Social enhancement and social inhibition of foraging behaviour in hatchery-reared Atlantic salmon

The results of two experiments showed that observation of a trained conspecific Atlantic salmon Salmo salar significantly increased the rate at which naïve hatchery-reared fish accepted novel, live prey items, whereas the presence of an untrained conspecific actually decreased learning rates due to social inhibition. Pre-release training involving exposure of hatchery-reared fish to live prey items in the presence of pre-trained demonstrators would result in a significant enhancement in their foraging success on release and help prevent starvation, which is thought to be one of the principal causes of post-release mortality in hatchery-reared fishes.     

Social learning of prey location in hatchery-reared Atlantic salmon

Naïve, hatchery-reared Atlantic salmon Salmo salar parr were paired with demonstrators that had been pre-trained to accept live prey from the surface or from the benthos. After 6 days of observing demonstrators through a clear perspex partition the naïve fish's benthic foraging skills were tested. The results revealed that hatchery-reared Atlantic salmon can be taught to target benthic prey items by observation alone and social learning protocols can be utilized to dramatically increase benthic foraging success. The results are discussed with reference to refining hatchery-rearing practices with a view to improving the post-release survival of hatchery fishes. The role of learning, and in particular social learning, in the development foraging behaviour is highlighted.     

Prey selection by experienced and naive juvenile Atlantic salmon

Both in foraging groups and in a sequential prey encounter context, learning had a visible effect on the pattern of selection for three live prey types ( Ecdyonurus larvae, Hydropsyche larvae, and Gammarus ) by juvenile Atlantic salmon Salmo salar . Compared to wild-caught fish, naive, hatchery-reared fish that had not been exposed to natural prey ate Hydropsyche larvae in a remarkably low proportion, and consumed a higher proportion of Gammarus. Ecdyonurus experienced a high and rather steady predation rate across the experience gradient, but after a short period of experience with live prey the consumption rate for Hydropsyche increased drastically, and that of Gammarus decreased, matching the selection pattern exhibited by wild fish. Individual fish offered prey in a sequential encounter context increased consumption rates of all the prey types as they gained experience, but the improvement was higher for the prey that were less consumed initially. Fish became more selective as they approached satiation, conforming to the prediction of optimal foraging theory that higher predator's energy requirements, as well as low food availability, result in reduced selectivity. The results also suggest that fish from distinct populations can differ in the degree of diet selectivity according to their energetic requirements for growth. The fast learning response of Atlantic salmon parr towards novel prey probably allows fish to maintain a high foraging efficiency when faced with frequent changes in the availability of different prey types.     

Determination of body shape variation in Irish hatchery‐reared and wild Atlantic salmon

Discriminant function analysis was used to distinguish morphologically between samples of parr, smolts and adult Atlantic salmon Salmo salar from several hatchery and river systems in Ireland. The effect of habitat shift was investigated in Atlantic salmon parr. Parr grown from the eyed‐egg stage with a non‐sibling group in a hatchery environment, came to resemble the mean body shape of their host hatchery Atlantic salmon stock more closely than that of a full sibling group grown at their natal hatchery. Wild Atlantic salmon smolts differed in shape from hatchery‐reared smolts. This difference was less pronounced, but still statistically significant when wild adults were compared with hatchery‐reared adults captured in the coastal drift‐net fishery after a year spent at sea. Rearing conditions had a significant impact on the production and growth of fish body shape. This in turn may have affected adaptability and survivorship of ranched Atlantic salmon in the marine environment.     

