Fin Damage Research Articles

Aggression‐induced fin damage modulates trade‐offs in burst and endurance swimming performance of mosquitofish

Locomotor performance is crucial to survival in many species. Swimming performance in fish depends on fin shape and size, and swimming performance may change with fin damage. The aim of this study was to investigate the relationship between fin size and swimming performance in male Eastern mosquitofish Gambusia holbrooki either with undamaged fins, or with fins that have sustained damage as a consequence of aggressive encounters. We show that in fish with undamaged fins burst swimming speeds increase with an increasing caudal fin size, while sustained swimming speeds (Ucrit) decrease with increasing fin size. In fish with damaged fins, Ucrit increases with an increasing caudal fin area, demonstrating a measurable cost of fin damage. The relationship between fin size and Ucrit is not linear but is best described by a Gaussian curve, where Ucrit decreases as fin size either increases or decreases from a central optimal value. We suggest that fish with large fins benefit because they can withstand more fin damage resulting from intraspecific aggression before experiencing detrimental effects such as reduced Ucrit.

Predation of stocked Brook CharSalvelinus fontinalisby Short-finned EelAnguilla australisand interactions with other salmonids in Wollondibby Creek in the high country of south eastern Australia

This study was initiated to investigate poor survival of stocked Brook Char Salvelinus fontinalis in Wollondibby Creek, a tributary of Lake Jindabyne in the south eastern highlands of NSW, Australia. Sampling of the creek following stocking was carried out on two separate occasions by poisoning 17.6 km, almost the entire length of the creek in 19 sections. Collection of fish was shown to be 78% effective. Five species of fish were present in the creek: Salvelinus fontinalis, Salmo trutta, Oncorhynchus mykiss, Anguilla australis and Galaxias sp. Population density in terms of numbers of fish per hectare and kilograms of fish per hectare of all species present was determined. Recovery of stocked S. fontinalis was low (0.38% in 1977 and 0.15% in 1979). Distribution of salmonid populations in the creek was significantly affected by physical barriers and eel predation, not cormorants as first thought. Eel predation was confirmed by the presence of salmonids in eel stomachs and extensive fin damage as a result of eel attack. Only five S. fontinalis in the first sampling event and one in the second sampling event had reached angling size (250 mm). It is suggested that stocking of S. fontinalis not be carried out in areas where A. australis occurs and that careful consideration be given to the location of physical barriers in the creek when stocking and the presence of native fish.

UV-induced fin damage in zebrafish as a system for evaluating the chemopreventive potential of broccoli and cauliflower extracts

This study applied broccoli and cauliflower extracts (whole, floret, and stem) to zebrafish larvae in parallel to receive 100 mJ/cm2 of UVB six times, and recorded their fin malformation phenotypes. Chemopreventive effects of each group, including UVB, whole-, floret-, and stem-extracts of broccoli and cauliflower on fin development were evaluated using Kaplan-Meier analysis, log-rank test, and Cox proportional hazards regression. Results showed that (1) zebrafish fins in the UVB + whole broccoli extract group are 6.20~9.32-times more likely to return to normal fins than ones in the UVB only group, but fins in the UVB + whole cauliflower extract group are only 5.13~11.10-times more likely to recover, indicated that whole broccoli and cauliflower extract had similar chemopreventive ability on fin development; and (2) the broccoli stem has the highest antioxidant capacity among other groups. In conclusion, zebrafish can be used as a system for evaluating the efficacy of other UVB protective compounds.

Social network analysis of behavioural interactions influencing fin damage development in Atlantic salmon ( Salmo salar) during feed-restriction

The role of behavioural interactions in the development of fin damage amongst Atlantic salmon ( Salmo salar) undergoing feed-restriction was investigated using social network analysis. Dorsal fin damage, particularly erosion, was seen only in groups subjected to feed-restriction. The amount of dorsal fin damage was positively correlated with aggression and fin-biting. Feed-restricted fish exhibited significantly lower weight gain, reduced growth-rate and body condition but no differences in total length. Social networks based on aggressive interactions in feed-restriction groups showed significantly lower distance, and higher density, higher clustering coefficient and higher in and out degree centrality. These findings indicated higher and more intense aggressive interactions in feed-restricted fish. A distinctive separation of roles according to aggression was found in feed-restriction groups where initiators had high out-degree centrality and receivers had high in-degree centrality. Fish initiating aggressive interactions had less fin damage, gained more weight and attained more central positions within the school. Fish receiving aggression had more fin damage and gained less weight. Association networks in the feed-restricted groups had significantly lower values for transitivity and distance with a tendency for higher centrality. These findings indicate higher levels of interaction and an imbalance in their relationships. We demonstrated the value of social network analysis in investigating behavioural interactions associated with aggression and the development of fin damage in Atlantic salmon.

