sign in sign up
<<
Home , Arctic char, Brown trout, Salvelinus

Limnologica 31 (2001) 281-288 LIMNOLOGICA http://www.urbanfischer.de/journals/limno © by Urban & FischerVerlag

Department of Biology, University of , ReykjavN, Iceland

Diel Changes in the Feeding Behaviour of Char ( alpinus) and Brown ( trutta) in Ellidavatn, a Small in Southwest Iceland

BJORN BJORNSSON

With 4 Figures and 2 Tables

Key words: Catch per effort, food intake, diel activity, foraging behaviour, stomach content

Abstract

In the lake Ellidavatn, southwest Iceland, the diel activity of Arctic of the year, (Salmo trutta L.), char and brown trout was studied in autumn (September) and spring [Salvelinus alpinus (L.)] and Atlantic (Salmo salar (April) by removing the from gillnets every 3 hours for three L.) parr may be more active at night (CnASTON 1968; LrNNgR days. Additional fish samples, evening and morning, were taken at et al. 1990; FRASER et al. 1995; VALDIMARSSONet al. 1997), other times of the year. The catch per unit effort was about 10 times when they take most of their food (JORGENSEN & JOBLING higher during the nighttime than daytime for both char and trout. The 1989; ALANARA& BR3,NNAS 1997). Field studies have shown median weight of char was lowest near the middle of night (80 g) and nocturnal activity of brown trout and brook char highest near the middle of day (140 g). The median weight of trout [Salvelinus was also lowest near the middle of night (110 g) and highest near the fontinaIis (M~TCnlLL)] (HEGGENES et al. 1993; BOURKE et al. middle of day (330 g). The main diet of char consisted of the clado- 1996). In a deep Norwegian lake in July-September, pelagic ceran Eurycercus lamellatus in September and chironomid larvae in Arctic char fed almost exclusively on both day April and the main diet of trout was always sticklebacks. In autumn and night while brown trout fed on zooplankton during the and spring, the feeding was nocturnal in the char but crepuscular in day and on surface insects and chironomid pupae during the the trout. The median stomach content of char was highest at dawn night (DERvO et al. 1991). (0.52% of body weight) and lowest at dusk (0.14% of body weight). In Ellidavatn, Arctic char fed throughout the year on small It is proposed that nocturnal char shift through the bottom substrate benthic , while brown trout fed on sticklebacks in search for prey. The median stomach content of trout was highest (BJORNSSON 2001). In an attempt to shed some light on the at 09.00-18.00 and 24.00-03.00 hours (0.13-0.18% of body weight) and lowest at 03.00-06.00 and 18.00-21.00 hours (0.02-0.07% of mechanism involved in the resource partitioning and niche body weight). Different feeding times may help to secure mutual co- shift of char and trout in Ellidavatn their diel feeding be- existence of char and trout in a small lake. In autumn and spring, the haviour were studied by analysing stomach contents from average daily consumption of char was about 0.8% of body weight, fish captured in gillnets throughout the 24-h period. Growth less than half their maximum food intake. rate of adult char in Ellidavatn appeared to be food limited (BJORNSSON 2001). The clear-cut diel changes in the stomach content of char in Ellidavatn made it possible to estimate Introduction their daily food intake and thus independently test whether these fish were food limited. Salmonid are generally considered to be visual feeders This paper is based on a B.Sc. honours program carried (NmSSON 1963; WARE 1972; ERIKSSON 1973; THORPE et al. out at the University of Iceland in the years 1974-76. The 1988; JORGENSEN 8,: JOBLING 1990; MALMQUIST 1992a, Municipality of Reykjavfk recently made plans to develop 1992b; MALMQUISTet al. 1992) and adaptations for high visu- the area around Ellidavatn and needed background informa- al acuity at daytime light intensities are generally incompati- tion about the ecology of the lake. This was seen as an oppor- ble with sensitive night vision (FRASER & METCALFE 1997). tunity to rework all the data for this study site and to publish Experimental studies indicate, however, that at certain times the most interesting material.