Food partitioning between coexisting Atlantic salmon and brook trout in the Sainte‐Marguerite River ecosystem, Quebec

Food resource partitioning between similar‐sized, sympatric Atlantic salmon Salmo salar and brook trout Salvelinus fontinalis was examined as a possible mechanism enabling their coexistence in a stream (Allaire) of the Sainte‐Marguerite River ecosystem, Quebec, Canada. Fish stomach contents and invertebrate drift were collected concurrently during three diel cycles in August to September 1996. The food and feeding habits of an allopatric brook trout population in a nearby stream (Epinette) were studied for comparison. The diel feeding rhythms of the two coexisting fish species were similar. The composition of their diet, however, showed significant differences. Atlantic salmon predominantly (60–90%) fed on aquatic insects, mainly Ephemeroptera (35–60% of the diet). The brook trout mostly (50–80%) fed upon the allochthonous terrestrial insects (mainly adults of Coleoptera, Hymenoptera and Diptera) which comprised 5–40% of the stream drift. The allopatric brook trout fed opportunistically on the more abundant aquatic insects and terrestrial insects rarely formed 25% of its diet. The allopatric trout fed nearly twice as much as the sympatric brook trout during a day. The results suggest that the differences in feeding by brook trout in the two streams (with and without Atlantic salmon) are the result of inter‐specific interaction with Atlantic salmon and are not related to the differences in food availability between the two streams. Food resource partitioning between Atlantic salmon and brook trout may be viewed as an adaptive response resulting in a greater exploitation of available resources and coexistence.     

Differences in the pattern of antipredator behaviour between hatchery‐reared and wild European sea bass juveniles

Shoals of hatchery‐reared and wild sea bass juveniles Dicentrarchus labrax were tested for differences in their antipredator responses towards a potential live predator, the eel Anguilla anguilla . Eight experimental shoals ( i.e . replicates), each composed of 15 individuals from the same stock of juveniles ( i.e . wild or hatchery), were video recorded for 5 min before and after predator exposure. A set of behavioural variables were measured during the pre‐stimulus and stimulus phases of each test and compared between the two groups of replicates. Results showed that in both hatchery‐reared and wild juveniles predator exposure elicited a significant increase in the mean level of shoal cohesiveness and mean shoal distance from the predator, and a significant decrease in the mean shoal distance from the bottom. Shoals of wild juveniles, however, aggregated more quickly and reached higher shoal cohesiveness within the first 20 s of the stimulus period than shoals of hatchery‐reared fish. During this period, the wild fish also reached the highest peak in shoal cohesiveness, which then decreased gradually towards the levels observed before predator exposure. Another component of the antipredator response, the predator inspection behaviour, was fully developed in both wild and hatchery fish. Wild fish, however, tended to inspect the predator at a closer distance than hatchery fish.     

Prey selection and efficiency of naïve and experienced juvenile sockeye salmon

Prey selection and growth efficiency of juvenile sockeye salmon Oncorhynchus nerka switched between live prey and pelleted diets were investigated. First feeding sockeye salmon fry were placed into one of three dietary treatments for 7 months prior to assessing potential differences with a growth and a behavioural assay. Dietary treatments were (1) adult Artemia franciscana for 1 month, followed by pelleted feed for an additional 6 months ( Art − BD), (2) pelleted feed from first feeding for 7 months (BD) and (3) adult A. franciscana for 1 month, and a combination of pelleted feed and live adult A. franciscana for 6 months ( Art + BD). Equal numbers from each treatment group were then tagged, pooled into replicate 'common garden' tanks and fed novel live prey items ( Daphnia sp. and mosquito Culex pipiens larvae) for an additional 3 weeks. No significant differences in the growth efficiency of sockeye salmon were found during the 3 week feeding trial on the novel prey items. Additional sockeye salmon from each dietary treatment were used in a behavioural assay to determine if the treatments had an impact on foraging efficiency (prey selection or time to capture prey). No significant differences in prey selection were found among treatment groups in time to capture pellets, A. franciscana or mosquito larvae. Also, no significant differences were found within treatment groups in time to capture different food sources. No substantive benefits in foraging efficiency of sockeye salmon associated with prior exposure to live prey were demonstrated. This suggests that altering existing hatchery practices for juvenile sockeye salmon by offering live food prior to release is unlikely to influence post-hatchery feeding behaviour or increase post-release survival.     

