Implications of domestication and rearing conditions for the behaviour of cultivated fishes

Farmed fishes are often selectively bred for desirable production traits, such as rapid growth, that brings with them behavioural differences. In addition, the striking differences in the environment experienced by wild and cultured fishes offers considerable scope both for unplanned, natural selection for different inherited behavioural phenotypes and for behavioural differences arising from differential experience. In this paper, the evidence that such processes have produced behavioural differences between wild and cultured fishes is reviewed in relation to feeding, antipredator responses, aggression and reproductive behaviour. The reported findings are discussed in relation to the concept of 'behavioural syndromes', or suites of co‐varying behavioural traits that adapt individuals of the same population to spatial and temporal variation in selection regimes. The implications of the behaviour of cultured fishes for their welfare in production cages, for the environmental impact of escapees on wild stocks and for the success of hatchery‐based restocking programmes are considered. The review inevitably concentrates on salmonids, in which such phenomena have been intensively researched.     

Environmental effects on behavioural development consequences for fitness of captive‐reared fishes in the wild

Why do captive‐reared fishes generally have lower fitness in natural environments than wild conspecifics, even when the hatchery fishes are derived from wild parents from the local population? A thorough understanding of this question is the key to design artificial rearing environments that optimize post‐release performance, as well as to recognize the limitations of what can be achieved by modifying hatchery rearing methods. Fishes are generally very plastic in their development and through gene–environment interactions, epigenetic and maternal effects their phenotypes will develop differently depending on their rearing environment. This suggests that there is scope for modifying conventional rearing environments to better prepare fishes for release into the wild. The complexity of the natural environment is impossible to mimic in full‐scale rearing facilities. So, in reality, the challenge is to identify key modifications of the artificial rearing environment that are practically and economically feasible and that efficiently promote development towards a more wild‐like phenotype. Do such key modifications really exist? Here, attempts to use physical enrichment and density reduction to improve the performance of hatchery fishes are discussed and evaluated. These manipulations show potential to increase the fitness of hatchery fishes released into natural environments, but the success is strongly dependent on adequately adapting methods to species and life stage‐specific conditions.     

Environmental variability in the early rearing environment generates behaviourally flexible cod: implications for rehabilitating wild populations

The release of hatchery-reared fishes for restoring threatened and endangered populations is one of the most controversial issues in applied ecology. A central issue has been to determine whether releases cause extinction of local wild populations. This may arise either through domesticated or non-local fishes hybridizing with wild fishes, or through inappropriate behavioural interactions; for example, many hatchery fishes show exaggerated aggressive and competitive behaviour and out-compete wild counterparts. The impact of the impoverished hatchery environment in shaping behaviour is only now receiving attention. Attempts to counteract hatchery-related behavioural deficiencies have utilized intensive training programmes shortly before the fishes are released. However, we show here that simple exposure to variable spatial and foraging cues in the standard hatchery environment generates fishes with enhanced behavioural traits that are probably associated with improved survival in the wild. It appears that fishes need to experience a varying and changeable environment to learn and develop flexible behaviour. Using variable hatchery rearing environments to generate suitable phenotypes in combination with a knowledge of appropriate local genotypes, rehabilitation of wild fishes is likely to succeed, where to date it has largely failed.     

Habitat use by a Midwestern U.S.A. riverine fish assemblage: effects of season, water temperature and river discharge

The hypothesis that temperate stream fishes alter habitat use in response to changing water temperature and stream discharge was evaluated over a 1 year period in the Neosho River, Kansas, U.S.A. at two spatial scales. Winter patterns differed from those of all other seasons, with shallower water used less frequently, and low‐flow habitat more frequently, than at other times. Non‐random habitat use was more frequent at the point scale (4·5 m²) than at the larger reach scale (20–40 m), although patterns at both scales were similar. Relative to available habitats, assemblages used shallower, swifter‐flowing water as temperature increased, and shallower, slower‐flowing water as river discharge increased. River discharge had a stronger effect on assemblage habitat use than water temperature. Proportion of juveniles in the assemblage did not have a significant effect. This study suggests that many riverine fishes shift habitats in response to changing environmental conditions, and supports, at the assemblage level, the paradigm of lotic fishes switching from shallower, high‐velocity habitats in summer to deeper, low‐velocity habitats in winter, and of using shallower, low‐velocity habitats during periods of high discharge. Results also indicate that different species within temperate river fish assemblages show similar habitat use patterns at multiple scales in response to environmental gradients, but that non‐random use of available habitats is more frequent at small scales.     