Growing amago and rainbow trout in duoculture with self-feeding systems: Implications for production and welfare

Some salmonids such as rainbow trout ( Oncorhynchus mykiss: Walbaum, 1792) can quickly learn to use self-feeding systems while others such as amago ( Oncorhynchus masou masou: Brevoort, 1856) can take longer. The current study examined the potential costs and benefits of growing each of these species in duoculture with self-feeding systems in terms of welfare and production. Three self-feeding treatments were evaluated: rainbow trout in monoculture (RM), amago in monoculture (AM), and rainbow trout and amago in duoculture (D). Self-feeder learning time for all RM and D groups varied from 1 to 9 days, while AM ranged from 2 to 25 days. Within the D treatment there was no indication that amago learned to use self-feeders by watching the rainbow trout. Initial food intake was significantly lower for AM than the other groups, but this did not result in significantly lower growth rates. Further, the growth performance of both amago and rainbow trout were similar, irrespective of being grown in either mono- or duoculture. There were no significant differences in feed wastage between treatments and feed wastage was < 6%. The welfare implications of duoculture for each species were quantified using four welfare indicators: aggression, fin damage, condition factor and mortality. No significant differences were found in any of these indices. The results of this study suggest that when using duoculture for growing rainbow trout and amago with self-feeding systems, amago benefits from reduced variability in self-feeder learning times, i.e. all duoculture groups learned within 5 days while amago-only groups took up to 25 days to learn. Further to this, there were no obvious detrimental effects upon production and welfare.

Immediate fate of angled-and-released Australian bass Macquaria novemaculeata

The prevalence of catch-and-release factors known to adversely affect the mortality or physical condition of Australian bass Macquaria novemaculeata were surveyed across a range of anglers throughout impoundments and rivers in New South Wales. Subsamples of tournament-caught fish were also monitored in tanks for 1 h after being weighed to quantify immediate mortalities and sublethal physical damage. Most fish were caught on actively fished artificial baits, and were mouth hooked (96%), which resulted in no immediate mortality amongst monitored fish and a low frequency of mortality-causing factors in surveys. However, fish that were weighed-in during tournaments often had fin damage (52% of fish) and/or barotrauma (25% of fish, in impoundments only), and were held in live wells with poor water-quality. The prevalence of these sublethal effects varied considerably according to the specific seasons, locations, rules and procedures of each tournament; but could not be reliably attributed to any of the recorded catch-and-release variables (except for dissolved oxygen, which was significantly influenced by live-well volume). While these results validate the current release of angled Australian bass as a means of conserving their stocks, the potential for adverse effects could nevertheless be minimised via simple changes to conventional handling.

Effects of temperature, stocking density and farming conditions on fin damage in European sea bass (Dicentrarchuslabrax)

This paper presents a non invasive, rapid and reliable way to quantitatively assess fin erosion in sea bass (Dicentrarchus labrax). The method is based on a visual assessment of fin profile and area loss of all fins except the anterior dorsal, on a scale from 0 to 4 in comparison with a perfect fin. The effects of stocking density (SD) and temperature on fin damage were investigated under experimental conditions (100-250 g fish). Over a 4-month period, mean erosion index (mean erosion level of all fins) was 10 times higher at 120 than at 20 kg m −3 ,w here most fi ns were undamaged. Damage was also dependent on time and oxygen concentration (35% lower at 53% O2 saturation than at 105%). Fin condition was also affected by temperature: mean erosion index was 0.22-0.25 at 13-16 ◦ C, but five times higher at 25 ◦ C. Caudal and dorsal fins were always the most eroded. Fin damage was then monitored in a large land-based farm using high SD, and in a small sea cage farm operating at low SD. At the first site, 6 batches of two market size groups were examined: L (850-930 g, 50-80 kg m −3 ) and S (375-400 g, 42-60 kg m −3 ). Fin condition was good in all batches (mean erosion index, 1.1-1.3) and lowest at the highest SD. At the second site, four batches of large fish (350-890 g, 26-24 kg m −3 ) and four other groups below market size (100-270 g, 8-16 kg m −3 ) were examined. Fin erosion was the highest in large fish (mean erosion index, 1.1-1.2) and in the sea cages most exposed to climatic disturbances. In both sites, the most eroded fins were the caudal and dorsal. Differences in other external injuries were also observed between the two sites (less necrosis and more scale injuries in sea cages). The causes of fin damage are discussed in relation to metabolic and/or behavioral adaptations to rearing conditions and the main actions that could be taken to improve fin condition are discussed.

