Volume 16 Number 7 Reports 673 artery blood flow and the retinal oxygen tension, however, remained elevated for 15 min. Discussion. The VER from anesthetized animals has been extensively studied by others, and our recordings are similar to those they have obtained. 7 It is not clear which cortical structures generate the different waves of the VER, but in the eye it appears to be primarily foveal in origin. s We have previously shown that elevation of the intraocular pressure to levels similar to those used in this study decrease the optic nerve discharge. 9 An increase in blood flow due to mannitol infusion should partially reverse the effects of elevation of the intraocular pressure on the visual evoked response if the effect of elevation of the intraocular pressure is on the intraocular vasculature. Thus the results of our study support the view that elevation of the intraocular pressure affects intraocular blood flow. The effect of mannitol infusion on axoplasmic flow in the optic nerve is not known. The intravenous infusion of mannitol increased the retinal oxygen tension. Tissue available oxygen may be increased by several mechanisms. 10 The increase in retinal oxygen tension we measured, however, is most likely due to an increase in intraocular blood flow. We measured an increase in carotid artery blood flow following the intravenous infusion of mannitol (Fig. 2). Moreover, the increase in carotid artery blood flow and retinal oxygen tension had similar latencies and response patterns. The intravenous infusion of mannitol caused a decrease in the systemic blood pressure but an increase in blood flow. This apparent contradiction is explained by vasodilation and accompanied decrease in peripheral resistance. There is thus an increase in venous return to the heart, resulting in an increased heart rate and stroke volume. The increase in cardiac output and decrease in peripheral resistance cause an increase in blood flow and more than compensate for the transient decrease in systemic blood pressure. The increase in blood flow occurred immediately and is thus probably due to a direct dilating effect on vessels rather than an effect mediated through the catecholamines. From the Eye Research Laboratories, Department of Ophthalmology, The University of Chicago, Chicago, 111. Supported in part by the United States Public Health Service grant EY- 00792 from the National Eye Institute, National Institutes of Health. Submitted for publication Sept. 1, 1976. Reprint requests: J. Terry Ernest, M.D., Department of Ophthalmology, The University of Chicago, 950 East 59th St., Chicago, 111. 60637. Key words: mannitol, visual evoked response, retinal oxygen tension, intraocular pressure, glau- REFERENCES 1. Weiss, D. I., Shaffer, R. N., and Wise, B. L.: Mannitol infusion to reduce intraocular pressure, Arch. Ophthalmol. 68:341, 1962. 2. Smith, E. W., and Drance, S. M.: Reduction of human intraocular pressure with intravenous mannitol, Arch. Ophthalmol. 68:734, 1962. 3. Adams, R. E., Kirschner, R. J., and Leopold, I. H.: Ocular hypotensive effect of intravenously administered mannitol, Arch. Ophthalmol. 69:55, 1963. 4. Dominguez, R., Corcoran, A. C, and Page, I. H.: Mannitol: kinetics of distribution, excretion, and utilization in human beings, J. Lab. Clin. Med. 32:1192, 1947. 5. Duke-Elder, W. S.: The reaction of the intraocular pressure to osmotic variation in the blood, Br. J. Ophthalmol. 10:1, 1926. 6. Ernest, J. T., Stern, W. H., and Archer, D. B.: Submacular choroidal circulation, Am. J. Ophthalmol. 81:574, 1976. 7. Brazier, M. A. B.: The Neurophysiological Background for Anesthesia, Springfield, 111., 1969, Charles C Thomas, Publisher, p. 104 8. Perry, N. W., and Childers, D. C: The Human Visual Evoked Response, Springfield, 111., 1969, Charles C Thomas, Publisher, p. 23. 9. Ernest, J. T., and Potts, A. M.: Pathophysiology of the distal portion of the optic nerve. III. Effect of intraocular pressure on optic nerve discharge, Am. J. Ophthalmol. 68:594, 1969. 10. Ernest, J. T.: Autoregulation of the optic-disk oxygen tension, INVEST. OPHTHALMOL. 13: 101, 1974. Retinal degeneration in cats fed casein. III. Taurine deficiency and ERG amplitudes. SUSAN Y. SCHMIDT, ELIOT L. BER- SON, GAIL WATSON, AND CAECILIA HUANG. Cats fed taurine-free casein diets develop pronounced 050% below normal) reductions in retinal taurine concentrations and decreases in rod and cone ERG amplitudes within 10 to 45 weeks. Supplementation of taurine-free casein diets with inorganic sulfate, vitamin B s, or vitamin B e with cysteine did not prevent development of retinal malfunction. A synthetic amino acid diet devoid of casein and taurine also resulted in retinal taurine deficiency and retinal malfunction. Only taurine-containing diets preserved retinal function. These findings have established a role for exogenous taurine in maintaining normal retinal function in the cat. In taurine-deficient cats, the decreases in retinal taurine concentrations and the reductions in ERG amplitudes were closely correlated. Pronounced reductions in retinal taurine concentrations occurred only when liver and plasma concentrations of taurine were near zero.