Does diet in Celtic Sea fishes reflect prey availability?

Feeding preferences of Celtic Sea fishes were investigated using a database of stomach content records, collected between 1977 and 1994. The diet of cod Gadus morhua , hake Merluccius merluccius , megrim Lepidorhombus whiffiagonis , whiting Merlangius merlangus and saithe Pollachius virens changed markedly as the animals grew larger, and although large predators generally chose larger bodied prey, the variability of prey sizes consumed also increased. Large predators continued to select small, low value, benthic prey ( e.g . Callionymus spp. and Trisopterus spp.) which were easier to catch, rather than larger, more energy lucrative pelagic prey ( e.g . mackerel Scomber scombrus ), even though these pelagic prey‐fishes were nearly always available and were often very abundant. Stock estimates of the International Council for the Exploration of the Sea and U.K. groundfish survey catches were used as indices of prey abundance. Blue‐whiting Micromesistius poutassou and other small pelagic fishes ( Argentina spp. and clupeoids) were identified as being particularly important, and were consumed by some predators more often than would be expected given the abundance of these prey in the environment. There was no evidence for density‐dependent feeding by predators on mackerel and only hake exhibited density‐dependent feeding on horse‐mackerel. Hake, cod and megrim consumed more blue‐whiting when this prey was at higher abundance in the environment. In choosing what prey to consume, predators must balance costs and benefits, considering the quality of prey and the energy expended during search, capture and handling.     

Evaluation of biochemical methods for the non‐destructive identification of sex in upstream migrating salmon and sea trout

Female‐specific markers of reproductive activity [plasma 17β‐oestradiol (E2), vitellogenin (VTG) and alkali‐labile phosphoprotein phosphorous (ALP)] were measured over 12 months in a captive population of brown trout Salmo trutta . During the early months of the reproductive season (February to May) and using the concentration of plasma E2 or plasma ALP as a marker for females the proportion of fish in which sex was misidentified was high (15–50%). The misidentification rate was considerably lower (1–8%) using plasma VTG. Preliminary evaluation of a commercial immunochromatographic VTG test system as a screen for the presence or absence of VTG in plasma from brown trout provided results that were consistent with those obtained from direct measurement of plasma VTG levels by enzyme‐linked immunosorbent assay (ELISA). These preliminary conclusions were verified by sampling upstream‐migrating anadromous brown trout, sea trout, and Atlantic salmon Salmo salar trapped over a 6 month period. Plasma E2 levels did not satisfactorily discriminate between male and female sea trout and Atlantic salmon. Plasma VTG levels in both species, however, were bimodally distributed and it was assumed that this divergence corresponded to male (plasma VTG levels <10 μg ml⁻¹) and female (plasma VTG levels >800 μg ml⁻¹) fishes. Plasma ALP provided a more accurate indication of sex in the wild Atlantic salmon and sea trout than was suggested by the pilot study on captive brown trout. The commercial immunochromatographic VTG test system provided results that were wholly consistent with the data obtained from the trapped fishes by direct measurement of plasma VTG.     

Gastric evacuation of mixed stomach contents in predatory gadoids: an expanded application of the square root model to estimate food rations

A mechanistic gastric evacuation model of mixed stomach contents was established. Its simple generic nature should prove useful for studying predator‐prey interactions in fish communities. Only prey lengths were required as additional input. Surface dependent digestive processes were considered to act on equal fractions of individual prey‐cylinder surfaces. Primary and interactive effects of size, energy density and resistance to digestion of individual prey in a stomach were described by the model. Model predictions of results from experiments on gastric evacuation of meals composed of different prey types demonstrated the capability of this model, as opposed to previously applied model principles, to predict evacuation of mixed stomach contents involving the three above‐mentioned prey characteristics. The study furthermore illustrated that estimates of food ration might be severely biased by using improperly formulated effects of prey characters on gastric evacuation, and demonstrated that the new model holds the potential to predict food ration and diet composition for wild populations of predatory fishes.     

