ReseaRch PaPeR Journal of Agricultural and Marine Sciences 2020, 25(1): 20–26 DOI: 10.24200/jams.vol25iss1pp20-26 Reveived 02 Aug 2019 Accepted 09 Jan 2020 Feed Intake in Atlantic Salmon Fed with or without Surface Spreading of Feed Turid Synnøve Aas1,4*, Trine Ytrestøyl1,4, Torbjørn Åsgård1,4, Kristoffer Rist Skøien2,4, Morten Omholt Alver2,3,4, Jo Arve Alfredsen2,4 Turid Synnøve Aas1,4*( ) synnove.aas@nofima.no, 1Nofima, Sjølsengveien 22, NO-6600 Sunndalsøra, Norway, 2NTNU Department of Engineering Cybernetics, NO-7491 Trondheim, Norway, 3SINTEF Ocean, NO-7465 Trondheim, Norway, 4Centre for Research Based Innovation in Aquaculture Technology (CREATE), SFI, SINTEF Sealab, NO-7645 Trondheim, Norway Introduction T he high growth rate of salmon depends on high feed intake, and feed utilization is highest at high feed intake (Einen et al., 1995; Einen et al., 1999; Grisdale-Helland et al., 2013). Thus, high feed intake is required for efficient production in salmon farming. A fish may only respond to a feed particle when it is within a certain distance from the fish, and to assure high feed intake, feed pellets must be available to each individual fish. In commercial salmon farming, the feed is common- ly spread over a large area of the surface, assuming this has a positive effect on feed intake. In Norwegian salmon farming, the early stages up to smolt of approximately 100 g have traditionally been kept in land based farms, whereas salmon in the grow- out phase up to slaughter size is kept in net pens in the sea. In later years the trend is to keep the fish longer in the land based farms, which implies an upscaling of the tanks which may reach 1000 m3 water volume or more (Gorle et al., 2019). Environmental conditions, such as استهالك سمك السلمون األطليس لالعالف مع أو بدون نرش االعالف عىل السطح توريد سينوف1,4 ،تورين ترستوي1,4، توربرن اسجارد1,4 ،كرستوفر رستو2,4 مورتن اوملت الفر2,3,4 و جو ارف الفردسن2,4 abstRact. In intensive salmon farming, it is common practice to spread the feed over a large surface area, assuming that spreading of the feed increases feed intake in the fish. However, the impact on the feed pellets during spreading results in feed loss due to pellet breakage. In this study, feed intake, growth and signs of aggressive behavior was compared in salmon fed without or with spreading of the feed on the surface area of the tanks. Atlantic salmon (Salmo salar) with initial body weight 0.6 kg were kept in 3.3 m 3 tanks supplied with sea water (salinity 32 %, mean temperature 11 °C) for one month. The salmon were fed one meal daily, either by dropping the feed from one point, or by spreading the feed over the water surface. Feed intake and growth was measured. Fin damage was given a score at termination of the trial as a measure of competitive behavior during feeding. The relative feed intake (i.e. percent of body weight per day) in salmon fed without spreading or with spreading of the feed was 0.63±0.05 and 0.64±0.02 %, respectively. The growth rate was identical in salmon fed without or with spreading of the feed, and no significant difference in variance in final weight was found. No difference in fin damage for salmon fed without or with spreading of the feed was revealed. The data showed that for the conditions used in this trial, spreading of feed had no influence on feed intake or growth of salmon. KeywoRds: Atlantic salmon; Spreading of feed; Feed intake; Feeding behavior. امللخــص: املســتخلص: مــن الشــائع يف االســتزراع املكثــف للســلمون نــرش االعــالف عــىل مســاحة كبــرة مــن الســطح، عــىل افــراض أن انتشــار األعالف يزيــد مــن اســتهالك األســاك لألعــالف. ولكــن هــذا االمــر يؤثــر عــىل حبيبــات األعــالف أثنــاء انتشــارها ويعطــي فقدانــا اكــر لألعــالف بســبب تكســر الحبيبــات. لقــد اجريــت هــذه الدراســة ملقارنــة تنــاول األعــالف والنمــو وعالمــات الســلوك العــدواين يف أســاك