Portal de Periódicos da CAPES

Low body fat content prior to declining day length in the autumn significantly increased growth and reduced weight dispersion in farmed Atlantic salmon Salmo salar L.

2018; Wiley; Volume: 49; Issue: 5 Linguagem: Inglês

10.1111/are.13650

1365-2109

Kjell‐Arne Rørvik, Jens‐Erik Dessen, Magnus Åsli, Magny S. Thomassen, Kjellrun G Hoås, Turid Mørkøre,

Aquaculture disease management and microbiota

Aquaculture ResearchVolume 49, Issue 5 p. 1944-1956 ORIGINAL ARTICLEOpen Access Low body fat content prior to declining day length in the autumn significantly increased growth and reduced weight dispersion in farmed Atlantic salmon Salmo salar L. Kjell-Arne Rørvik, Kjell-Arne Rørvik Nofima, Ås, Norway Department of Animal and Aquaculture Sciences, Norwegian University of Life Sciences, Ås, NorwaySearch for more papers by this authorJens-Erik Dessen, Corresponding Author Jens-Erik Dessen jens-erik.dessen@nofima.no orcid.org/0000-0002-0667-4183 Nofima, Ås, Norway Department of Animal and Aquaculture Sciences, Norwegian University of Life Sciences, Ås, Norway Correspondence Jens-Erik Dessen, Nofima, Ås, Norway. Email: jens-erik.dessen@nofima.noSearch for more papers by this authorMagnus Åsli, Magnus Åsli Nofima, Ås, Norway Department of Animal and Aquaculture Sciences, Norwegian University of Life Sciences, Ås, NorwaySearch for more papers by this authorMagny S Thomassen, Magny S Thomassen Department of Animal and Aquaculture Sciences, Norwegian University of Life Sciences, Ås, NorwaySearch for more papers by this authorKjellrun G Hoås, Kjellrun G Hoås Nofima, Ås, NorwaySearch for more papers by this authorTurid Mørkøre, Turid Mørkøre Nofima, Ås, Norway Department of Animal and Aquaculture Sciences, Norwegian University of Life Sciences, Ås, NorwaySearch for more papers by this author Kjell-Arne Rørvik, Kjell-Arne Rørvik Nofima, Ås, Norway Department of Animal and Aquaculture Sciences, Norwegian University of Life Sciences, Ås, NorwaySearch for more papers by this authorJens-Erik Dessen, Corresponding Author Jens-Erik Dessen jens-erik.dessen@nofima.no orcid.org/0000-0002-0667-4183 Nofima, Ås, Norway Department of Animal and Aquaculture Sciences, Norwegian University of Life Sciences, Ås, Norway Correspondence Jens-Erik Dessen, Nofima, Ås, Norway. Email: jens-erik.dessen@nofima.noSearch for more papers by this authorMagnus Åsli, Magnus Åsli Nofima, Ås, Norway Department of Animal and Aquaculture Sciences, Norwegian University of Life Sciences, Ås, NorwaySearch for more papers by this authorMagny S Thomassen, Magny S Thomassen Department of Animal and Aquaculture Sciences, Norwegian University of Life Sciences, Ås, NorwaySearch for more papers by this authorKjellrun G Hoås, Kjellrun G Hoås Nofima, Ås, NorwaySearch for more papers by this authorTurid Mørkøre, Turid Mørkøre Nofima, Ås, Norway Department of Animal and Aquaculture Sciences, Norwegian University of Life Sciences, Ås, NorwaySearch for more papers by this author First published: 05 March 2018 https://doi.org/10.1111/are.13650Citations: 7 The copyright line for this article was changed on 11th October 2018 after original Online publication. AboutSectionsPDF ToolsRequest permissionExport citationAdd to favoritesTrack citation ShareShare Give accessShare full text accessShare full-text accessPlease review our Terms and Conditions of Use and check box below to share full-text version of article.I have read and accept the Wiley Online Library Terms and Conditions of UseShareable LinkUse the link below to share a full-text version of this article with your friends and colleagues. Learn more.Copy URL Share a linkShare onFacebookTwitterLinkedInRedditWechat Abstract Based on the regulatory effects of body fat on appetite and seasonal variations in fat deposition and growth of Atlantic salmon, the present study tested the hypothesis that body fat content prior to declining day length in the autumn can significantly modulate growth rate. The growth rate of salmon (mean initial body weight, BW = 2.3 kg) with different muscle fat content prior to autumn, subjected to natural photoperiod and temperature, during a 7-month period (mean final BW = 6.6 kg) was studied. In August, three fish groups (HF, LF and 0.5LF