Effects of global warming on sex ratios in fishes

In fishes, sex is determined by genetics, the environment or an interaction of both. Temperature is among the most important environmental factors that can affect sex determination. As a consequence, changes in temperature at critical developmental stages can induce biases in primary sex ratios in some species. However, early sex ratios can also be biased by sex-specific tolerances to environmental stresses that may, in some cases, be amplified by changes in water temperature. Sex-specific reactions to environmental stress have been observed at early larval stages before gonad formation starts. It is therefore necessary to distinguish between temperature effects on sex determination, generally acting through the stress axis or epigenetic mechanisms, and temperature effects on sex-specific mortality. Both are likely to affect sex ratios and hence population dynamics. Moreover, in cases where temperature effects on sex determination lead to genotype–phenotype mismatches, long-term effects on population dynamics are possible, for example temperature-induced masculinization potentially leading to the loss of Y chromosomes or feminization to male-biased operational sex ratios in future generations. To date, most studies under controlled conditions conclude that if temperature affects sex ratios, elevated temperatures mostly lead to a male bias. The few studies that have been performed on wild populations seem to confirm this general trend. Recent findings suggest that transgenerational plasticity could mitigate the effects of warming on sex ratios in some populations.     

Abscisic acid involvement in the induction of summer‐dormancy in Poa bulbosa, a grass geophyte

Summer‐dormancy occurs in geophytes that inhabit regions with a Mediterranean climate (mild, rainy winters and hot, dry summers). The environmental control of summer‐dormancy and the involvement of phytohormones in its induction have been little studied. Poa bulbosa L. is a perennial grass geophyte in which summer‐dormancy is induced by long days and by high temperature. Prolonged treatment with ABA (0.1‐1.0 m M ) under non‐inductive 8‐h short days (SD) resulted in cessation of leaf and tiller production and in the development of typical features of dormancy: bulbing at the base of the tillers and leaf senescence. Short‐term applications of ABA had similar effects but dormancy was transient, i.e. after a short while, leaf growth from the formed bulbs was resumed. ABA treatment of plants growing under an inductive 16‐h photoperiod (LD) enhanced the onset of dormancy. Endogenous levels of ABA in leaf blades and at the tiller base (where the bulb develops) increased markedly after the plants were transferred from SD to LD. This increase was greater in the tiller base, and concomitant with bulb maturation. High temperature (27/22 vs 22/17°C) accelerated both bulb development and ABA accumulation in leaf blades.
These results suggest that ABA plays a key role in the photoperiodic induction and development of summer‐dormancy in P. bulbosa .     

Widespread hybridization among species of Indian major carps in hatcheries, but not in the wild

Twenty‐one allozyme loci in samples of wild‐caught and hatchery‐reared Indian major carps from Bangladesh were analysed. Bayesian model‐based clustering analysis revealed the presence of four taxa, corresponding to the three known species along with a fourth unknown taxon present in two hatchery samples. Individual admixture coefficients showed that 24% of all hatchery‐reared fishes were hybrids, whereas a single hybrid was observed in the wild‐caught samples. Only catla Catla catla × rohu Labeo rohita and mrigal Cirrhinus cirrhosus × rohu hybrids were observed, the vast majority of which were F₁ hybrids, though five individuals represented putative backcrosses. Mitochondrial DNA analysis revealed that catla × rohu hybridization primarily involved catla males and rohu females, whereas mrigal × rohu hybrids primarily resulted from rohu males and mrigal females. Despite the high percentage of F₁‐hybrids in hatchery samples, reproductive barriers among species have so far precluded widespread introgression. Continued hybridization may eventually lead to a breakdown of species barriers, thereby compromising the genetic integrity of the species in the wild, and leading to production losses in aquaculture.     