674 Reports Invest. Ophthalmol. Visual Set. July 1977 The large range (i.e., 35 weeks) in time of occurrence of pronounced reductions in retinal taurine could be explained in part by the large range in taurine content in cat livers (170 to 1,476 Htnol per whole liver) prior to feeding them taurine-free diets. Studies with labeled taurine revealed that the half-lives of taurine in retina and liver were about 30 to 88 and 2 to 5 days, respectively. These findings are consistent with the observation that pronounced reductions in retinal taurine concentrations can occur as early as 10 weeks or as late as 45 weeks after feeding cats taurine-free diets. Cats fed a taurine-free casein diet develop retinal taurine deficiency and photoreceptor cell malfunction and subsequently photoreceptor cell death. 1 " 3 Supplementation of the casein diet with taurine prevented the development of retinal degeneration, whereas supplementation with taurine precursors, cysteine and methionine, did not "prevent degeneration even though the concentrations ot these precursors were elevated above normal in plasma. 2 These findings pointed to a role for exogenous taurine in maintaining photoreceptorcell function in the_cat. Taurine (NH 2 CH,CH 2 SOOOH) has been previously considered to be a nonessential component in the diet; the carbon skeleton can be derived from cysteine, and the sulfur group can be supplied by methionine or inorganic sulfate. 4 Vitamin B 6 has been shown to be the cofactor for decarboxylation of taurine precursors (cysteic acid or cysteine sulfinic acid) to taurine. 4 Previous studies have not resolved whether or not addition of inorganic sulfate, vitamin B 6, or a combination of vitamin B o and cysteine to the taurine-free casein diet would result in enough endogenous taurine synthesis to prevent retinal malfunction. Furthermore, the possible toxic effect of the casein diet on the cat had not been excluded. The present study has been done to define further the effects of taurine-free diets on retinal function in the cat and to study the relationship of retinal taurine concentrations to plasma and liver taurine concentrations. The half-lives of labeled taurine have been determined in retina and liver to help understand some factors that influence the rate of development of retinal degeneration in taurine-deficient cats. Methods. Fifty-three cats weighing 2 to 4 kg. were used for dietary studies as follows. Twenty-four cats were fed a taurine-free casein diet; two groups of three cats each were fed the casein diet supplemented with vitamin B c (4 mg./kg. diet) with or without added cysteine (8 gm./kg. diet); three cats were fed the casein diet supplemented with sodium sulfate (7 gm./ kg. diet); five cats were fed a completely synthetic amino acid diet devoid of taurine; ten cats were fed a taurine-containing Cat Chow diet (Ralston Purina Co., St. Louis, Mo.); and five cats were fed casein plus taurine (8 gm./kg. diet). The casein diets with supplements had 125 mmol. sulfur per kilogram which was equivalent to the sulfur content in the Chow diet. A synthetic amino acid diet was formulated to contain 125 mmol. sulfur per kilogram and to be identical to the casein diet, except that the casein component was replaced by an amino acid mixture. All cats were maintained on respective diets for 10 to 45 weeks in individual cages in cyclic light (12 hr. on/12 hr. off). Full-field electroretinogram (ERG) testing, taurine analyses in retina, plasma, and liver, and DNA analyses in retina were done as previously described. 1 " 2 Four Chow-fed cats weighing 1 to 2 kg. were used to compare liver taurine concentrations with those observed in Chow-fed cats weighing 2 to 4 kg. These cats were maintained in the laboratory for 1 week prior to sacrifice. Ten Chow-fed control cats and 14 slightly taurine-deficient cats were used to determine the half-lives for taurine in retina and liver. The slightly taurine-deficient cats were maintained on a taurine-free casein diet for 4 to 10 weeks and had nearly normal ERG amplitudes, 10% to 15% reductions in retinal taurine concentrations, and about 80% reductions in liver taurine concentrations. Cats were injected intravenously via the femoral vein with either 3 H-taurine (400 nci, 0.55 jumol), 3 5 S-taurine (528 MCi, 10.23 /xmol)! or 14 C-taurine (100 fid, 22.0 /«nol). These three isotopes of taurine were used to determine whether or not taurine was metabolized in the retina, liver, and plasma. Plasma samples were collected over a 2 hr. period to monitor the disappearance of labeled taurine from the plasma. The cats were sacrificed at 2, 24, 48, 72, 96, 120, or 180 hr. after injection. Because of the rapid disappearance of labeled taurine from plasma (see Results), the specific activities (disintegrations/minute/nanomole of taurine) in the plasmas immediately after injection had to be calculated, taking into account the endogenous plasma taurine concentration, the dose of labeled taurine injected, and the plasma volume (i.e., 6% of the body weight). Taurine accumulation in the retina or liver was calculated from the radioactivity in a given tissue divided by the specific activity of the corresponding plasma. The data for accumulation of taurine in the retina and liver were in turn used to calculate the half-lives of taurine in these tissues. 5 The radioactivity of extracts of retina, liver, and plasma was measured in a liquid scintillation counter. These extracts were also used for thin-layer chromatography to see if taurine was metabolized and to determine if the radioactivity in the extracts corresponded to taurine. Cellulose-coated
Volume 16 Number 7 Reports 675 CHOW-NORMAL < 80 2 DNA CASEIN <* 60 Q "J 40 CASEIN+CYSTEINE +VIT B6 CASEIN+SULFATE CASEIN+TAURINE M \ v v v v v v AJ v\. ^ Fig. 1. Full-field ERG responses from cats on taurine-containing and taurine-free diets for 20 to 45 weeks. Rod-dominated responses were obtained to single flashes of white light (30 ft.-l.) in the dark-adapted state (left column) and cone responses to 40 c.p.s. flickering white light stimuli (middle column) and to single flashes of white light in the presence of full-field white background (right column). Responses in the middle and right columns represent computer summation of 256 and 128 sweeps, respectively. Calibration symbol (lower right corner) represents vertically 50 /tv for left column, 2 /*V for middle column, and 4 /tv for right column; horizontally, 20 msec, for all tracings. Stimulus onset is vertical hatched line for left and right columns, vertical shock artifacts for middle column. thin-layer sheets and a pyridine:dioxane:25% ammonia:water (35:35:7:23) solvent system 6 were used to separate taurine from isethionic acid, cysteine, cysteic acid, cysteine-sulfinic acid, cysteamine, and hypotaurine. Results. All cats fed taurine-containing diets (either Chow or casein plus taurine) maintained normal ERG amplitudes (Fig. 1) and had normal taurine concentrations (expressed as naromoles per milligram dry weight ± S.E.M.) in retina (201.0 ± 13.5) and liver (40.5 ± 10.5). Cats fed taurinefree diets showed reductions in ERG amplitudes within 20 to 45 weeks (Fig. 1) and decreases in concentrations of taurine in retina (98.5 ± 18.5) and liver (0.5 ± 0.4). Cats fed a synthetic amino acid diet devoid of casein and taurine 100 80 60 40 20 RETINAL TAURINE CONCENTRATION % OF NORMAL Fig. 2. Peak-to-peak ERG amplitudes and retinal DNA concentrations related to retinal taurine concentrations in 10 control and 38 taurine-deficient cats. Controls were considered to have 100% retinal taurine concentration, ERG amplitude, and retinal DNA concentration. Taurine-deficient cats were divided into six groups of four to eight according to the amount of taurine in their retinas. For each group, average amplitudes (mean t S.E.M.) for rod ( ) and cone (O) responses and average DNA values ( ) are presented. The coefficients of correlation for rod-erg amplitude and cone-erg amplitude to retinal taurine concentration were respectively 0.90 and 0.84. showed reductions in ERG amplitudes and retinal and liver taurine concentrations that were comparable to those fed taurine-free casein diets. Amplitudes of rod-dominated dark-adapted responses and cone responses in the light-adapted state were closely related to the retinal taurine concentrations (Fig. 2). In all taurine-deficient cats, retinal DNA as a measure of cell viability became significantly reduced when retinal taurine decreased below 55% of normal (Fig. 2). Pronounced decreases (>50% below normal) in retinal taurine concentrations were observed only in cats when liver taurine concentrations were near zero (Fig. 3, A). Taurine concentrations in the retina declined gradually, if at all, as the liver concentrations decreased from normal to about 2% of normal and were at various stages of depletion in those cats whose liver concentrations were below 2% of normal. Plasma taurine concentrations remained at a plateau of 10% to 20% of normal as liver taurine concentrations decreased from 55% to 5% of normal (Fig. 3, B). In Chow-fed cats weighing 2 to 4 kg., liver weights varied from 76 to 144 gm., and total liver content of taurine ranged from 246