0075-9511/01/31/04-281 $ 15.00/0 281 Study site was used to compare means and a Mann-Whitney U-test to compare medians (CAMPBELL 1967). A Bezier curve in the computer program Corel Draw 8 was selected to describe the continuous diel changes in Ellidavatn is a small and shallow lake close to Reykjavl"k, stomach content of char (two nodes) and trout (four nodes). Iceland, 64°06 N and 21°47 W and 75 m above sea level. The lake has a surface area of 2 km 2 and is 1-3 m deep. Approxi- mately 2/3 of the bottom is soft and muddy and 1/3 turf with scattered rocks. The shore is rocky with a strip of hard bottom Results surface a few meters wide. There are two tributaries, H61ms~ and Sudurfi, and one outlet , Ellidafi. The average renew- In 19-22 September 1975 the nighttime was 10 hours with al time of the lake was about one week. The water tempera- full moon and average cloud cover 72, 22 and 100% during ture was 0 °C from November to February and 8-12 °C from the three nights, respectively. In 26-28 April 1976 the night- June to August. time was 51/2 hours with new moon and 100% cloud cover About half of the bottom was covered with macrophytes, during all three nights (Icelandic Weather Bureau, data bank). mostly Myriophyllum alterniflorum Cand., Potamogeton During all nights of the study except one, the mean wind- spp., Nitella sp., Eleocharis palustris (L.) and Littorella uni- speed was low (3-5 m s~l). During the third sampling night in flora (L.). On mud and turf bottom, oligochaetes and chi- September there was a strong breeze (12 m s-I) sufficient to ronomid larvae were the dominant taxa followed by bivalves increase the turbidity in the lake. There was no snow cover on Pisidium spp., trichopteran larvae, water snails Lymtzaea the ground to increase light intensity in either study. In the peregra (MOLL.) and leeches Hirudinea. On hard bottom, L. September and the April study the mean temperature was peregra, chironomid larvae and trichopteran larvae were the 5.2 °C and 6.5 °C, respectively; the mean temperature change dominant taxa. During late summer and autumn, the clado- within the 24-h period being 1.5 °C and 0.9 °C in September ceran Eurycercus lamellatus MOLL. was numerous in the and April, respectively. lake. The dominant fish in the lake were Arctic char, In September 1975 and in April 1976 the catch per unit ef- brown trout and threespine stickleback (Gasterosteus aculea- fort of char and trout in Ellidavatn was approximately 10 tus L.). Eel [AnguiIla anguilla (L.)] and also fold in the nighttime compared to the daytime (Fig. 1). In all occupy the lake. For a more detailed description of the study cases the differences in catch between night and day were site see BJORNSSON (2001). statistically different (t-test assuming unequal variances, n -- 6 to 14, P < 0.05). Usually, 2-9 char and 0.5-2.0 trout were caught per net per hour at night, with a maximum catch near Methods sunset, but only 0.2-1.0 char and 0.0-0.2 trout in the day (Fig. 1). In the combined fish catch from autumn (September and The sampling site was at the east shore of the lake near the farmhouse Ellidavatn (BJORNSSON 2001). Fish samples were obtained by two or October) and spring (April and May) the median weight of three monofilament bottom gillnets, 24 m long and 1.15 m deep, with the char and trout changed gradually during the 24 hours, a mesh size of 22 mm (knot-to-knot). The nets were tied together and reaching a minimum of 81 and 113 g near the middle of the set from a small boat at a right angle to the shore, starting 3 m from night and a maximum of 141 and 328 g near the middle of the the shoreline. Water temperature was measured at the sampling site day for char and trout, respectively (Fig. 2). The diel changes before and after deployment of the nets. Measurements of cloud were highly significant for the char, and in 5 out of 8 cases cover at the weather station in Reykjavfk 8 km away from the sam- there was a significant difference between adjacent 3-h peri- pling site in Ellidavam were used to estimate the light conditions dur- ods (Mann-Whitney, n = 15 to 138, P < 0.05). The data for ing nighttime. Nighttime was defined as the dark period from the end trout were much more limited with no significant differences of twilight until the beginning of twilight. In two periods, 19-22 September 1975 and 26-28 April 1976 the between the individual periods (Mann-Whitney, n = 13 to 66, nets were deployed for nearly 72 hours and the entangled fish re- P > 0.05). moved every 3 hours. Additional fish samples were taken at different The stomach content of the char was significantly higher times of the year. Those were paired samples, one in the evening and in the morning (03.00-09.00) than in the evening one in the morning, deploying the nets for 1-3 hours. The fish were (18.00-24.00 hours) (Table 1) (t-test, P < 0.001 for 20 measured and dissected within two hours from sampling. The maxi- September, 21 September, 27 April). The mean stomach con- mum length (cm) and ungutted weight (g) were recorded and the tent in the morning was 0.53% in September and 0.49% in stomach content emptied into glass vials with a mixture of iso- April, not a significant difference (P > 0.05). The mean stom- propanol (70%), water (20%) and glycerol (10%). A magnifying ach content in the evening was 0.17% in September and lamp (x2) or stereoscope (xl0) was used to identify the content to 0.23% in April (P < 0.05). There was not a significant differ- easily identifiable groups of prey and indigestible plant remains and gravel. Each group was drained briefly on tissue paper and weighed ence between adjacent mornings (20-21 September: P>0.05) with an accuracy of 0.01 g. and usually not between adjacent evenings (19-20 Septem- Both means and medians with 95% CL were used to study diel ber, 20-21 September, 21-22 September, 26-27 April: P > changes in stomach content expressed as % of fish weight; a t-test 0.05; 27-28 April: P < 0.05).

