Aeromonas spp. suggested as the causative agents of red spot disease in northern Vietnamese grass carp Ctenopharyngodon idella.

In northern Vietnam, a disease called 'red spot disease' has been causing high morbidity and mortality in populations of farmed grass carp Ctenopharyngodon idella for about 2 decades. The name 'red spot disease' refers to a condition characterised by haemorrhagic lesions, reddening and ulceration of the skin. Eight different bacterial isolates, namely Aeromonas hydrophila, A. sobria, Staphylococcus epidermidis, Vibrio alginolyticus, Pseudomonas fluorescens, P. luteola, Citrobacter freundii and P. putida, were isolated from diseased grass carp and used for experimental infection of the same species. Fish were challenged with the different bacterial isolates both by immersion and intramuscular injection. Different concentrations of bacteria were tested to evaluate their pathogenicity. Injection with 1 × 105 CFU of A. hydrophila and A. sobria resulted in clinical signs identical to those of red spot-diseased grass carp in Vietnam. None of the other bacterial isolates tested caused any morbidity or mortality in fish challenged either intramuscularly (1 × 106 CFU) or by bath immersion (1 × 106 or 1 × 108 CFU ml-1).

Leaves from banana (Musa nana) and maize (Zea mays) have no phyto-prophylactic effects on the susceptibility of grass carp (Ctenopharyngodon idella) to Aeromonas hydrophila infection.

BACKGROUND: The ubiquitous and opportunistic bacterial pathogen Aeromonas hydrophila has been associated with ulcerative dermatitis in fish, especially under stressful conditions. It can cause severe losses in fresh water aquaculture and is particularly prevalent in tropical and subtropical regions. Fresh leaves from maize and bananas have been used as feed supplement by fish farmers in Vietnam and it has been reported that they may have phyto-prophylactic benefits. In the present study, a feeding trial was conducted to investigate the benefits of providing maize and banana leaves as feed supplement: to determine if they were taken up and digested by grass carp (Ctenopharyngodon idella), if this uptake resulted in improved growth performance, and if leaf supplementation protected fish when challenged with A. hydrophila by intramuscular injection. RESULTS: All fish were fed an identical ratio of commercial pelleted feed relative to biomass. However, in 12/18 tanks, this diet was supplemented with either fresh banana leaves or fresh maize leaves; offered ad libitum. Addition of leaves increased the overall feed conversion ratio (FCR) significantly. However, if only the pellet were taken into account, then no difference was found between treatments. Changes to the isotopic composition of the fish showed leaf nutrient uptake occurred. No prophylactic effects of feeding banana or maize leaves were detected against infection with A. hydrophila, and the diet did not induce changes in the fish haematocrit. However, addition of the maize leaves was associated with significantly reduced severity of the skin lesions, which could improve the market value of the fish. CONCLUSIONS: Addition of the leaf supplement did not result in significantly improved growth performance. Similarly, the effect of the supplement on the fish survival to infection was not significant.

Engineered maize as a source of astaxanthin: processing and application as fish feed.

Astaxanthin from a transgenic maize line was evaluated as feed supplement source conferring effective pigmentation of rainbow trout flesh. An extraction procedure using ethanol together with the addition of vegetal oil was established. This resulted in an oily astaxanthin preparation which was not sufficiently concentrated for direct application to the feed. Therefore, a concentration process involving multiple phase partitioning steps was implemented to remove 90 % of the oil. The resulting astaxanthin raw material contained non-esterified astaxanthin with 12 % 4-keto zeaxanthin and 2 % zeaxanthin as additional carotenoids. Isomeric analysis confirmed the exclusive presence of the 3S, 3'S astaxanthin enantiomer. The geometrical isomers were 89 % all-E, 8 % 13-Z and 3 % 9-Z. The incorporation of the oily astaxanthin preparation into trout feed was performed to deliver 7 mg/kg astaxanthin in the final feed formulation for the first 3.5 weeks and 72 mg/kg for the final 3.5 weeks of the feeding trial. The resulting pigmentation of the trout fillets was determined by hue values with a colour meter and further confirmed by astaxanthin quantification. Pigmentation properties of the maize-produced natural astaxanthin incorporated to 3.5 µg/g dw in the trout fillet resembles that of chemically synthesized astaxanthin. By comparing the relative carotenoid compositions in feed, flesh and feces, a preferential uptake of zeaxanthin and 4-keto zeaxanthin over astaxanthin was observed.

