Variations in spawning ground area and egg density of the Japanese sardine in Pacific coastal and oceanic waters

The spawning ground of the Japanese sardine, Sardinops melanostictus (Schlegel), was distributed over the oceanic waters as well as the coastal waters along the Pacific coasts of western and eastern Japan during 1978–1992. The area of the spawning ground in the coastal waters on the continental shelf has ranged from 95 000 km² in 1992 to 143 000 km² in 1988, constituting 44–77% of the total area of the spawning ground. The area of the coastal spawning ground was relatively constant in spite of the large fluctuations in egg abundance, i.e. size of the spawning population, from 88 trillion (1987) to 668 trillion (1989) in the waters. Spawning adults seemed to extend over the coastal waters irrespective of the size of the spawning population. In contrast to the coastal waters, the spawning area in the oceanic waters offshore of the continental shelf increased from 31 000 km² in 1978 to 183 000 km² in 1988 and then shrank to 83 000 km² in 1992, as a function of the spawning population size. The egg distribution density in the coastal waters stayed less than 6000 m⁻² mo⁻¹, but it reached as high as 27 400 m⁻² mo⁻¹ in the expanded spawning ground in the oceanic waters. The oceanic waters seemed to function as a reserve spawning ground for the sardine in years of extremely high spawning population.     

Effects of protein hydrolysate in weaning diets for Atlantic cod (Gadus morhua L.) and Atlantic halibut (Hippoglossus hippoglossus L.)

The study aims to test whether predigested dietary protein enhances the utilization of formulated diets at weaning, and also whether it stimulates intestinal maturation. In this study, Atlantic cod [ Gadus morhua L.; 41 days posthatch (dph)] and Atlantic halibut [ Hippoglossus hippoglossus L.; 63 days postfirst feeding (dpff)] were weaned onto diets with graded levels of protein hydrolysate. By increasing the inclusion of dietary protein as hydrolysate from 0 to 400 g kg⁻¹, cod increased the rate of survival from 7 ± 1% to 18 ± 2% (82 dph; regression, P =  4*10⁻⁷). In halibut, the survival rate decreased from 57 ± 9% to 22 ± 7% as the inclusion of protein in the form of hydrolysate increased from 0 to 450 g kg⁻¹ (119 dpff; regression, P =  8*10⁻⁵). Growth was not affected in any of the species. Results in specific activities of the intestinal enzymes leucine aminopeptidase (LAP) and alkaline phospatase (AP) supported the results in survival in halibut and partly also in cod, by showing increased activities in groups with increased survival ( anova , P <  0.05). The lower optimal level of hydrolysed protein in halibut than in cod is suggested mainly because of a slower feeding practice in halibut, which allows more extensive nutrient leaching before ingestion.     

Effects of dietary protein and lipid level, and water temperature, on the post-feeding oxygen consumption of Atlantic cod and haddock

Tank respirometry was used to study the interactive effects of protein:lipid level (55%:11% vs. 42%:16%; both diets isoenergetic) and temperature (11, 6 and 2 °C) on the magnitude and duration of specific dynamic action (SDA) in juvenile Atlantic cod and haddock. The protein:lipid level did not affect any measured variable. However, numerous temperature and species effects were observed. For example, although the maximum post-feeding oxygen consumption (30–50% above routine metabolic rate; RMR) and SDA duration (∼55–85 h; SDA_DUR) were not affected by temperature, SDA_DUR g⁻¹ of food increased from 11 to 2 °C (from ∼3 to 12 h g food⁻¹). While absolute SDA (mg O₂) decreased by ∼60–65% in cod and ∼75% in haddock from 11 to 2 °C, due to a concomitant decrease in food consumption from ∼2.0% to 0.6% body mass, SDA comprised between 3.3% and 5.2% of the dietary energy content at all temperatures. Finally, RMR at 11 and 2 °C and SDA_DUR at 2 °C were 25–35% and 25% greater in cod, respectively, as compared with haddock. These results suggest that feeding reduced protein diets at low water temperatures is unlikely to improve the growth of these species.     