282 Limnologica 31 (2001) 4 DAY DAY I~ ~11~[~|iII

19-21 September 1975 26-28 April 1976 '~"1 9 O x: 8

t- 7 -~ 6 v 5 Char Char n- 4 O IJ_ __.IJJ 2 z 1 --'1 re 0 U.l l~ii,i,,~ I///i/itili~ ...... iiililiii a_ 1 iiii'ii~iilil/ii~ [,[:L;I=I;L;" -i- Trout o 2 Trout i,!!i!!!!!~} Fig. 1. Diel changes in the catch per unit effort of char and trout in Ellidavatn in September 1975 and 3 6 9 1) 1'5 18 2'1 24 3 3 ; ; 12 1; 1; 2'1 24 3 April 1976. TIME OF DAY (hours) TIME OF DAY (hours)

N= 27 27 13 30 66 33 43 63 24 17 15 107 138 58

500

32 co 400

U_ Trout O 300 I- "l- Char iTi 200,

Fig. 2. Diel changes in the weight of char and trout in autumn and 100 spring in Ellidavatn 1975-76 (me- dian with 95% CL). The number of fish is given at the top of the 3 6 9 1'2 1'5 1'8 2'1 24 3 3 6 9 12 1'5 1'8 21 24 3 graphs. TIME OF DAY (hours) TIME OF DAY (hours)

Table 1. The stomach content of char (% of fish weight) in Ellidavatn in the morning (03.00-09.00) and evening (18.00-24.00 hours) in 19-22 September 1975 and 26-28 April 1976, number of stomachs (n), mean and 95% CL.

Morning (03.00-09.00) Evening (18.00-24.00)

Date n mean 95% CL n mean 95% CL

19 September 1975 29 0.179 0.052 20 September 1975 31 0.501 0.088 43 0.188 0.060 21 September 1975 26 0.564 0.094 29 0.148 0.055 22 September 1975 2 0.841 20 0.168 0.068 26 April 1976 28 0.294 0.092 27 April 1976 18 0.491 0.107 32 0.246 0.105 28 April 1976 1 0.438 19 0.122 0.049

There was a low number of char stomachs during the day September and in April (Fig. 3). The mean stomach content and fewer stomachs in the morning than in the evening and was lowest at dusk (0.14-0.19% of body weight) increasing thus the data were too limited to show distinctly the diel rapidly throughout the night, reaching a maximum near dawn changes for each 24-h period. However, by grouping the data (0.49-0.54% of body weight) and then decreasing throughout according to time of day, clear diel changes were seen both in the day. In both studies, the largest increase in stomach con-