group) with significantly different muscle fat content (HF = 16.4%, LF = 13.2% and 0.5LF = 11.3%), individually marked with PIT-tag, were mixed into the four net-pens and fed a standard high-energy diet until March the following year. The muscle fat content prior to the autumn had a highly significant (p < .0001) effect on growth during the 7-month main-dietary period, even after identical fat stores among the groups were restored, indicating a more complex explanation than just a lipostatic regulation mechanism. Mean thermal growth coefficients were HF = 2.9, LF = 3.4 and 0.5 LF = 3.9, resulting in increased final weight gain for LF and 0.5LF of 590 g and 980 g, respectively, compared to the HF group. The LF groups obtained a significantly higher homogeneity in BW and shape than HF-fed fish in March, optimizing automatic gutting and filleting at slaughter. The improved growth response among the LF groups by reducing lipid levels can potentially be utilized in closed and semi-closed production units where photoperiod can be manipulated. 1 INTRODUCTION Fish that encounter setbacks induced by nutritional deficit, feed deprivation or sub-optimal conditions often display increased feed consumption (hyperphagia) and compensatory growth (CG) when circumstances are normalized (Ali, Nicieza & Wootton, 2003; Foss & Imsland, 2002; Metcalfe & Monaghan, 2001). The degree of CG in fish varies and is often categorized based on the growth catch-up ability (Ali et al., 2003). Feed restriction or deprivation induces changes in body energy by depleting lipid stores, and during the course of CG and hyperphagia, body weight and lipid reserves are gradually restored (Ali et al., 2003; Bull & Metcalfe, 1997; Jobling & Miglavs, 1993; Metcalfe & Thorpe, 1992). The lipostatic model is often discussed within the circumstances of CG responses in fish (Jobling & Johansen, 1999; Johansen, Ekli, Stangnes & Jobling, 2001). The lipostatic regulation hypothesis identifies adipose tissue and stored lipids to have an important role in governing appetite (Jobling & Johansen, 1999; Keesey & Corbett, 1984; Kennedy, 1953). The model implies that the amounts of stored fat has a negative feedback control on feed intake and is important for the regulation of energy homeostasis. Hence, CG is not only a response to recover body weight, but also a strong response to restore lipid levels and thereof CG will cease once this is achieved (Ali et al., 2003; Jobling & Johansen, 1999; Johansen, Ekli, & Jobling, 2002). Johansen, Ekli and Jobling (2002) showed that altering body lipids of juvenile salmon by dietary administration of low-fat feeds yield similar growth responses as deprivation or feed restriction per se. In modern high-fat diets for salmonids, lipids of marine and vegetable origin are the main sources of energy and support growth efficiently if essential fatty acids requirements are met (Bell et al., 2001; Thomassen & Røsjø, 1989; Torstensen, Lie & Frøyland, 2000). Because salmonids have a high ability to utilize large amount of lipids efficiently for growth, high-fat diets with up to 380 g/kg of fat are commonly used in intensive salmon farming (Torrissen et al., 2011). However, salmonids also have the capacity to store large amounts of excess fat as triacylglycerols mainly in the muscle and visceral cavity (Aursand, Bleivik, Rainuzzo, Leif & Mohr, 1994). Body lipid content of farmed salmonids correlates with fish size, dietary fat level and feed intake (Aksnes, 1995; Hemre & Sandnes, 1999; Torstensen, Lie & Hamre, 2001). Like other anadromous species, Atlantic salmon display seasonal changes in feed intake, growth and lipid deposition during the seawater phase (Mørkøre & Rørvik, 2001). Farmed Atlantic salmon display elevated deposition of lipids in muscle and increased retention of lipids in whole body during declining day length in autumn, with a concomitant increase in feed intake, somatic growth and condition factor (CF) (Alne, Oehme, Thomassen, Terjesen & Rørvik, 2011; Dessen, Weihe, Hatlen, Thomassen & Rørvik, 2017; Mørkøre & Rørvik, 2001; Rørvik et al., 2010). This is particularly pronounced for salmon reared at high latitudes that experience long winters and late spring, which results in reduced lipid levels and