Investigations into the Causes of Early Larval Mortality in Cultured Summer Flounder (Paralichthys dentatus L.)

Experiments were conducted to investigate larval mortality in cultured summer flounder during the first two weeks after hatch. The importance of feeding success, parentage, addition of algae, water quality, and the microbial community to mortality during this period were investigated. Larvae were raised in 2-L bowls at initial densities of 50 and 75/L with light aeration, 12L:12D photoperiod, and regular 1-L water changes. In all experiments mortalities were recorded and removed daily. In the first two experiments daily samples of larvae were taken to assess feeding success and to relate that to survival. The second experiment investigated the effects of both feeding success and the addition of algae to larval culture bowls on larval survival. The third experiment investigated the effects of water quality and bacterial load on survival during the experimental period. The first two experiments indicated that failure to establish .feeding is probably not the cause of catastrophic mortality of the larvae, although a statistical relationship existed between feeding incidence and survival in two of six cases. High variability (34 ± 38% n=82) in survival was seen in the first two experiments (both within and between parental crosses) suggesting that catastrophic mortalities were due to rearing conditions rather than gamete quality. The addition of algae to larval cultures increased survival from 13 ± 24% (n=33) during the

experiments mortalities were recorded and removed daily. In the first two experiments daily samples of larvae were taken to assess feeding success and to relate that to survival.
The second experiment investigated the effects of both feeding success and the addition of algae to larval culture bowls on larval survival. The third experiment investigated the effects of water quality and bacterial load on survival during the experimental period. The first two experiments indicated that failure to establish .feeding is probably not the cause of catastrophic mortality of the larvae, although a statistical relationship existed between feeding incidence and survival in two of six cases. High variability (34 ± 38% n=82) in survival was seen in the first two experiments (both within and between parental crosses) suggesting that catastrophic mortalities were due to rearing conditions rather than gamete quality. The addition of algae to larval cultures increased survival from 13 ± 24% (n=33) during the ii ·f i rst experiment to 46 ± 39% (n=49 ) during the second experiment.
The final experiment indicated that larval mortality was not linked to the measured microbial or water quality conditions. The relationship between the percentage of floating eggs at time of fertilization and survival at 10 DAH was found to be not significant, providing further evidence that gamete quality was not as important as rearing conditions in these experiments.
iii and didn't when I didn't think I did, and in general was a very good friend.

ACKNOWLEDGMENTS
Over the two years at the University of Rhode Island, I was fortunate to have the assistance of many people.
Special thanks go to Sheila Polof sky for teaching me iv histology. "Given the opportunity , any system will undergo spontaneous change in that direction resulting in an increase in entropy, entropy being a measure of randomness or disorder." The second law of thermodynamics.
"Everything existing in the universe is the fruit of chance and necessity." Democritus "What a long strange trip it's been." Grateful Dead.  The culture of larval marine fish has long been a problem.
In the Northwestern Atlantic, a pleuronectiform flatfish, the summer flounder (Paralichthys dentatus L.) is a popular species for commercial and sport fishing. As with many species, the population of this fish has declined (NOAA/NMFS, 1993) to the point where severe restrictions on the allowed catch have been put in place. These restrictions may make both commercial aquaculture and/or stock enhancement economically feasible. Either of these ventures would benefit from increased hatchery efficiency in production of juvenile fish.
A period of high mortality during the larval stage occurs from hatch through first feeding. Successful first feeding, in which the larvae make the transition from endogenous to exogenous nutrient supply, is critical to survival. Smigielski (1975) found that in summer flounder, 90-95% of mortalities occurred within one week of hatch.