Improving estimates of trophic shift (Δδ(trophic)) for diet reconstruction studies using enzyme activities.

For diet reconstruction studies using stable isotopes, accurate estimates of trophic shift (Δδtrophic) are necessary to get reliable results. Several factors have been identified which affect the trophic shift. The goal of the present experiment was to test whether measurements of the activities of enzymes could improve the accuracy of estimation of trophic shift in fish. Forty-eight Nile tilapia (Oreochromis niloticus) were fed under controlled conditions with two diets differing in their protein content (21 and 41%) each at four different levels (4, 8, 12 and 16gkg(-0.8)d(-1)). At the end of the feeding experiment, proximate composition, whole body δ(13)C and δ(15)N as well as the activities of enzymes involved in anabolism and catabolism were measured. Step-wise regression specified contributing variables for Δδ(15)N (malic enzyme, aspartate aminotransferase and protein content) and Δδ(13)Clipid-free material (aspartate aminotransferase and protein content). Explained variation by using the significant main effects was about 70% for Δδ(15)N and Δδ(13)Clipid-free material, respectively. The results of the present study indicate that enzyme activities are suitable indicators to improve estimates of trophic shift.

The effect of dietary amino acid abundance and isotopic composition on the growth rate, metabolism and tissue δ13C of rainbow trout.

The aim of the present study was to test whether the dietary non-essential/conditionally essential amino acid composition has an effect on growth and protein utilisation and on δ13C of individual amino acids in rainbow trout (Oncorhynchus mykiss). Trout were reared on six purified diets containing only synthetic amino acids in place of protein. Diet 1 mimicked the amino acid composition of fishmeal, in diet 2, cysteine (Cys), glycine (Gly), proline (Pro) and tyrosine (Tyr) were isonitrogenously replaced by their precursor amino acids serine (Ser), glutamic acid (Glu) and phenylalanine (Phe), and in diet 3, alanine (Ala), asparagine and aspartate, Cys, Gly, Pro, Ser and Tyr were isonitrogenously replaced by Glu. Diets 4, 5 and 6 resembled diets 1, 2 and 3 except that Glu contained 0·1 % 13C-enriched Glu. A control group was reared on a fishmeal-based diet. A total of forty-two trout (4·7 (sd 0·57) g) were fed one of the diets at a level of 3·5 % body mass for 10 weeks in a flow-through system. Dietary non-essential amino acid composition significantly influenced protein gain (P < 0·025) and δ13C of Ala, arginine (Arg), Gly, histidine (His), Phe and Tyr. Non-enriched Glu was predominantly found in trout fed 13C-enriched Glu, which is consistent with the fact that Glu has been shown to be used extensively in the gut as an energy source but is less consistent with the enrichment of Pro in fish fed diet 6 compared with fish fed diet 3. Further research is required to better understand the mechanisms that lead to the alteration of amino acid δ13C between diet and body tissues.

Microplastic fiber diet-Fiber-supplemented pellets for small fish.

Ingestion of microplastic particles and fibers is frequently reported for aquatic organisms collected in the field. At the same time, only few studies investigate potential effects of ingestion of microplastic fibers due to handling issues in the laboratory. Exposure studies, which provide organisms with microplastic fibers via the diet, are a necessary step to analyze impact thresholds of vital and fitness parameters of aquatic organisms. Based on the limited number of studies providing fish with fiber-supplemented pellets, the following protocol presents a way to prepare a diet for fish that is supplemented with homogeneous distributed microplastic fibers for exposure studies. Produced pellets are suitable for small experimental fish, such as sticklebacks (2-5 cm), and can be manufactured up to amounts of several hundred grams and even few kilograms. The method can be adapted to different commercial fish feeds and microplastic fiber types due to manual preparation.•Low-cost, manual preparation of microplastic fibers•Preparation of a pelleted fish diet with uniformly distributed fibers•Adaptable to different commercial fish feeds and microplastic fiber types.

The effect of different fish feed compositions on δ13C and δ15N signatures of sea bass and its potential value for tracking mariculture-derived nutrients.