CF prior to summer and autumn. The recent increase in automation of fish processing at slaughter requires uniform body weight (BW) and shape among the salmon for optimal efficiency and quality of products such as gutted fish and fillets. Increased uniformity of BW and CF reduces the need for manual gutting/filleting of very small or large individuals. Due to this, the homogeneity in body shape and mass of salmonids are important parameters in salmon farming industry and low dispersion in BW and CF are beneficial at the time of harvest. The homogeneity of BW may be strongly influenced by events occurring during the production cycle, that is, disease outbreaks, handling stress, reduced seawater tolerance or competition of feed (McLoughlin, Nelson, McCormick, Rowley & Bryson, 2002; Ryer & Olla, 1996; Taksdal et al., 2007; Usher, Talbot & Eddy, 1991). The dispersion in the distribution of BW, length and CF are often assessed by calculating the coefficient of variation (CV). The CV of BW for farmed salmon grown from 70 until 300 g and from 60 until 500 g fed either in excess or restrictively for period followed by unrestricted feeding, are reported to vary from 9% to 13% and 16% to 21%, respectively (Johansen et al., 2001). In the latter study, no significant differences were observed in the CV of BW between fish fed in excess and fish fed restrictively. The majority of studies regarding growth responses related to lipid content are based on in-house laboratory experiments with small juvenile salmonids under constant conditions. To our knowledge, few have investigated grow out salmon with different lipid content subjected to seasonal environmental changes in photoperiod and temperature. Due to the regulatory effects of body fat on appetite and the observed fat storage in salmon linked to the seasonal cues, the present study tested the hypothesis that lipid status prior to declining day length in the autumn functions as a significant growth regulator. Accordingly, the growth rate for three groups of salmon with different muscle fat content prior to autumn, subjected to natural photoperiod and temperature, was studied throughout a 7-month period. About each second month, weight samplings and analysis of muscle fat content were conducted to investigate any relationship between fat accumulation and periodic growth rate, and to identify the duration of a potential lipostatic regulatory effect. Changes in visceral fat, CF, length and the dispersion in BW and CF were further assessed. 2 MATERIALS AND METHODS This experiment was conducted in accordance with laws and regulations that control experiments and procedures in live animals in Norway, as overseen by the Norwegian Animal Research Authority. Stunning and sampling of fish were performed in accordance with the Norwegian Animal Welfare act. Fish were treated as production fish up to the point of tissue sampling which was only conducted after the fish were put to death. The experiment was conducted in seawater on the Norwegian west coast (Ekkilsøy, Norway 3°03′N, 7°35′E) at Nofima research centre from August 2011 to March 2012. In July 2010, the fish were transferred to seawater as S1 smolt, at which time the BW was 62 g. From the 10–12 of May 2011, the post-smolt were re-stocked into three net-pens (343 m3) with 650 fish per pen. Prior to this, all individual fish were measured for weight and length, and tagged using passive integrated transmitter tags (PIT-tags) placed in the body cavity just posterior to the gut. The average BW per pen was 1,085 g (SD = 79 g) and each pen received different dietary treatments: a high-fat diet (HF), a low-fat high-protein diet (LF) or half the ration of the low-fat high-protein diet (0.5LF). The 0.5LF group was given half the amount of the feed provided to fish administrated the LF diet the day before. Skretting (Averøy, Norway) produced the feeds and the composition of the HF diet was (wet weight, as is basis): dry matter 93.4%, crude protein 33.5%, crude lipid 34.1%, nitrogen-free extract (NFE) 21.2%, ash 4.6% and gross energy of 25.1 MJ/kg. The composition of the LF diet was (wet weight, as is basis) as follows: dry matter 91.7%, crude protein 49.9%, crude lipid 17.5%, NFE 17.1%, ash 7.2% and gross