A major consideration in early larval feeding is the re l ationship between larval mouth gape and prey size (Houde , 1 9 7 8 ; Beck & Bengtson, 1982;Appelbaum , 1985;Leger et a l ., 1987van der Meeren, 1991Watanabe & Kiron, 1994;Lavens et al., 1995;). This relationship is critical in the hatchery setting where it is usual practice to provide a single prey species for the cultured larvae. An associated factor would be developmental problems of the jaw apparatus, which would affect ingestion of prey. Abnormal jaw development has been a concern in halibut culture (Pittman et al., 1987;Morrison & MacDonald, 1995;) and has been commented on in summer flounder culture (Bisbal, 1993).
The addition of algae to larval culture systems (the so-called green-water method, as opposed to the clear-water method) seems to have become an acce~ted practice (Eda et al., 1990;Reitan et al.~ 1993;Naas et al., 1992;Tamaru et al., 1994;Stottrup et al., 1995). The advantages of the addition of algae to the larval fish culture include nutrition (rotifers in tanks maintain nutritional values via continued uptake of algae) (Reitan et al., 1993), antibacterial properties of algae (Kellam & Walker, 1989, Strottrup et al., 1995, and enhanced feeding with increased turbidity (Boehlert & Morgan, 1985). However Dhert et al. 5 (1994) came to the conclusion that the addition of algae was not necessary during the rotifer feeding stage in turbot culture. In our laboratory, it has become de facto practice to add algae to larval culture tanks. One study in our laboratory (Ainley, unpublished data) showed that addition of algae significantly increased survival of summer flounder from 5-42 days after hatch (DAH) .
For the last six years we have been investigating the potential of summer flounder for aquaculture, with emphasis on the larval stages through metamorphosis. We routinely placed thousands of newly hatched larvae from each individual male X female cross into a 150 L aquarium. Some of these batches survived and grew well, while others did not.
Because we did not rear and examine replicate batches from each cross we do not know whether early larval survival rates were being determined by gamete quality (e.g., due to nutritional provisioning of eggs or genetics) or tank conditions (e.g., water quality factors or bacterial contamination), or a combination of the two. While large variability has been reported in larval culture survival (Smigielski, 1975;Klein-MacPhee, 1981;Eda et al., 1990;Buckley et al., 1991;Reitan et al., 1993;Stottrup et al., 1995), generally few authors in the aquaculture literature report inter-replicate variability, or they have had too few replicates to determine if there is a significant variance.
6 These experiments, conducted over a two year period, were designed to investigate larval first feeding mortality.
The first experiment, consisting of two trials, was designed to investigate the variability within and between crosses and determine the degree to which larval mortality at the critical first feeding was a result of a failure of the larvae to initiate feeding. Such failure might be due to a mismatch in larval mouth gape and prey size, to a jaw development abnormality which affected the larvae's ability to ingest prey or to a digestive tract problem which interfered with the digestion and assimilation of the prey.
The second experiment, consisting of two trials, was designed to elaborate on the findings of the first experiment.
In these trials we continued the quantification of mortality and initiation of first feeding. Additionally, this experiment was designed to investigate whether the addition of algae to the culture medium and rinsing of the rotif ers before being offered significantly affected survival or variability.
The third experiment, consisting of a single trial, investigated whether the inter-replicate variability in survival was associated with bacterial flora, water quality, or some combination of the two.  (Smigielski, 1975) over a two week period. Eggs and milt were collected separately in dry containers. The milt was activated with a small amount of seawater, added to the dry eggs, and allowed to stand for three to five minutes. Seawater (100 ml) was added and the fertilized eggs were poured into a graduated cylinder and allowed to stand for five to ten minutes, after which total volume of eggs and volume of floating eggs were determined. The floating eggs were assumed to be of good quality, whereas sinking eggs were assumed to be of poor quality. The floating eggs were then poured into 37-L aquaria with seawater filtered to On days three and ten (initiation of first feeding and the end point of these trials) larvae were measured for total length and fixed in neutral buffered formalin for histological analysis. Samples were embedded in paraffin blocks and serial sagittal sections of 6 µn were prepared.