Increasing demand for fish and seafood calls for an expansion of aquaculture production. At the same time, the status of the marine environment must not be jeopardised. Stable isotopes are potential markers for tracking feed-based nutrient flows from aquaculture into marine biota. Here, we demonstrate how four experimental diets (main protein components: fishmeal, soya protein concentrate, wheat gluten, and Jatropha kernel meal) and a commercial diet induce characteristic δ13C and δ15N signals in sea bass (Dicentrarchus labrax) during nine weeks of laboratory feeding under replicate conditions. The plant-protein-based diets containing wheat gluten and soya, and the commercial feed consistently induced the largest isotopic differentiation of the fish, both from the feed source and from the pre-experimental condition. The large difference of the fish on plant-protein-based diets compared to the range of natural isotopic variation in the macrozoobenthos of the North Sea lends support to the idea that plant-based feeds are suitable for tracing mariculture-derived organic matter under practical conditions. The commercial feed had a similar effect as the experimental feeds and would be a cost-effective option for an offshore aquaculture experiment.

Effect of different ratios of wheat to corn flour in the diet on the development and isotopic composition (delta13C, delta15N) of the red flour beetle Tribolium castaneum.

The back-calculation of the diet is a common application of stable isotopes in animal ecology. The method is based on a predictable relation between the isotopic signature of the diet and the animal's tissues. Frequently, the assumption of a constant difference in isotopic signatures (trophic shift) is made. Carbon isotopic ratios of C(3) and C(4) plants differ by approximately 10 per thousand, making wheat (C(3)-plant) and corn (C(4)-plant) ideal materials for isotopic studies in nutritional ecology and especially for testing the back-calculation method. In this experiment, red flour beetles, Tribolium castaneum, were reared on wheat flour, corn flour and three different mixtures thereof, either in pure flour or with the addition of yeast inoculum or yeast grains. Development of T. castaneum on these experimental diets was monitored, and isotopic signatures of carbon and nitrogen in emerging adults were analysed. The values of trophic shift of C and N isotopes for wheat and corn flour were different, and the values for the mixtures did not correspond to those expected from a linear mixing model. The latter can be taken as an indication that the tiny larvae of T. castaneum may be capable of differentiating between particles of wheat and corn flour, making this animal model unsuitable for testing the back-calculation method.

Dietary back-calculation using stable isotopes: can activities of enzymes involved in amino acid metabolism be used to improve estimates of trophic shifts in fish?

The aim of this study was (1) to assess the effects of dietary protein content and feeding level on trophic shifts of C and N isotopes (Delta delta(13)C(tissue-diet) and Delta delta(15)N(tissue-diet)) and (2) to test whether the measurement of the activities of two enzymes involved in the metabolism of amino acids could improve the accuracy of estimation of the trophic shifts of C and N isotopes. For this, 36 Nile tilapia (Oreochromis niloticus) were kept under controlled conditions for 8 weeks and fed at three different levels (2, 4 and 8 g kg(-0.8) d(-1)) with three diets differing in their protein content only (20, 29 and 39 %). For each fish, food to fish body trophic shifts of C and N isotopes were measured as well as the hepatic activities of aspartate aminotransferase (ASAT) and glutamate dehydrogenase (GDH). The feeding level affected the activities of ASAT and GDH as well as the trophic shifts of C and N isotopes significantly but the dietary protein content had no significant effect except on the specific activity of ASAT. Fish fed at the lowest level had significantly higher trophic shifts of C and N isotopes than fish fed at higher levels. The trophic shifts were significantly lower in fish with a high protein utilisation. Values of the 'goodness-of-fit' for linear regressions between enzyme activities and trophic shifts were low. Thus, activities of ASAT and GDH are not suitable for predicting estimates of trophic shifts in situations where the amount of food consumed or the dietary protein content is not known. In further studies, activities of enzymes involved in the metabolism of amino acids combined with measurements of the activities of other enzymes should be used to try and improve the accuracy of estimates of trophic shifts.

Plant protein-based feeds and commercial feed enable isotopic tracking of aquaculture emissions into marine macrozoobenthic bioindicator species.

Brittle stars (Ophiura spp.) and other benthic macrofauna were collected in a prospective mariculture area in the North Sea to determine if these taxa could be used as indicator species to track nutrients released from future offshore aquaculture sites. We analysed natural carbon and nitrogen stable isotopic signatures in tissues from macrofauna and compared these to six feed ingredients and four experimental diets made thereof, as well as to a commercial feed with and without lipid and carbonate removal. Our data suggest practicability of using isotopic signatures of Ophiura spp. to track aquaculture-derived organic material if plant-based fish diet ingredients and commercial feed were used for fish farming in the German Exclusive Economic Zone. Diets with high fish meal content would not be detected in Ophiura spp. using isotopic measures due to the similarity with the marine background. Our data provide valuable baseline information for studies on the impact of offshore aquaculture on the marine environment.