Further development of the “Fin Index” method for quantifying fin erosion in rainbow trout

Fin erosion is common in farmed salmonids and has recently been highlighted as a fish welfare issue. Methodology is needed to quantify the severity of fin erosion to support research, and this paper examines Kindschi's [Kindschi, G.A. 1987. Method for quantifying degree of fin erosion. Prog. Fish-Cult. 49, 314–315.] method which expresses fin length measurements relative to fish length. Naturally spawned wild rainbow trout Oncorhynchus mykiss ( n = 26; 92–335 mm standard length ( L S)) were sampled from a UK river to establish reference values for intact fin length. The relationships between intact fin size and L S were isometric for the anal, pectoral and pelvic fins, but negatively allometric for the dorsal and caudal fins, indicating that relative fin size decreases with increasing fish size for the latter two fins. Predictive relationships were derived between fin length and L S for the five rayed fins. These relationships showed good correspondence to maximum fin lengths observed in samples of the UK farmed rainbow trout metapopulation, and a UK feral rainbow trout population. The potential errors associated with assuming a constant relative fin size and not allowing for allometric fin growth, are illustrated. The equations predicting intact fin length from L S were formatted to enable estimation of the proportion of fin length lost from measures of observed fin length and L S in farmed fish. Such equations provide an intuitive fish welfare indicator (higher values representing greater fin loss), and represent an advance on Kindschi's “Fin Index” method for assessing the severity of fin damage in farmed rainbow trout.

Marine survival of reared Atlantic salmon in the Baltic Sea: The effect of smolt traits and annual factors

We examined the factors underlying the deteriorating marine survival of a hatchery stock of Atlantic salmon, Salmo salar. The stock originated from the salmon population of the Neva River, which flows into the Gulf of Finland in the Baltic Sea. During 1981–2005, about 134,000 Carlin-tagged Neva salmon smolts were released in the Gulf of Finland and the Bothnian Sea. We investigated the effects of fin damage, smolt length, sexual maturity (precocious maturation), smolt year, the time (days) between tagging and release and the release date on survival. The survival of the released smolts clearly decreased towards present. In earlier years, larger smolts tended to survive better than smaller ones, but the advantage of a large smolt size decreased in the later years of the study. No significant effect of dorsal fin damage on survival was detected. On average, immature smolts survived better than precocious males, and the proportion of precocious males increased over time. However, the proportion of precocious males among the tagged fish did not explain the decreasing trend in marine survival. Release year was the most important predictor of the long-term decreasing trend in marine survival. The optimal tagging time varies: in some years smolts tagged in the autumn survived better, while in other years spring taggings were more advantageous. For most years, a late release (in May or June) was more favourable than an early one (in April). As the releases were mainly carried out quite early in the spring in the last 13 years of the study, an unfavourable timing of release may partly explain the lower marine survival recorded during this period.

Submergence of Atlantic salmon (Salmo salar L.) in commercial scale sea-cages: A potential short-term solution to poor surface conditions

Submergence of Atlantic salmon (Salmo salar L.) in commercial scale sea-cages (1600–2000 m3) affected their behaviour, but did not alter growth rates, food conversion ratios, appetite, condition factor or fin condition in comparison with control cages held under similar environmental conditions. Four sea-cages each held 3300–4200 Atlantic salmon of 0.45 kg; two cages acted as controls, while two were submerged for 22 days with the roof held at 3 m depth. Salmon in the control cages mainly swam at similar depths to those maintained by submerged fish, resulting in both treatments experiencing similar temperature, light, salinity and dissolved oxygen levels. Submerged fish swam an average of 1.6 times faster (0.88 BL s−1) and in more structured schools than control fish (0.55 BL s−1). Specific growth rates (1.1–1.3% day−1), appetite (0.8–1.2% body weight day−1) and condition factor (1.10–1.15) were high, while food conversion ratios (0.6–1.0) and the incidence of fin damage were low in all cages, with no significant differences between the control and submerged treatments. Our results highlight the potential for submergence of salmon in sea-cages for short periods to avoid negative surface events.

