10166
J. Agric. Food Chem. 2005, 53, 10166−10178
Tailoring of Atlantic Salmon (Salmo salar L.) Flesh Lipid Composition and Sensory Quality by Replacing Fish Oil with a Vegetable Oil Blend BENTE E. TORSTENSEN,*,† J. GORDON BELL,§ GRETHE ROSENLUND,# R. JAMES HENDERSON,§ INGVILD E. GRAFF,† DOUGLAS R. TOCHER,§ ØYVIND LIE,† AND JOHN R. SARGENT§ National Institute of Nutrition and Seafood Research, N-5817 Bergen, Norway; Institute of Aquaculture, University of Stirling, Stirling FK9 4LA, Scotland, U.K.; and Nutreco Aquaculture Research Centre A/S (ARC), Stavanger, Norway
Atlantic salmon (Salmo salar L.) juveniles were fed either 100% fish oil (FO), 75% vegetable oil (VO), or 100% VO throughout their life cycle to harvest weight followed by a finishing diet period when all groups were fed 100% FO. The two experimental VO diets were tested at two different locations (Scotland and Norway) against the same control diet (100% FO). The VO blend was composed of rapeseed oil, palm oil, and linseed oil using capelin oil as a control for fatty acid class compositions. Flesh fatty acid profiles were measured regularly throughout the experiment, with the times of sampling determined by changes in pellet size/lipid content and fish life stage. Growth and mortality rates were not significantly affected by dietary fatty acid compositions throughout the life cycle, except during the seawater winter period in Norway when both growth and protein utilization were increased in salmon fed 100% VO compared to 100% FO. Flesh fatty acid composition was highly influenced by that of the diet, and after the finishing diet period the weekly intake recommendations of very long chain n-3 polyunsaturated fatty acid (VLCn-3 PUFA) for human health were 80 and 56% satisfied by a 200 g meal of 75% VO and 100% VO flesh, respectively. No effect on flesh astaxanthin levels was observed in relation to changing dietary oil sources. Sensory evaluation showed only minor differences between salmon flesh from the dietary groups, although prior to the finishing diet period, flesh from 100% VO had less rancid and marine characteristics and was preferred over flesh from the other dietary groups by a trained taste panel. After the finishing diet period, the levels of typical vegetable oil fatty acids in flesh were reduced, whereas those of VLCn-3 PUFA increased to levels comparable with a 100% FO fed salmon. No differences in any of the sensory characteristics were observed between dietary groups. By blending VOs to provide balanced levels of dietary fatty acids, up to 100% of the fish oil can be replaced by the VO blend without compromising growth or flesh quality. At the same time, 75% of the dietary fish oil can be replaced without compromising flesh VLCn-3 PUFA content, thereby providing a beneficial nutritional profile for human consumption. KEYWORDS: Fatty acids; lipid level; rapeseed oil; capelin oil; palm oil; linseed oil; sensory analysis; tailoring; flesh quality
INTRODUCTION
Atlantic salmon (Salmo salar L.) flesh has a naturally high content of very long chain n-3 polyunsaturated fatty acids (VLCn-3 PUFA) such as EPA (20:5n-3) and DHA (22:6n3). It has been suggested that these fatty acids are essential for * Address correspondence to this author at the National Institute of Nutrition and Seafood Research (NIFES), P.O. Box 2029 Nordnes, N-5817 Bergen, Norway (telephone +47 55 90 51 45; fax +47 55 90 52 99; e-mail
[email protected]). † National Institute of Nutrition and Seafoood Research. § University of Stirling. # Nutreco Aquaculture Research Centre A/S (ARC).