energy 21.7 MJ/kg. The three dietary treatments were fed from 12 of May until 9 of August (pre-dietary phase). May 12th, the fish were sampled for analysis of initial muscle fat content and biometric data. The analysis showed the following (mean ± SE, n = 30): BW: 1087 ± 97 g, initial muscle fat: 12.2 ± 1.1% and initial CF: 1.10 ± 0.06. After ending the pre-dietary phase, the PIT-tag, BW and length of all individual fish in the three pens were recorded. In addition, fish from each pen were sampled for analysis of muscle and visceral fat content. The pre-dietary feeding phase generated three fish groups with significantly different (p < .05) muscle fat content, visceral fat and visceral mass (Table 1). During the pre-dietary phase, 2.5%, 0.6% and 0.3% fish died in the HF, LF and 0.5LF group, respectively. The majority of mortality occurred from May until mid-June and was not related to any disease outbreak (non-specific morality). Table 1. Biometrics and fat content of Atlantic salmon in August 2011 fed a diet high-fat diet (HF), low-fat high-protein diet (LF) or half ration of the low-fat diet high-protein diet (0.5LF) from May until August 2011, referred to as pre-dietary feeding phase Dietary treatment HF LF 0.5LF Biometric parameters, all fish Number of fish, n 584 584 602 Bodyweight, g 2651 ± 335 2506 ± 287 1865 ± 253 Fork length, cm 59.1 ± 2.3 59.1 ± 2.1 55.8 ± 2.3 CF 1.28 ± 0.09 1.21 ± 0.07 1.07 ± 0.08 Biometric parameters, sampled fish, n = 20 Bodyweight, g 2619 ± 70a 2515 ± 63a 1881 ± 47b Fork length, cm 59.0 ± 0.5a 59.0 ± 0.4a 55.7 ± 0.5b CF 1.22 ± 0.02a 1.18 ± 0.02a 1.03 ± 0.01b VSI 11.3 ± 0.4a 9.6 ± 0.2b 8.5 ± 0.1c Fat content, sampled fish, n = 20 Muscle fat, % 16.4 ± 0.3a 13.1 ± 0.2b 11.3 ± 0.3c Visceral fata, % 39.0 29.0 26.6 CF, condition factor; VSI, Visceral-somatic index. a The analysis of visceral fat content was conducted on pooled samples in August 2011 (n = 1). Values in the same row with different letters are significantly different (p ≤ .05) determined by one-way ANOVA followed by Duncan's multiple range test. Biometric parameters for all fish are presented as means ± SD, whereas biometric parameters and fat content for sampled fish are presented as means ± SEM together with indications of significant differences. At the 10–11 of August, the fish were restocked from the three original pens used in pre-dietary phase into four new pens (125 m3). Each of the four pens contained 50 fish from each of the three pre-dietary treatments (HF, LF and 0.5LF), 150 fish in total (Figure 1). During the period from 11 of August until termination at 20 of March 2012 (main-dietary phase), the pens were fed isonitrogenous and isoenergetic diets produced by Ewos (Bergneset, Norway) (Table 2). The current experiment was an integrated part of a large study were potential effects of dietary oil source were investigated. Therefore, two pens in the main-dietary phase were fed a diet with a marine oil profile (MO), whereas the two other pens were fed a diet with a rapeseed oil profile (RO). The MO diet had an inclusion of 70% South American fish oil and 30% of rapeseed oil. The RO diet had an inclusion of 70% rapeseed oil and 30% South American fish oil. During the main-dietary phase, the pellet size was changed from 7 to 9 mm in December due to the increase in BW of the fish. Figure 1Open in figure viewerPowerPoint Schematic overview of the experimental design during the pre- and the main-dietary phase. The squares during the pre-dietary phase represent net-pens fed different diets; HF, high-fat diet (black filled square); LF, low-fat diet (grey filled square); 0.5LF, half ration of the low-fat diet (white filled square). The large squares in the main-dietary phase represent the net-pens and the squares within the net-pens are the pre-dietary groups Table 2. Chemical compositions (wet weight, as is basis) and fatty acid composition (% of total fatty acids) of the diets used in the main-dietary phase Diet code 7 mm pellet 9 mm pellet MO RO MO RO Chemical composition (wet weight, as is basis) Dry matter, % 93.2 94.0 93.8 93.9 Crude protein (N × 6.25), % 41.2 41.7 34.4 34.6 Crude Lipid, % 31.2 31.4 37.0 35.7 Starch, % 6.2 6.1 6.7 6.8 Ash, % 4.8 4.8 5.1 5.1 NFEa, % 16.0 16.1 17.3 18.5 