Rapid and non-invasive measurements of fat and pigment concentrations in live and slaughtered Atlantic salmon (Salmo salar L.)

The aim of this study was to non-invasively determine fat and pigment concentrations in salmon muscle based on visible and near infrared (VIS/NIR) spectroscopy measurements of live/whole fish and fillets, and by means of digital photography (DP) of fillets. The fish used were two populations of farmed Atlantic salmon (Salmo salar L.) consisting of 46 salmon averaging 0.7 kg (range 0.17–1.7 kg, Group S) and 30 salmon averaging 2.3 kg (range 1.4–4.1 kg, Group L). Chemical analyses (fat and pigment content) and computerized tomography, CT (fat content) were used as reference methods. Group L was analysed in the live state (VIS/NIR), after gutting (VIS/NIR and CT), and as fillets (VIS/NIR and DP). Group S was analysed in the gutted state (VIS/NIR) and as fillets (VIS/NIR and DP). VIS spectroscopy predictions of pigment in whole salmon from Group S were obtained with a root mean square error of prediction (RMSEP) of 0.9 mg kg^− 1 astaxanthin, and a correlation between VIS spectroscopy predicted and chemically measured pigment of r = 0.85 (p < 0.0001). The fat concentration was determined by the NIR spectroscopy in live fish with RMSEP = 1.0 fat% unit, and a correlation with chemical reference values of r = 0.94 (p < 0.0001). Fat predictions from NIR spectroscopy correlated also well with predictions from CT analyses, r = 0.95 (p < 0.0001). VIS spectroscopy and DP were equally well suited to determine pigment concentrations in salmon fillets, with prediction errors of only 0.4 mg kg^− 1 astaxanthin, and a correlation with chemically determined pigment of r = 0.92 (p < 0.0001). The results obtained in the present study are the first to demonstrate successful non-invasive pigment predictions in whole salmon using VIS/NIR spectroscopy, and the possibility for simultaneous, rapid and non-destructive quantification of fat and pigment concentrations.     

Temporal changes in texture, gaping, composition and copper status of Atlantic salmon (Salmo salar, L) fed moist feed or extruded dry feed

Temporal development in product quality and copper status was investigated in Atlantic salmon (initial weight 2.3 kg) fed alginate-based moist feed (MF) or commercial extruded dry feed (DF). The fieldwork was conducted from January to May in sea-cages (800 m³) off the Norwegian west coast. The dry matter content (DM) in the muscle of the MF group was stable at about 32% throughout the experiment, while the DM of DF group increased linearly from 32% to 35%. Salmon were sampled for analyses of Cu status, fillet gaping, texture and pH in March and April. The serum Cu content of the MF group was highest in March, but no significant effect of dietary treatment was observed on Cu status of muscle or liver. The Cu level in serum and muscle decreased from March to April in both fish groups, but the hepatic Cu level showed no temporal changes. Gaping incidence and pH were higher in the MF group in April compared with the DF group and, for both fish groups, fillet gaping was higher and hardness was lower in April than in March. Variation in serum Cu level explained 32% of the variation in fillet hardness, and serum and muscle Cu level related negatively with gaping. At the end of the experiment the condition factor, and fillet and visceral fat contents were higher in the DF group, while sensory evaluated hardness and tastiness were similar for both dietary treatments. In conclusion, sampling time had a greater impact on Cu status, texture and gaping than dietary treatment during springtime in adult Atlantic salmon. Dietary treatment significantly affected body shape, and fat and dry matter contents.     

Anisakid larva in the viscera of a farmed Atlantic salmon (Salmo salar)

Parasitic nematodes of the family Anisakidae occur in the visceral cavity and surrounding tissues of many marine fish species at a prevalence as high as 100% in wild salmon samples. Human consumption of fish products containing these parasites can result in the zoonotic disease anisakiasis, and Anisakis simplex is most commonly associated with human disease. Previous studies of farmed salmon have found no anisakids in viscera or muscle, presumably because formulated feed production renders parasite larvae nonviable. However, among farmed Atlantic salmon (Salmo salar) examined by histopathology as part of a provincial government auditing program in British Columbia, Canada, 1 of 894 (0.11%) had an anisakid larva partly embedded in the wall of an intestinal cecum. Skeletal muscle is not examined as part of the government's program, but other studies have correlated anisakids in the viscera and muscle. Using anisakid prevalence in viscera as an estimate of its prevalence in muscle, the risk ratio of anisakid parasites in commercial product is 570 times less in farmed than in wild Atlantic salmon.     