Metabolism of Seriola lalandi during Starvation as Revealed by Fatty Acid Analysis and Compound-Specific Analysis of Stable Isotopes within Amino Acids.

Fish starvation is defined as food deprivation for a long period of time, such that physiological processes become confined to basal metabolism. Starvation provides insights in physiological processes without interference from unknown factors in digestion and nutrient absorption occurring in fed state. Juveniles of amberjack Seriola lalandi were isotopically equilibrated to a formulated diet for 60 days. One treatment consisted of fish that continued to be fed and fish in the other treatment were not fed for 35 days. The isotopic signatures prior to the beginning of and after the starvation period, for fish in the starvation and control treatments, were analysed for lipid content, fatty acid composition and isotopic analysis of bulk (EA-IRMS) and of amino acids (compound specific isotope analysis, CSIA). There were three replicates for the starvation group. Fatty acid content in muscle and liver tissue before and after starvation was determined to calculate percent change. Results showed that crude lipid was the most used source of energy in most cases; the PUFAs and LC-PUFAs were highly conserved. According to the protein signature in bulk (δ15N) and per amino acid (δ13C and δ15N), in muscle tissue, protein synthesis did not appear to occur substantially during starvation, whereas in liver, increases in δ13C and δ15N indicate that protein turnover occurred, probably for metabolic routing to energy-yielding processes. As a result, isotopic values of δ15N in muscle tissue do not change, whereas CSIA net change occurred in the liver tissue. During the study period of 35 days, muscle protein was largely conserved, being neither replenished from amino acid pools in the plasma and liver nor catabolized.

Microcystins in natural blooms and laboratory cultured Microcystis aeruginosa from Laguna de Bay, Philippines.

Laguna de Bay, the largest freshwater lake in the Philippines, experiences periodic blooms of the cyanobacteria Microcystis aeruginosa. Blooms of these cyanobacteria in 1996, 1998 and 1999 were sampled. HPLC and MALDI-TOF mass spectrometry were used to analyze for microcystins. A total of 16 structural variants of the toxin were isolated from the samples with microcystin LR (MC-LR) as the most abundant variant in the samples from 1996 and 1999 making up 77 to 85% of the total, respectively. MC-RR was the dominant variant in the 1998 bloom making up 38%. The samples from 1996 had the highest total toxin concentration (4049 microg g(-1)) followed by those from 1998 (1577 microg g(-1)) and 1999 (649 microg g(-1)). A strain of M. aeruginosa previously isolated from the lake was also cultured in the laboratory under different nitrogen concentrations (1, 3 and 6 mg L(-1)) and elevated phosphorus concentration (0.5 mg L(-1)) to determine the influence of these factors on toxin production. A total of 9 different structural variants of microcystin were isolated from the laboratory cultures with MC-LR consisting more than 75% of the total in all treatments. No significant differences in the total toxin concentration as well as the % distribution of the different variants among treatments were observed. However, the strain of M. aeruginosa cultured in the laboratory had from 3 to 20 times higher total microcystin than those harvested from the lake.

Dietary lipid content influences the activity of lipogenic enzymes in the liver and on whole body delta13C values of Nile tilapia, Oreochromis niloticus (L.).

The use of stable isotope techniques for the reconstruction of diets has increased over the last decade. However, isotopic ratios in an animal are not only affected by the composition of the feed, but also by the amount of feed consumed. An uncertainty of up to 1 per thousand for both delta13C and delta15N values has been observed when the feeding level is unknown. This may have substantial effects on the results of back-calculation. As the feeding level of animals is unknown in nature, an additional indicator for their nutritional status is needed. High feeding levels and a consequent surfeit of dietary energy lead to the synthesis of lipids. In order to test whether the level of lipogenesis could be used as an indicator, Nile tilapia (Oreochromis niloticus) were fed four isonitrogenous and isoenergetic wheat-based semi-synthetic diets with different lipid contents (2.0 %, 4.5 %, 9.5 % and 13.3 %) for eight weeks. Body composition, gross energy content and delta13C values in the lipids and the lipid-free material were determined in diets and fish bodies. The livers of three fish per feeding group were assayed for the activity of two lipogenic enzymes, ATP-citrate lyase and malic enzyme. There was a strong negative correlation between delta13C values in the lipids of the individual fish and the apparent lipid conversion. The activities of lipogenic enzymes decreased with rising lipid content in the diet. The delta13C values in the lipids decreased significantly with increasing specific activity for both enzymes. In this experiment where lipogenesis was influenced by the composition of the diet, it was possible to determine the exact value for the trophic shift in relation to the enzyme activities. Further experiments to investigate the use of enzyme activities in situations where the feeding level of an animal is unknown are recommended.