Acute LD50 and kidney histopathology following injection of erythromycin (Erythro-200) and its carrier in spring chinook salmon, Oncorhynchus tshawytscha (Walbaum)

We evaluated the toxicity of commercial injectable erythromycin (Erythro-200) and the injectable drug carrier alone (PEG-400, ethyl alcohol and ethyl acetate) in sub-adult spring chinook salmon, Oncorhynchus tshawytscha (Walbaum). Both compounds were toxic to salmon when administered in high dosages. The 50% lethal dose (LD₅₀) of the erythromycin within its carrier (Erythro-200) at 12 °C was estimated at 429 mg erythromycin kg^–1 or 2.145 mL kg^–1 when measured by volume. The LD₅₀ of the drug carrier alone was 2.95 mL kg^–1 by volume or equivalent to 3.21 g kg^–1 of the carrier. We evaluated and scored the histopathology in kidney sections removed from fish 96 h after injection with Erythro-200, the drug carrier alone, or sterile phosphate buffered saline (PBS) that served as a reference control. Kidneys from fish injected with saline had lowest average pathology scores; those injected with the carrier had moderate scores; and the pathology score was highest in tissues from erythromycin-injected fish. In tests at 12 °C, vacuoles were more prominent in proximal and distal tubules of fish injected with higher dosages of 403 and 545 mg erythromycin kg^–1 than in tissues from fish injected with 299 mg kg^–1.     

Nutrient values of dietary ascorbic acid (l-ascorbyl-2-polyphosphate) on growth, survival and stress tolerance of larval shrimp, Litopenaeus vannamei

l -ascorbyl-2-polyphosphate (ApP) was used as a vitamin C source to investigate the ascorbic acid (AsA) requirements on growth performance and stress resistance of the larval white shrimp, Litopenaeus vannamei . Five isoenergetic and isonitrogenous fish meal-fish protein hydrolysate-based diets with five levels of ApP, AsA equivalent to 91.8, 188, 271, 360 and 436 mg kg⁻¹ diet were fed to triplicate groups of L. vannamei (mean initial wet weight 1 mg) for 32 days. The diet with AsA 91.8 mg kg⁻¹ showed high cumulative mortality after 10 days of feeding. After the 32-day trial, the shrimp that fed the diet had significantly lower survival and weight gain (WG, %) than those that fed 188, 271, 360 and 436 mg AsA kg⁻¹ diets. Specific growth rate (SGR, % day⁻¹) and final body wet weight (FBW, mg) showed the same pattern as WG (%). There were no significant differences in growth performance (FBW, WG and SGR) among the groups that fed 188, 271, 360 and 436 mg kg⁻¹ of AsA at the termination of feeding trial. Broken-line regression analysis on WG indicated that 191 mg AsA kg⁻¹ in the diet was the optimum for larval L. vannamei . On the contrary, dietary level of more than 360 mg AsA kg⁻¹ was needed to ensure high resistance to stressful conditions such as low dissolved oxygen stressors.     

Metabolic scope, swimming performance and the effects of hypoxia in the mulloway, Argyrosomus japonicus (Pisces: Sciaenidae)

The culture of the mulloway (Argyrosomus japonicus), like many other Sciaenidae fishes, is rapidly growing. However there is no information on their metabolic physiology. In this study, the effects of various hypoxia levels on the swimming performance and metabolic scope of juvenile mulloway (0.34 ± 0.01 kg, mean ± SE, n = 30) was investigated (water temperature = 22 °C). In normoxic conditions (dissolved oxygen = 6.85 mg l^− 1), mulloway oxygen consumption rate (M·o₂) increased exponentially with swimming speed to a maximum velocity (U_crit) of 1.7 ± < 0.1 body lengths s^− 1 (BL s^− 1) (n = 6). Mulloway standard metabolic rate (SMR) was typical for non-tuna fishes (73 ± 8 mg kg^− 1 h^− 1) and they had a moderate scope for aerobic metabolism (5 times the SMR). Mulloway minimum gross cost of transport (GCOT_min, 0.14 ± 0.01 mg kg^− 1 m^− 1) and optimum swimming velocity (U_opt, 1.3 ± 0.2 BL s^− 1) were comparable to many other body and caudal fin swimming fish species. Energy expenditure was minimum when swimming between 0.3 and 0.5 BL s^− 1. The critical dissolved oxygen level was 1.80 mg l^− 1 for mulloway swimming at 0.9 BL s^− 1. This reveals that mulloway are well adapted to hypoxia, which is probably adaptive from their natural early life history within estuaries. In all levels of hypoxia (75% saturation = 5.23, 50% = 3.64, and 25% = 1 .86 mg l^− 1), M·o₂ increased linearly with swimming speed and active metabolic rate (AMR) was reduced (218 ± 17, 202 ± 14 and 175 ± 10 mg kg^− 1 h^− 1 for 75%, 50% and 25% saturation respectively). However, U_crit was only reduced at 50% and 25% saturation (1.4 ± < 0.1 and 1.4 ± < 0.1 BL s^− 1 respectively). This demonstrates that although the metabolic capacity of mulloway is reduced in mild hypoxia (75% saturation) they are able to compensate to maintain swimming performance. GCOT_min (0.09 ± 0.01 mg kg^− 1 m^− 1) and U_opt (0.8 ± 0.1 BL s^− 1) were significantly reduced at 25% dissolved oxygen saturation. As mulloway metabolic scope was significantly reduced at all hypoxia levels, it suggests that even mild hypoxia may reduce growth productivity.     