Managing the Impacts of SCUBA Divers on Thailand's Coral Reefs

While dive tourism enjoys continued growth worldwide, concern exists that it is contributing to the degradation of coral communities, biologically and aesthetically. This study examined the effect of SCUBA diver contacts with coral and other substrates. Ninety-three percent of divers made contact with substrata during a 10-minute observation period with an average of 97 contacts per hour of diving. Two-thirds of the divers caused some coral damage by breaking fragments from fragile coral forms with an average of 19 breakages per hour of diving. Fin damage was the major type of damage. Underwater photographers caused less damage per contact than non-photographers; as did male divers, compared with females. Diver-induced damage decreases with increasing number of logged dives and attendance at pre-dive briefings. Park managers can help reduce impact by identifying and directing use to sites that are resistant to damage, matching diver competence and site preferences, and alerting operators to dive conditions. Minimising impact requires dive operators to be proactive in promoting minimal impact diving behaviour. This includes selecting sites that match diver expectations and experience, and providing pre-dive briefings in the context of diver activities and physical capacity, and site susceptibility to impact and current strength.

Effects of Two Different Landing Nets on Injuries to the Barramundi Lates calcarifer, an Iconic Australian Sport Fish

AbstractUse of catch and release, a growing aspect of recreational angling, is promoted for most recreational fisheries, including the fishery for the barramundi Lates calcarifer, an iconic sport fish of northern Australia. Landing nets are commonly used to remove fish from the water and to help hold and control fish while hooks or lures are removed before release. We investigated damage to mucus, scales, skin, and fins of 1.0–3.5‐kg barramundi that were caught and held for 60 s in two different types of landing net. The two landing nets represented the commonly used knotted‐mesh net and the newer knotless net, which has a smaller mesh size and a flat bottom. Both types of landing net caused some injury, such as mucus loss, scale loss, and fin tears. However, use of the knotless, flat‐bottomed landing net considerably reduced the occurrence of these injuries. Holding in the knotted net was associated with fin damage scores and dermal abrasion scores that were over four and five times greater, respectively, than the scores associated with holding in the knotless net. No fish died as a result of being handled in either net. These results are similar to those of other studies that investigated smaller fish. The effects of these injuries on fish survival are discussed.

Managing the Impacts of SCUBA Divers on Thailand's Coral Reefs

While dive tourism enjoys continued growth worldwide, concern exists that it is contributing to the degradation of coral communities, biologically and aesthetically. This study examined the effect of SCUBA diver contacts with coral and other substrates. Ninety-three percent of divers made contact with substrata during a 10-minute observation period with an average of 97 contacts per hour of diving. Two-thirds of the divers caused some coral damage by breaking fragments from fragile coral forms with an average of 19 breakages per hour of diving. Fin damage was the major type of damage. Underwater photographers caused less damage per contact than non-photographers; as did male divers, compared with females. Diver-induced damage decreases with increasing number of logged dives and attendance at pre-dive briefings. Park managers can help reduce impact by identifying and directing use to sites that are resistant to damage, matching diver competence and site preferences, and alerting operators to dive conditions. Minimising impact requires dive operators to be proactive in promoting minimal impact diving behaviour. This includes selecting sites that match diver expectations and experience, and providing pre-dive briefings in the context of diver activities and physical capacity, and site susceptibility to impact and current strength.

Mortality and physical damage of angled-and-released dusky flathead Platycephalus fuscus

We completed 2 experiments to quantify the mortality and physical damage (fin, blood and scale loss) of angled dusky flathead Platycephalus fuscus during a live weigh-in tournament (Expt 1) and after being immediately released by anglers (Expt 2). In each experiment, 84 and 79 angled P. fuscus were placed into up to 6 replicate tanks and, along with appropriate numbers of controls, monitored for mortalities over 5 d. Five and 7 of the angled fish died within 12 h, providing mortalities of 3.6 and 8.9% in Expts 1 and 2, respectively. One control fish died, which was attributed to an incorrectly inserted tag. None of the continuous or categorical variables collected during angling and subsequent release, or the experimental design, could explain the few observed mortalities to treatment fish. However, knotted large-mesh landing nets caused significantly greater fin damage than knotless fine-mesh designs in both experiments (p < 0.05). Furthermore, although water quality had no measured impact on the confined fish, the samples taken from anglers' live wells (mostly for Expt 1) were significantly poorer than the environment from which the fish were caught (p < 0.05). These latter differences may have had a cumulative negative impact on fish that manifested as the observed deaths. Simple changes to the post-capture handling of P. fuscus, involving appropriate landing nets and live wells, should reduce some angling impacts on this species.