protecting humans against cardiac diseases (1) and that these VLCn-3 PUFAs cannot be replaced by the shorter and less unsaturated n-3 fatty acid, 18:3n-3 (LNA) (1). Furthermore, salmon flesh lipid is naturally low in n-6 fatty acids and consequently has a high n-3/n-6 ratio, which is recommended for promoting human health (2). As capture fisheries decline, an increasing proportion of human fish consumption is provided by aquaculture (3-5). In 2000 the global production of Atlantic salmon was ∼885 000 metric tonnes, and the production is expected to exceed 1 million metric tonnes by 2005 (6). The aquaculture of salmonids [Atlantic salmon and rainbow trout
10.1021/jf051308i CCC: $30.25 © 2005 American Chemical Society Published on Web 12/03/2005
J. Agric. Food Chem., Vol. 53, No. 26, 2005
Tailoring of Atlantic Salmon Flesh
10167
Table 1. Composition of the Different Basal Diets (According to Pellet Size; Grams per Kilogram) pellet size FWb ingredient fish meal (LT Nordsildmel, Norway) fish meal (Consortio, Peru) corn gluten (Cargill, USA) soybean meal extracted (Denofa, Norway) wheat (Statkorn, Norway) oild vitamins and mineralse
0.3−2.0 672
164 139 25
mma
SWc
2.0 mm
3.0 mm
3.0 mm
4.0 mm
6.0 mm
567
553
147 407 100 54 60 208 25
506 100 100 46 223 25
473 100 100 80 223 25
100
100
155 153 25
166 153 25
9.0 mm 386 100 100 99 291 25
a Includes five different pellet sizes. Here composition of the largest two diet sizes (1.2 and 2.0 mm) is given. b Freshwater stage. c Seawater stage. d Oil is capelin oil (Nordsildmel, Norway) for fish oil based diet or a mixture of 55% rapeseed oil (Oelmuhle, Germany), 30% palm oil (Denofa, Norway), and 15% linseed oil (NV Oliefabriek, Belgium) for vegetable oil based diets. e Vitamin and mineral supplementation is estimated to cover requirements according to NRC, 199.
(Oncorhynchus mykiss)] currently accounts for 60% of the total fish oil used in aquaculture feeds, with 70% of world supply being used for all aquafeeds (6). By 2010, with continuing growth in aquaculture estimates suggest 97% of the global fish oil supply will be used in aquafeeds (6). Thus, the decline in fisheries and the increase in aquaculture will make using fish oil (FO) and fishmeal nonsustainable as feed ingredients in the very near future (7, 8). A sustainable alternative to fish oil is vegetable oils (VO), and these are all devoid of VLCn-3 PUFAs, whereas the levels of 18:2n-6 and monoene fatty acids are usually high, giving a low n-3/n-6 ratio. A number of studies have shown that complete or partial replacement of FO with single VOs, such as rapeseed oil, palm oil, linseed oil, or soy oil, in parts of the salmon growth phase does not affect growth but does affect the fatty acid composition of the edible portion (9-16). Other quality aspects of Atlantic salmon fillet, such as color, liquid holding capacity (17), and moisture levels (14), have been reported to be affected by dietary VO. However, organoleptic properties of salmonid flesh after VO feeding have been contradictory with reports of both significant effects (1820) and no effects (21-23) of dietary VO feeding on sensory parameters. The visual color and carotenoid concentrations of salmon flesh have also been reported to be both affected (17) and unaffected (9, 13) by replacing dietary FO with VO. Although salmon flesh fatty acid compositions have been shown to be closely correlated with dietary fatty acids (9, 12, 14, 15, 24), other metabolic regulators may influence flesh fatty acid composition. This includes fatty acid digestibility (15, 25, 26), preferential catabolism of certain fatty acids (13, 27, 28), and desaturation and elongation of dietary fatty acids (29-32). Thus, if possible, a FO replacement should aim toward a fatty acid composition favoring high digestibility and retention of VLCn-3 PUFA, high levels of fatty acids preferred for catabolism, and high levels of precursor fatty acids for elongation and desaturation to VLCn-3 PUFA. Concomitantly, n-6 fatty acid levels should be kept low to maintain a high n-3/ n-6 ratio and high VLCn-3 PUFA levels beneficial to human