METHODS AND MATERIALS
Prepared slides were stained with hematoxylin and eosin, or every other slide in a series was stained with a Mallory-Heidenhain trichrome stain, chosen to investigate cartilage development of the jaw apparatus (Humason, 1962;Bisbal & Bengtson, 1995a) . intestine (signs of pinocytosis in healthy larvae which was absent in starved specimens) (Bisbal & Bengtson, 1995c EXPERIMENT 2, Green-water trials. Experiment two consisted of two trials. The first trial was conducted using a single male X female cross. A 2 X 3 factorial design with 5 replicate bowls per treatment was used. The first factor was culture medium (algae added to the seawater, or not) and the second factor was feeding condition (larvae fed rinsed rotifers, larvae fed unrinsed rotifers, or larvae not fed). Rinsed rotifers consisted of the rotifers being sieved and rinsed· with clean seawater before being offered to larvae. Unrinsed treatments consisted of rotifers added directly from the rotifer culture to the treatments. Algae, a mixture of equal volumes of T. suecica and I. galbana, was added to the appropriate culture bowls at a rate of 50 ml per day.
Density of rotifers was maintained at 5000/L throughout the trial. This trial began on 3 DAH, as in the previous trials, but was extended to last until 14 DAH to allow for 12 the possible extended survival due to any nutritional value (Van der Meeren, 1991;Stottrup, 1994) of algae in the unfed , algae added, controls. The culture methodology of experiment 1 was followed with some changes: a) the number of larvae per bowl at the start of the trial was increased to 150 to account for sampling during the increased length of the trial, and b)larvae were measured on days 3, 10, and 14 after hatch. Daily samples were fixed in neutral buffered formalin for possible future histological examination.
In trial 2 of this experiment the milt from one male was used to fertilize separate batches of eggs from three females. Trial 2 was designed as a 2 x 2 x 3 factorial with factor one being culture medium (algae added to the seawater, or not), factor two being feeding condition (larvae fed rinsed rotifers, or larvae unfed), and factor three being cross (cross one, cross two, or cross three).
Each cell of the experiment had five replicate bowls.
A single male X female cross was cultured using ten bowls with 150 larvae each.  Keppel, 1991) which is a procedure for measuring the strength of association.
All analyses were done using the SYSTAT statistical program.
All analyses had, a priori, the significance level set at et=0.05 (Cowles & Davis, 1982).
In trial 1 survival ranged from 0-80% (mean 28% ± 32% at 10 DAH, n=l2) among replicate bowls (Fig. 1). When the average percentage of larvae with food in gut (for days when there were larvae alive) for each replicate was regressed against the survival in that replicate at 10 DAH the relationship was not significant (r 2 =0.24, P>0.05). Low levels of jaw or skeletal deformities were noticed in both the daily samples and mortalities, (totals in the first two experiments, four trials, were 84 and 90, respectively, out of 17,150 total larval observations, 0.48% and 0.52% respectively) . Complete mortality was observed in some fed replicates beginning at 5 DAH, whereas complete mortality was not observed in the unfed replicates until 9 DAH.
ANOVA at 10 DAH showed no significant effects on survival from cross, food in gut, hatching mortality, or length at 3 DAH of larvae.
Survival in trial 2 ranged from 0-60% (mean 5% ± 15% at 10 DAH, n=21) ( Fig. 2, A & B). When the average percentage of larvae with food in gut (for days when there were larvae alive) for each replicate was regressed against the survival in that replicate at 10 DAH the relationship was weak and not significant (r 2 =0.14, P>0.05). ANOVA at 10 DAH showed no significant effect of food in gut, cross, hatching mortality, or length at 3 DAH of larvae on survival.
It is noteworthy that the unfed controls from crosses B,C,E and F ( Fig. 2, A & B) survived longer than did the fed treatments.
Histological analysis of the larvae showed that development of the digestive tract, and mucosal epithelium appeared to proceed normally, as did cartilage development in the jaw apparatus.