The growth, feed conversion ratio and fin damage of rainbow troutOncorhynchus mykissunder self-feeding and hand-feeding regimes

The growth, feed conversion ratio and fin damage of rainbow trout<i>Oncorhynchus mykiss</i>under self-feeding and hand-feeding regimes

Growth, production and fin damage in cage-held 0+ Atlantic salmon pre-smolts (Salmo salar L.) fed either a) on-demand, or b) to a fixed satiation–restriction regime: Data from a commercial farm

This experiment investigated the impact of feeding regime upon the growth performance, production efficiency and fin damage of cage-held 0+ Atlantic salmon pre-smolts. Six groups of pre-smolts (n=61847±2620 fish group−1) were held in six 12×12×4 m production cages for 64 days (23rd August until 26th October) at initial and final densities of 4.6±0.2 kg m−3 and 9.5±0.5 kg m−3, respectively (mean±SEM). Fish were subjected to ambient photoperiod (11.67–16.05 h min–max) until 16th October, before being held under 24 h light. Fish were fed throughout the light phase and comparisons were made between three groups fed to an imposed regime (scheduled fixed ration feeding every 10 min) and three groups fed on-demand using commercial interactive feedback systems. The study was divided into primary and secondary phases: during the primary phase (part I) each regime was fed to satiation, whereas during the secondary phase (part II) fish under the imposed regime received a restricted ration that was ca. 20% less than those fed on-demand. During part I there were no differences in ration (g fish−1), fish condition and Specific Growth Rate (Gw) between regimes (Gw: 1.45±0.06 vs. 1.44±0.04 for imposed vs. on-demand fed fish, respectively). During part II underfeeding in the imposed regime had a significant and deleterious impact upon fish condition, FCR and growth (Gw: 0.45±0.04 vs. 0.84±0.09 for imposed vs. on-demand fed fish, respectively). Fish under the imposed regime also had significantly more dorsal fin splitting than those fed on-demand for the duration of the study, and significantly more dorsal fin erosion at the end of Parts I and II. In addition, the smallest fish within each cage suffered the greatest amount of fin splitting and erosion at the end of Parts I and II, irrespective of feeding regime. Daily ration amongst the on-demand cages decreased as winter approached; stepwise multiple regression analyses demonstrated that this was significantly related to a seasonal decrease in temperature.

Effect of electrical stunning at slaughter on the carcass, flesh and eating quality of farmed sea bass (Dicentrarchus labrax)

Food quality aspects of farmed sea bass (Dicentrarchus labrax) were compared following two methods of slaughter: the normal commercial method of killing, by immersion in an ice slurry, or by first electrically stunning the fish, before immersion in an ice slurry. Quality was assessed for up to 10 days of storage on ice after slaughter. No differences were found between the slaughter methods in terms of an overall sensory evaluation of cooked fillets, or in terms of overall carcass quality: overall appearance, internal and external haemorrhage, fin damage, burst gall bladder, staining of the body cavity by leakage from the gut or damage to the spine. Using objective measurements of colour, no differences between fish from either treatment were found in terms of external colour or colour of the fillets. A chemical analysis of flesh nucleotide breakdown products as well as the freshness indicator Ki value did not differentiate the two treatments nor did the industry standard freshness scoring technique (QIM, quality index method), over 10 days of storage on ice. Flesh pH was marginally lower in electrically stunned fish at 4 h post mortem (6.42 cf 6.56) but by 24 h, pH in fish from both treatments had decreased to a similar level (6.22). Humane electrical stunning of sea bass at slaughter neither measurably improved nor decreased product quality for between 1 and 10 days of storage on ice. Electrical stunning accelerated the pattern of onset and resolution of rigor mortis. If electrical stunning were to be widely adopted, re-education of buyers would be necessary as rigor mortis is currently used by buyers as a proxy measure of fish freshness.