676 Reports Invest. Ophthalmol. Visual Sci. July 1977 s &80 I 260 150 100 50 - A O o * ^,/raurina -deficient / v- s o A 005 0.10 015 020 025 0.30 PLASMA TAURINE to 3 M 520 a! LIVER TAURINE CONCENTRATION % OF NORMAL 2 18 12 06 0 B LIVER TAURINE CONCENTRATION % OF NORMAL 2 t8 12 08 0 Fig. 3. A, Retinal taurine concentrations ( ) in 10 control and 38 taurine-deficient cats related to the respective liver taurine concentrations. Controls were considered to have 100% liver (40.5 ± 10.5 nmol./mg. dry weight) and 100% retinal taurine (201 ± 13.5 nmol./mg. dry weight) concentrations. B, Plasma taurine concentrations (A) in 10 control and 38 taurine-deficient cats related to the respective liver taurine concentrations. Controls were considered to have 100% liver and 100% plasma taurine (0.109 mm ± 0.015) concentrations. The values represent the mean - S.E.M. for four to six measurements. to 1,476 fimol per whole liver. In Chow-fed cats weighing 1 to 2 kg., liver weights were 34 to 76 gm., and total liver content ranged from 170 to 450 Mmol per whole liver. In both Chow-fed control cats and slightly taurine-deficient cats, trace doses of radioactive taurine disappeared rapidly from plasma; 60% of the injected dose disappeared within 1 min., 98% by 10 min., and 99.8% by 2 hr. Taurine accumulation by the retina could be correlated with plasma taurine concentrations; accumulation was higher in slightly taurine-deficient retinas 24 48 72 96 120 HOURS AFTER INJECTION OF LABELLED TAURINE Fig. 4. A, Taurine accumulation (nmol./gm. wet weight) by the retina of control (A,, ) and slightly taurine-deficient cats ( A t Q ( o) related to plasma taurine concentration. The cats were injected with either 3 H-taurine (A, A), 35 S- taurine (, O) or "C-taurine (, O). See Methods. The coefficients of correlation are 0.81 for taurine-deficient cats and 0.90 for the control group. B, Taurine accumulation in liver (O) and calculated values for taurine accumulation in retina ( ) related to the time interval following the injection of labeled taurine. The data from control and taurine-deficient cats have been pooled, and each point represents the mean ± S.E.M. for three to four individual measurements. than in normal retinas for a given plasma concentration. For the controls, the equation of the line (Fig. 4, A) permitted calculation of accumulation of taurine in the retina as a function of average normal plasma taurine concentration (i.e., 0.109 mm); a slow decay in taurine radioactivity could be seen (Fig. 4, B). For the taurinedeficient cats, retinal taurine accumulation comparable to normal occurred at a plasma concentration of 0.040 mm; data points from these
Volume 16 Number 7 Reports 677 cats corrected for a plasma concentration of 0.040 mm were similar to those from controls (Fig. 4, B). Labeled taurine accumulation by the liver in normal and slightly taurine-deficient cats was comparable for a given time period; accumulation was maximal in animals sacrificed 2 hr. after injection of radioactive taurine and decreased thereafter. The rate of decrease was faster in the liver than in the retina (Fig. 4, B). The decay of radioactive taurine in the retina and liver during the observed period could be described by simple exponential functions; these data expressed in terms of natural logarithms were used to determine that the half-life of taurine was about 30 to 88 days in retina and about 2 to 5 days in liver. Thin-layer chromatograms of retina, liver, and plasma extracts from control and slightly taurinedeficient cats injected with 3 H-, 35 S-, and relabeled taurine demonstrated that 96% to 98% of the radioactivity in each extract migrated as authentic taurine. No label was found in spots corresponding with isethionic acid. Discussion. These studies, together with previous work, 1-3 have