Carbonic Anhydrase Activities from the Rainbow Trout Lens Correspond to the Development of Acute Gas Bubble Disease

Abstract

Dissolved gas supersaturation is hazardous to fish and can result in gas bubble disease (GBD). Signs of GBD typically include bubbles in the eyes, fins, skin, lateral line, and gill filaments. Ocular abnormalities in diseased salmonids typically occur after aberrant gas production in the eyes. In this study, freshwater rainbow trout Oncorhynchus mykiss were exposed experimentally to percent total gas pressure (TGP%) levels of 104% (control) and 115%. No mortalities occurred during the 7-d experimental period. Effects of GBD were observed externally as a darkened skin, exophthalmia, localized hemorrhage in the eye, and gas bubbles on the operculum. Additional signs included increased swimming activity and, more frequently, panic episodes. Carbonic anhydrase (CA) enzyme activities from the lens and retina were determined at days 0, 1, 3, 5, and 7 of the study. Venous blood gases were also measured on day 7. Retinal pH did not differ between normal and affected fish, but blood characteristics such as the partial pressure of O₂, partial pressure of CO₂, carboxyhemoglobin level, and bicarbonate ion concentration were significantly elevated in affected fish relative to normal fish. Venous blood pH and oxyhemoglobin levels were not significantly different between affected and normal fish. Patterns of response to total dissolved gas levels differed between the lens and the retina. Mean CA activities in the lenses of fish exposed to a TGP% level of 115% were significantly below those of control fish. However, retinal CA activities did not significantly differ between the two groups over the course of the experiment. These findings show that dissolved gas supersaturation reduces CA activity in the rainbow trout lens.

Received June 17, 2010; accepted March 20, 2011     

Clinical and pathological observations on Streptococcus sp. infection on South African trout farms with gas supersaturated water supplies

Infection with a faecal Streptococcus species belonging to Lancefield group D has been a major cause of mortalities on certain trout farms on the escarpment of the Mpumalanga Province in the Republic of South Africa. Evidence is given of a connection between subclinical gas bubble disease and infection with this organism in rainbow trout, Oncorhynchus mykiss. The condition causes serious losses during the summer months when water temperatures rise above 18 degrees C. The larger table-size fish are affected first and present with exophthalmos before developing signs of a typical bacterial septicaemia. The design of the water supply to the fish ponds on these farms allowed supersaturation of the water with atmospheric gas. Vortex formation at water extraction points appears to have been the main cause of air entrainment into the water supply leading to the ponds. By correcting the faults responsible for the development of hyperbaric gas pressure in the water, it was possible to control this disease on two badly affected farms. The overlapping similarities in the clinical signs and pathology of gas bubble disease and that seen in fish suffering from Streptococcus D infection are discussed, as well as the diagnostically significant changes in the gill tissues that are specific to subclinical exposure to gas supersaturation in the water. An explanation is given for the increased susceptibility to infection with this organism.     

Pathological effects of Arcobacter cryaerophilus infection in rainbow trout (Oncorhynchus mykiss Walbaum)

Arcobacter cryaerophilus was isolated from naturally infected rainbow trout (Oncorhynchus mykiss Walbaum), and its pathogenicity was tested by intramuscular injection into 40 healthy 1-year-old rainbow trout at 16 degrees C. The lethal dosage of 50% end point (LD50) for A. cryaerophilus was calculated 2.25 x 10(4) viable cells. Experimental infection caused deaths with gross clinical abnormalities such as degenerated opercula and gills, liver damage, haemorrhagic kidney and serous fluid in swollen intestines. The counts of A. cryaerophilus in kidney, liver and gills of experimentally infected fish ranged from 1.59 x 10(10) colony forming units (cfu)/g to 7.41 x 10(12) cfu/g. The means of erythrocyte (RBC) count, haematocrit level, serum cholesterol, triglyceride, albumin and total protein concentrations in the blood of the experimentally infected rainbow trout group were significantly lower than in the healthy fish. Leukocyte (WBC) counts of the experimentally infected rainbow trout were significantly higher than those of healthy fish. The present work shows that the selected blood characteristics may be good indicators of response to infections in rainbow trout.     