health (33). Capelin oil, which is a typical northern hemisphere FO most frequently used in Atlantic salmon aquaculture feeds, generally has high levels of the long-chain monounsaturated and saturated fatty acids and VLCn-3 PUFAs. By using capelin oil as a control, a fatty acid composition formulated by mixing different VOs to provide similar levels of the different fatty acid classes (saturated, monounsaturated, and polyunsaturated n-3 fatty acids) may be better physiologically for salmon health and welfare. Thus, even at 75 or 100% replacement a balanced fatty acid composition should be more beneficial compared to the more extreme fatty acid compositions obtained by replacing FO
with a single VO as demonstrated in previous studies (9-16). The aim of the current study was to investigate the effects of 75 and 100% replacement of FO by a VO blend, based on capelin oil as a control, on Atlantic salmon growth, feed utilization, and final flesh product quality. The dietary trial was conducted through a complete Atlantic salmon life cycle followed by a period with a FO finishing diet and tested at two different locations (Scotland and Norway). MATERIALS AND METHODS Animals and Diets. The effect of replacing FO with VO at two replacement levels (75 and 100%) was investigated in Atlantic salmon in a trial conducted over an entire two-year production cycle. As the trial was both large scale and long term, it was carried out as a collaboration between the Institute of Aquaculture, University of Stirling, Scotland, and the National Institute of Nutrition and Seafood Research, Bergen, Norway, with the 75% replacement diet tested in Scotland and the 100% replacement tested in Norway, with the control FO diet replicated at each site. The diets were fed to triplicate tanks/ cages from April 2002, and the experiments were performed using identical culture conditions other than the obvious environmental differences such as water temperature and day length. In Scotland, the trial was carried out at Marine Harvest Ltd. facilities at Invergarry, Highland (freshwater), and Loch Duich, Highland (seawater); and in Norway, the entire trial was conducted at the Nutreco Aquaculture Research Centre, Lerang Research Station, Stavanger. At each site, Atlantic salmon fry were distributed randomly into six tanks (3 m × 3 m, depth ) 0.5 m) at a stocking level of 3000/tank and weaned onto extruded feeds containing 20% added oil, which was either FO [capelin oil (Norsildmel, Fyllingsdalen, Norway)] or a VO blend, containing rapeseed oil (Oelmuhle, Hamburg, Germany), palm oil (Denofa, Gamle Fredrikstad, Norway), and degummed linseed oil (NV Oliefabriek, Lichterveld, Belgium) in a 3.7:2:1 ratio, replacing 75 or 100% of the FO. Fish were fed the diets described above for around one year until seawater transfer, at which point fish (average weight of ∼50 g in Scotland and ∼120 g in Norway) were transferred into 5 m × 5 m × 5 m net pens at 700 fish/pen. The fish were fed the same diet in seawater as in freshwater, although the dietary oil levels were increased to 25% (3 mm pellet) rising to 32% (9 mm pellets) through the yearlong seawater phase (22 months in Norway and 25 months in Scotland) including the finishing diet phase lasting for an additional 5 months in Norway and 6 months in Scotland, when all dietary groups were fed a 100% FO based diet. The diets aimed to be practical and therefore were formulated according to current practices in the salmon feed industry (Table 1) and were manufactured by Nutreco ARC, Stavanger, Norway. All diets were formulated to satisfy the nutritional requirements of salmonid fish (34). The measured proximate and fatty acid compositions of the diets are given in Table 2. In Norway, the diets were fed to satiation by hand and the exact amounts consumed recorded. In Scotland, the diets were supplied by automatic feeders controlled by an automated feeding sensor system (Arvo-tec, Sterner AquaTech UK, Inverness, Scotland). Mortalities were recorded and dead fish removed daily.
10168
J. Agric. Food Chem., Vol. 53, No. 26, 2005
Torstensen et al.