Furunculosis in Atlantic salmon (Salmo salar L.) is not readily controllable by bacteriophage therapy

The potential of bacteriophage therapy to control bacterial disease in farmed fish was tested using, as an example, furunculosis of Atlantic salmon, caused by Aeromonas salmonicida subsp. salmonicida.

In vivo testing with Atlantic salmon and rainbow trout (Oncorhynchus mykiss Walbaum) showed no adverse effects, with bacteriophage generally cleared within 96 h of administration by either intraperitoneal (i.p.) injection or oral in-feed.

Juvenile Atlantic salmon were administered a combination of bacteriophage O, R and B (1.9 × 10⁸ pfu fish^− 1) by i.p. injection, after they had been challenged with A. salmonicida subsp. salmonicida 78027, also by i.p. injection. The fish that were injected with bacteriophage immediately after challenge died at a significantly slower rate then those that were either not treated with bacteriophage, or treated 24 h post-challenge. However, the end result (100% mortality) was not affected.

In further experiments the effects of oral (1.88 × 10⁵ pfu g^− 1 fish^− 1 daily for 30 days), bath (1.04 × 10⁵ ml^− 1 daily for 30 days) and i.p. (6.25 × 10⁷ pfu fish^− 1) phage treatment to control furunculosis in experimentally infected Atlantic salmon were compared with antibiotherapy (treatment with 10 mg kg^− 1 bw^− 1 day^− 1 oxolinic acid for 10 days), using an indirect cohabitation challenge. No protection was offered by any of the bacteriophage treatments, compared to the positive challenge group, although significant protection was offered by the oxolinic acid treatment. Analysis of samples taken from the trials demonstrated that bacteriophage were correctly administered to the fish and, on occasion, were isolated from fish that had succumbed to furunculosis. It was also shown that bacteriophage resistant A. salmonicida subsp. salmonicida isolates could be recovered from mortalities in all the treatment groups.

The results suggest that, although there were no safety problems associated with the approach, furunculosis in Atlantic salmon is not readily controllable by application of bacteriophage.     

The mineral content affects vertebral morphology in underyearling smolt of Atlantic salmon (Salmo salar L.)

This study investigated the individual variation in vertebral morphology and mineral content (as % of bone dry weight) in rapid growing underyearling smolt (n = 33) twelve weeks after transfer to seawater. The smolt, which were reared in triplicate tanks, had a large individual variation in vertebral bone mineral content (Tank A: 36.2% min, 45.2% max, 39.6% mean. Tank B: 32.6% min, 46.6% max, 40.0% mean. Tank C: 38.5% min, 47.9% max, 42.4% mean), and were subdivided into two groups based on mean within each tank: high mineral content (HMC, above tank mean, n = 15) and low mineral content (LMC, below tank mean, n = 18). Vertebrae of smolt with a low mineral content were more compressed in the anterior–posterior direction (HMC vertebral length/dorso-ventral diameter ratio 0.83, LMC 0.80). This was related to a change in the growth direction of the vertebral cone. This change occurred around transfer to seawater and was characterised by an increased angle between the wall of the cone of the vertebrae and the cranial–caudal axis, and gave the vertebrae a compressed appearance. Smolt with a high mineral content were significantly longer (HMC 28.1 cm, LMC 26.5 cm) and heavier (HMC 270 g, LMC 210 g) twelve weeks after transfer to seawater.     

Lipid absorption in different segments of the gastrointestinal tract of Atlantic salmon (Salmo salar L.)