Olfactory nerve response of masu salmon (Oncorhynchus masou Brevoort) and rainbow trout (O. mykiss Walbaum) to clove oil and MS-222

Two anaesthetics, clove oil and methane sulphonate (MS-222), were examined for their effects on the olfactory nerve response of masu salmon ( Oncorhynchus masou Brevoort) and rainbow trout ( O. mykiss Walbaum). Exposing both species to clove oil for 3 min at concentrations of 50 and 100 mg L⁻¹, or for 10 min at 50 mg L⁻¹, did not significantly reduce their olfactory response. Directly applying clove oil anaesthesia to the olfactory epithelium significantly reduced olfactory response though after 20 min, olfactory response recovered to 70% and 52% of pre-treatment levels in masu salmon and rainbow trout respectively. Compared with the post-anaesthetic recovery of responses after clove oil (50 mg L⁻¹), buffered MS-222 (100 mg L⁻¹) with NaHCO₃ (100 mg L⁻¹), and unbuffered MS-222 (100 mg L⁻¹) treatment for 3 min, the response after MS-222 treatment declined gradually and significantly, but not after clove oil and MS-222+NaHCO₃ treatments. Clove oil appears to be an effective and relatively safe anaesthetic for salmonids with little long-term impact on their olfactory response, which plays a crucial role in their life history.     

A pharmacokinetic study of flumequine in sea bass, Dicentrarchus labrax (L.), after a single intravascular injection

The pharmacokinetic properties of flumequine following a single intravascular injection (10 mg kg^–1 fish) were studied in sea bass, Dicentrarchus labrax (L.), 120 g held at 18 °C. The absorption half life ( t _1/2α) and the elimination half life ( t _1/2β) of the drug were calculated to be 1.05 and 10.71 h, respectively. Tissue penetration of flumequine seemed to be moderate because both the apparent volume of distribution of the drug at steady-state ( V _d(ss)) and the extensive apparent volume of the central compartment ( V _c) were found to be small (1.51 and 0.626 L kg^–1). The mean residence time ( MRT ) was short (09.73 h) and the total clearance (CL_T) of the drug was rapid (0.156 L kg^–1 h^–1).     

The influence of predation risk on habitat selection and food intake by Arctic charr, Salvelinus alpinus (L.)

Abstract— Large piscivorous fish are assumed to affect habitat selection and food intake of prey fish. To study the effects of cannibalistic Arctic charr, Salvelinus alpinus (L.), on smaller stunted charr, we sampled the prey fish in littoral and pelagic habitats using gill nets, before and shortly after the release of large charr in a small lake (0.52 km²). In the habitats where the risk of predation was highest, the catch per unit effort de creased from 13.3 to 4.8 fish per 100 m² of gillnet after release of pred ators. The large decrease in numbers of charr < 18 cm corresponded with the predicted vulnerable prey sizes, according to a model based on the size distribution of predators. The occurrence of planktivorous fish and weight-specific food intake decreased in the high risk habitat and remained unaffected in the low risk habitats. Changes in the food intake of prey fish could not be explained in terms of fish length, indicating that prey fish changed diet when the risk of predation was high.     