established a role for exogenous taurine in maintaining normal retinal function in the cat. Cats fed taurine-free diets have shown a large range (10 to 45 weeks) in time of occurrence of pronounced (>50% below normal) reductions in ERG amplitude. This range could be explained by the relationship of retinal taurine concentrations and ERG amplitudes to liver stores of taurine. A sixfold to eightfold range in total liver content of taurine was observed in the Chow-fed control cats. Liver stores of tau-_ rine were depleted to near zero before significant, retinal taurine deficiency and reductions in ERG amplitudes were seen in cats on taurine-free diets^ Since the halt-lite of taurine in the liver was found to be about 2 to 5 days, it would be expected that liver content would be near zero only after taurine was eliminated from the diet for 4 to 12 weeks. Since the half-life of taurine in the retina was found to be about 30 to 88 days and since taurine is not metabolized in the cat or rat 7 retina, it would be expected that a greater than 50% reduction in retinal taurine concentration would occur some 10 to 25 weeks after liver taurine was zero, assuming that the liver is one of the major stores 4 for retinal taurine. These half-lives of taurine in the liver and retina help explain why cats of 2 to 4 kg. have developed severe retinal malfunction 15 to 45 weeks after elimination of taurine from the diet. The relatively low liver content of taurine in cats less than 2 kg. compared with those found in larger cats is consistent with the fact that some cats 1-2 on taurine-free diets have developed severe retinal malfunction within 10 weeks after elimination of taurine from the diet. Since it is well known that taurine is stored in. muscle and brain as well as the liver, 4 it is possible that plasma levels were affected by release of taurine from these tissues. Decreases in liver and plasma taurine concentrations were not linearly related; plasma levels remained at about 10% to 20% of normal, while liver concentrations decreased from 55% to 5% of normal. Studies with tritiated taurine of high specific activity showed that uptake of taurine into the retina could be detected when plasma levels of taurine were reduced to 10% of normal (Fig. 4, A); this could explain why retinal levels of taurine were maintained near normal while plasma concentrations of taurine were decreasing. Plasma taurine therefore proved to be a less reliable indicator of the level of retinal taurine depletion than ' ERG amplitudes. ^ The close relationship of ERG amplitudes to retinal taurine deficiency has been demonstrated in this study, even prior to detectable cell death as measured by changes in retinal DNA concentrations. Therefore the ERG amplitude decreases are not simply an indication of loss of photoreceptor cells. Since generation of the ERG a wave and b wave are known to reflect, respectively, hyperpolarization of photoreceptor cells and depolarization of Miiller cells, the fall in peakto-peak ERG amplitudes could be explained by some effect of taurine deficiency on the ionic fluxes around these cells. This idea gains further support from biochemical 8 ' 9 and autoradiograrjhjc studies 10 that have shown that in the retina taurine is maximally concentrated in the outer nuclear layer where photoreceptor cell bodies and Miiller cell processes are located. The mechanism by which taurine deficiency eventually affects photoreceptor cell viability is still unknown. The authors thank Jennifer Moulton for typing the manuscript and Andrew Levin for photography of charts. From the Berman-Gund Laboratory for the Study of Retinal Degenerations, Harvard Medical School, Massachusetts Eye and Ear Infirmary, Boston. This work was supported by grants from the National Retinitis Pigmentosa Foundation, Baltimore, Md.; the George Gund Foundation, Cleveland, Ohio; the William F. Milton Fund, Harvard Medical School, Boston; and Research grants EY00169 (E. L. B.) and EY01687 (S. Y. S.) and Career Development Award EY70800 (E. L. B.) from the National Eye Institute. Submitted for publication Dec. 29, 1976. Reprint requests: Susan Y. Schmidt, Ph.D., Berman-Gund Laboratory, 243 Charles St., Boston, Mass. 02114. Key words: taurine, casein, diet, retinal degeneration, cat, photoreceptor cell, retina, electroretinogram, amino acids.