Crude protein/lipid ratio 1.32 1.33 0.93 0.97 Calculated valuesb Gross energy, MJ/kg 24.8 25.1 25.7 25.5 DP, g/kg 354 359 296 298 DE, MJ/kg 21.4 21.5 22.2 21.9 DP/DE ratio, g/MJ/kg 16.6 16.6 13.3 13.6 Fatty acid composition (% of total fatty acids) C 16:0 12.7 8.5 14.3 9.3 C 18:0 3.2 2.7 3.7 2.9 ∑SFAc 22.6 15.1 24.0 15.9 C 18:1 n-9 26.8 42.1 23.3 42.5 ∑MUFAd 38.1 49.8 36.2 52.8 C 18:2 n-6 8.1 13.9 7.4 13.9 C 18:3 n-3 3.4 6.5 2.9 6.0 C 20:5 n-3 10.1 4.6 11.1 4.0 C 22:5 n-3 1.3 0.6 1.4 0.5 C 22:6 n-3 7.2 3.5 7.5 3.6 ∑PUFAe 34.3 30.4 32.7 29.0 SUM EPA + DHA 17.4 8.1 18.6 7.5 n-6/n-3 ratio 0.4 0.9 0.4 1.0 MO, Marine oil profile; RO, Rapeseed oil profile; N, Nitrogen; NFE, Nitrogen-free extracts; DP, digestible protein; DE, digestible energy; MJ, Mega joule; SFA, Saturated fatty acids; MUFA, monounsaturated fatty acids; PUFA, polyunsaturated fatty acids. a NFE was calculated as = 100 − (protein + lipids + ash + water). b Gross energy, DP and DE were estimated assuming 23.7, 39.5 and 17.2 MJ/kg as the gross energy content of protein, lipids and carbohydrates, respectively. The apparent digestibility coefficients (ADCs) for protein and lipids used were 0.86 and 0.94, respectively (Einen & Roem 1997), whereas 0.50 was used for NFE (Arnesen & Krogdahl 1993). c SFA; C14:0. C15:0, C16:0, C18:0 and 22:0. d MUFA; C16:1n-9, C16:1n-7, C17:1n-7, C18:1n-7, C:18:1n-9, C20:1n-7, C20:1n-9,C20:1n-11, C22:1n-9,C22:1n-11,C24:1n-9. e PUFA; C16:2n-3, C16:3n-4, C18:2n-6, C18:3n-6, C18:3n-3, C18:4n-3, C20:4n-3, C20:2n-6, C20:3n-6, C20:4n-6,C20:5n-3, C22:5-n-3, C22:6n-3. In both periods, feed was administrated using automatic feeders (Betten Maskinstasjon AS, Vågland, Norway) and uneaten feed was collected as described in Einen, Mørkøre, Røra and Thomassen (1999) and corrected for the recovery of dry matter as described by Helland, Grisdale-Helland and Nerland (1996). The fish groups (except the 0.5LF group during the pre-dietary phase) were fed to satiation and the feed ration was set at 5%–10% in excess (ad libitum feeding). The fish were fed four times a day until October 2011; after this, the fish were fed three times a day until termination in March 2012. Adjustments of the feed ration were done according to the daily amount of uneaten feed collected. Due to the stocking of 50 fish from each of the pre-dietary treatments into each net-pen, it was not possible to determine the feed intake or feed utilization of the different pre-dietary groups during the main-dietary phase. The pens were checked for mortalities daily and the dead fish were collected and weighed. The fish were exposed to natural variations in photoperiod and sea temperature during the experiment (Figure 2). Figure 2Open in figure viewerPowerPoint Ambient daily sea temperature (°C, y-axis) and hours of daylight (hours, z-axis) during the pre-dietary phase (May to August 2011) and the main-dietary phase (August 2011 to March 2012). The length of the different periods are indicated by the different grey colours (light grey, pre-dietary phase; dark grey, main-dietary phase) Three samplings were performed during the main-dietary phase; from 9 to 11 October 2011, from 6 to 9 December 2011 and the final sampling and termination of the experiment was conducted from 20 to 22 March 2012. At each sampling, all fish were anaesthetized (MS-222 metacaine 0.1 g/L, Alpharma, Animal Health, Hampshire, UK) and the PIT-tag, fork length and weight of each individual fish were recorded. All fish were starved 2 days prior to the samplings in August and October, and 3 days prior to the samplings in December and March to avoid feed matter in the gastrointestinal system. At each sampling, 10 fish from each pre-dietary group in all the pens were sampled. The sampled fish at each sampling point were selected so that the mean weight corresponded to the mean weight of all the fish in the respective fish group within each pen (as all possible fish were weighted and PIT-tag read). After anesthetization, a blow to the head was used to kill fish sampled for analysis. Then, the gill arches were cut and the fish were bled out in ice seawater. Length and weight of each individual fish sampled for analysis were recorded after bleeding and the fish visually tagged. The fish