Production and predation rates in a cannibalistic Arctic char (Salvelinus alpinus L.) population

Abstract – Production of cannibalistic Arctic char ( Salvelinus alpinus (L.)) and their prey conspecifics were estimated in an allopatric population, situated in an alpine ultraoligotrophic lake in central Norway. Estimates of population abundance were obtained by mark recapture and successive removal experiments. Assuming stable age distribution, we estimated mean yearly biomass (±95% CL) of char in sampled size groups vulnerable to predation (60 to 150 mm) to 7.93 (5.11–14.30) kg · ha⁻¹. Similarly, mean yearly biomass (±95% CL) of cannibalistic char (> 250 mm) was estimated to 0.62 (0.50–1.06) kg · ha⁻¹. Annual production (±95% CL) of char in length groups 60 to 150 mm was estimated to be 4.31 (2.74–8.03) kg · ha⁻¹ · year⁻¹, and production (±95% CL) of cannibalistic char to 0.19 (0.15–0.33) kg · ha⁻¹ · year⁻¹. Depending on the food conversion factor (set to vary from 0.1 to 0.4), the cannibalistic char removed from 10% to 40% of the production of char in sampled size-classes vulnerable to predation yearly. The overall ecological efficiency in energy transfer between the prey and predator population was 4.4%.     

Winter distribution and movements of northern Atlantic cod (Gadus morhua) along the Newfoundland—Labrador continental shelf edge derived from observations on commercial trawlers

Catch-effort data recorded by observers aboard commercial trawlers fishing the Newfoundland-Labrador continental shelf during the winters of 1980–1991 were used to define the distribution and movements of pre-spawning and spanning Atlantic cod, Gadus morhua. Although cod were widely distributed over the outer continental shelf in the months of January to April, commercial concentrations were consistently located in three areas along the shelf edge. These three areas, constituting only 16% of the 190000 km² fishing ground for northern cod, were centred north-east of Belle Isle Bank, north-east of Funk Island Bank, and between Funk Island Bank and the northern Grand Bank Tow positions with catch rates 5000 kg h^-1 were plotted on a weekly basis to define movements of cod schools during the winters of 1986–1988. It was assumed that change in the collective position of trawlers taking large quantities of cod reflected shoal movements. Analyses indicate that cod moved southeast along the shelf edge in January and February of each year, crossing from NAFO Division 2J into 3K Each March, cod in Division 3K apparently moved northwest against prevailing ocean currents into the Hamilton Bank region of 2J. These movements do not appear to be a direct response to ocean temperature or pack ice conditions, although cod moving along the shelf break were generally in bottom waters of 24°C. We suggest these patterns reflect the migratory behaviour of pre-spawning and spawning cod.     

Remotely sensed spawning habitat of Pacific sardine (Sardinops sagax) and Northern anchovy (Engraulis mordax) within the California Current

We use trivariate kernel density estimation to define spawning habitat of northern anchovy ( Engraulis mordax ) and Pacific sardine ( Sardinops sagax ) in the California Current using satellite data and in situ egg samples from the Continuous Underway Fish Egg Sampler (CUFES) deployed during surveys in April by the California Cooperative Oceanic Fisheries Investigations (CalCOFI). Observed egg distributions were compared with monthly composite satellite sea surface temperature (SST) and surface chlorophyll a (chl a ) data. Based on the preferred spawning habitat, as defined in SST and chl a space, the satellite data were used to predict potential spawning habitat along two areas of the west coast of North America. Data from the southern area (21.5 to 39°N) were compared to observations from the CUFES data for the period 1998–2005. Northern anchovy and Pacific sardine exhibited distinctly different spawning habitat distributions. A significant relationship was found between satellite-based spawning area and that measured during surveys for sardine. CUFES area estimated for sardine was similar in magnitude to that estimated from satellite data (∼60 000 km²). In contrast, spawning habitat of anchovy averaged between 1000 and 200 000 km² for the period 1998–2005, for CUFES and satellite estimates, respectively. Interannual variability in the area (km²) and duration (months) of estimates of suitable habitat varied between species and between the northern (39 to 50.5°N) and southern portions of the California Current. Long-term monitoring of habitat variability using remote sensing data is possible in the southern portion of the California Current, and could be improved upon in the northern area with the addition of surveys better timed to describe relationships between observed and estimated spawning habitats.     

An experimental investigation on aspects of infectious salmon anaemia virus (ISAV) infection dynamics in seawater Atlantic salmon, Salmo salar L.