678 Reports Invest. Ophthalmol. Visual Sci. July 1977 REFERENCES 1. Schmidt, S. Y., Berson, E. L., and Hayes, K. C: Retinal degeneration in cats fed casein. I. Taurine deficiency, INVEST. OPHTHALMOL. 15:47, 1976. 2. Berson, E. L., Hayes, K. C, Rabin, A. R., Schmidt, S. Y., and Watson, C: Retinal degeneration in cats fed casein. II. Supplementation with methionine, cysteine, or taurine, INVEST. OPHTHALMOL. 15:52, 1976. 3. Hayes, K. C, Rabin, A. R., and Berson, E. L.: An ultrastructural study of nutritionally induced and reversed retinal degeneration in cats, Am. J. Pathol. 78:505, 1975. 4. Jacobsen, J. G., and Smith, L. H., Jr.: Biochemistry and physiology of taurine and taurine derivatives, Physiol. Rev. 48:424, 1968. 5. Shipley, R. A., and Clark, R. E.: Tracer Methods for In Vivo Kinetics: Theory and Applications, New York, 1972, Academic Press, Inc., pp. 1-9. 6. Kraffczyk, F., Helger, R., Lang, H., and Bremer, H. J.: Thin layer chromatographic screening test for amino acid anomalies in urine without desalting using internal standards, Clin. Chim. Acta 35:345, 1971. 7. Starr, M. S., and Voaden, M. J.: The uptake, metabolism and release of 14 C-taurine by rat retina in vitro, Vision Res. 12:1261, 1972. 8. Orr, H. T., Cohen, A. I., and Lowry, O. H.: The distribution of taurine in the vertebrate retina, J. Neurochem. 26:609, 1976. 9. Keen, P., and Yates, R. A.: Distribution of amino acids in subdivided rat retinae, Br. J. Pharmacol. 52:118P, 1974. 10. Young, R. W.: The organization of vertebrate photoreceptor cells. In Straatsma, B., Allen, R., Hall, ML, et al., editors: The Retina: Structure, Function and Clinical Characteristics, Berkeley, 1969, University of California Press, pp. 177-209. Fusing human rod outer segments from an eye enucleated for choroidal melanoma. BESSIE BORWEIN, JOHN A. MEDEIROS, AND J. WILLIAM MCGOWAN. In the retina from an eye enucleated for choroidal melanoma opposed rod outer segments were seen by transmission electron microscopy to be fusing together in groups by the confluence of their plasma membranes and the formation of cytoplasmic bridges between adjacent outer segments. These fusions were seen in areas distant from the melanoma but where the retina was not normal in appearance. Human eyes enucleated for choroidal melanoma have been used for a wide variety of investigations on the retina, and there has been an explicit or implicit assumption that the areas not immediately bordering the melanoma are normal in every way. 1 " 7 As a preliminary to a study of the effects of threshold laser lesions on the human retina, a survey was made of the so-called normal areas of the retina of a human eye enucleated for choroidal melanoma. We found that there were areas away from both the laser lesions and the melanoma that appeared typically normal as seen by transmission electron microscopy but that there were also areas that displayed many abnormal features. 7 * 1 This report deals only with the clumping and fusing of rod outer segments, which have not been reported previously. Methods. The eye of a 47-year-old woman with normal 6/6 vision was enucleated by the snare method for a choroidal melanoma, inferotemporal in position. At 3 to 4 hr. before enucleation, argon laser lesions ranging from 50 to 1,000 n spot size were made in the posterior fundus around the disc and macula. Following enucleation the pathologist cut out the area of the melanoma, and within 30 min. after excision the rest of the retina was placed in fixative. The delay was due to unforeseen procedural requirements by the hospital. Fixation was in phosphate-buffered 2.5% glutaraldehyde + 0.5% paraformaldehyde (0.1M, ph 7.4) for 4 hr., and postfixation in 1% buffered osmium tetraoxide for 2 hr. The material was taken through graded alcohols and embedded in Epon 812. Sections were cut on a Reichert 0mU2 ultramicrotome, stained with uranyl acetate and lead citrate, and examined in AE1 800 electron microscope. Samples reported here were taken from three areas of the fundus: (1) circum-macular (of the samples taken, this area is the nearest to the site of the melanoma but is not adjacent to it), (2) nasal to the optic disc, and (3) inferior to the optic disc. Results. Rod outer segments are easily identified in cross-section by their lobulated outlines. Normally, rod outer segments seen in cross-section are single and separated from each other by the interphotoreceptor matrix and also by processes from the pigment epithelium. They do not abut on each other. We found zones in which the rod outer segments appeared to be closely apposed and clumped together in groups (Figs. 1 to 4). At higher magnifications many of these apposed rod outer segments were seen to be enclosed in one continuous plasma membrane, with the formation of small cytoplasmic bridges where the rod outer segments abut on each other (Figs. 1 to 3). In some instances, the outer segment discs looked normal (Figs. 1 and 3); in other cases, there was some disorganization and destruction of discs (Figs. 2 and 4). Clumping and fusing groups of two (Fig. 2),