were then gutted and filleted by hand during the pre-rigor state. Norwegian Quality Cut, NQC (NS9401, 1994) from the left fillet was photographed and the fat content was predicted by digital image analyses (PhotoFish, AKVAgroup, Bryne, Norway), as described by Folkestad et al. (2008). The visceral mass of the sampled fish were pooled on group level, homogenized and frozen at −20°C for later analyses of total lipid content as described by Folch, Lees and Stanley (1957). The proximate composition of crude protein, lipid (acidic-hydrolysis method), starch and moisture of the diets were analysed according to the methods described by Oehme et al. (2010). To determine the fatty acid (FA) composition of the diets, lipids were first extracted according to Folch et al. (1957), and a sample of 2 ml from the chloroform–methanol phase was dried under N2 gas, then the residual lipid extract was trans-methylated overnight with 2′,2′-dimethoxypropane, methanolic HCl and benzene at room temperature according to Mason and Waller (1964). Finally, the methyl esters were separated by gas chromatography and individual FA were identified as described in Røsjø et al. (1994). The growth rates of the fish are presented as the thermal growth coefficient (TGC), and are calculated as described by Iwama and Tautz (1981): , where M0 and M1 are the initial and final BW, respectively, and ΣT is the sum of day degrees during the period (feeding days × average temperature, °C). The mean TGC for the total main-dietary phase was calculated as the weighted arithmetic mean of the periodical TGC to balance these values in relation to their relative contribution to the weight gain. All fish sampled and killed for analysis were starved and bled. The calculation of visceral-somatic index is therefore based on BW with minimal blood content and no feed material in the gastrointestinal system. Visceral-somatic index (VSI) was calculated as follows: Y (g) × BW (g)−1 × 100, where Y is the measured visceral mass. The visceral mass was defined as all mass in the abdominal cavity except liver, heart, kidney and swim bladder. The CF was calculated as follows: 100 × BW (g) × fork length (cm)−3. The dispersions in the distribution of BW, length and CF were assessed by calculating the CV: (standard deviation × mean value−1) × 100. The results were analysed by the General Linear Model (GLM) procedure in the SAS 9.4 computer software (SAS Institute Inc., Cary, NC, USA). Mean results per fish group in each pen were initially subjected to a two-way analysis of variance (ANOVA) to evaluate the effects of muscle fat content due to the pre-dietary phase (0.5LF, LF and HF), main-dietary treatment (oil source; MO diet and RO diet) and their interaction (pre-diet × main diet). As the statistical analysis showed that neither oil source nor the interaction term has significant effects on the traits studied, the data were analysed using pre-dietary treatment as the only experimental factor (one-way ANOVA). Significant differences among experimental groups within treatments were indicated by Duncan's multiple range test. Least-square means (lsmeans) comparison was also used to identify differences among variables within treatments. The Pearson product-moment correlation coefficient was used to describe the association between two variables. Linear regression analysis was conducted using Microsoft excel. The proportion of total variance explained by the model was expressed by R2 and the level of significance was chosen at p ≤ .05. Tendencies were identified at p = .05–.1. The results are presented as means ± SEM, if not otherwise stated. No significant effects of the main-dietary treatment (RO diet and MO diet) or interaction term (main × pre-diet) per se were detected on the traits examined during the main-dietary phase. Thus, only the effects of body fat content due to the pre-dietary treatment are presented in the section "Results." No significant differences in mortality among the pre-dietary groups were observed during the main-dietary phase (24 out of 650 fish, 3.6%). 