Does behaviour explain the larval dispersal of coral reef fishes better than oceanography?– An individual based model simulation approach

We housed offspring from northern (70° N) and southern (60° N) coastal cod ( Gadus morhua ) together in a 'common garden' rearing experiment at a temperature and light regime representative of the southern population. Through a more active feeding behaviour and a higher success, the northern cod achieved a larger food share and a higher growth rate and condition than their southern conspecifics. This is contrary to what was demonstrated by field data of fish from their natural habitats. The northern cod also allocated more energy to the liver throughout the experiment. Our results agree with the theory of countergradient variation, suggesting that genetic influences on growth and condition have been opposed by environmental constraints in their natural habitat. The observation that the offspring from these populations differ in behavior and growth when housed together support the idea that the growth response to selection would be through a behavioral response.
The field data suggest that density‐dependent population process and high juvenile density relative to prey limit the growth and condition in the wild and not necessarily the length of the growth season per se as assumed in the literature. The topographic distance (over 2000 km) limit mixing of early life stages of cod from the northern and southern population, and the different environmental stimuli (seasonality, temperature, food‐web interactions and habitat heterogeneity) in north and south are likely to evolve genetic differences.     

Bitterling as models for studies of sperm competition

The plight of the world fish stocks is all too well documented. As part of an ongoing attempt to bolster fish stocks for both commercial and conservation purposes, many fish are reared in captivity and released into the wild. It is well known that hatchery‐reared fish have low post‐release survival compared with wild fish of similar age. Part of the reason for this high mortality is that hatchery fish show deficits in virtually all aspects of their behaviour, including prey selection and predator avoidance. Much behaviour requires repeated experience so that it may become fine‐tuned to prevailing circumstances via learning during development. It has been suggested that inappropriate behaviour is encouraged when fish are reared in the unnatural surroundings of the hatchery. However, hatchery fish can be taught to recognise live, novel prey items and predators and the rate of learning is increased in the presence of a more knowledgeable conspecifics. Here we present data showing how social learning protocols can be used to dramatically increase foraging success in juvenile Atlantic salmon. We also outline related aspects of our ongoing research and discuss some of the possibilities for altering hatchery practices to maximize post‐release survival.     

Effects of temperature on free‐embryonic diapause in the autumn‐spawning bitterling Acheilognathus rhombeus(Teleostei: Cyprinidae)

In bitterling Acheilognathus rhombeus , developmental arrest always occurred at stage D of the free‐embryonic phase, regardless of incubation temperature. Developmental arrest was terminated only by a cold treatment at 4° C for 60–90 days, initiated 10 days post‐hatching. After the termination of developmental arrest, free‐embryos became larvae c . 6 months after hatching, regardless of the time of initiation and duration of the cold treatment. In hybridization experiments between A. rhombeus and several species of spring‐spawning bitterlings, free‐embryos became free‐swimming larvae within 60 days after hatching in all experiments. Developmental arrest was not observed in any of the hybrids, regardless of parental sex. These results suggest that free‐embryonic diapause in A. rhombeus is not induced by environmental factors, such as cold, but by genetic factors, which are recessive to those in spring‐spawning bitterlings. Free‐embryonic diapause in A. rhombeus appears to be an adaptation to winter, which might have evolved with reproduction in autumn among autumn‐spawning bitterling species. This is the only report of free‐embryonic diapause after hatching in fishes, and only the second example of diapause in fishes, along with annual killifishes (Rivulidae).     