Gastrointestinal lipid absorption in Atlantic salmon was studied following force-feeding with a standard grower diet to which isotope-labelled free fatty acid (FFA) [1-¹⁴C]-decanoic acid (¹⁴C-10:0) and triacylglycerol (TAG) [9,10-³H(N)]-triolein (trioleic acid; ³H-18:1) were added. Following force-feeding, the 15 fish, with an average weight of 704 g, were kept in five circular tanks, three fish per tank, supplied with seawater (salinity, 32 g l⁻¹). At 3, 6, 12, 18 and 48 h, the fish from one tank were killed, and the gastrointestinal (GI) tract was removed and divided into six segments: stomach (ST), anterior, mid, and distal pyloric ceca (PCA, PCM, and PCD, respectively), mid intestine (MI), and distal intestine (DI). The levels of radioactivity (dpm [disintegrations per minute] μg⁻¹ tissue) from the ¹⁴C-10:0 and the ³H-18:1 substrates were measured in tissue samples from each segment, allowing an estimation of the amounts of fatty acid substrates absorbed from the lumen into the intestinal cells. The highest levels of radioactivity from both fatty acid substrates were measured in tissue from pyloric ceca (PC) 12 to 18 h after force-feeding. The level of radioactivity from the ¹⁴C-10:0 substrate was significantly higher in the PCA than in the PCM and PCD, while there was no clear difference in the level of radioactivity from the ³H-18:1 substrate measured among the different PC tissue segments and the level measured in MI. The level of radioactivity in tissue from DI was negligible to low for both fatty acid substrates. Total lipid accumulation in tissue from PCM, MI and DI was quantified morphometrically at each sampling point using light microscopy and image analysis. These results confirmed the findings from the isotope study, where the highest levels of fat were present at 12 and 18 h following force-feeding. In conclusion, PC was the most important site for fatty acid absorption in the GI tract of Atlantic salmon.     

Microsatellite genetic variation between and within farmed and wild Atlantic salmon (Salmo salar) populations

Genetic diversity between three farmed and four wild populations of Atlantic salmon from Ireland and Norway were analysed using 15 microsatellite markers. High levels of polymorphism were observed over all populations with the average number of alleles and average heterozygosity at 17.8 and 0.70, respectively. Farmed salmon showed less genetic variability than wild salmon in terms of allelic diversity but not necessarily in terms of overall heterozygosity. Between farmed populations significant differences were observed in expected heterozygosity suggesting that more intensive breeding practices may have resulted in a further erosion of genetic variability. Phylogenetic analysis using either populations or individuals as nodes show a clustering of populations into two groups, farmed and wild. This suggests that founder effects and subsequent selection have had more effect on the genetic differentiation between these strains than geographical separation. This technology has great potential for use in aquaculture situation where levels of genetic variation could be monitored and inbreeding controlled in a commercial breeding progra.     

The effect of slaughtering procedures on blood spotting in rainbow trout (Oncorhynchus mykiss) and Atlantic salmon (Salmo salar)

Rainbow trout (Oncorhynchus mykiss) and Atlantic salmon (Salmo salar) were submitted for various slaughter and bleeding procedures to see what effect this would have on blood drainage of the muscles. Results show that the bleeding method is of less importance, while it is the timing that is important. No significant difference in bloodspotting was observed between fish that were bled live by a gill cut or percussive killed and bled by gutting. Most of the drainage of blood in the fish muscle seems to occur within the first hours postmortem, so rigor mortis is of little importance. The visual appearance of the fillet was influenced by number and size of the bloodstains. Colour measurements with Hunter L*, a*, b* did not reveal this. We conclude that a gill cut is not necessarily to obtain bleeding, so the industry can omit this phase and go directly to gutting.     