Overall, experiment 2, trial 1, was characterized by high survival (0-93%, mean 75 ± 30% at 10 DAH, mean 48 ± 37% at 14 DAH, n=19) ( In the replicates which did not have algae added, average percent of larvae with food in gut (for days when larvae were alive) regressed on survival at 10 DAH exhibited a relationship that was not significant (r 2 =0.39, P>0.05).
In the replicates which did have algae added, average percentage of larvae with food in gut (for days when larvae were alive) regressed on survival at 10 DAH exhibited a weak relationship that was not significant (r 2 =0.05, P>0.05).
Trial two in experiment 2 (mean survival 28 ± 32% at 10 DAH, 4 ± 12% at 14 DAH, n=30) revealed a much different picture than trial one ( Fig EXPERIMENT 3, Bacterial-water quality trial. Survival ranged from 0-85% (mean 81 ± 14% at 10 DAH, mean 59 ± 35% at 14 DAH, n=5) in the fed replicates, with only one replicate exhibiting complete mortality before the end of the experiment (Fig. SA). Colony forming units enumerated on the marine agar showed a trend in all replicates to increase towards the end of the experiment.
Presumed Vibrio spp. appeared early in the experiments, but then disappeared by 10 DAH.
Pseudomonas were never detected on the cetrimide agar in any of the larval (fed or control ) , rotifer, algal cultures, or in seawater alone. In the unfed controls ( Fig. 5B) the same trends were evident: an initial Vibrio presence which then decreased and an initially low CFU on the marine agar followed by a increase.
In the rotifer culture there was a low but consistent presence of presumed Vibrio spp. The CFU on marine agar was consistently higher than the Vibrio CFU on the TCBS agar.
The algal culture never sh9wed CFU on TCBS agar, but showed relatively high levels of CFU on the marine agar. Survival 9f individual replicates ranged from 0-98%.
Mean survival for all replicates in a given cross treated in the same manner ranged from 3-81%. Coefficients of variation (CV) ranged from 20-430 for all replicates (n=87) .
Plots of data points relating average daily percentages of food in gut with survival at 10 DAH for each replicate bowl in experiments 1 & 2 indicate interesting differences between bowls with and without algae (Fig. 6).
In bowls without algae, if average food in gut was below about 40%, survival was 0%, whereas, if average· food in gut was above about 40%, survival varied from 0-90%. In bowls with algae, only one replicate had average food in gut below about 60%, but those above about 60% had survival levels from 0-90%.
It appears that some aspect of algae addition may have increased the average percentage of larvae with food in the gut.
Statistical source tables, regression equations, and graphs are located in Appendix III.

DISCUSSION
This series of experiments has yielded data that l)quantifies the variability in survival within and between crosses and treatments, 2)indicates that inability to initiate first feeding is probably not the sole cause of mortality, 3)demonstrates that there was a significant statistical relationship between feeding and survival in only two cases out of six examined, 4)suggests that some as yet unidentified, factor(s) in the rearing environment is(are) the cause of catastrophic mortality, 5)suggests that green water can sometimes improve survival, and 6)demonstrates that there is no relationship between percentage of floating eggs at time of fertilization and larval survival through the critical first feeding period.
The fact that the results are equivocal (sometimes green water results in higher survival, sometimes not; sometimes feeding was correlated with survival·, sometimes not) demonstrates the complexities of larval rearing. Bromage et al. (1994)  These experiments showed that inability to establish first feeding on prey, Brachionus plicatilis, by larvae was Histological examination did not detect signs of starvation as described by Bisbal & Bengtson(1995c). Some observations of skeletal deformities were observed but these deformities never reached the proportions (27%) reported by Andrades et al. (1996) (Kellam & walker, 1989), b)increased· feeding due to turbidity (Boehlert & Morgan, 1985), c)maintenance of rotifer nutritional value to the larvae (Lubzens et al., 1989), and d)therapeutic properties (Austin et al., 1992).