The effect of trigger depth on self-feeder utilisation, growth and the vertical distribution of tank-held rainbow trout Oncorhynchus mykiss (Walbaum)

This study investigated the effect of placing a self-feeding trigger at either a) 1 cm below the water surface or b) mid-water (35 cm below the surface), on self-feeder usage, growth, feed wastage and fin damage of juvenile rainbow trout, Oncorhynchus mykiss (Walbaum). The vertical distribution of trout was also measured for the first 31 days of self-feeding. Group-held trout (n=10 fish group−1, 5 groups treatment−1), weighing 51.06 g±8.09 g (mean±SD) were held in 200 l cylindroconical tanks (depth=75 cm) for 62 days. Trigger placement had no significant effect upon self-feeder conditioning, and upper trigger group (UTG) and lower trigger group (LTG) fish took an average of 16±7.4 days and 14±9.8 days to reach a stable level of self-feeding, respectively. Specific growth rate was significantly higher in UTG fish (1.83±0.20 vs. 0.95±0.21 for UTG vs. LTG fish, respectively) whilst condition factor and self-feeder utilisation (ration size, 1.21±0.18 vs. 0.52±0.15 kg tank−1 for UTG vs. LTG fish) were also significantly higher in UTG fish. Trigger depth had no significant effect on size heterogeneity, mortality, feed wastage, feed conversion ratio (FCR) or the incidence of caudal and dorsal fin splitting and fin erosion. Dorsal fin splitting decreased with time in UTG but not in LTG fish. Caudal fin erosion increased with time in UTG (upper lobe only) and LTG (both lobes). The vertical distribution of fish also differed between treatments and more LTG fish preferred the lower part of the tank. Placing a self-feeding trigger at a mid-water depth can significantly limit growth, condition factor and self-feeder utilisation in juvenile rainbow trout, in addition to hindering recovery from dorsal fin erosion.

Effects of Rearing Density and Structural Complexity on the Pre‐ and Postrelease Performance of Atlantic Salmon

AbstractHatchery fish released for supplementation purposes often have difficulties adapting to wild conditions and, therefore, perform poorly in the wild. This can, at least partly, be explained by differences between hatchery and wild conditions. Our objective was to evaluate the effects of rearing density and structural complexity on the culture and postrelease performance of age‐0 Atlantic salmonSalmo salar. By using a 2 × 2 factorial design, we manipulated density (standard density and one‐third of standard density) and structure (standard rearing tanks with and without added rocks and plastic plants) in a conventional hatchery. After 3 months of rearing, 300 fish/treatment were individually tagged with passive integrated transponders, released in a nearby stream, and recaptured in late autumn (November) and summer (June). Fish not released were retained in the hatchery until smolting occurred during the next spring. In the stream, Atlantic salmon reared at reduced density grew faster during the first period after release, but there was no difference in June. The treatments had no effect on postrelease survival estimated by the recapture rates. In the hatchery, fish kept at low density with structure grew faster than conventionally reared fish. At smoltification, fish kept at low density had higher levels of insulin‐like growth factor I than those reared at standard density. Independent of size, fish kept at low density were more silvery (smolt‐like) and had a lower mortality rate than fish reared at high density. There was also a density effect on dorsal fin damage; Atlantic salmon at reduced density had less‐damaged fins than those at standard density. These results collectively indicate that reduced rearing density may be more important than structural complexity for improving postrelease performance of juvenile Atlantic salmon.

Influence of feeding regime on intraspecific competition, fin damage and growth in 1+ Atlantic salmon parr (Salmo salar L.) held in freshwater production cages

The feeding behaviour, growth and feed conversion ratio (FCR) of cage-held Atlantic salmon parr (Salmo salar L.) were studied when in 576 m3 (12 m × 12 m × 4 m) commercial freshwater cages under ambient water temperature (8.84±3.53°C) and photoperiod (11.02±2.05 h) for 205 days. The effect of feeding regime on fin damage was also investigated. Six groups (n=31 234±2051 fish group−1, initial stocking density 1.25±0.14 kg m−3) were fed to satiation using either (a) an imposed regime involving scheduled, fixed ration feeding every 10 min from dawn till dusk or (b) on demand from dawn till dusk using commercial interactive feedback systems. During feeding, there were no significant differences in aggression although swimming speeds and turning angles were significantly higher in fish under the imposed regime. On-demand feeding significantly reduced the incidence of dorsal fin damage. There was no clear relationship between fish size, feed regime and the incidence of fin damage until 1 week before the fish were transferred to marine cages, when the smallest fish under each feeding regime had the highest incidence of fin damage. Interestingly, growth did not differ between regimes, but fish under the imposed regime were significantly overfed and achieved higher FCRs.