Seawater adaptation and growth of post-smolt Atlantic salmon (Salmo salar) of wild and farmed strains

Atlantic salmon smolts (Salmo salar L.) of wild (Namsen) and farmed (AquaGen) strains were transferred to full-strength seawater (33‰) for 0 (initial control group), 0.5, 1, 2, 4, 8, 14, 30, 42 and 60 days at three different temperatures (4.3, 9.4 and 14.3 °C). Freshwater temperature in each tank was adjusted to test conditions 10 days prior to transfer. Physiological adaptation was monitored as changes in plasma growth hormone levels, gill Na⁺,K⁺-ATPase activity, plasma chloride levels and survival in seawater. Overall, smolts from the wild strain were better able to tolerate transfer to seawater than smolts from the hatchery strain. A delay in the osmotic disturbance and a prolonged period of osmotic stress were observed at the low temperature. Circulating GH levels increased transiently in all groups during the first 12–48 h in seawater and long-term GH levels were positively correlated with seawater temperature. Growth rates were influenced by temperature and strain, with the farmed smolts showing a higher growth than the wild smolts. Food Conversion Efficiency (FCE) was higher in smolts from the farmed strain, whereas no differences in daily food consumption were observed. Optimum temperature for FCE was calculated to be 10.5 °C, whereas the optimal temperature for growth in seawater was calculated to be 13.0 °C. We suggest that the observed differences in seawater tolerance, growth and food conversion probably are genetic and may reflect the fact that the hatchery fish have been bred for rapid growth for several generations.     

Long-term separate and combined effects of environmental hypercapnia and hyperoxia in Atlantic salmon (Salmo salar L.) smolts

Atlantic salmon (Salmo salar L.) parr (mean start weight 50 g) were reared in freshwater (FW) and exposed to three levels of oxygen saturation measured in effluent water; control group (93% O₂, LO₂), medium (111% O₂, MO₂) and high (123% O₂, HO₂). Further three groups were exposed to similar water oxygen levels in combination with elevated carbon dioxide levels (17–18 mg L^– 1 CO₂), named LO₂–CO₂, MO₂–CO₂ and HO₂–CO₂, respectively. The experiment was run in duplicate tanks for 42 days, and the fish were subsequently transferred to the same seawater (SW) regime for 45 days for an assessment of post-smolt growth. As a consequence of the CO₂ addition, tank pH levels in the FW period were reduced from 6.7 to 5.9 for the hypercapnia groups compared to for the normcapnia groups. Water temperature in FW ranged between 6.4 and 9.0 °C. Citrate was added to the water to complex labile aluminium.In the CO₂ groups observed ventilation frequencies were significantly increased compared to the control (p < 0.05). This difference declined towards the end of the FW period, suggesting acclimation to elevated CO₂. The degree of oxygenation appeared to contribute to the acclimation as the lowest mean ventilation frequency on day 36 was found in the HO₂–CO₂ group and the highest in the LO₂–CO₂ group. Lower plasma chloride and sodium levels were observed in the CO₂ groups relative to the respective oxygenation groups during the FW period, while plasma chloride and sodium levels were normalised to equal levels for all groups after 44 days in SW. No significant differences were found among treatments for blood concentrations of red blood cells, haemoglobin, potassium and glucose during the experiment.By termination of the FW period, the HO₂ group had significantly higher body weight than all other groups (p < 0.05), with specific growth rate significantly higher than the CO₂ groups (p < 0.05). Further, the condition factor was significantly lower in all the CO₂ groups at the end of the FW period compared to the control and normcapnia groups (p < 0.05). Although variable among replicates, occurrence of nephrocalcinosis was 10 times higher in the hypercapnia groups than in the control and normcapnia groups. Mortality was negligible (< 2.0%) during the trial, and most of the mortality occurred following SW transfer.     

Effects of Carlin tagging and fin clipping on survival of Atlantic salmon (Salmo salar L.) released as smolts

During the period 1976–1979, in total 2567 wild migrating smolts of Atlantic salmon were Carlin-tagged, 2528 were adipose fin-clipped and 2309 individuals were released unmarked into the River Imsa, S.W. Norway. The smolts were caught in a permanent Wolf trap. The adult return rate to the river was highest for unmarked and lowest for Carlin-tagged fish. The main reason for this is probably a mortality of smolts due to handling, anaesthesia and tagging.     