Clinical studies of fish blood: importance of sample collection and measurement techniques

Effects of blood sample collection and measurement techniques were assessed for blood gas tensions, acid-base status, and hematologic and plasma biochemical values of rainbow trout. Blood samples were collected via intraaortic cannulae from immersed, unrestrained fish and from emersed, restrained fish. The samples were analyzed at either fish body temperature (10 to 14 C) or clinical blood analyzer temperature (37 C); results obtained at 37 C were back-adjusted to fish body temperature, using standard mammalian temperature-correction factors. Fish emersion and handling for 30 seconds significantly (P less than 0.05) altered blood PCO2, acid-base status, and hematologic and plasma biochemical values. The results were consistent with respiratory acidosis and hemoconcentration. The use of mammalian temperature-correction factors for determination of fish blood gas tensions and acid-base status yielded values that were significantly (P less than 0.05) different from those measured directly at fish body temperature.     

Effects of Salinity on Yersinia ruckeri Infection of Rainbow Trout and Brown Trout

Abstract

Juvenile rainbow trout Oncorhynchus mykiss and brown trout Salmo trutta acclimated to freshwater or salinities of 9.0‰ or less were exposed to Yersinia ruckeri, the bacterial pathogen that causes enteric redmouth disease (ERM). Both species of fish were kept in the same recirculating systems after bacterial exposure. Rainbow trout mortality was significantly (P < 0.05) different in each salinity: 96.5% in freshwater, 89.5% in water of 1.1‰ salinity, 81.3% in 3.0‰ salinity, and 75.0% in 9.0‰ salinity (model SE = 1.0). All deaths occurred between 3 and 12 d after exposure to Y. ruckeri. Only 2.3% of brown trout in all salinities died, and differences among treatments were not significant. For both fish species, Y. ruckeri was isolated from liver, spleen, and trunk kidney of fish dying during this experiment, and lesions of rainbow trout were consistent with ERM. Yersinia ruckeri was not isolated from brown trout surviving for 21 d after bacterial exposure but was isolated from 3 of 24 surviving rainbow trout; a polymerase chain reaction assay detected the DNA of Y. ruckeri in 3 additional rainbow trout survivors. Neither the lesions of fish with ERM nor the percentage of surviving fish subclinically infected with Y. ruckeri was affected by salinity. Bacterial growth in vitro was not affected by low (≤9.0‰) salinity; however, bacterial adhesion to polystyrene was significantly reduced as salinity increased. Although mortality caused by Y. ruckeri was significantly lower for rainbow trout in water with slightly increased salinity, none of the salinities tested was effective in preventing serious losses caused by this pathogen in recirculating systems.     

Improved husbandry to control an outbreak of rainbow trout fry syndrome caused by infection with Flavobacterium psychrophilum

CASE DESCRIPTION: A cohort of 35,200, 13-week-old, female rainbow trout at a fish farm was evaluated because of a 2-week history of anorexia and lethargy and a mortality rate of approximately 100 fish/d. CLINICAL FINDINGS: Affected fish were lethargic and thin and had disequilibrium, bilateral exophthalmia, pale red gills and kidneys, red-tinged coelomic fluid, and pale brown livers. Some fish were differentially pigmented bilaterally. The presumptive diagnosis was bacterial or viral septicemia. The definitive diagnosis was rainbow trout fry syndrome caused by infection with Flavobacterium psychrophilum. TREATMENT AND OUTCOME: A strategy for controlling the outbreak based on reducing pathogen numbers in affected tanks and reducing pathogen spread among tanks was developed. The option of treating with antimicrobial-medicated feed was discussed with the farmer, but was declined. After changes were made, mortality rate declined quickly, with no more deaths within 10 days after the initial farm visit. CLINICAL RELEVANCE: Bacterial coldwater disease is the most common manifestation of infection with F psychrophilum in fingerling and adult rainbow trout. However, the organism can also cause rainbow trout fry syndrome. This condition should be included on a list of differential diagnoses for septicemia in hatchery-reared rainbow trout fry.     

A Comparative Investigation of Arcobacter cryaerophilus Infection among Albino Crosses and High- and Low-Body-Weight Rainbow Trout

Abstract

Arcobacter cryaerophilus was isolated from naturally infected rainbow trout Oncorhynchus mykiss, and its pathogenicity was tested by intramuscular injection into healthy 1-year-old high-body-weight (HBW) and low-body-weight (LBW) normally pigmented rainbow trout and albino crosses. Experimental infections caused deaths with gross clinical abnormalities such as exophthalmia, liver damage, bloody hemorrhagic kidney and heart, and swollen intestines. No significant differences in deaths were observed among the three infected fish groups. Hematocrit levels in blood of the experimentally infected HBW rainbow trout were significantly less than in healthy fish. No significant decreases were observed in the serum total protein of both the experimentally infected albino crosses and the high weight groups. Albumin and creatinine concentrations in serum were not significantly different among the three treatments.     