There was no indication of direct nutritional value from the addition of algae to the larvae. The lack of significant differences between the unfed control replicates in experiment 2 trial 1 & 2, with and without the addition of algae, provide evidence of this.
A similar result was also reported by Qasim (1955) . We did notice that summer flounder larvae did ingest algae at low levels. The possibility of algal nutrient value (Naas et al., 1992), the possible presence of enzymes appropriate for algae digestion in larval fish (Baragi & Lovell, 1986), and physical stimulation of digestive enzyme release even due to inert particles (Hjelmeland et al., 1988) has been reported. The difference in patterns of percent food in gut between the clear and green water treatments (Fig. 6) suggests that the addition of algae does enhance the feeding response in larval summer flounder.        (1) (.) ...     APPENDIX I

LITERATURE REVIEW
In the western North Atlantic, summer flounder, Paralichthys dentatus L., a pleuronectiform flatfish , is a popular target species for sport and commercial fishing. A concise summary of the habitat, spatial and temporal distribution of larvae, juvenile, and adult stages is found in Able & Kaiser(1994). Morse (1981) found that the males are generally smaller than the females.
Smith & Fahay that were landed at a Boston pier. He stated that summer flounder seem to be particularly susceptible to abnormal coloration, and that this is often found coupled with eye migration and fin ray abnormalities. The problem with abnormal coloration is seen often in cultured flounder and is apparently related to a nutritional deficiency of highly unsaturated fatty acids (HUFA) in larval stages of Japanese flounder Paralichthys olivaceus (Kanazawa, 1993), turbot Scophthalmus maximus (Dhert et al. 1994) and summer flounder Paralichthys dentatus (Baker & Bengtson, 1996) In contrast to the sparse literature on summer flounder culture, literature on turbot culture is quite extensive.
Anthony (1910)  Al-Maghazachi & Gibson (1984) divided the process in turbot into 5 distinct phases, each sub-divided into substages, based on gross morphological changes. During this time period, the digestive tract also undergoes functional and morphological changes, described for summer flounder by . Segner et al. (1994) describe this developmental sequence in turbot. The latter authors proposed a division of the development of organs found in larvae into two groups, l)those found in the larvae at hatch, differentiated into functional organs, and 2)those which are not present in the larvae, but develop during metamorphosis. Padros et al. (1993) followed the histopathological events during the critical first feeding stage and noted that progressive bacterial colonization of the intestine was seen in turbot larvae, especially in the more mature larvae of the cohort. These authors suggested that the immune system of the larval flatfish is less well developed than that of other teleost which might account for the increases susceptibility to bacterial infections.
Cousin & Baudin-Laurencin (1987) and Cousin et al. (1986) examined development of the turbot in a pair of histological studies. Govoni et al. (1986) Fukuhara (1986) had looked at t~e Japanese flounder with the same outlook two years earlier, adding ecological changes as well. One of the concerns noted by Bisbal (1993) in earlier experiments was jaw apparatus maldevelopment. Morrison & MacDonald (1995) looked at this in halibut, and came to the conclusion that, at least in halibut, it was due to a secondary bacterial infection. Pittman et al. (1990) described the morphological and behavioral development of halibut larvae. Appelbaum et al. (1983) (1993).
An important component to larval culture, one that has to a large extent not been investigated, is egg quality and broodstock nutrition. In more mature animal husbandry fields, the importance of broodstock management has been explored . As marine fish culture is a relatively new endeavor , compared to land animal culture , other concerns have been considered more pressing than broodstock management. Kjorsvik et al. (1990) Devauchelle et al. (1987) reported on the same parameters, also over a 12 year period, on the spawning of sole in the laboratory. Berlinsky et al. (1996) reported on the induced spawning of southern flounder 67 using gonadotropin analogues. Suquet et al. (1995 ) reported on optimal time and ratio of sperm:egg interaction, with a time of 3 minutes recommended for sperm:egg interaction.
Howell & Scott (1989) discussed the ovulatory cycle and egg deterioration.