Relevance of dietary oil source for contraction and quality of pre-rigor filleted Atlantic cod, Gadus morhua

Rigor contraction and quality development were studied during storage of pre-rigor fillets of farmed Atlantic cod (body weight 2 kg). For eight weeks prior to slaughtering, the fish were fed diets containing either 100% fish oil (M-group) or 60% fish oil and 40% soybean oil (S-group). Filleting was performed 0.5 h after slaughtering and the fillets were stored at 6 °C for 48 h post-mortem. The contraction rate was faster for the S-group, but after 48 h storage the contraction was 21% of the initial fillet length for both fish groups. The ATP content analysed 1 h post-mortem was lower of the S-group (4.5 μm g^− 1 ww) compared with the M-group (6.1 μm g^− 1 ww). After 16 h, the ATP content was similar for both fish groups (1.7-2.2 μm g^− 1 ww). The muscle pH analysed 1 h post-mortem was 0.2 unit lower of the S-group (pH 6.9) than of the M-group (pH 7.1). From 8 to 12 h, the pH fell significantly to 6.3 and 6.6 for the S-group and M-group, respectively. From 24 h and onwards, the pH was similar of both fish groups (pH 6.2). Texture development differed significantly due to dietary treatment. The S-group had highest initial breaking force, demonstrating that the texture of this group was firmer immediately after slaughtering. The breaking force of the S-group dropped to a stable level after 4 h, whereas the breaking force of the M-group decreased continuously during the storage period. The breaking force was similar of both groups after 16 h storage. The L*-value (lightness) increased during the first 4-6 h storage. From 8 to 16 h, the L*-value was stable or slightly decreasing. The L*-value was consistently higher of the M-group. In conclusion, dietary inclusion of soybean oil resulted in faster energy depletion post-mortem, faster rigor contraction, faster reduction of the breaking strength and lower L*-values.     

Differences in sea migration between wild and reared Atlantic salmon (Salmo salar L.) in the Baltic Sea

The effect of origin, smolt size and year of release on the sea migration pattern of Atlantic salmon (Salmo salar L.) in the Baltic Sea was examined by tagging experiments conducted in 1991–1993 on wild and reared smolts of the Simojoki river salmon stock. The tag recovery data analysed by log-linear models revealed significant differences in both spatial and temporal sea migrations between the wild and reared salmon; the variation was attributed to the year of release and to the origin of the fish. Grilse accounted for the majority of reared returners (76%) but for a smaller proportion (46%) of the wild fish. The effect of smolt size could be studied only in the smolt groups tagged in 1991. Wild fish were more frequently (71%) caught in the Baltic Main Basin than were reared fish (51%) during their second sea year, and the size variation between wild and reared smolts did not explain the recovery site. No such differences in spatial distribution were found during the third sea year. The tagging place (hatchery/trap) of the reared fish did not affect their later sea migration. The differences in sea migration patterns suggest that the wild salmon are more vulnerable to the intensive salmon fishery in the Baltic Main Basin than are reared fish.     

Microsatellite analysis in domesticated and wild Atlantic salmon (Salmo salar L.): allelic diversity and identification of individuals

Samples of wild and domesticated salmon in Norway were genotyped at 12 microsatellite loci to compare allelic variability and investigate the potential of microsatellite markers for identification of individuals. The following loci were amplified: Ssa20, Ssa62NVH, Ssa71NVH, Ssa90NVH, Ssa103NVH, Ssa105NVH, SsaF43; Ssa20.19; Ssa13.37; SsOSL85; Ssa197; Ssa28. All domesticated strain samples displayed reduced variability compared to wild salmon. On average 58% of the allelic richness observed within the four wild stocks were present in the samples taken from domesticated strains. No systematic differences in heterozygosity were observed between samples representing the two groups.

Pairwise genetic distances, as estimated by Fst values and Nei [1978] was 2–8 times higher among domesticated strains than among wild strains. Among the wild stocks, the highest genetic distances were observed between the river Neiden, located in northern Norway, and the other wild stocks located in the southwest of Norway.

Assignment tests indicated that the wild and domesticated salmon could be distinguished with high precision. Less than 4% of domesticated salmon were misassigned as wild salmon, and less than 3% of wild fish were misassigned as domesticated salmon. Fish from individual domesticated strains were identified with similarly high precision. Assignment to wild salmon stocks was less accurate, with the exception of the sample taken from the river Neiden, for which 93% of the individuals were correctly assigned.     