ســلمون تــم تغذيتهــا بــدون أو مــع نــرش األعــالف عــىل ســطح الخزانــات. تــم تنفيــذ التجــارب عــىل ســمك الســلمون األطلــيس ) ســلمو ســالر( مــع وزن جســم أويل 0.6 كجــم. تــم حفــظ االســاك يف خزانــات )3.3 م3( مــزودة مبيــاه البحــر ) امللوحــة 32 ٪ ، متوســط درجــة الحــرارة 11 درجــة مئويــة( ملــدة شــهر واحــد. تــم تغذيــة الســلمون وجبــة واحــدة يوميــا، إمــا عــن طريــق إســقاط التغذيــة مــن نقطــة واحــدة كقطــارة، أو عــن طريــق نــرش األعــالف عــىل ســطح املــاء. تــم قيــاس كميــة التغذيــة والنمــو. تــم رصــد درجــة تــرر زعنفــة الســلمون كمقيــاس للســلوك التنافــيس أثنــاء التغذيــة. كانــت كميــة التغذيــة النســبية )النســبة املئويــة مــن وزن الجســم يوميــا( يف ســمك الســلمون بــدون انتشــار أو مــع انتشــار األعــالف 0.63 ± 0.05 و 0.64 ± 0.02٪ ، عــىل التــوايل. وكان معــدل النمــو متطابقــا يف تغذيــة الســلمون بــدون أو مــع انتشــار ااالعــالف، ومل يوجــد هنــاك فــرق كبــر يف التبايــن يف الــوزن النهــايئ للســلمون. باالضافــة اىل ذلــك، مل يوجــد فــرق يف أرضار الزعانــف التــي تتغــذى عــىل الســلمون بــدون أو مــع انتشــار العلــف. أظهــرت نتائــج البيانــات النهائيــة أنــه بالنســبة للظــروف املســتخدمة يف هــذه التجربــة، مل يكــن النتشــار العلــف تأثــرا عــىل تغذبــة او منــو ســمك الســلمون. الكلات املفتاحية: سمك السلمون، األطليس، انتشار األعالف، استهالك العلف، سلوك التغذية 21ReseaRch PaPeR Aas, Ytrestøyl, Åsgård, Skøien, Alver, Alfredsen light, temperature, oxygen levels, salinity and water current affect behavior in salmon in sea cages, and feed- ing induces the changes in the behavior (reviewed by Oppedal et al., 2011). In tanks, the behavior is restrict- ed by the volume of the tank. Volume and design of the tanks have been shown to affect feed intake and growth in salmon (Espmark et al., 2017; Føre et al., 2018). In salmon farming in sea cages as well as in land based farms, the feed is usually transported from a stor- age unit to the cages or tanks with a pneumatic system where the feed is carried by air through a pipe system, and a spreader may be mounted on the outlet of the pipe. Spreading of the pellets can be controlled by ad- justing the air stream in the system (Alver et al., 2016; Oehme et al., 2012). Air stream is a main factor for pellet breakage in the feeding system (Aas et al., 2011a), and moderate spreading of feed is therefore advantageous to avoid losses due to pellet breakage. Feeds with phys- ical pellet quality that is optimal for the feeding systems may not be optimal for the fish (Aas et al., 2017; Aas et al., 2011b; Oehme et al., 2014) and losses due to pellet breakage may be weighed against losses due to subop- timal fish growth. Feed loss also occurs when uneaten feed pellets sink to the bottom of the tank. It is difficult to quantify un- eaten feed, but 7% of the total feed in salmon farming has been suggested (Gjøsæter et al., 2008). Clearly, good feeding routines produce minimal feed spill. Re- duced feed utilization and growth also represent losses. As high feed intake is a prerequisite for high feed utili- zation in salmon (Einen et al., 1995; Einen et al., 1999; Grisdale-Helland et al., 2013), some overfeeding may be necessary to achieve maximum feed intake and feed uti- lization. Optimal feeding routines in salmon farming implies minimal pellet breakage and minimal feed spill while maximal feed intake is assured. At fish farms, control of feeding is assisted by camera systems. Models intending to optimize feeding are also derived (Alver et al., 2004; Alver et al., 2016; Skøien et al., 2016). Spreading of feed across the surface is assumed to increase feed intake in fish by making feed available to all individuals and min- imize hierarchical behavior during feeding. The manner of feed dispersal influences the equality of access of feed among individuals (reviewed by Attia et al., 2012). At re- stricted feeding, the aggression