Post-ovulatory deterioration is a concern in my work , as the determination of optimum spawning stage has not been investigated in summer flounder.
The variability of larval survival in summer flounder is a major finding of my work. While this variability is critical to experimental design and findings, it is often glossed over in the literature. Smigielski (1975) found survival to metamorphosis for summer flounder to be between 0 and 5%, with a mean of 1.3% ± 2.0%. He did not report survival during the critical first feeding stage, nor did he report parentage. Klein-MacPhee (1981), using two replicates per treatment, reported mean survival to 30 days after hatch of 0, 11.8, 37.4, 0.9, and 0.6% for stocking densities of 5, 10, 20, 40, and 80 fish per liter respectively.  report survival (pooled mean of three replicates) of up to 40% at a culture temperature of 12.5°c, and 90% at 21°c. This is the extent of summer flounder survival data that I found.
Data for other species also indicate that first feeding mortality is extremely high. Buckley et al. (1991) found that larval winter flounder survival ranged from 0.07 to 6~ 0 • Shelbourne (1964) reported survival through metamorphosis of plaice larvae to be 0 . 1 to 6.6% over the span of five years, 1957years, -1961years, . Qasim (1955 reported between 0 and 40% survival of Banius pholis L. at 32 days after hatch. Planas (1994), in his review of different production systems for turbot larval culture, reported a survival range of 1 to 37% for 18 experiments. Minkoff & Broadhurst (1994 ) , in their discussion of intensive turbot fry production in Europe, stated that while survival of larvae can be 40-50% in the first month, rearing success is unpredictable. These authors reported that, up to 9 days after hatch, egg and larval quality have the largest impact on survival. They reported mortalities of 25-80% during the critical early larval stages from hatch to first feeding. Dhert et al. (1994) reported survival up to 20%, but did not report parentage or the number of repetitions used. Although Olesen & Minck (1983) reported in the abstract of their article survival of turbot larvae of 40%, they actually showed results from 7 experiments, with survival ranging from 9 to 40%. I assume that the survival rates are a pooled mean, but no standard deviation was reported. In one of the early works of the "modern era", Jones (1973) reported survival of turbot larvae to be very low, less than 1% overall. In a later research effort, Jones et al. (1981) reported mortality to be greatest 5-12 days after hatch.
Overall survival ranged from 3-6%, with individual batches ranging from 0-25%. In other species of fish survival rates vary, but the early larval period during the transition from endogenous to exogenous feeding remains a critical time. N~ss et al. (1996) reported 69% survival during the first 15 days after the initiation of feeding for halibut, with 20% of the mortality occurring between days 3 and 6. Appelbaum (1985) reported survival rates between 20 and 90% for sole larvae during the critical first-feeding stage. Eda et al. (1990), working with striped mullet, reported larval survival of 11.5 ± 6.3% and 34.3 ± 11.1% during two years of experiments.
Larval nutrition, and the development of an artificial feed for larvae is long term goal of research into the early larval stages of fish culture. We currently rely on the culture of live prey, rotifers and Artemia. Lubzens et al. (1989) reviewed the culture of rotifers and their suitability as first prey for larval marine fishes. Scott & Baynes (1978) reported on the nutritional value of rotifers when they were cultured on different · algae and at different temperatures. Leger et al. (1987) reviewed the use of Artemia in larval culture. The development of an artificial diet for larval culture is an active area of research. In reviews by Dabrowski (1986), Watanabe & Kiron (1994) and Lavens et al. (1995) it was noted that this goal is still not at hand.