Lower prevalence of fin erosion in mature than in immature Atlantic salmon (Salmo salar) parr

Samples of hatchery-reared Atlantic salmon parr (age 0+ and 1+) from two different Norwegian river strains were investigated for connections between sexual maturity and prevalence of dorsal fin ray erosion. Mature males showed consistently lower prevalence as well as better recovery than immature males. Among the immature fish, no differences were detected between males and females. The results are discussed relative to morphological and physiological changes in maturing salmon parr.     

Influence of high content of dietary soybean oil on quality of large fresh,smoked and frozen Atlantic salmon (<Emphasis Type="Italic">Salmo salar</Emphasis>)

Atlantic salmon (Salmo salar) was grown in sea cages from 700 g to a market size of 3.2 kg on diets containing either 29% Peruvian high polyunsaturated fatty acids (PUFA) fish oil (FO) or 29% soybean oil (SO) as oil source. Chemical analyses and a triangular consumer test were performed on fresh muscle, while colour, texture and liquid holding capacity (LHC) analyses were performed on both fresh muscle, frozen muscle (stored for 3 months) and smoked salmon. The growth and chemical composition of flesh was not affected by the dietary treatment. The muscle fatty acid (FA) profile was reflected by the dietary oil source, and the amount of malondialdehyde (MDA) was threefold higher in the salmon fed FO than SO. Muscle pigment concentration was significantly different (p < 0.01) with 7.9 mg kg^–1 for FO and 5.6 mg kg^–1 for SO fed salmon, respectively. This result was also significantly (p < 0.05) reflected in the difference between the instrumentally measured colour of fresh, frozen and smoked muscle, and visual impressions of fresh and frozen muscle. Gaping, texture and liquid holding capacity was not affected by the dietary treatment, and the consumer panel did not detect any differences between the dietary groups. SO can be used as a dietary oil source in the grow-out phase of salmon production without sacrificing product quality in terms of texture, liquid holding capacity and consumer preference. However, a total substitution of high PUFA fish oil by SO in diets for salmon grown to market size, affects muscle colour and the FA profile significantly. (p < 0.05).     

Feed intake, growth and feed utilization of offspring from wild and selected Atlantic salmon (Salmo salar)

The aim was to assess whether selection for increased growth rate in Atlantic salmon (Salmo salar) is associated with increased feed intake and/or better feed utilization. Growth responses of offspring from selected and wild lines of Atlantic salmon (initial weight 814±14 g and 533±12 g, respectively) were tested in a 14-week experiment. Selected and wild salmon increased body weight by 79 and 39%, respectively, during the experiment. Relative feed intake (DFI), thermal growth coefficient (TGC) and feed efficiency ratio (FER) were significantly higher in the selected (DFI: 0.67% BW d⁻¹, TGC: 2.96×10⁻³ and FER: 1.16), than in the wild, line (DFI: 0.48% BW d⁻¹, TGC: 1.39×10⁻³ and FER: 0.93). FER was positively correlated with mean growth rate (r=0.90, n=6, P<0.05), and differences between the lines indicated a 4.6% increase in FER per generation of selection. Fish of the selected line had a significantly lower intake of protein and energy per kilogram gain, so the higher growth rate of the selected line was the result of both greater feed consumption and more efficient feed utilization for growth. This implies that selection for increased growth rate in Atlantic salmon may improve both of these traits.     

Identification of differential broodstock contribution affecting genetic variability in hatchery stocks of Atlantic salmon (Salmo salar)

Supportive breeding of Atlantic salmon (Salmo salar) is commonly employed to maintain numbers of fish where the species has become locally endangered. Increasingly, one of the main aims of population management is the preservation of natural genetic diversity. If the stocks employed in supportive breeding exhibit reduced variation they can alter the natural pattern of genetic variation observed in wild populations. In northern Spain, wild adult salmon are caught every year from local rivers and artificially crossed in order to create supportive stocks. The offspring are hatchery reared until the juvenile stage, then released into the same river where their parents were caught. In the current study, our findings demonstrate that although adult broodstock exhibit a pattern of variation similar to the wild populations, variability at microsatellite loci was drastically reduced in the juveniles released into one of three rivers analyzed. The contribution of broodstock to this juvenile stock was examined by pedigree analysis. A restricted number of females contributing to the hatchery stock was identified as the main cause of loss in genetic variation, possibly due to overmaturity of some multi-sea-winter females. We suggest that better monitoring and control of parental contribution will help in solving the problem of loss of genetic diversity in hatchery populations.     