level may be high (Jones et al., 2010), and localized feeding may result in larger growth variation among individuals than when feed is dispersed. Growth and aggression levels related to feed disper- sion is poorly documented. As spreading is correlated to pellet breakage (Aas et al., 2011a), such information is crucial in order to optimize feeding routines. In the present study, the effect of spreading of feed on feed in- take was tested in 3.3 m3 tanks. Atlantic salmon were fed either from one single point over the tank with no spreading on the water surface, or the feed was spread over a large area of the water surface. Feed intake, growth and signs of competitive behavior (fin erosion) was measured. Materials and Methods Fish Trial A tank experiment with Atlantic salmon with mean ini- tial body weight 607 g was run in for 30 days with ap- proximately 100 fish per tank in triplicate at Nofima’s Research Station for Sustainable Aquaculture at Sunn- dalsøra, Norway. The fish were kept in octahedral tanks (quadratic with the corners ‘cut’; 2×2 m surface and vol- ume 3.3 m3) supplied with sea water (salinity 32‰) in a flow through system with mean water temperature 11.0 °C (SD 1.6, range 7.5-13.4 °C, logged every 5 min) and at continuous light. The feed was distributed with Poro EX 04 automatic feeders (Poro AB, Kåge, Sweden). For tanks with spread- ing of the feed on the surface, a Poro EX 06 spreader was connected to the feeder, and the spreading area was as- sumed to cover the whole surface of the tank. For tanks without spreading of the feed on the water surface, the spreader was dismounted, so that the feed fell into the water from the opening of the feeding automat in a small area of a few cm2. Prior to the trial, the tanks were standardized with regard to water flow (80 L·min-1), water velocity and spreading of feed. The water current was measured in all tanks just below the feeder, 40 cm from the tank wall and at 30 cm depth. The overall mean current was 23 cm·s-1 (Table 1). In the three tanks with spreading of the feed, each spreader was manually adjusted so that a largest possible area of the tank was covered, but without losing pellets out of the tank. As pellets met the water surface, they started to sink and followed the circular water current until they settled on the tank bottom. During this movement through the Table 1. Measurement of water current (cm s-1) measured below the feeder, 40 cm from the tank wall and at 30 cm depth. Data are given as mean and SD (n=10) No spreading Spreading Tank number 208 209 214 207 213 215 Water current, cm s-1 22 23 23 24 23 23 SD=4 SD=3 SD=3 SD=6 SD=3 SD=2 22 sQU JoURnal of agRicUltURal and MaRine sciences, 2020, VolUMe 25, issUe 1 Feed Intake in Atlantic Salmon Fed with or without Surface Spreading of Feed water volume, the feed was available for the fish. After some circular movements on the tank bottom, the pel- lets followed the water flow out of the outlet in the cen- ter in the bottom of the tank. The spreading of pellets in time was measured by running the feeding system for 3 s in one tank without and one tank with spreader, with no fish in the tanks. The time (s) from start of feeding to the first and the last pellet reached the bottom of the tank was record- ed (n=10, Table 2). The difference between first and last pellets, which expresses the time the feed was in the wa- ter column, was 12 and 15 s in a tank without and with spreading, respectively. The difference was not signifi- cant with ANOVA (P<0.05). The salmon were fed a commercial feed with pellet size 4.5 mm (Skretting Supreme, Skretting, Stavanger, Norway). The physical properties of the feed are shown in Table 3. The fish was fed one meal daily, lasting from 9 to 10 AM. The feed intake was estimated by collecting and weighing uneaten feed and measuring dry matter. The recovery (%) of uneaten feed