This study investigated infection dynamics of infectious salmon anaemia virus (ISAV) by conducting two experiments to examine minimum infective dose and viral shedding of ISAV. In terms of minimum infective dose, the high variability between replicate tanks and the relatively slow spread of infection through the population at 1 × 10¹ TCID₅₀ mL⁻¹ indicated this dose is approaching the minimum infective dose for ISAV in seawater salmon populations. A novel qPCR assay incorporating an influenza virus control standard with each seawater sample was developed that enabled the quantity of ISAV shed from infected populations to be estimated in values equivalent to viral titres. Viral shedding was first detected at 7 days post-challenge (5.8 × 10⁻² TCID₅₀ mL⁻¹ kg⁻¹) and rose to levels above the minimum infective dose (4.2 × 10¹ TCID₅₀ mL⁻¹ kg⁻¹) on day 11 post-challenge, 2 days before mortalities in ISAV inoculated fish started. These results clearly demonstrate that a large viral shedding event occurs before death. Viral titres peaked at 7.0 × 10¹ TCID₅₀ mL⁻¹ kg⁻¹ 15 days post-infection. These data provide important information relevant to the management of ISA.     

Using Efinol®L during transportation of marbled hatchetfish, Carnegiella strigata (Günther)

The objective of this experiment was to test the efficacy of a probiotic (Efinol^®L) during transportation of marbled hatchetfish, Carnegiella strigata . Wild specimens were captured from a small stream and transported for 24 h in plastic fish boxes with a probiotic (10 mg L⁻¹) and probiotic-free water. The boxes were sampled at 3, 12 and 24 h of transport. At the end of the experiment, the survival rate was close to 100% in both treatments. Dissolved oxygen diminished with time in both treatments, but the probiotic group had significantly higher levels. Conductivity, pH and ammonia increased significantly during the transport, demonstrating higher levels in the probiotic-free group. Fish from both treatments presented very high net Na⁺ and K⁺ effluxes after 3 h of transport. At 24 h, net K⁺ effluxes in fish of the probiotic treatment reached values close to zero and a significantly lower Na⁺ efflux was observed. Cortisol levels in both treatments at 3 and 12 h were significantly higher than that in control samples. Higher body cortisol levels were observed in the probiotic-free group than that in the probiotic group at 3 and 12 h. The results demonstrate that addition of a probiotic during fish transport improves water quality and leads to fish presenting a lower stress response intensity.     

A preliminary evaluation of physiological filtration variables for Crassostrea corteziensis (Hertlein, 1951) and Anadara tuberculosa (Sowerby, 1833) in shrimp aquaculture effluents

This study examined the main filtration variables [filtration rate ( FR ), clearance rate ( CR ) and assimilation efficiency ( AE )] of the Cortez oyster, Crassostrea corteziensis (Hertlein, 1951), and the mud cockle, Anadara tuberculosa (Sowerby, 1833), in shrimp aquaculture effluents at three different flux velocities (1.5, 3 and 4.5 L h⁻¹) using a 36-respirometer system, each with a 0.5 L capacity. Under inverted photoperiod conditions, free variations in the environmental parameters were allowed to mimic effluent conditions during a pair of 12-h trials. The FR s for both species (0.124, 0.328 and 0.402 L h⁻¹ for the Cortez oyster; 0.093, 0.189 and 0.345 L h⁻¹ for the mud cockle) were relatively low as compared with those reported for similar or related species. The CR s were higher for the Cortez oyster (20.04, 52.92 and 64.70 L h⁻¹) than for the mud cockle (10.96, 22.95 and 42.12 L h⁻¹); in both cases, the values were in the range reported previously for the last species. The AE for both mollusks (over 92% for the Cortez oyster and over 95% for the mud cockle) was very high and greater than that found by other authors for the same or related species. The three filtration variables were better at higher effluent flux velocities. These preliminary results strongly suggest that both species are good candidates to be considered for bioremediation of aquaculture effluents.     

Experimental study of the susceptibility of Atlantic cod, Gadus morhua (L.), to infection with an IPNV strain pathogenic for Atlantic salmon, Salmo salar L

Intraperitoneal (IP) injection, cohabitation and immersion routes of infection were used to determine if Atlantic cod, Gadus morhua (L.), of 1 and 3 g are susceptible to infectious pancreatic necrosis (IPN). Mortalities of cod injected IP were significantly higher when challenged with infectious pancreatic necrosis virus (IPNV) than with phosphate buffered saline. This is the first report of Atlantic cod mortalities caused by IPNV. Fish challenged by cohabitation had significantly higher mortalities than the controls, but mortalities of Atlantic cod challenged with IPNV by immersion were not significantly different from controls. Titres of IPNV in the tissues of infected fish were sometimes very high (range 10²–10¹⁰ infectious units per gram of tissue) suggesting virus replication and titres of fish that died were generally higher than those of fish which survived. However, the relatively low mortality rates when challenged by cohabitation and immersion (20% and 17%, respectively), compared to the IP injection challenge (100%) suggest that 1 and 3 g cod have low susceptibility to IPN when challenged by more natural routes. These data strongly suggest that the cause of death of experimentally challenged cod was IPNV and further histological evidence for this came from 1 g cod challenged IP with IPNV in which the pancreas showed severe necrosis and heavy immunostaining for IPNV coincidentally with the peak of mortalities.     