The Effects of Whirling Disease on Growth and Survival of Snake River Cutthroat and Colorado River Rainbow Trout Fingerlings

Abstract

Two sizes of fingerling Snake River cutthroat trout Oncorhynchus clarkii behnkei and Colorado River rainbow trout O. mykiss were raised at hatcheries testing negative for Myxobolus cerebralis and stocked into the Dolores and Cache la Poudre rivers from 1999 to 2001. Populations were resampled over a 2-year period to determine which species and size combination had the highest growth and survival rates. Fish were tested for M. cerebralis via polymerase chain reaction and pepsin?trypsin digest analyses. Growth and survival rates between the species and size groups were not significantly different in either river. In the Dolores River, annual survival for both species and sizes of fish combined ranged from 0.063 to 0.12. In the Cache la Poudre River, survival for both sizes of rainbow trout was 0.004; survival for cutthroat trout ranged from 0.182 to 0.53. Larger fish had higher growth rates than smaller fish, and cutthroat trout had higher rates than similar sizes of rainbow trout. In both rivers, a higher percentage of the rainbow trout sample was infected than in the cutthroat trout sample. Rainbow trout also had a higher mean number of spores per head than cutthroat trout, and small rainbow trout had higher spore counts than large rainbow trout. Survival rates for cutthroat trout in the Cache la Poudre River were the highest of any of the groups, suggesting a difference that is biologically significant. Raising fingerlings to sizes greater than 100 mm can improve poststocking survival. If rainbow trout are stocked into contaminated waters, raising fingerlings to a larger size does not appear to improve growth or survival rates. Stocking rainbow trout in the spring could maximize growth rates but will expose fish to greater triactinomyxon densities, resulting in higher intensities of infection.     

Communications: Tolerance of Rainbow Trout to Dissolved Oxygen Supplementation and a Yersinia ruckeri Challenge

Abstract

Dissolved oxygen accounted for a significant proportion of the variation in mortality rate of rainbow trout Oncorhynchus mykiss acclimated for 10 weeks to a range of dissolved oxygen levels and challenged with Yersinia ruckeri. Moderate levels of oxygen supersaturation (150%) resulted in greater cumulative mortality (17.9%) among fish exposed to a 0.5-h static bath challenge with Y. ruckeri (1 × 10⁷ colony-forming units/mL). In contrast, hypoxic (70%) and normoxic (100%) levels of dissolved oxygen resulted in 12.8% and 10% cumulative mortality. Higher oxygen levels provided no additional “protection” against infection by Y. ruckeri in this study.     

Rate of Disappearance of Gas Bubble Trauma Signs in Juvenile Salmonids

Abstract

To assess the rate of disappearance of gas bubble trauma (GBT) signs in juvenile salmonids, we exposed spring chinook salmon Oncorhynchus tshawytscha and steelhead O. mykiss to water containing high levels of dissolved gas supersaturation (DGS) for a time period sufficient to induce signs of GBT, reduced the DGS to minimal levels, and then sampled fish through time to document changes in severity of GBT. Because of the tendency of GBT signs to dissipate at different rates, we conducted trials focusing on emboli (bubbles) in the gill filaments and lateral line and separate trials that focused on bubbles in the external surfaces (fins, eyes, and opercula). Bubbles in gill filaments dissipated almost completely within 2 h after transfer of fish to water of nearly normal DGS (104%), whereas bubbles in the lateral line dissipated to negligible levels within 5 h. Bubbles on external surfaces were more persistent through time than they were in gill filaments and the lateral line. Although typically dissipating to low levels within 48 h, external bubbles sometimes remained for 4 d. Assuming a direct relation exists between easily observable signs and direct mortality, our results suggest that fish can recover quickly from the potentially lethal effects of DGS once they move from water with high DGS to water of almost normal gas saturation. These results should be of fundamental importance to fishery managers interpreting the results of monitoring for the severity and prevalence of GBT in juvenile salmonids in the Columbia River system and perhaps elsewhere.     