was estimated by fol- lowing the same routine as in the trial, but with no fish in the tanks. The recovery value was used to correct the amount of uneaten feed, and daily feed intake was calculated as feed given minus corrected uneaten feed (Helland et al., 1996). All tanks were fed the same amount of feed, and the size of the ration was adjusted daily based on the last three days’ feed intake aiming at 20 % overfeeding in the tanks with highest feed intake. Sampling At start and end of the trial, biomass was recorded and the fish counted. At termination of the trial, the individ- ual weight of 30 fish from each tank was also registered. The fin damage of these 30 fish was evaluated by a scor- ing system where dorsal, caudal, pelvic and pectoral fins were given an integer score from 0 (no visible damage) to 4 (severe damage), and damages were classified as fin erosion, split fin, deformed fin, hemorrhage or asym- metric fins. At handling and weighing, the fish were se- dated with Aqui-S® (clove oil, isoeugenol 2-5 mg L-1). Measurement of Physical Feed Quality Diameter and length of the pellets were measured with an electronic caliper. Bulk density was measured by loosely pouring the feed from a funnel into a 1000 ml measuring cylinder. Sinking velocity was measured in a 1.3 m high cylinder with tight bottom and filled with 34‰ sea water at 10 °C and start and end of 1 m distance marked on the outside. The time the pellets used for sinking 1 m was recorded for one pellet at a time. Calculation Feed intake, given on dry matter (DM) basis, was esti- mated according to Helland et al. (1996). (1) Recovery was estimated by following the experimental feeding routines, but with no fish in the tanks: (2) (3) The relative feed intake (RFI, % of body weight per day) and specific growth rate (SGR, %) and thermal growth coefficient (TGC) are calculated from the following equations as: (4) (5) (6) where, Sum daydegrees = Number of days in trial×Mean temperature (˚C). Statistical Analysis Tank data were analyzed with one-way ANOVA (t-test). Individual data were compared with a hierarchical (nest- ed) ANOVA using the ‘Nested’ procedure in SAS. The score data (individual data) were also analyzed with a nested ANOVA after arcsine transformation of the score data divided by 4 (to obtain data in the range 0-1). ANOVA of original data and transformed data gave cor- responding results. Variance among treatments was an- alyzed by comparing the standard deviations with one- way ANOVA. A significance level of α=0.05 was used for all statistical analyses. Statistical analyses were per- 23ReseaRch PaPeR Aas, Ytrestøyl, Åsgård, Skøien, Alver, Alfredsen formed with the SAS 9.4 computer software (SAS, USA). Results There was no mortality in the trial and the fish appeared to be at good health. There were no significant differ- ences in feed intake or growth. The total feed intake was 132±10 and 127±1 g (dry matter basis) per individual in salmon fed without or with spreading of the feed, re- spectively. The relative feed intake (% of body weight per day) in salmon fed without spreading or with spreading of the feed was 0.63±0.05% and 0.64±0.02%, respective- ly. The growth rate was 0.97% per day for both groups (Table 4). Comparing the body weight of 30 individual fish from each tank did not reveal any effect of spreading the feed on variance in body weight (Table 5). No signifi- cant effect of spreading the feed was found on scoring of fin damage (Table 6). The damage on the dorsal fins was mainly fin erosion and some split fins and deformed fins. The damage on the caudal fins was also mainly classified as fin erosion, but some split fins and red spots were also present. For pectoral fins, split fins were the most com- mon damage, followed by fin erosion and red spots. For pelvic fins, there were very little damage except some split fins. In