Long-term culture of Atlantic salmon (Salmo salar L.) in submerged cages during winter affects behaviour, growth and condition

In the search for alternative farming methods, we investigated whether large salmon submerged below 10 m in winter conditions behaved normally and performed as well as control fish held in standard surface cages. On average, 2345 salmon of ~ 3.5 kg were kept in each of six 2000 m³ sea-cages for 6 weeks; three of which were submerged to 10–24 m depth and three acted as surface controls (0–14 m). Behaviour during both day and night was studied with echo-sounders, and underwater video cameras fitted with infra-red lamps. A sub-sample of fish from each cage was weighed, measured and assessed for fin and snout condition prior to and after the experimental period. In addition, the vertebral column of 50 fish from the control and submerged treatments were dissected and X-rayed to assess vertebral deformities. The submerged salmon seemed unable to re-fill any gas into the swim bladder, as a linear decrease in echo reflection to < 5% of pre-submergence levels after 22 days of submergence indicated loss of almost all gas from the physostomous swim bladders and negatively buoyant fish. Around day 22, submerged salmon swam at night time with a distinct ‘tail-down, head-up’ tilt (26°) compared to the horizontal swimming position of control fish (− 3°). Average swimming speed (body length per second) of submerged salmon were 1.3–1.4 times faster (day: 0.77 ± 0.02; night: 0.46 ± 0.02, (mean ± SE)) than control fish (day: 0.54 ± 0.01; night: 0.37 ± 0.02) both during day and night. Almost no mortality was seen, and the submerged salmon maintained similar diurnal vertical migrations as the surface fish, indicating that deep submergence did not exhaust the fish. However, submerged fish fed less efficiently, resulting in lower growth and reduced feed utilization. Fins and snouts of the submerged fish had small, but significantly more erosion than the control fish. Vertebrae in the tail region were significantly compressed in the submerged fish compared to control fish. This could be an early symptom of development of vertebral deformities. The results suggest that continuous submergence below 10 m for longer than 2 weeks reduces the welfare and performance of Atlantic salmon.     

Studies of limiting factors governing the waterflow requirement for Atlantic salmon (Salmo salar L.) in landbased seawater systems

In the first part of the investigation two different models are derived to obtain the direct ratio between (a) total ammonia and oxygen and (b) unionized ammonia and oxygen as limiting factors for the waterflow requirement for salmonids in landbased seawater systems. In both models the mass related ammonia quotient, expressed as the ratio between total ammonia excreted (mg NH₄---N per kilogramme fish per minute) and oxygen consumed (mg O₂ per kilogramme fish per minute) is an important factor.

The second part of the study reports measurements of the mass related ammonia quotient (diurnal cycles) of Atlantic salmon kept in a landbased tank farm and calculations of waterflow requirements. The total variation find of the mass related ammonia quotient values were between 0·03 and 0·08. Based on two of the most extreme diurnal cycles with respect to temperature and fish size, the specific waterflow requirement applying ammonia as a limiting factor was in the range 0·04–0·17 litres/kg min.

The third part of the study applies the measurements obtained for the ammonia quotient to the models. The results show that un-ionized ammonia may become the first limiting factor by combinations of a high difference in the dissolved oxygen concentration between the inlet water and the discharge, a high pH and a low threshold level for un-ionized ammonia. However, oxygen is usually found to be the first limiting factor when compared to un-ionized ammonia.

If the value 0·5 mg NH₄---N/litre is used as the limiting criterion for total ammonia, the difference in the dissolved oxygen concentration between the inlet water and the discharge should not exceed about 7 mg/litre. However, in seawater oxygenation is necessary to provide and maintain this difference. Therefore both models result in the same conclusion for moderate levels of oxygen addition.