Table 2. Proximate Compositions (Percentage) and Fatty Acid Compositions (Percentage of Weight of Total Fatty Acids) of Representative Diets Used in Freshwater (3 mm Pellet) and Seawater (9 mm Pellet) freshwater FO
seawater
75% VO
100% VO
FO
75% VO
100% VO
42.1 ± 0.2 31.4 ± 0.8 7.0 ± 0.3 7.0 ± 0.1
41.2 ± 0.4 32.8 ± 0.3 6.4 ± 0.2 7.1 ± 0.0
42.9 ± 0.3 31.3 ± 0.1 6.5 ± 0.1 6.6 ± 0.1
6.2 ± 0.1 14.5 ± 0.4 2.4 ± 0.6
2.2 ± 0.2 16.1 ± 0.2 3.0 ± 0.5
0.6 ± 0.1 15.3 ± 0.3 2.7 ± 0.0
protein lipid moisture ash
46.5 ± 0.4 19.6 ± 0.1 8.4 ± 0.0 7.8 ± 0.0
45.8 ± 0.3 18.2 ± 0.2 6.8 ± 0.1 7.9 ± 0.0
Proximate Composition 46.7 ± 0.2 18.8 ± 0.1 6.6 ± 0.0 7.5 ± 0.0
14:0 16:0 18:0
6.1 ± 0.1 12.4 ± 0.2 1.5 ± 0.0
2.7 ± 0.1 15.5 ± 0.5 2.4 ± 0.1
Fatty Acid Composition 1.1 ± 0.1 16.9 ± 0.4 2.7 ± 0.0
total saturateda
20.3 ± 0.3
20.8 ± 0.6
21.8 ± 0.4
23.6 ± 0.9
21.5 ± 0.5
19.4 ± 0.3
16:1n−7b 18:1n−9 18:1n−7 20:1n−9c 22:1n−11d 24:1n−9
7.9 ± 0.1 11.9 ± 0.4 3.3 ± 0.1 19.9 ± 0.4 15.8 ± 0.3 0.7 ± 0.0
3.2 ± 0.0 30.6 ± 0.7 2.6 ± 0.1 7.4 ± 0.1 6.4 ± 0.1 0.4 ± 0.0
1.1 ± 0.0 40.4 ± 0.5 2.5 ± 0.1 2.7 ± 0.1 2.3 ± 0.1 0.3 ± 0.0
4.9 ± 0.2 13.2 ± 0.4 2.4 ± 0.1 11.1 ± 1.0 16.5 ± 1.9 0.7 ± 0.0
1.8 ± 0.3 35.2 ± 0.0 2.5 ± 0.2 3.8 ± 0.3 5.1 ± 0.4 0.2 ± 0.0
0.5 ± 0.1 43.0 ± 0.2 2.4 ± 0.0 1.3 ± 0.1 0.8 ± 0.0 0.0 ± 0.0
total monoenes
59.4 ± 1.3
50.6 ± 0.8
50.2 ± 0.5
48.8 ± 2.5
48.6 ± 0.6
48.2 ± 0.5
18:2n−6 20:4n−6
3.9 ± 0.1 0.2 ± 0.0
11.7 ± 0.3 0.2 ± 0.0
13.5 ± 0.2 0.2 ± 0.0
3.6 ± 0.6 0.5 ± 0.1
12.7 ± 1.2 0.2 ± 0.1
17.1 ± 0.2 0.0 ± 0.0
total n−6 PUFAe
4.4 ± 0.1
12.2 ± 0.3
13.7 ± 0.2
4.6 ± 0.7
13.0 ± 1.1
17.1 ± 0.2
18:3n−3 18:4n−3 20:4n−3 20:5n−3 22:5n−3 22:6n−3
0.6 ± 0.0 1.9 ± 0.1 0.3 ± 0.0 5.8 ± 0.6 0.4 ± 0.1 5.9 ± 0.6
6.8 ± 0.3 1.0 ± 0.1 0.2 ± 0.0 3.6 ± 0.3 0.2 ± 0.0 4.2 ± 0.4
8.0 ± 0.2 0.4 ± 0.1 0.1 ± 0.1 2.1 ± 0.0 0.2 ± 0.0 3.4 ± 0.2
1.2 ± 0.1 2.5 ± 0.1 0.7 ± 0.0 6.5 ± 0.2 0.9 ± 0.2 10.0 ± 0.6
9.0 ± 0.7 0.8 ± 0.0 0.2 ± 0.0 2.4 ± 0.5 0.3 ± 0.2 3.7 ± 1.1
13.4 ± 0.5 0.2 ± 0.0 0.0 ± 0.0 0.6 ± 0.0 0.0 ± 0.0 1.0 ± 0.0
total n−3 PUFAf
14.9 ± 1.4
15.9 ± 1.1
14.3 ± 0.6
21.8 ± 0.8
16.5 ± 1.1
15.2 ± 0.5
total PUFAg
20.3 ± 1.6
28.6 ± 1.3
28.0 ± 0.8
27.6 ± 1.6
29.8 ± 0.1
32.3 ± 0.7