one tank (tank number 215, with spreading of feed), several fish had red spots in the skin. The reason for this is unknown. In the other tanks, the fish generally appeared normal for salmon of this size kept in tanks. Discussion A tank experiment was chosen to test whether spreading of the feed affects feed intake in Atlantic salmon. Com- pared to experiments in sea cages, a tank experiment has the advantages that spreading of feed, feed intake and growth can be measured with high accuracy, and it can be run at a relatively low cost. Data from small scale studies are not necessarily valid for large scale con- ditions (Espmark et al., 2017). As for all trials, the data from the present study are only representative for the conditions used in this study. There was large difference in spreading pattern on the water surface depending on whether the feed was spread or not, which represents spreading in space. Feed is also spread in time, and the feed is available for the salmon while sinking through the water. Numerically, spreading of the feed on the water surface resulted in a longer time in the water (Table 2), although this was not significantly different (P < 0.05) from the time the pellets were in the water when feeding from one point. There was some variation in these measurements as all pellets follow different routes through the water column. There was a difference among pellets in sinking velocity (Skøien et al., 2016). The 10 replicates did not reveal any significant difference in spreading in time. The overall mean SGR was 0.97% per day, which is in accordance with (Austreng et al., 1987) or just below (Skretting, 2011) expected growth of salmon of this size at this temperature. According to Skretting (2011), At- lantic salmon of 600 and 800 g are expected to grow 1.28 and 1.14 % per day, respectively, at 11 °C. A period of one or two weeks for acclimation to new conditions is nor- mal in salmon trials. This was also observed in the pres- ent trial, where the feed intake was moderate, but grad- ually increasing, during the first ten days. Thereafter, the feed intake was as expected, and the overall growth in the trial was only slightly below expected values. It can therefore be assumed that the feed intake and growth was at normal levels during the last 20 days of the trial. The feed intake was very similar in both treatment groups, showing that in this trial, spreading of the feed on the water surface did not affect the mean feed intake in salmon. Spreading the feed on the water surface is generally believed to reduce variance in feed intake and correspondingly, body weight, in fish (Attia et al., 2012; Ryer and Olla, 1996). This is assumed to be due to that feeding from one point may favor the most dominant individuals whereas when spreading the feed, it is avail- able to all individuals. The variation in individual body weight when salmon were fed without spreading versus with spreading the feed was measured by comparing the standard deviations. An ANOVA of these standard devi- ations did not indicate any effect of spreading of the feed on variation in final body weight. The individual body weight at start was not measured because this require extra handling of the fish, which imply extra stress to the fish, which again may lead to reduced feed intake in the trial. The variation in body weight at start of the trial was assumed to equal in both treatment groups as fish was allocated randomly to the experimental tanks. The salmon was fed full ration in the present trial, whereas at restricted feeding, competitive behavior may result in larger differences in feed intake when the fish is fed Table 2. Time(s) from start of feeding till first and last pellet reached the tank bottom, and the difference, at 3 s feeding in one tank without and one tank with spreading and with no fish in the tank. Data are given as