15N tracer application to evaluate nitrogen dynamics of food webs in two subtropical small-scale aquaculture ponds under different managements.

Small, semi-intensively managed aquaculture ponds contribute significantly to the food security of small-scale farmers around the world. However, little is known about nutrient flows within natural food webs in such ponds in which fish production depends on the productivity of natural food resources. (15)N was applied as ammonium at 1.1 and 0.4 % of total nitrogen in a traditionally managed flow-through pond and a semi-intensively managed stagnant pond belonging to small-scale farmers in Northern Vietnam and traced through the natural food resources over 7 days. Small-sized plankton (1-60 µ m) was the dominant pelagic biomass in both ponds with higher biomass in the stagnant pond. This plankton assimilated major portions of the applied tracer and showed a high sedimentation and turnover rate. High re-activation of settled nutrients into the pelagic food web was observed. The tracer was removed more quickly from the flow-through pond than from the stagnant pond. A steady nutrient supply could increase fish production.

Determination of underivatized amino acid delta(13)C by liquid chromatography/isotope ratio mass spectrometry for nutritional studies: the effect of dietary non-essential amino acid profile on the isotopic signature of individual amino acids in fish.

This study provides data for the effect of dietary non-essential amino acid composition on the delta(13)C values of individual amino acids in rainbow trout (Oncorhynchus mykiss) using liquid chromatography coupled to isotope ratio mass spectrometry (LC/IRMS). In this experiment, trout were reared either on a control diet or on three experimental diets, differing in the composition of non-essential/conditionally essential amino acids, for a period of 6 weeks. The control diet was a commercial trout starter feed with fish meal as the main protein source. The experimental diets contained no protein, only synthetic amino acids. Diet 1 resembled the composition of fish meal in both essential and non-essential amino acids, Diet 2 had all essential amino acids, but cysteine, glycine, proline and tyrosine were replaced by the corresponding amounts of their precursors, and in Diet 3 all non-essential amino acids were replaced by glutamate. LC/IRMS was used for the determination of delta(13)C values of individual amino acids from diets and tissues without derivatization. Diet affected the delta(13)C of individual amino acids in fish. For fish on Diets 1-3 amino acid delta(13)C values showed a similar trend: phenylalanine showed very little change from diet to body tissue. Arginine, lysine, tyrosine and proline showed strong depletion from diet to body tissue and glycine, alanine, aspartate and serine all showed variable but strong enrichment in (13)C. Improvements are necessary before all amino acid delta(13)C values can be determined; however, this study demonstrates that measuring amino acid isotopic signatures by LC/IRMS is a promising new technique for nutritional physiologists.

Comparison of metabolic rates and feed nutrient digestibility in conventional, genetically improved (GIFT) and genetically male (GMNT) Nile tilapia, Oreochromis niloticus (L.).

Various aspects of energy metabolism and feed digestibility were evaluated in two reportedly improved strains of Nile tilapia (Oreochromis niloticus) namely GIFT (genetically improved farmed tilapia) and GMNT (genetically male Nile tilapia) and compared with those of CNT (conventional Nile tilapia). Fish were stocked individually in a computer-controlled respirometer system at 27+/-0.1 degrees C for 10 weeks. Metabolic rates were measured at three different feeding levels: starved, maintenance (3.0 g kg(-0.8) day(-1)) and growth (7.5 g kg(-0.8) day(-1)) using a fishmeal based feed containing TiO2 marker (41% crude protein, 9% crude lipid and 19 kJ (g DM)(-1) gross energy). The standard metabolic rate (SMR), measured at the beginning of the experiment (45.4+/-4.6, 52.4+/-7.7 and 46.8+/-4.6 mg O2 kg(-0.8) h(-1) respectively for GIFT, GMNT and CNT), did not differ significantly between the groups (p<0.05). Similarly, non-significant differences were also observed in the routine metabolic rates under starved, maintenance and growth conditions but the variability was higher in the case of GMNT and CNT than in GIFT. The latter group showed a significantly lower active metabolic rate (145 mg O2 kg(-0.8) h(-1)) compared to GMNT and CNT (232 and 253 mg O2 kg(-0.8) h(-1), respectively) at maintenance feeding level. The specific dynamic action (% offered feed energy) showed no significant differences among the groups. Digestibility coefficients of feed dry matter, protein, lipid and energy for the three tilapia groups also did not differ significantly. Therefore, we concluded that the genetic improvement or modification in the GIFT or GMNT might not upgrade the inherent physiological potential compared to CNT as far as energy metabolism and digestion efficiencies are concerned.