Limnologica 31 (2001) 4 283 tent occurred in the period 00.00-06.00 hours. In the Septem- tributions. The results indicate that the trout in Ellidavatn ate ber study, about 70% of the diet was E. lamellatus and in the mainly during the evening (21.00-01.00) and morning April study, about 60% of the diet was chironomid larvae. (05.00-11.00 hours) (Fig. 4). About 80% of the diet of trout Thus, it seems clear that char in Ellidavatn consumed their were sticklebacks. There was a significant difference in stom- food mainly at night in September and April. Most of the ach content between the periods 18.00-21.00 and samples taken at other times of the year also showed larger 00.00-03.00; 09.00-18.00 and 18.00-21.00; 06.00-09.00 stomach content in the morning than in the evening (Table 2). and 18.00-21.00 hours (Mann-Whitney, P < 0.05) and almost The feeding of the char in Ellidavatn, therefore seems to be a significant difference between the periods 03.00-06.00 and nocturnal, at least in the spring and autumn. 09.00-18.00 hours (Mann-Whitney, P = 0.06). Due to limited sample size, all data for the trout in autumn A comparable analysis was carried out for the char. The (September and October) and spring (April and May) were total number of fish was 465 (55% of them from September pooled according to time of day. The total number of fish was and 35% from April). The rapid increase in stomach content 196 (49% of them from September and 45% from April). In from 21.00-06.00 hours indicates that the char consumed this analysis, medians were used due to skewness of the dis- most of their food at night (Fig. 4) and the continuous de-

DAY Ii[i~l#l[~--il September 1975 26-28 April 1976 0.6 k~ "6 0,5

O9 0.4' -r-

O 0.1 O Fig. 3. Diel changes in the stomach I.L content (wet weight of food as % of 3 6 9 1'2 15 1'8 2'1 2'4 3 3 6 9 1'2 1'5 1'8 2'1 2'4 3 fish weight) of the char in Ellida- vatn in September 1975 and April TIME OF DAY (hours) TIME OF DAY (hours) 1976 (mean with 95% CL).

Table 2. The stomach content of char (% of fish weight) in evening and morning samples from gillnets set for 1-3 hours: date, time of net de- ployment, water temperature (°C), number of stomachs (n), mean, standard deviation (STD), t-value and significance at the 0.05 level.

Date Time °C n Mean STD t Sign.

23 October 1975 22 6.0 5 0.366 0.228 24 October 1975 8 6.1 11 0.381 0.237 0.122 NS 14 November 1975 21 2.0 6 0.140 0.188 15 November 1975 10 1.0 18 0.469 0.420 1.836 NS 11 April 1976 23 1.4 17 0.248 0.120 12 April 1976 6 0.5 5 0.528 0.125 4.540 * 25 May 1976 24 8.4 14 0.161 0.072 26 May 1976 10 8.4 15 0.421 0.183 4.960 * 30 June 1976 24 8.0 14 0.052 0.063 1 July 1976 11 8.2 14 0.240 0.128 4.913 * 27 July 1976 24 10.5 11 0.220 0.136 29 July 1976 8 9.5 8 0.328 0.122 1.785 NS 24 September 1976 21 8.2 17 0.089 0.068 25 September 1976 8 7.6 14 0.296 0.151 5.066 * 2 November 1976 20 3.0 19 0.387 0.292 3 November 1976 8 2.2 21 0.730 0.523 2.522 *