mean and SD (n=10) Without spreading With spreading First pellet 7 SD=1 7 SD=1 Last pellet 19 SD=5 22 SD=4 Difference 12 SD=4 15 SD=5 Table 3. Pellet length and diameter, bulk density and sinking velocity of the feed used in the trial (Mean and SD) Physical properties of the feed Pellet diameter (mm) 4.6 SD=0.3 (n=20) Pellet length (mm) 7.5 SD=1.0 (n=20) Bulk density (g L-1) 650 SD=2.1 (n=3) Sinking velocity (m s-1) 11.2 SD=0.9 (n=20) 24 sQU JoURnal of agRicUltURal and MaRine sciences, 2020, VolUMe 25, issUe 1 Feed Intake in Atlantic Salmon Fed with or without Surface Spreading of Feed without spreading of feed than when the feed is spread (Juell, 1995; Ryer and Olla, 1996). The trial was not designed as a growth trial. To re- veal significant effects of growth in salmon, a doubling in weight during the experiment is often used as a rule of thumb. In the present trial, the growth was 32.5%. There was a slight difference in body weight in the two treatment groups, but the TGC, which is independent of body size, was also identical in salmon fed with or with- out spreading of feed. Since the growth rate was exactly the same in both treatment groups, it can be concluded that spreading of the feed did not affect growth. Scoring of fin damage was used as a measure of hi- erarchical, aggressive behavior. Restricted feeding has been shown to increase fin damage in Atlantic salmon (Noble et al., 2008). With the two feeding patterns used in the present trial, no differences in score of fin dam- age were found. Spatially concentrated feed delivery is believed to increase competition (Juell, 1995; Symons, 1971) but this effect may not be seen due to the over- feeding in the present study. This, together with feed intake and body weight data, indicates that there is no need for using the feed spreader under the conditions used in this trial if feed is sufficiently available. As all handling and spreading of feed increases the risk of pel- let breakage (Aas et al., 2011a), the feed should rather be distributed from one point, without the spreader. The spreading of feed in time in the water volume is related to the sinking velocity of the feed. Furthermore, pellet size will affect spreading of the feed in the water volume, since small pellets will be scattered in the water compared to larger pellets where the feed is concentrat- ed in larger particles. Such factors can be taken into ac- count to adjust spreading of feed, in addition to spread- ing on the surface. The response distance in fish, which is the maximum distance fish responds to a feed particle, depends on several factors, such as fish species, fish size, swimming speed, light conditions, water turbidity, characteristics of the feed and experimental conditions. Response distanc- es in the range 5-25 cm have been measured for large fish (60 cm and larger) of different species under variable experimental conditions (summarized by Richmond et al., 2004). In a fish tank of limited volume and with feed following the movement of the swirl as in the present study, the fish has a high probability of being within this distance of some of the feed particles during a meal. In a large sea cage, even when spreading the feed over a large area, the density of feed particles is considerably smaller in most of the cage volume (Alver et al., 2004; Alver et al., 2016; Skøien et al., 2016), and feeding behavior with high swimming activity is probably necessary for the fish to be able to feed to satiation. In a sea cage with cir- cumference 150 m or more and variable wind and water Table 4. Body weight, growth and feed intake in Atlantic salmon fed without or with spreading of the feed on the water surface for 30 days Data are given as mean ± standard error of mean (SEM; n=3) Without spreading With spreading Initial body weight (g) 618 ± 4 596 ± 17 Final body weight (g) 818 ± 15 789 ± 19 