Characteristics of Peripheral Blood Cells from Rainbow Trout Evaluated by Particle Counter,Image Analysis,and Hemocytometric Techniques

Abstract

Hematological assessments of peripheral blood cells from rainbow trout Oncorhynchus mykiss were compared by using a hemocytometer, image analysis, and an electronic particle counter. Although statistical differences were noted, the diameters of white blood cells (WBCs) were comparable among the methods used, but lengths and widths of red blood cells (RBCs) were remarkably different among the three methods. Red blood cells were significantly smaller when the particle counter was used compared with sizes obtained by using the hemocytometer or image analysis. Counts of whole blood and isolated RBC concentrations were similar between the hemocytometer and the particle counter in the automatic calculation mode; yet they were significantly different with the manual calculation method, which used the raw data from the particle counter. White blood cell counts from peripheral whole blood could not be obtained with the particle counter; it converted all particle volumes into spherical equivalents and, thus, interpreted rainbow trout RBCs and WBCs to be similar in size, resulting in an inability to accurately count the number of WBCs in mixed-cell populations. In addition, a lysing agent, which is commonly used to count WBCs in mammalian peripheral blood, did not completely remove all RBCs from samples of trout peripheral whole blood because of the presence of nuclei within the fish RBCs. These results indicate that assessing fish hematological characteristics with the particle counter may have limitations.     

Developmental competence of eggs produced by rainbow trout Doubled Haploids (DHs) and generation of the clonal lines

Poor quality eggs produced by the fully homozygous doubled haploids (DHs) may impair generation of clonal lines in fish species. In the present research, gynogenetic development of rainbow trout (Oncorhynchus mykiss) was induced in eggs originated from the DH females. Eggs were activated with the UV‐irradiated grayling (Thymallus thymallus) spermatozoa and subjected to the high hydrostatic pressure (HHP) shock to provide diploid clonal individuals. Only two of four DH females produced eggs that were successfully activated by the irradiated spermatozoa and subsequently developed into the gynogenetic embryos. Survival rates of rainbow trout from the clonal lines equalled 21.5% and 19.8% during embryogenesis and decreased after hatching to 18.6% and 14.9%, respectively. Some of the dead rainbow trout clones collected between hatching and swim‐up stage were emaciated and exhibited spinal deformities including scoliosis. Provided results confirmed limited developmental competences of eggs produced by rainbow trout DH females. Clonal rainbow trout developing in such eggs exhibited reduced survival and increased frequency of the body abnormalities.     

Distribution of Myxobolus cerebralis in Salmonid Fishes in Montana

Abstract

Over an approximately 2-year period, 20,974 fish (trout and other salmonid species) from 230 separate waters (creeks, rivers, lakes, reservoirs, ponds, hatcheries, and irrigation ditches) within 21 of the 22 major drainages in Montana were examined for Myxobolus cerebralis. Nine of the major river drainages have waters containing infected fish: Beaverhead, Big Hole, Blackfoot, Clark Fork above the Bitterroot River, Flathead below the south fork of the Flathead River, Jefferson, Madison, Missouri above the Marias River, and Yellowstone above the Bighorn River. The Beaverhead, Clark Fork above the Bitterroot River, Jefferson, Madison, and Missouri above the Marias River have the greatest number of waters containing fish infected with M. cerebralis. Comparisons of infection levels (number of pooled samples that contain fish infected with M. cerebralis) between species among these drainages show significantly lower levels of infection in brown trout Salmo trutta and rainbow trout Oncorhynchus mykiss in the Missouri above the Marias River drainage and significantly higher levels of infection of rainbow trout in the Jefferson. Comparisons of differences in infection levels between drainages among species show that, in the Beaverhead, Clark Fork above the Bitterroot River, and Madison, infection levels in brown trout are significantly higher than in rainbow trout. This is partially attributed to losses of juvenile rainbow trout because of M. cerebralis infection, leading to biased samples. Histopathologic studies showed lesions were consistently less severe in brown trout than other species and occurred in a different location (gill arches versus ventral calvarium). In six of the nine affected drainages (Beaverhead, Blackfoot, Clark Fork above the Bitterroot River, Flathead below the South Fork, Jefferson, and Madison), infected fish were found at or near the time that intensive sampling was initiated in the spring of 1995. In the three remaining affected drainages (Missouri above the Marias River, Yellowstone above the Bighorn River, and the Big Hole), infected fish were not identified until at least 15 months after the initiation of widespread testing. This indicates that in the first six drainages listed above, the infection was well established prior to 1995 but spread into the last three drainages in the ensuing months. Methods of transmission and the sources of infection are unknown, although the absence of infected fish in state, private, and federal hatcheries in Montana indicates hatchery fish from these sources are not likely to be responsible.     