Survival and behaviour of post‐smolt Atlantic salmon in coastal habitat with extreme tides

The early marine migration of 55 Atlantic salmon post‐smolts tagged with acoustic transmitters was automatically monitored using 13 to 25 km long arrays of receivers deployed inside the Bay of Fundy, a coastal system on the east coast of Canada. The survival of post‐smolts from groups with short‐ and long‐term transmitters in coastal habitat up to 10 km from the river was 92 to 100%, indicating a successful transition to salt water and departure. Migration for 68 to 77% of post‐smolts followed a direct route and it was rapid (transit time usually <12 h). Post‐smolts initially migrated in a south to south‐west direction ( i.e . orientation towards the mouth of the bay) and they were aggregated near the coast. Post‐smolts with long‐term transmitters were monitored 20 km from the river where they continued to be aggregated, moving near the coast through a 'common corridor', and their survival to that point was at least 84%. Post‐smolts from both groups travelled heading out of the coastal system during ebb tides. Flood tides interrupted migration, and they caused changes in travel direction and delays in departure for post‐smolts not leaving by a direct route. Monitoring of coastal habitat inside the Bay of Fundy intercepted 62% of migrating post‐smolts with long‐term transmitters returning after an initial absence of 2 to 22 days. Returning post‐smolts displayed a resident behaviour, using the habitat monitored inside the Bay of Fundy during July and August.     

Individual‐level analysis of pre‐ and post first‐feed growth and development in Atlantic salmon

Atlantic salmon Salmo salar were followed from egg to smoltification using genetic analysis to identify individuals and to link observations from pre‐ and post first feeding. Egg size and hatch timing significantly influenced alevin size at first feed but neither egg size, hatch timing or alevin sizes were correlated to size, condition factor or smolt status post first feed. In a hatchery environment the potential advantage gained by early hatching, larger alevin does not persist after first feeding. Different physiological and genetic complexes appear to influence growth in these two distinct phases of the Atlantic salmon's life‐cycle.     

Migration of hatchery‐reared Atlantic salmon and wild anadromous brown trout post‐smolts in a Norwegian fjord system

Hatchery‐reared Atlantic salmon Salmo salar ( n  = 25) and wild anadromous brown trout (sea trout) Salmo trutta ( n  = 15) smolts were tagged with coded acoustic transmitters and released at the mouth of the River Eira on the west coast of Norway. Data logging receivers recorded the fish during their outward migration at 9, 32, 48 and 77 km from the release site. Seventeen Atlantic salmon (68%) and eight sea trout (53%) were recorded after release. Mean migratory speeds between different receiver sites ranged from 0·49 to 1·82 body lengths (total length) per second (bl s⁻¹) for Atlantic salmon and 0·11–2·60 bl s⁻¹ for sea trout. Atlantic salmon were recorded 9, 48 and 77 km from the river mouth on average 28, 65 and 83 h after release, respectively. Sea trout were recorded 9 km from the release site 438 h after release. Only four (23%) sea trout were detected in the outer part of the fjord system, while the rest of the fish seemed to stay in the inner fjord system. The Atlantic salmon stayed for a longer time in the inner part than in the outer parts of the fjord system, but distinct from sea trout, migrated through the whole fjord system into the ocean.     

Effect of density on competition between wild and hatchery‐reared Atlantic salmon for shelter in winter

The effect of varying the density of hatchery‐reared Atlantic salmon Salmo salar on the ability of single wild fish to occupy a shelter is assessed. Although there was strong density‐dependence on sheltering overall, the ability of wild Atlantic salmon parr to occupy a shelter was not affected by the presence of hatchery‐reared fish even when outnumbered by four to one. These findings illustrate a competitive asymmetry for shelter in favour of the wild fish at the densities tested.     