Feeding fish with diets of different ratios of C3- and C4-plant-derived ingredients: a laboratory analysis with implications for the back-calculation of diet from stable isotope data.

Back-calculation of the diet is one of the most frequent applications of stable isotope techniques in animal ecology. These calculations are often based on two assumptions: a constant trophic shift for all dietary items and a linear response of the isotopic ratios to different mixtures of two isotopically distinct feeds. In a laboratory experiment, fish (Nile tilapia, Oreochromis niloticus) were fed semi-synthetic diets prepared either from wheat or corn ingredients, or from three blended diets (25, 50, 75% wheat components). Isotopic analysis of the lipid-free and lipid fraction of the fish after the experiment revealed that the trophic shift was not constant for wheat- and corn-based diets. The isotopic response to the mixed diets was not linear, leading to a statistically significant over-estimation of the corn component in the back-calculation. Both effects are in agreement with published data on the isotopic effects of C3- and C4-plant materials in the diet.

Individual protein balance strongly influences delta15N and delta13C values in Nile tilapia, Oreochromis niloticus.

Although stable isotope ratios in animals have often been used as indicators of the trophic level and for the back-calculation of diets, few experiments have been done under standardized laboratory conditions to investigate factors influencing delta(15)N and delta(13)C values. An experiment using Nile tilapia [ Oreochromis niloticus (L.)] was therefore carried out to test the effect of different dietary protein contents (35.4, 42.3, and 50.9%) on delta(15)N and delta(13)C values of the whole tilapia. The fish were fed the isoenergetic and isolipidic semi-synthetic diets at a relatively low level. delta(15)N and delta(13)C values of the lipid-free body did not differ between the fish fed the diets with different protein contents, but the trophic shift for N and C isotopes decreased with increasing protein accretion in the individual fish, for N from 6.5 per thousand to 4 per thousand and for C in the lipid-free body from 4 per thousand to 2.5 per thousand. This is the first study showing the strong influence of the individual protein balance to the degree to which the isotopic signature of dietary protein was modified in tissue protein of fish. The extrapolation of the trophic level or the reconstruction of the diet of an animal from stable isotope ratios without knowledge of the individual physiological condition and the feeding rate may lead to erroneous results.

Feeding level and individual metabolic rate affect delta 13C and delta 15N values in carp: implications for food web studies.

Stable isotope analyses are often used to calculate relative contributions of multiple food sources in an animal's diet. One prerequisite for a precise calculation is the determination of the diet-tissue fractionation factor. Isotopic ratios in animals are not only affected by the composition of the diet, but also by the amount of food consumed. Previous findings regarding the latter point are controversial. As stable isotope analyses have often been used to investigate aquatic food webs, an experiment with carp (Cyprinus carpioL.) was carried out to test the influence of the feeding level and individual metabolic rate on delta(13)C and delta(15)N values of the whole body. After an initial phase, 49 carp were assigned randomly to four groups and fed the same diet at different levels for 8 weeks. For 15 fish, the energy budget was determined by indirect calorimetry. Feed and individual fish were analysed for their proximate composition, gross energy content and delta(13)C and delta(15)N values. delta(13)C and delta(15)N values differed significantly at different feeding levels. While delta(13)C values of the lipids and delta(15)N values decreased with increasing feeding rate, delta(13)C values of the lipid-free matter showed a non-linear pattern. Data obtained from fish held in the respirometric system revealed a relationship between delta(13)C values and the percentage retention of metabolizable energy. Our results show that reconstructing the diets of fish from the isotopic ratios when the feeding level and individual metabolic rates are unknown would introduce an error into the data used for back-calculation of up to 1 per thousand for both delta(13)C and delta(15)N values and may have substantial effects on the results of calculated diets. As other workers have pointed out, the development and application of stable isotopes to nutritional ecology studies is a field in its infancy and gives rise to erroneous, misleading results without nutritional, physiological and ecological knowledge.