Piscivory in larval perch (Perca fluviatilis): mechanisms structuring larval roach (Rutilus rutilus) cohorts

Abstract – Perch ( Perca fluviatilis ) can act as a piscivore from larval stage VI (body size 10.3 mm) on newly hatched larval roach ( Rutilus rutilus ), bream ( Abramis brama ) and smaller siblings of its own cohort. Consumption rates at this stage were approx. 0.5 prey/perch*h at 21°C. Larval perch predation was strongly gap-limited, and the maximum size of roach consumed by perch (perch length interval 10.3–62.0 mm) under experimental conditions followed the linear regression, P_prey-max.=0.478*L_Pred.+1.829 ( r ²=0.99, P <0.001, n =12). Under experimental conditions, predatory 0⁺ perch substantially affected the size distributions of 0⁺ roach prey cohorts, since smaller prey individuals were predated more frequently than larger ones. In both unimodal and bimodal size distributions of prey roach, the distributions changed according to the maximum prey size consumed by the added predatory perch. Unimodal prey distributions were positively skewed when piscivorous perch were added compared with controls without predators. According to the size distributions of lake-living 0⁺ roach and 0⁺ perch and the relative size difference between prey and predator, the vulnerability of 0⁺ roach cohort to 0⁺ perch predation changed from June to September. Prey vulnerability was extremely sensitive to the relative size difference between predator and prey. Therefore differences in hatching time and growth rates between the two species will strongly influence the potential for predator-prey interactions. ^Note     

Feeding and growth of glass eels, Anguilla anguilla L.: the effect of feeding stimulants on feed intake, energy metabolism and growth

Abstract. Two experiments were conducted in which glass eels, Anguilla anguilla L., were fed with six different diets. Two natural diets, cod roe and minced bovine spleen, and four artificial diets, an unflavoured trout fry crumble and three trout fry crumble diets with added feeding stimulants, were used. The feeding stimulants tested were cod roe extract, spleen extract and a mixture of synthetic L-amino acids. In the first experiment the glass eels were fed at a restricted feeding level and in the second experiment at an ad libitum feeding level. The intake of metabolizable energy (ME) was determined as the sum of retained energy and heat production. Metabolizability of the diets decreased with increasing feeding level and was 75-95% for cod roe, 35-50% for spleen and 40-60% for the artificial diets. Since utilization of ME for maintenance and growth was not affected by feeding level and diet, differences in growth rates could be explained by differences in ME intake. Group ME intake in turn was determined by the fraction of the fish which were actually eating. The uniform feeding response of the glass eels fed the natural diets resulted in the highest ad libitum ME intake (80-90 kJ/kg^0.8/day) and growth rate (4-4-4-9g/kg^0.8/day). Addition of feeding stimulants significantly increased the ad libitum ME intake of the artificial diets from 45 kJ/kg^0.8/day to 50-60 kJ/kg^0.8/day, which resulted in an increase in growth rate from 1·3 g/kg^0.8/day to 24 g/kg^0.8/day.     

Water resources management of the southern rivers of Russia with reference to fisheries requirements

Water consumption from Russia's southern rivers has increased substantially in recent years as a consequence of intensified economic activity. The volume of irreversible river-flow withdrawal from the Azov Sea basin varies between 11 and 13 km³ year^1. The majority of the abstracted water is used for irrigation. By 1990, the decrease in discharge into the Caspian Sea was 41 km³ year^1. Withdrawals of water from the Kura, Terek, Ural and other rivers have so far exceeded the ecologically admissible limits and become critical. The paper discusses methodological approaches for the elaboration of standards for water withdrawal and ecologically safe discharges into the southern seas. New data are presented on the dynamics of fish abundance as a function of the hydrological characteristics of rivers and water salinity. The results of the system analysis and regressive modelling formed the basis of a programme to develop policies for water management for fisheries in the lower Don. The principles and criteria for water withdrawal under variable hydrological regimes for the rivers entering the Azov and the Caspian Seas should be tested for their applicability for other water bodies (e.g. rivers, water reservoirs and lakes).