Embryo survival and smolt to adult survival in second‐generation outbred coho salmon

Outbreeding depression was not detectable in observations of embryonic survival and of survival of smolts to adulthood relative to controls in coho salmon Oncorhynchus kisutch experimentally outbred over two generations by crossing fish from three widely separated populations. Survival in outbred coho salmon was not detectably less than that in second‐generation hybrid controls or second‐generation parental controls. Variation of embryonic survival was affected by the maternal source population and by individual females but not by the paternal source population or by individual males. Survival of smolts to adulthood at sea was greater in one second‐generation control group than in two others, but survival in second‐generation outbred groups was no less than in parental controls or hybrid controls.     

Evidence for co‐dominant allelic expression of a phosphoglucomutase regulatory locus polymorphism in the Atlantic salmon

Starch gel electrophoresis of the phosphoglucomutase isozyme, PGM‐1, in liver tissue in Atlantic salmon Salmo salar shows three phenotypes – full, partial and no expression. The genetic basis of the variation has not been established. Studies of the inheritance of the variation and of liver enzyme levels carried out indicate that the variation, as in rainbow trout Onchorynchus mykiss , reflects co‐dominant allelic variation at a cis ‐acting regulatory locus, designated PGM‐1r *, with one allele promoting and the other suppressing expression.     

Long‐term declines in body size and shifts in run timing of Atlantic salmon in Ireland

Records extending back from three to seven decades from Atlantic salmon Salmo salar fisheries on the Blackwater, Newport and Owenduff River systems in Ireland, were used to test the hypothesis that there have been progressive declines in size and associated shifts in run timing. They showed identical patterns in four respects: 1) peaks of catch, in spring (March to April) and summer (June to July), 2) the largest fish were caught in the spring, with declines in size over the summer and a slight increase in autumn, 3) marked declines over time in average size and in the proportion of the Atlantic salmon taken in spring and 4) reductions in size in most months at all three sites. Correlations were also detected in average fish mass among the rivers over time (especially the Owenduff and Newport systems), showing a slight rise from the late 1960s until the late 1970s, followed by a steep decline. Data from an electronic counter on the Blackwater showed a reduction in the abundance of early migrants. The early fish were apparently much more vulnerable to fishing than later migrants, so the catch records overstated the abundance of early fish. These long‐term declines in size, which have been noted elsewhere, may be due to other factors in addition to changes in marine productivity. Specifically, reduction in the abundance of large, spring‐running Atlantic salmon may be due to selection resulting from heavy angler exploitation.     

The survival of semi‐wild, wild and hatchery‐reared Atlantic salmon smolts of the Simojoki River in the Baltic Sea

The recapture rate and survival of hatchery‐reared Atlantic salmon Salmo salar stocked as 1 year‐old parr (semi‐wild) with that of hatchery‐reared Atlantic salmon stocked as 2 year‐old smolts and wild smolts of Atlantic salmon in the northern Baltic Sea were compared. This was done through tagging experiments carried out in 1986–1988 and 1992. The recapture rate of the semi‐wild groups varied from 1·0 to 13·1%, being similar in 3 tagging years and lower in 1 year than that of the wild groups (1·7–17·0%). The recapture rate of the semi‐wild groups was similar (in 2 years) or higher (in 2 years) than that of the hatchery‐reared groups stocked as smolts (1·3–6·3%). The survival of semi‐wild smolts during the sea migration was as high as that of wild Atlantic salmon of an equal size and two to three times higher than hatchery‐reared Atlantic salmon stocked as smolts. The survival rate was positively associated with smolt size. The suitability of hatchery‐reared parr and smolts in the management of reduced Atlantic salmon stocks is compared.     

Predator avoidance behaviour in wild and hatchery‐reared brown trout: the role of experience and domestication

Juvenile brown trout Salmo trutta from natural populations reacted to the presence of piscivorous brown trout by increasing the use of refuges. In contrast, second‐generation hatchery fish and the offspring of wild fish raised under hatchery conditions were insensitive to predation risk. The diel pattern of activity also differed between wild and hatchery brown trout. Second‐generation hatchery fish were predominantly active during daytime regardless of risk levels. Wild fish, however, showed a shift towards nocturnal activity in the presence of predators. These findings emphasize the potential role of domestication in weakening behavioural defences. They support the idea that the behavioural divergence between wild and domesticated individuals can arise from a process of direct or indirect selection on reduced responsiveness to predation risk, or as a lack of previous experience with predators.     