Sampling. Samples were taken from all diets and stored at -20 °C. Fish were not fed during the 24 h prior to sampling. In Norway 10 fish were sampled randomly from each tank and anesthetized with MS222 (7 g/L, Norsk Medisinaldepot, Oslo, Norway). In Scotland, four fish were sampled per tank/pen until the final and wash out points, at which time 6 fish per pen were sampled. Samplings were done when pellet size and hence dietary lipid level changed and during smoltification, which was, respectively, 6, 9, 14, 16, and 22 months after start of feeding in Norway and 6, 9, 12, 14, 17, 21, and 25 months after start of feeding in Scotland. For total lipid, fatty acid composition, and astaxanthin analysis prior to the finishing diet period, the Norwegian quality cuts (NQC) from salmon sampled from each tank were pooled, homogenized, and frozen on dry ice. During the finishing diet period pooled samples of NQC from 3 fish per net pen (6 fish/pen in Scotland, pooled as two groups of 3) were sampled for analysis. All samples were stored at -80 °C. Proximate Composition of Diets. Dry matter in the diets was measured gravimetrically after freeze-drying of homogenized samples for 48 h. Total nitrogen was determined on homogenized, freeze-dried samples using a nitrogen determinator (LECO, FP-428 system 601700-500; Perkin-Elmer Corp., Norwalk, CT). Protein was calculated as N × 6.25. Total lipid of the diets and fillet was measured gravimetrically as described by Lie and co-workers (35) after ethyl acetate extraction and after acid hydrolysis of the diet samples. Lipid Extraction and Fatty Acid Analysis. Total lipid was extracted from salmon flesh NQC by homogenization in chloroform/methanol (2:1, v/v) 19:0 methyl ester as internal standard (Norway), basically according to the method of Folch et al. (36). Fatty acid methyl esters (FAME) were prepared from total lipid either by acid-catalyzed transesterification, with FAME extracted and purified as described previously (37) (Scotland), or by boron trifluoride following saponification essentially as described by Lie et al. (38) and Torstensen et al. (28) (Norway). In Scotland FAME were separated and quantified by gas-liquid chromatography with on-column injection using a Thermo Finnegan Trace 2000 GC (Thermoquest, Hemel Hempstead, U.K.)