Individual weight gain (g) 201 ± 14 193 ± 6 Weight gain (%) 32.5 ± 2.2 32.5 ± 1.3 SGR (% per day) 0.97 ± 0.06 0.97 ± 0.03 TGC 2.6 ± 0.2 2.6 ± 0.1 Feed intake (g per individual, dry matter) 132 ± 10 127 ± 1 Relative feed intake (% of body weight per day) 0.63 ± 0.05 0.64 ± 0.02 Table 5. Final body weight of 30 randomly selected individuals from each tank of Atlantic salmon fed without or with spreading of the feed on the water surface for 30 days No spreading Spreading Tank number 208 209 214 207 213 215 Mean weight (g) 782 771 851 779 765 806 SD 207 188 197 163 148 210 Maximum (g) 1132 1358 1289 1052 1255 1436 Minimum (g) 464 409 498 412 479 306 25ReseaRch PaPeR Aas, Ytrestøyl, Åsgård, Skøien, Alver, Alfredsen current, the effect of spreading versus no spreading of feed may be different from the data obtained in a tank experiment. Since feed pellets are prone to breaking upon spreading (Aas et al., 2011a) the feed should not be spread needlessly. To optimize feeding routines, it is therefore necessary to measure the effect of spreading the feed in large-scale sea cages also. Conclusion No significant differences in feed intake or growth were found in salmon fed from one point or feed spread over the water surface in 3.3 m3 experimental tanks. Neither was there any effect of spreading of the feed on varia- tion in body weight, or in fin damage. In the present tri- al thus, spreading the feed on the water surface did not improve any of the measured parameters compared to feeding the fish from one point. Acknowledgements The staff at Nofima’s Research Station for Sustainable Aquaculture at Sunndalsøra, and especially Frode Ner- land, is acknowledged for running the fish trial, and Rog- er Selset for technical advice. This study was funded by the Centre for Research-Based Innovation in Aquacul- ture Technology (CREATE) and their partners. References Aas TS, Oehme M, He G, Sørensen M, Lygren I, Åsgård T. (2011a). Analysis of pellet degradation of extruded high energy fish feeds with different physical quali- ties in a pneumatic feeding system. Aquacultural En- gineering 44(1): 25-34. Aas TS, Sixten HJ, Hillestad M, Sveier H, Ytrestøyl T, Hatlen B, Åsgård T. (2017). Measurement of gas- trointestinal passage rate in Atlantic salmon (Salmo salar) fed dry or soaked feed. Aquaculture Reports 8: 49-57. Aas TS, Terjesen BF, Sigholt T, Hillestad M, Holm J, Refstie S, Baeverfjord G, Rørvik K-A, Sørensen M, Oehme M . (2011b). Nutritional responses in rainbow trout (Oncorhynchus mykiss) fed diets with different physical qualities at stable or variable environmental conditions. Aquaculture Nutrition 17: 657-670. Alver MO, Alfredsen JA, Sigholt T. (2004). Dynamic modelling of pellet distribution in Atlantic salmon (Salmo salar L.) cages. Aquacultural Engineering 31(1-2): 51-72. Alver MO, Skøien KR, Føre M, Aas TS, Oehme M, Al- fredsen JA. (2016). Modelling of surface and 3D pellet distribution in Atlantic salmon (Salmo salar L.) cag- es. Aquacultural Engineering 72-73: 20-29. Attia J, Millot S, Di-Poï C, Bégout ML, Noble C, San- chez-Vazquez FJ, Terova G, Saroglia M, Damsgård B. (2012). Demand feeding and welfare in farmed fish. Fish Physiology and Biochemistry 38(1): 107-118. Austreng E, Storebakken T, Åsgård T. (1987). Growth rate estimates for cultured Atlantic salmon and rain- bow trout. Aquaculture 60(2): 157-160. Einen O, Holmefjord I, Åsgård T, Talbot C. (1995). Au- diting nutrient discharges from fish farms: theoret- ical and practical considerations. Aquaculture Re- search 26: 701-713. Einen O, Mørkøre T, Thomassen MS. (1999). Feed ra- tion prior to slaughter - a potential tool for manag- ing product quality of Atlantic salmon (Salmo salar). Aquaculture 178(1-2): 149-169. Espmark ÅM, Kolarevic J, Åsgård T, Terjesen BF. (2017). Tank size and fish management history matters in experimental design. Aquaculture Research 48(6): 