Eye colour in juvenile Atlantic salmon: effects of social status, aggression and foraging success

Status-based differences in sclera colour in small groups of juvenile Atlantic salmon Salmo salar , held in a semi-natural environment over a 20 day period, became obvious 3 days after the start of the study and persisted for the 20 days. Dominant fish had pale sclera, and this pattern was very stable over the experimental period. In contrast, the sclera colour of subordinate fish (ranks 2–5) fluctuated from day to day. Median sclera colour offish ranked 4–5 darkened on days they received more aggression, and sclera of rank 2 fish were lighter on days on which they initiated more attacks. Changes in sclera colour of fish ranked 2–4 were more frequent during feeding periods than non-feeding periods. This study confirms that the relationship between eye colour and status described in tanks is also seen in more natural environments, and also that colour change in juvenile salmonids is a complex response to local events.     

Imperfect optimal foraging and the paradox of enrichment

We show that the paradox of enrichment can be theoretically resolved in a flexible predator–prey system in which the predator practices imperfect optimal foraging. A previous study showed that perfect optimal foraging can mitigate increases in the amplitude of population oscillations associated with enrichment, but it did not show a stabilization pattern. Our results show that imperfect optimal foraging can stabilize the system and resolve the paradox of enrichment under nonequilibrium dynamics. Furthermore, the degree of stabilization with enrichment was stronger when the imperfection of optimal foraging was larger.     

Genetic implications of hatchery rearing in Atlantic salmon: effects of rearing environment on genetic composition

In an experiment to investigate genetic consequences of hatchery rearing in salmon, allozyme variation at five polymorphic loci was examined in Atlantic salmon of known initial genetic composition, which were reared throughout freshwater life in the hatchery or stocked into the wild as swim-up fry. The genetic composition of the juveniles in the hatchery remained homogeneous from fertilization up to stocking, and from stocking to 2+ in the wild, however, those remaining at the hatchery developed genetic differences among smolting and nonsmolting 1+ parr. These differences were attributed to conditions leading to early smolting at 1+ among the hatchery fish, with 1+ smolts diverging from the gene pool from which they were derived, whereas those stocked into the wild did not smolt until a year later and retained the original genetic composition. The results are discussed in relation to hatchery rearing of salmon and implications for the use of reared fish in stocking and enhancement programmes.     

Effects of early experience on the reproductive performance of Atlantic salmon

To examine the link between early experience and subsequentreproductive performance, we experimentally manipulated theearly experience of Atlantic salmon (Salmo salar). Salmon ofa common genetic background were reared as juveniles eithernaturally in the river or artificially in a hatchery (sea ranched),depriving them of river experience, and then allowed to growto maturity naturally in the ocean. At maturity, the competitiveand reproductive abilities of these fish were compared in fourexperiments quantifying reproductive success. Although levelsof aggressive behavior were similar, the experience-deprived,sea-ranched males were involved in more prolonged aggressiveencounters and incurred greater wounding and mortality thanwild males. Furthermore, sea-ranched males were less able tomonopolize spawnings and as a result obtained 51% the reproductivesuccess of wild males across the experiments. This reproductiveinferiority varied directly with the male density and bias inthe sex ratio, reflecting the intensity of male breeding competition.A lower intensity of female than male competition was likelyresponsible for the lack of differences in breeding performancebetween sea-ranched and wild females. Sea-ranched females, however,produced smaller eggs than wild females, apparently in responseto their higher juvenile growth rate. Differences in migratorybehavior were also apparent, as sea-ranched males and femalesascended the River Imsa later than wild fish. Our results indicatethat early experience has implications for subsequent adultreproductive performance, affecting the development of specializedskills and traits important not only for early life, but alsolater life.