equipped with a fused silica capillary column (ZB wax; 30 m × 0.32 mm i.d.; Phenomenex, Macclesfield, U.K.) with helium as carrier gas; temperature programming was from 50 to 150 °C at 40 °C/min and then to 195 °C at 1.5 °C/min and finally to 220 °C for 2 min. In Norway, a Thermo Finnegan Trace 2000 GC equipped with a fused silica capillary column was used (CP-sil 88; 50 m × 0.32 mm i.d.; Chrompak Ltd.) with temperature programming of 60 °C for 1 min, 160 °C for 28 min, 190 °C for 17 min, and finally 220 °C for 10 min with all intervening temperature ramps being at 25 °C/min. Individual methyl esters were identified by comparison to known standards and by reference to published data (39). Data were collected and processed using Chromcard for Windows (version 1.19) computer package (Thermoquest Italia S.p.A., Milan, Italy) (Scotland) or using Totalchrom software (ver. 6.2, Perkin-Elmer) (Norway). Astaxanthin Analysis. Norway. all-trans- and 9,13-cis-astaxanthin were measured in fish muscle using a procedure modified from that of Bligh and Dyer (40) by Bjerkeng et al. (41). Briefly, homogenized fish muscle was extracted using a water/methanol/chloroform (1:1:3) solution. An aliquot of the astaxanthin-containing chloroform phase was evaporated under N2 and redissolved in hexane. Quantitative analysis of astaxanthin from flesh was performed using an HPLC system consisting of a Spectra-Physics 8810 LC pump, a Gilson 234 Autoinjector, a Spectra-Physics 770 UV-vis detector coupled with a SpectraPhysics Wavelength drive SFA 339, a Supelcosil 5 µm LC-CN column (25 cm × 4.6 mm), and Xtra Chrom software (HP/Nelson analytical) and eluted isocratically using n-heptane/isopropyl acetate/acetone/ ethanol (84.5:10:4:1.5) at 2.75 mL min-1. Quantification of astaxanthin was performed using an external standard method by which standard solutions prepared from crystalline all-trans-astaxanthin (Hoffman-La Roche, Oslo, Norway) were measured with a spectrophotometer (UV260, Shimadzu, Kyoto, Japan) using a molar absorptivity in hexane of E1cm 1% ) 2100 at λmax ) 472 nm for astaxanthin. The retention time for all-trans-astaxanthin was 2.9 min, and the detection wavelength was 470 nm.
J. Agric. Food Chem., Vol. 53, No. 26, 2005
Tailoring of Atlantic Salmon Flesh Scotland. Astaxanthin was extracted from salmon muscle largely by following the method of Barua et al. (27). Tissue samples were homogenized in 5 mL of absolute ethanol and 5 mL of ethyl acetate using an Ultra-Turrax tissue disrupter. The homogenate was centrifuged (1000g, 5 min) and the supernatant removed to a stoppered glass tube. The pellet was rehomogenized in 5 mL of ethyl acetate and recentrifuged, and the supernatant was combined with the first supernatant. Finally, the pellet was rehomogenized in 10 mL of hexane and recentrifuged, and the supernatant was combined with the pooled supernatant. The pooled supernatant was dried under N2 and vacuum desiccated for 2 h before the residue was redissolved in 2 mL of hexane containing 0.2% (w/v) BHT. Measurement of astaxanthin (Ax) was carried out using a 5 µm Luna ODS2 column (4.6 × 150 mm, Phenomenex, Macclesfield, U.K.). The chromatographic system was equipped with a Waters model 501 pump, and astaxanthin was detected at 470 nm using a Waters 490E multiwavelength UV-vis detector [Millipore (U.K.), Watford, U.K.]. An isocratic solvent system was used containing ethyl acetate/methanol/water (20:72:8 v/v/v) at a flow rate of 1 mL/min. Ax was detected at 470 nm and quantified using an external standard of Ax obtained from Roche (Heanor, U.K.). Sensory AnalysissObjective and Subjective Tests Using Taste Panels. Fish was sampled for sensory analysis in January and June 2004 (Norway), that is, before and after the finishing diet period. Furthermore, fish from both Scotland and Norway were sampled in February 2004 (after 23 months of feeding; before the finishing diet period) for sensory analysis performed in Scotland. For the sensory analysis in Scotland fresh-cooked fillets were assessed from the two locations, and smoked salmon was assessed for 100% FO (S) and 75% VO. Two fish from each dietary group were chilled in ice slurry and killed by a blow to the head and cutting the gills. Gutted, chilled salmon were then shipped on ice to the Food Industry Forum (FIF), Queen Margaret University