of off sluggish transient cells. However, picrotoxin did not affect the inhibition by a radial grating for on centre Y or on sluggish transient cells. Assuming that this inhibition is mediated by amacrine cells in the rabbit, as it may be in mudpuppy, the present experiments suggest that GABAnergic amacrine cells inhibit off centre Y cells, and non-gabanergic amacrine cells inhibit on centre Y and on centre sluggish transient cells. For off centre Y cells picrotoxin caused an enhanced off transient response to a spot, large increases in spontaneous activity, and a relative increase in the centre with respect to the surround for diffuselight. All of the effects of picrotoxin upon the off Y. including spontaneous activity and the effects upon centre surround tests, could result from the block of a single type of GABAnergic amacrine cell, a block which is evident only with the radial grating test. The means of generating transient responses of retinal ganglion cells is unknown. Possible mechanisms are presynaptic effects such as synaptic fatigue of an excitatory synapse (e.g. transmitter depletion), post-synaptic membrane changes, or the convergence of excitation and inhibition with different time courses. Transients of centre surround cells and complex cells (Caldwell et al. 1978) were affected in much the same way by the drugs that we tested. Picrotoxin enhanced and prolonged both on and off transients. Strychnine considerably shortened the transients and even abolished the transient of some centre surround cells. One possible explanation for the strychnine effect upon transients is that some feedback inhibition of bipolar cells and inhibition of ganglion cells by amacrine cells is blocked. Block of this amacrine cell might lead to a sustained response in the ganglion cell, but this does not explain why there is no large off response for the off centre Y (see Fig. 6). A second explanation for the loss of the transient is that the time course of excitation and/or inhibition has been changed. A third explanation for the decrease of both the size and the time course of the transients is that a glycinergic amacrine cell inhibits another class of amacrine cell (perhaps GABAnergic) which in turn inhibits the ganglion cell. There is a morphological basis for this suggestion in the serial amacrine synapses seen with the electron microscope (Dowling, 1970), which could be serial inhibitory synapses. Although this appears complicated, it simply suggests that strychnine blocks the inhibition of a neurone which is itself inhibitory. Thus the net effect of strychnine would be to release the GABA amacrine from inhibition and thus provide more inhibition of the ganglion cell. This model of serial inhibition could also explain the hybrid cell which became direction selective when strychnine was applied but had no antagonistic surround when picrotoxin was applied. These results with picrotoxin and strychnine agree with those of Ames & Pollen (1969), although direct comparisons are not possible because their methods did not allow detailed receptive field characterizations. They observed the effects of picrotoxin and strychnine upon the isolated rabbit retina. Picrotoxin (1-5,/M in the bathing fluid) increased the spontaneous activity and evoked responses of the off centre cells and of some on and on-off centre cells. Strychnine (2#M) increased spontaneous activity and caused bursting activity in some cells. The results agree in some respects, and differ in others, from those obtained by Kirby & Enroth-Cugell (1976) in the cat. They applied GABA antagonists (picro-

11 PICROTOXIN AND STRYCHNINE IN RETINA 309 toxin and bicuculline) while recording from X and Y cells. Both centre and surround responses in the cat were reduced, but the surround was reduced to a larger extent; the net effect of the GABA antagonist was to make the centre stronger with respect to the surround. However, in the rabbit, rather than a reduction of the responses, the on and off centre Y cells have centre and surround responses that were either unchanged or were increased, which agrees with the results in the isolated rabbit retina (Ames & Pollen, 1969). The cat and rabbit results agree in that picrotoxin affects Y cells rather than X