3 RESULTS The muscle fat content increased by 8.1% for 0.5LF fish, 5.6% for the LF group and 3.6% for HF group from August to October (Figure 4A1). Thus, during an 8-week period of declining day length, the initial significant differences in muscle fat content were equilibrated. TGC was highest for the 0.5LF group, intermediate for the LF group and lowest for the HF group (Figure 5A). The growth rate and the increase in muscle fat content from August to October showed a significant positive linear relationship, and the increase in muscle fat explained 81% of the variation in growth (Figure 3). From August to October, the growth rates were therefore highly affected by the pre-dietary treatment (ANOVA: R2 = 0.97, p < .001). The muscle fat did not differ significantly between the pre-dietary treatments in October or December (Figure 4A1), but pre-diet still significantly influenced the growth rates (ANOVA: p < .05, R2 = .51) and the TGCs were similar, relatively, to the period from August to October (0.5LF > LF > HF) although no significant differences was found between LF and HF group. In the period December to March, the TGC for the 0.5LF and LF group were significantly higher (p < .05) than the HF group (Figure 5A). At the end of the main-dietary phase, the muscle fat content of the LF group was significantly lower (p < .05) than the 0.5LF group, and tended to be lower (p < .1) than the HF group (Figure 4A2). Figure 3Open in figure viewerPowerPoint Regression line between thermal growth coefficients (TGC) and the increase in muscle fat (%) from August to October in Atlantic salmon fed three different pre-dietary treatment from May to August 2011; HF, high-fat diet (black filled squares); LF, low-fat diet (grey filled triangles), 0.5LF, half ration of the low-fat diet (white filled circles). Each point represents average per fish group/experimental unit (n = 12) Figure 4Open in figure viewerPowerPoint Muscle fat content (a1) and body weight (b1) development of Atlantic salmon fed three different pre-dietary treatment from May to August 2011. Values are means ± SEM, n = 4 (n = 1 at termination of the pre-dietary phase). Values not sharing common superscript letters within each sampling period are significantly different (p ≤ .05). a2 and b2 present the final muscle fat and BW of the groups, respectively. The values 11.3%, 13.2% and 16.4% represent the obtained fat content at the beginning of the main-dietary phase (August 2011) for the 0.5LF, LF and HF groups, respectively. ns, not significant. *trend (p < .1) Figure 5Open in figure viewerPowerPoint Thermal growth coefficients (TGC) (a) and weight gain (kg) (b) of Atlantic salmon fed three different pre-dietary treatment from May to August 2011. Values are means ± SEM, n = 4. Values not sharing common superscript letters within each sampling period are significantly different (p ≤ .05). The values 11.3%, 13.2% and 16.4% represent the obtained fat content at the beginning of the main-dietary phase (August 2011) for the 0.5LF, LF and HF groups, respectively The BW of the LF group reached a similar BW as the HF fish in October, whereas the 0.5LF group reached a similar BW as the HF group in December (Figure 4B1). At the end of the trial in March, the LF group (6.87 ± 0.07 kg) had a significantly higher (p < .05) BW than the HF group (6.40 ± 0.16 kg) (Figure 4B2). The 0.5LF group (6.62 ± 0.12 kg) had numerical higher BW than the HF group; however, no statistically significant difference was detected. From August 2011 to March 2012, the 0.5LF group gained 980 g and the LF group gained 590 g more relative to the BW of the HF group (Figure 5B). The overall weighted mean TGC during the main-dietary phase were 3.9 for the 0.5LF group, 3.4 for the LF group and 2.9 for the HF group. Hence, the pre-dietary treatment and consequently the fat status in August 2011 had a clear and significant effect on growth, weight gain and the changes in BW throughout the whole main-dietary phase, with a total duration of 7 months. No significant differences in length between LF and HF group were detected during the trial (Figure 6B1). The strong growth spurt of the 0.5 LF group resulted in no significant differences in length between the 0.5 LF (75.9 ± 0.2 cm) and HF group (76.4 ± 0.8 cm) at the trial termination in March. However, the LF (77.9 ± 0.1 cm) group was significantly longer (p < .05) than the 0.5LF group (Figure 6B2).