An area of active research is the bacterial milieu of larval culture. Levin et al. (1972) reported that Vibrio anguillarum was isolated from winter flounder and found to be the cause of disease. Austin (1983) reported on the bacterial microflora found in a coastal fish farm and isolated 30 different bacteria, including Vibrio and Pseudomonas species. Tanasomwang & Muroga (1988) investigated the intestinal flora of Japanese flounder larvae and found that the levels of bacteria decreased with the transition from live to artificial diets, with the two largest groups represented being Vibrio and Pseudomonas. Angulo et al. · (1988) found that, of the bacteria associated with turbot culture tanks in Spain, Vibrio and Pseudomonas represented the largest percentage. Perez Benavente & Gatesoupe (1988) found that when rotifers were disinfected before being presented to larval turbot, survival rates improved. Iida et al. (1989) found that a viral disease was responsible for mass mortality in Japanese flounder culture.
They isolated it to the point they hypothesized that a herpes virus was responsible. Nicolas et al. (1989) examined the bacteria associated with the trophic chain of algae, rotifers and turbot larvae and concluded that Vibrio found in the guts of larval turbot were probably introduced by the rotifers. Kellam & Walker (1989) studied the antibiotic activity associated with marine microalgae, and found that Tetraselmis suecica, a species that I use in my experiments, has antibacterial properties. Gatesoupe (1990) found that, by rinsing rotifers and offering them in pulses, rather than all at once, he reduced the bacteria associated with larval culture and improved survival and growth in turbot. Although he reported survival between 22 and 82%, he did not report the number of repetitions or the parentage of the larval cultures. Toranzo et al. (1993) investigated the bacterial differences in three Spanish turbot farms.
Their finding that all farms had high levels of Vibrio and Pseudomonas species led to their conclusion that good husbandry is the most cost effective way of controlling bacterial disease. Skjermo & Vadstein (1993) investigated the bacterial levels associated with enrichment of rotifers.
They found that the bacterial levels increased, and the species composition shifted, with addition of enrichment, then decreased and returned to the original composition with passage of time. Hernandez-Cruz et al. (1994) found that the addition of antibiotics, to the culture vessels or to the rotifers before feeding, did not significantly improve survival and growth of sea bream larvae. The authors also found that rotifers and larvae that were treated with antibiotics had lower Omega-3 HUFA levels than those that were not treated.
Since Qasim (1955) workers have investigated the possibility that algae is a contributor to early larval nutrition. Van der Meeren (1991) concluded that cod do indeed ingest algae, possibly through a filter feeding mechanism, although he did not test whether larvae fed algae had better survival rates than larvae without algae. Austin et al. (1992 ) tested the use of Tetraselmis suecica as an antibacterial preparation in the culture of fish, using various disease-causing bacteria from salmonid culture.
They found that T. suecica did reduce bacteria numbers in culture tanks and, when used therapeutically, reduced mortalities in already infected fish. Naas et al. (1992) found that the use of green water led to increased feeding rates in halibut larvae cultures. Finding that both growth and survival were enhanced, the authors concluded that there was no indication that the larvae were feeding on the algae; the improvement was likely due to turbidity effects. Boehlert & Morgan (1985) found that turbidity increased feeding in larval herring, a possible advantage in the addition of algae to the culture medium. The authors postulated that larvae might be able · to pick out prey better with the additional contrast provided by algae. Reitan et al. (1993) looked at the nutritional effects of the addition of algae to larval turbot culture. They found that the culture of larvae together with rotifers and algae was better than just the enrichment of rotifers with algae prior to the addition of rotifers to the larval tanks. The authors concluded that two effects were at work: l)that rotifer HUFA levels were maintained in the larval culture vessels when algae was maintained, an indirect nutritional effect.

DAH = Days After
Hatch. To make these tables easier to read, exponential notation lOx was left out, e.g., 1.5 -5 equals 1.5 x 10-5. TCBS is thiosulfate-citrate-bile salts-sucrose agar which selects for Vibrio spp. No entry for any particular day under the agars indicates that there were no CFU for that days sample ..   Experiment 2 trial 1, clear water treatments (upper graph) and green water treatments(lower graph). Average food in gut during daily sampling (when survival was greater than zero) on the X axis, survival percentage at 10 DAH on the Y axis.