cells, and possibly changes the centre surround balance for the Y cells in favour of the centre. General conclusionr. It is clear that the drugs have much less dramatic effects on the centre surround cells studied in this paper than on the more complex ones studied in the previous paper (Caldwell et al. 1978). In general, picrotoxin abolishes a number of effects believed to be due to lateral interactions. These include the inhibition by a rotating radial grating for off centre Y and possibly off sluggish transient cells; the inhibition by the surround from certain 'hybrid' units; the directional sensitivity of directionally sensitive units; the orientation sensitivity of orientation sensitive units; inhibition by both centre and surround in uniformity detectors; and specificities for speed and size in various types of cell. If there is any change in the centre surround balance in centre surround units, picrotoxin alters it in favour of the centre. Strychnine, on the other hand, affects the transients in the response, either to shorten them or abolish them, and does not affect the lateral interactions. The one exception to this is the local edge detectors, where strychnine affects the size specificity. It seems likely, therefore, that GABA containing amacrines make lateral inhibitory connexions in the inner plexiform layer, to cause the specificity of several different types of ganglion cell. Whether it is the same amacrine cell responsible for the different types of ganglion cell receptive field remains to be determined. Amacrines containing glycine, taurine, or whatever transmitter is antagonized by strychnine, probably make more local connexions, perhaps feed-back connexions, to affect the transient nature of the response. Quite possibly the GABA amacrines are the stratified ones, and the glycine amacrines are the diffuse ones. REFERENCES AMiEs, A. III & POLLEN, D. A. (1969). Neurotransmission in central nervous tissue: a study of isolated rabbit retina. J. Neurophy8iol. 32, CALDWELL, J. H. & DAW, N. W. (1978). New properties of rabbit retinal ganglion cells. J. Phy~iol. 276, CALDWELL, J. H., DAW, N. W. & WYATT, H. J. (1978). Effects of picrotoxin and strychnine on rabbit retinal ganglion cells: lateral interactions for cells with more complex receptive fields. J. Phy8iol. 276, CHAN, R. Y. & NAKA, K.-I. (1976). The amacrine cell. Vision Res. 16, CLELAwD, B. G., LEVICK, W. R. & SANDERSON, K. J. (1973). Properties of sustained and transient ganglion cells in the cat retina. J. Physiol. 228, DOWLING, J. E. (1968). Synaptic organisation of the frog retina: an electron microscopic analysis comparing the retinas of frogs and primates. Proc. R. Soc. B 170, DOWLING, J. E. (1970). Organization of vertebrate retinas. Investve Ophth. 9, KANEKO, A. (1970). Physiological and morphological identification of horizontal, bipolar, and amacrine cells in the goldfish retina. J. Physiol. 207,

12 310 J. H. CALDWELL AND N. W. DAW KANEKO, A. (1973). Receptive field organization of bipolar and amacrine cells in the goldfish retina. J. Phy8i0l. 235, KIRBY, A. W. & ENROTH-CUGELL, C. (1976). The involvement of gamma-aminobutyric acid in the organization 6f cat retinal ganglion cell receptive fields. A study with picrotoxin and bicuculline. J. gen. Phyqiol. 68, MATSUMOTO, N. & NAKA, K.-I. (1972). Identification of intracellular responses in the frog retina. Brain Re8. 42, MILLER, R. F. & DACHEUX, R. F. (1976). Synaptic organization and ionic basis of on and off channels in mudpuppy retina. I. Intracellular analysis of chloride sensitive electrogenic properties of receptors, horizontal cells, bipolar cells, and amacrine cells. J. gen. Phy8iol. 67, NAKA, K.-I., MAIARELIS, P. Z. & CHAN, R. Y. (1975). Morphological and functional identification of catfish retinal neurons. III. Functional identification. J. Neurophysiol. 38, NELSON, R. (1973). A comparison of electrical properties of neurons in Necturua retina. J. Neurophysiol. 36, WERBLIN, F. S. (1972). Lateral interactions at inner plexiform layer of vertebrate retina: antagonistic responses to change. Science, N.Y. 175, WERBLIN, F. S. & DOWUNG, J. E. (1969). Organization of the retina of the mudpuppy, Necturuw maculoaue. II. Intracellular recording. J. Neurophy8iol. 32,