2876-2894. Føre M, Alver MO, Alfredsen JA, Senneset G, Espmark Å, Terjesen BF. (2018). Modelling how the physical scale of experimental tanks affects salmon growth performance. Aquaculture 495: 731-737. Gjøsæter J, Otterå H, Slinde E, Nedreaas K, Ervik A. (2008). Effekter af spillfôr på marine organismer. Bergen, In- stitute of Marine Research, Norway. Annual series. No. 1.11 (in Norwegian). Gorle JMR, Terjesen BF, Summerfelt ST. (2019). Hydro- dynamics of Atlantic salmon culture tank: Effect of inlet nozzle angle on the velocity field. Computers and Electronics in Agriculture 158: 79-91. Grisdale-Helland B, Lemme A, Helland SJ. (2013). Thre- onine requirement for maintenance and efficiency of utilization for threonine accretion in Atlantic salm- on smolts determined using increasing ration levels. Aquaculture 372-375: 158-166. Helland SJ, Grisdale-Helland B, Nerland S. (1996). A simple method for the measurement of daily feed in- take of groups of fish in tanks. Aquaculture 139(1-2): 157-163. Table 6. Scoring of fin damage in 30 randomly selected indi- viduals from each tank of Atlantic salmon fed without or with spreading of the feed on the water surface for 30 days. Dam- age of each fin was given a score from 0 (no visible damage) to 4 (severe damage). Data are given as mean±SEM., where SEM is calculated from the pooled standard deviation of data from all individuals (n=90) within each treatment Without spreading With spreading Dorsal fin 2.0 ± 0.1 1.8 ± 0.1 Caudal fin 1.8 ± 0.1 1.7 ± 0.1 Pectoral fins 1.6 ± 0.1 1.5 ± 0.1 Pelvic fins 0.6 ± 0.1 0.5 ± 0.1 26 sQU JoURnal of agRicUltURal and MaRine sciences, 2020, VolUMe 25, issUe 1 Feed Intake in Atlantic Salmon Fed with or without Surface Spreading of Feed Jones HaC, Hansen LA, Noble C, Damsgård B, Broom DM, Pearce GP. (2010). Social network analysis of behavioural interactions influencing fin damage de- velopment in Atlantic salmon (Salmo salar) during feed-restriction. Applied Animal Behaviour Science 127(3): 139-151. Juell JE. 1995. The behaviour of Atlantic salmon in rela- tion to efficient cage-rearing. Reviews in Fish Biology and Fisheries 5(3): 320-335. Noble C, Kadri S, Mitchell DF, Huntingford FA. (2008). Growth, production and fin damage in cage-held 0+ Atlantic salmon pre-smolts (Salmo salar L.) fed ei- ther a) on-demand, or b) to a fixed satiation–restric- tion regime: Data from a commercial farm. Aquacul- ture 275(1): 163-168. Oehme M, Aas TS, Olsen HJ, Sørensen M, Hillestad M, Li Y, Åsgård T. (2014). Effects of dietary moisture content of extruded diets on physical feed quality and nutritional response in Atlantic salmon (Salmo sal- ar). Aquaculture Nutrition 20: 451-465. Oehme M, Aas TS, Sørensen M, Lygren I, Åsgård T. (2012). Feed pellet distribution in a sea cage using pneumat- ic feeding system with rotor spreader. Aquacultural Engineering 51: 44-52. Oppedal F, Dempster T, Stien LH. (2011). Environmen- tal drivers of Atlantic salmon behaviour in sea cages: A review. Aquaculture 311(1-4): 1-18. Richmond HE, Hrabik TR, Mensinger AF. (2004). Light intensity, prey detection and foraging mechanisms of age 0 year yellow perch. Journal of Fish Biology 65(1): 195-205. Ryer CH, Olla BL. (1996). Growth depensation and ag- gression in laboratory reared coho salmon: the effect of food distribution and ration size. Journal of Fish Biology 48(4): 686-694. Skretting. Fiskeføde [Internet]. Stavanger, Norway: Skretting; (2011) Available from: http://www.Fwww. skrettingguidelines.com (In Norwegian) (accessed 12 july 2019). Skøien KR, Aas TS, Alver MO, Romarheim OH, Alfred- sen JA. (2016). Intrinsic settling rate and spatial dif- fusion properties of extruded fish feed pellets. Aqua- cultural Engineering 74: 30-37. Symons PEK. (1971). Behavioural adjustment of popu- lation density to available food by juvenile Atlantic salmon. Journal of Animal Ecology 40(3): 569-587.