Effect of vitamin E on carbonic anhydrase enzyme activity in rainbow trout (Oncorhynchus mykiss) erythrocytes in vitro and in vivo

Considering that the excessive usage of vitamin E causes hypervitaminosis and thus reduces blood erythrocyte concentrations, therefore it is worth studying how its pharmacological dosage affects the activity of carbonic anhydrase (CA) enzyme found in erythrocytes of rainbow trout (Oncorhynchus mykiss) in vitro and in vivo. Vitamin E inhibited CA enzyme and the IC50 value of the vitamin was 0.039 mM in vitro. Similarly, it was seen that vitamin E inhibited CA enzyme activity after the first hour following vitamin E injections in vivo. The activities of CA in groups of trout given vitamin E injection were measured at 1, 3 and 5 h and the corresponding activities were found to be 772.7 +/- 290.5 (P < 0.05), 1286.4 +/- 378.2 and 1005.7 +/- 436.1 enzyme units (EU) g Hb(-1). The difference over the control was significant (P < 0.05) in the first hour and insignificant at 3 and 5 h (P > 0.05). The activity of CA in the control, which did not contain vitamin E, was determined as 1597.7 +/- 429.0 EU g Hb(-1).     

Effects of Acute Stress on Blood Clotting and Yeast Killing by Phagocytes of Rainbow Trout

Abstract

The effects of different levels of acute stress on blood parameters in rainbow trout Oncorhynchus mykiss were examined. Experimental fish were exposed to a combination of stressors, including handling, confinement, and thermal shock, and were bled 30 s to 30 min later. Control fish were sampled quickly without intentional stress. Two activities of blood were investigated: (1) whole-blood clotting (coagulation) times and (2) the ability of heparinized plasma to influence the survival of yeast soaked in this plasma and subsequently exposed to trout phagocytes. The speed of blood clotting increased with increasing levels of stress, whereas blood from many resting fish failed to clot. Yeast incubated in plasma from stressed fish before exposure to pronephric phagocytes experienced a greater reduction in colony-forming units (cfus) than yeast incubated in plasma from resting fish. Different levels of stress did not induce significant differences in reduction of yeast cfus. These data support the idea that acute stress may actually stimulate primary defense mechanisms rather than suppress them. Furthermore, whole-blood clotting times seemed to be reliable indicators of the severity of stressors to which fish were exposed. We postulate that factors appearing in blood very quickly, such as catecholamines and specific plasma proteins, play important roles in both the killing activities of phagocytes and the blood-clotting process.     

Susceptibility of Selected Inland Salmonids to Experimentally Induced Infections with Myxobolus cerebralis,the Causative Agent of Whirling Disease

Abstract

Laboratory exposures to the infectious stages (triactinomyxons) of Myxobolus cerebralis demonstrated a range of susceptibility to whirling disease among four species of inland salmonids. Replicate groups of each species were exposed to two concentrations of triactinomyxons, a low dose (100–200 per fish) and a high dose (1,000–2,000 per fish). Exposed fish were evaluated for clinical signs, for severity of microscopic lesions at 35 d, 2 and 5 months, and for spore concentrations in the head cartilage at 5 months. A standard strain of rainbow trout Oncorhynchus mykiss matched for age served as a susceptible species control. Rainbow trout, westslope cutthroat trout O. clarki lewisi, Yellowstone cutthroat trout O. clarki bouvieri, and bull trout Salvelinus confluentus were susceptible to M. cerebralis infections. Clinical signs, including radical swimming (“whirling”) and black tails, were observed at 7 weeks postexposure among rainbow and cutthroat trout challenged at 3 weeks of age. Clinical signs were rare among bull trout exposed at an age of 4 weeks and absent among rainbow and cutthroat trout exposed at 3 months posthatch. Most rainbow, cutthroat, and bull trout were found to be infected when examined at 5 months postexposure. The most severe microscopic lesions among infected fish at 5 months postexposure were found among rainbow trout. Cutthroat trout had less severe lesions, bull trout had mild infections, and no evidence of infection was found among Arctic grayling Thymallus arcticus. Mean spore concentrations among infected fish correlated with the severity of microscopic lesion scores. Rainbow trout had mean concentrations of spores in head cartilage reaching 10⁶, whereas more resistant species such as bull trout had 10⁴ spores; no spores were found among Arctic grayling at 5 months postexposure.     