College, Edinburgh. A trained panel performed quantitative descriptive analysis (QDA) methods to profile each salmon in terms of sensory attributes related to smell, appearance, texture, and flavor. For each of the 12 sensory attributes of this QDA test, a twoanchored (e.g., soft and firm) linear scale (0-10) was used, in which the score of 5 is the midpoint. This scale is objective and has nothing to do with the likes or dislikes of a given panelist. The trained panel was also asked to be subjective and score their degree of like or dislike (preference) for each attribute of each sample (two fish from each dietary group from both Norway and Scotland; see Table 6) on another linear scale (0-10). On this scale, scores have the following meanings: 0 ) extremely disliked, 10 ) extremely liked; 5 ) midpoint (which indicates that an attribute was neither liked nor disliked). For example, all scores of >5 are on the liked side of the preference scale. The trained panelists have been taught to be more discerning than the average consumer. In Norway three salmon from each group of 100% FO and 100% VO (nine in total from each dietary treatment) were chilled in an ice slurry and killed by cutting the gills. Gutted, chilled salmon were then shipped on ice to Matforsk, Ås, Norway, and evaluated on the following day for QDA, ISO 6564:1985 E, by testing 23 sensory attributes with 11 judges. On arrival, the salmon was filleted, and six fillet samples with a thickness of 2.5 cm from each fish fillet were put in coded diffusion-tight plastic bags and vacuum sealed. Prior to the sensory evaluation the fillet samples were heated in a water bath at 75 °C for 30 min. The samples were served to the sensory panel in steel containers on a heated plate at 65 °C. The cooked fillet samples were served to the sensory panel in a randomized manner regarding dietary group of salmon, judge, and replicates. Six samples were served in each session. For each sensory parameter the grading is from 1.0 ) no intensity to 9.0 ) distinct intensity. Statistics. The sensory attributes of salmon flesh were analyzed using Sirius for Windows (version 6.5). The purpose of principal component analysis (PCA) is to express the main information in the variables by a lower number of variables, the so-called principal components (PC1, PC2, ...). A high positive or negative loading reveals a significant variable in the actual PCA model. Score plots from the PCA explore the main trends in the data, and their respective loading reveals variables with a significant loading. Samples with similar sensory attributes are located in the same area in the score plot. These classes are indicated
10169
in Figure 5 by circles drawn freehand. Because samples with the same sensory attributes will be located on top of each other, to ease interpretation of the samples the classes contained are written beside the circle. Differences between the dietary treatments through the feeding experiment and after the finishing diet period were analyzed by Breakdown and one-way ANOVA followed by Tukey’s HSD test, using CSS:Statistica (version 6.1; StatSoft Inc.). The significance level was set to P e 0.05, and data are presented as mean ( standard deviation (SD) (n ) 3 or otherwise stated). Calculations. Specific growth rate (SGR%) ) {exp[ln W2 - ln W1)/ (T2 - T1)] - 1} × 100, where W1 and W2 are initial and final body mass in g and T2 - T1 is the sum of experimental days. Feed conversion ratio (FCR) ) g of dry feed eaten/g of live weight gain. Thermal growth coefficient (TGC) ) [(W20.333 - W10.333)/(∑ °C)] × 1000, where W1 and W2 are initial and final body mass in g and ∑ °C is the sum day - degrees in the experiment (43). Protein productive value (PPV) ) g of retained protein (g of protein intake)-1. RESULTS
In freshwater, the control diet (FO), formulated with 100% FO, contained ∼20% of the total fatty acids as saturated fatty acids, mainly 16:0 and 14:0, and almost 60% as monounsaturated fatty acids, over half of which were the long-chain monounsaturated fatty acids 20:1 and 22:1. Polyunsaturated fatty acids of the n-6 and n-3 series accounted for 4.4 and 15%, respectively, of the total diet fatty acids. 18:2n-6 was the predominant n-6 fatty acid, and within the n-3 fatty acids 20: 5n-3 and 22:6n-3 were present in approximately equal amounts and