Genetic risk involved in stock enhancement of fish having environmental sex determination

Environmental sex determination is reported in various fish species, including some fishes subject to stock enhancement. We studied the influence of stock enhancement on the sex-determining system of fish with both genotypic and environmental sex determination. We constructed and analyzed recurrence formula models for the dynamics of gene frequency in the male heterogametic (XX female and XY male) sex-determining system of the Japanese flounder (Paralichthys olivaceus). In this species, an XX individual can develop as a phenotypic male (sex-reversed male) depending on the conditions experienced by a juvenile. We show that the release of sex-reversed males may result in the extinction of the Y gene that determines sex. We also studied how the risk depends on the kind of hatchery broodstock used (e.g., fish collected from the wild or fish from a lineage established in a hatchery), sex-reversal rates in the hatchery and in the wild, and the relative contribution of released fish to reproduction in the wild. We propose some ways to reduce the risk. Received: October 10, 2001 / Accepted: December 17, 2001     

Patterns of growth and sexual size dimorphism in two species of box turtles with environmental sex determination

An adaptive explanation for environmental sex determination is that it promotes sexual size dimorphism when larger size benefits one sex more than the other. That is, if growth rates are determined by environment during development, then it is beneficial to match developmental environment to the sex that benefits more from larger size. However, larger size may also be a consequence of larger size at hatching or growing for a longer time, i.e., delayed age at first reproduction. Therefore, the adaptive significance of sexual size dimorphism and environmental sex determination can only be interpreted within the context of both growth and maturation. In addition, in those animals that continue to grow after maturation, sexual size dimorphism at age of first reproduction could differ from sexual size dimorphism at later ages as growth competes for energy with reproduction and maintenance. I compared growth using annuli on carapace scales in two species of box turtles (Terrapene carolina and T. ornata) that have similar patterns of environmental sex determination but, reportedly, have different patterns of sexual size dimorphism. In the populations I studied, sexual size dimorphism was in the same direction in both species; adult females were, on average, larger than adult males. This was due in part to males maturing earlier and therefore at smaller sizes than females. In spite of similar patterns of environmental sex determination, patterns of growth differed between the species. In T. carolina, males grew faster than females as juveniles but females had the larger asymptotic size. In T. ornata, males and females grew at similar rates and had similar asymptotic sizes. Sexual size dimorphism was greatest at maturation because, although males matured younger and smaller, they grew more as adults. There was, therefore, no consistent pattern of faster growth for females that may be ascribed to developmental temperature. Received: 20 March 1996 / Accepted: 10 March 1998     

Genetic architecture of growth and early life‐history transitions in anadromous and derived freshwater populations of steelhead

Heritabilities of growth, precocious maturation and smolting were measured in 75 families of juvenile steelhead or rainbow trout Oncorhynchus mykiss , progeny of within and between line matings (crosses) of wild, anadromous steelhead and wild, resident (lake) rainbow trout originally derived from the same anadromous stock 70 years earlier. The tagged yearling progeny were combined by line in common freshwater rearing containers and graded into three categories: mature, smolt or rearing (undifferentiated) at age 2 years. Heritabilities of precocious male maturity, smolting and growth were moderate to high, and the genetic correlation between growth and smolting was low. Smolting and precocious male maturity were highly variable among families within lines and significantly different between lines. Each of the four lines produced significant numbers of smolts at age two. Smolting and maturation were negatively genetically correlated, which may explain the persistence of smolting in the lake population despite strong selection against lake smolts; balancing selection on male maturation age may help to maintain variation for smolting. The high heritability of smolting, coupled with the inability of smolts that leave the lake to return to it indicates that the genetic potential for smolting can lie dormant or be maintained through a dynamic interaction between smolting and early maturation for decades despite complete selection against the phenotype. The results have significant implications for the preservation of threatened anadromous stocks in fresh water and the inclusion of resident fish of formerly anadromous populations, currently trapped behind long‐standing barriers to migration, as one component of the same population.     

Tractable models for testing theories about natural strategies: foraging behaviour and habitat selection of free‐ranging sharks