The leech Batracobdelloides tricarinata (Blanchard, 1897) (Hirudinea: Glossiphoniidae) as a possible reservoir of the rainbow trout pathogenic Streptococcus species

A Streptococcus species biochemically and serologically identical to the rainbow trout pathogenic Streptococcus species was isolated from the internal organs of the fish specific leech, Batracobdelloides tricarinata. These leeches were obtained from Roodeplaat Dam, near Pretoria, in which rainbow trout do not occur. This is the first isolation of this bacterium from an environmental source not related to rainbow trout and it is proposed that this leech is a possible reservoir of the rainbow trout pathogenic Streptococcus sp. in South Africa.     

Phagocytic activity of macrophages of rainbow trout against Vibrio anguillarum and the opsonising effect of antibody and complement

The phagocytosis of Vibrio anguillarum by peritoneal macrophages from normal rainbow trout was enhanced by antibody and complement. Treatment of either macrophages or the bacteria by antibody also enhanced opsonisation. Five weeks after immunisation with V anguillarum, the phagocytic activity of macrophages from rainbow trout was increased significantly compared with the activity of those from unvaccinated fish. Although agglutinin titres did not increase until three weeks after immunisation, seven out of 10 fish challenged one week after immunisation survived, indicating that immunised fish had developed resistance to vibrio infection before significant levels of antibody or phagocytic activity became detectable.     

The effects of Myxobolus cerebralis on the physiological performance of whirling disease resistant and susceptible strains of rainbow trout

The development of rainbow trout Oncorhynchus mykiss strains that are resistant to whirling disease has shown promise as a management tool for populations in areas where Myxobolus cerebralis is present. However, the physiological effects of the disease on characteristics necessary for fish survival in natural river conditions have not been tested in many of these strains. Five rainbow trout strains were evaluated for their swimming ability and growth characteristics in relation to M. cerebralis exposure: the resistant German rainbow trout (GR) strain (Hofer strain), the susceptible Colorado River rainbow trout (CRR) strain, and three intermediate (hybrid) strains (F1 = GR x CRR; F2 = F1 x F1; B2 = backcross of F1 x CRR). Three broad response patterns among strain and exposure were evident in our study. First, exposure metrics, growth performance, and swimming ability differed among strains. Second, exposure to the parasite did not necessarily produce differences in growth or swimming ability. Exposure to M. cerebralis did not affect batch weight for any strain, and critical swimming velocity did not differ between exposed and unexposed families. Third, although exposure did not necessarily affect growth or swimming ability, individuals that exhibited clinical deformities did show reduced growth and swimming performance; fish with clinical deformities were significantly smaller and had lower critical swimming velocities than exposed fish without clinical deformities. Research and management have focused on GR x CRR hybrid strains; however, given the performance of the GR strain in our study, it should not be discounted as a potential broodstock. Additional field trials comparing the GR and F1 strains should be conducted before wholesale adoption of the GR strain to reestablish rainbow trout populations in Colorado.     

Intraocular oxygen tensions in normal and diseased eyes of farmed halibut

We have previously documented ocular abnormalities in farmed halibut and reported that the production of gas-filled posterior segment cysts appears to be central in the pathogenesis of many, if not all of the signs in these eyes. In a number of fish, gas bubbles may be seen in the anterior chamber of the eye, especially after trauma or strenuous exertion associated with handling. A knowledge of the composition of this gas is important in understanding fully the mechanism of gas and cyst production. Here we report investigations into the composition of gas in the globes of normal fish, fish in which intraocular cysts have been documented previously by ultrasonography and fish where post-mortem examination demonstrated both gas- and fluid-filled choroidal cysts. High levels of oxygen were demonstrated in samples of aqueous from all fish but fish with affected eyes had a statistically significantly higher partial pressure of oxygen than did fish with normal eyes.