Marine and terrestrial environments differ fundamentally in space‐time scales of both physical and ecological processes. These differences will have an impact on the animals inhabiting each domain, particularly with respect to their spatial ecology. The behavioural strategies that underpin observed distributions of marine species are therefore important to consider. Comparatively little is known, however, about how wild fishes actually respond to gradients in food supply and temperature, and to potential mates. This paper describes how behavioural theory is being used to elucidate the strategies and tactics of free‐ranging sharks in three specific areas of study, namely, foraging on zooplankton, behavioural energetics and sexual segregation. The studies discussed are novel because shark movements were tracked in the wild using electronic tags in relation to simultaneous measurements of prey densities and thermal resources. The results show that filter‐feeding (basking shark, Cetorhinus maximus ) and predatory (dogfish, Scyliorhinus canicula ) sharks have relatively complex behaviour patterns integrally linked to maximizing surplus power, often through making short and longer term 'trade‐off' decisions between optimal foraging and thermal habitats. Interestingly, female S. canicula exhibit alternative behavioural strategies compared to males, a difference resulting in spatial segregation by habitat. Sexual segregation in this species occurs primarily as a consequence of male avoidance by females. Studies on free‐ranging sharks provide a useful model system for examining how a predator's strategy is shaped by its environment. More theory‐based studies of the behavioural processes of sharks are required however, before critical comparisons with other vertebrate predators are possible. Suggestions for further research to address this knowledge gap are given.     

Vulnerability of seamount fish to fishing: fuzzy analysis of life‐history attributes

Despite rather broad definitions, global analysis showed that seamount fishes, particularly seamount‐aggregating fishes, had higher intrinsic vulnerability than other groups of fishes. The pattern was similar when considering only commercially exploited species. Biological characteristics leading to greater vulnerability included a longer life span, later sexual maturation, slower growth and lower natural mortality. The results supported the contention that seamount fishes, especially those that aggregate on seamounts, are highly vulnerable to exploitation and that fishing on seamounts may not be sustainable at current levels and with current methods. A number of seamount populations have already been depleted; more depletion, extirpations, and even species extinctions may follow if fishing on seamounts is not reduced.     

Life‐history traits of the common pandora Pagellus erythrinus L., interpreted using information from aquaculture experiments

A study was conducted in in the Dale River as a part of a stock enhancement programme. The aim was to compare growth and the incidence of precocious maturation between offspring from precocious and large maturing males, and to study genotype‐environment interactions. River and hatchery performance was compared for 5 × 2 maternal half‐sib family groups, which were stocked as 0+ juveniles or maintained in the hatchery throughout. To identify the offspring, the broodfish were characterized by DNA‐fingerprinting (eight microsatellite markers). Smolt size of 1+ hatchery‐reared smolt and fish caught in a smolt trap, and the size and incidence of precocious maturation among 1+ hatchery‐reared parr and 1+ and 2+ parr caught in the river are compared between the families.     

Fasting modulates metabolic responses to cortisol, GH and IGF‐I in Arctic charr hepatocytes

Hepatocytes in primary culture from fed and 2 month fasted Arctic charr Salvelinus alpinus were exposed to physiological doses of either cortisol, salmon growth hormone (GH), salmon insulin‐like growth factor‐I (IGF‐I) or a combination of salmon GH and salmon IGF‐I. Fasting significantly lowered medium glucose levels compared to the fed fish, but had no significant effects on hepatocyte glycogen content or on the activities of enzymes involved in the intermediary metabolism. Cortisol treatment had no effect on hepatocyte glycogen content or on the enzyme activities investigated, but resulted in a significant increase in medium glucose concentration in hepatocytes isolated from fasted, but not fed fish. GH and IGF‐I treatments, both singly and in combination, significantly increased the glycogen content of hepatocytes isolated from fed fish, with less pronounced effects on hepatocytes isolated from fasted fish. The combination of GH and IGF‐I significantly increased lactate dehydrogenase activity regardless of the feeding state and significantly reduced the phosphenolpyruvate carboxykinase activity and medium glucose concentration in hepatocytes isolated from fed fish. Further, GH and IGF‐I significantly increased the activities of alanine aminotransferase and aspartate aminotransferase in hepatocytes isolated from fasted fish, but not fed fish. There were no effects of GH, IGF‐I, or their combination, on glucose 6‐phosphate dehydrogenase or 3‐hydroxyacyl‐CoA dehydrogenase activities. The results demonstrated that nutritional status of the animal modulates hepatocyte responsiveness to metabolic hormones, and suggested a role for GH and IGF‐I in hepatic glycogen conservation.