284 Limnologica 31 (2001) 4 a = 27 27 13 30 66 33 43 63 24 17 15 107 138 58 t"

0 m 0.6 Trout Char r- O3 0.5 "6

09 0.4 -1- 0 < 0.3 0 Fig. 4. Diel changes in the stomach I-- content (wet weight of food as % of o3 0.2 fish weight) of the char and trout in Z autumn and spring in Ellidavatn CI 0.1 0 1975-76 (median with 95% CL). 0 The number of fish is given at the 1.1_ top of the graphs and a Bezier 9 1'2 1'5 1; 2'1 24 3 ; 9 12 1'5 1; 21 24 3 curve fitted to the results. TIME OF DAY (hours) TIME OF DAY (hours) crease in stomach content from 06.00-21.00 hours indicates juvenile char and trout were rarely seen during the day that little or no feeding occurred during the daytime. The whereas at night many could be spotted from the shore with a stomach content in the period 03.00-06.00 was significantly pocketlight and caught by a handnet (BJORNSSON 2001). It higher than in all the other periods except 06.00-09.00; and was noticed that the smaller fish (< 15 cm) swam towards the in the period 18.00-21.00 the stomach content was signifi- shore and the larger ones away from the shore when spotted, cantly lower than in all the other periods except 21.00-24.00 indicating that the rocky shore provides a shelter for the hours (Mann-Whitney, P < 0.05). smallest fish. The diel changes in stomach content of char were used to The median weight of char and trout was larger during the estimate the average daily food intake. The difference be- day than at night, indicating that during the day smaller fish tween maximum and minimum stomach content gives the are less active than larger fish, perhaps due to risk of preda- minimum intake, 0.40% of body weight per 24 hours. By as- tion (VALDIMARSSON& METCALFE 1998; METCALFE et al. suming that there is no food intake from 06.00 to 18.00 hours, 1999). The predation risk of trout and char in Ellidavatn by the average stomach evacuation rate was 0.04% of body large trout must decrease with their size (BJORNSSON 2001). weight per hour. Assuming the same evacuation rate for ap- An alternative explanation of the diel changes in median proximately 9 hours of feeding, 0.36% of body weight per 24 weight is that net avoidance during the day may be signifi- hours can be added to the daily food consumption. The aver- cantly affected by fish size, as large fish with higher swim- age duration of each fishing session was 3 hours and thus the ming speed and larger momentum may have less chance of char were in the nets for 1.5 hours on average. This factor af- avoiding entanglement than small fish. fects the stomach content more at dawn than at dusk since the The observed did changes in stomach content show that evacuation rate increases with stomach content (ELLIOTT char in Ellidavatn mainly eat at night in September and April. 1972; ELLIOTT & PERSSON 1978) and thus approximately In other less extensive observations, a significant increase in 0.04% can be added to the daily food intake. Accordingly, the stomach content from evening till morning in April, May, daily food consumption of char in Ellidavatn in autumn and June, September and November was also found, indicating spring may have been approximately 0.8% of body weight at that nocturnal feeding of the char prevails in spring and au- a mean water temperature of 5.6 °C, weighted by sample size. tumn. In Ellidavatn, the diet of char throughout the year is small benthic invertebrates (BJORNSSON 2001) most of them living on soft bottom and many of them are found below the Discussion bottom surface in the mud. The present data do not indicate that changes in light in- Although net avoidance was probably higher during the day tensity from night to night had any effects on the diel changes than at night, it is clear that the large catches at night show in the stomach content of char. In the September study, the high swimming activity of both char and trout after dark. light intensity was highest during the second night and lowest These results are consistent with direct activity measure- during the third night but there was no indication of day-to- ments of brown trout in (CHASTON 1968, 1969; day changes in stomach content. In the April study, it was HEGGENES et al. 1993) and brook char in [Salvelinus overcast with low light intensity during all three nights, but fontinalis (MITCHILL)] (BoURKE et al. 1996). In Ellidavatn, the stomach content was significantly lower during the third

Limnologica 31 (2001) 4 285 night compared to the first two nights. The foraging efficien- flocculent substrate. For 24 cm char the average distance be- cy of juvenile Atlantic salmon has, however, been found to tween the gillrakers is about 1 mm (DERVO et al. 1991; decline rapidly through the range of light levels experienced MALMQUIST1992b), thus, retaining prey as well as plant re- at night (METCALFEet al. 1997; FRASER & METCALFE 1997). mains and gravel >1 mm in diameter. It is clear that salmonids can feed under conditions where The primary function of nocturnal activity of salmonids visual perception is not possible. Juvenile Atlantic salmon may be to hide from diurnal predators (VALDIMARSSON& were able to feed to a certain extent in total darkness (MET- METCALFE 1998). During the day, a moving, foraging fish is CALFE et al. 1997) and juvenile chinook salmon [On- more likely to be detected by a predator than a stationary, vig- corhynchus tshawytscha (WALBAUM)] were able to feed on ilant one (DONNELLY& DILL 1994). Potential diurnal preda- surface, planktonic and benthic prey in highly turbid water tors of char and trout in Ellidavmn, beside man (sportsfish- (turbidity: 230-810 NTU) where reaction distance to prey ing), are trout, American mink (Mustela vision SCHREBER), was 0 cm (GREGORY & NORTHCOrE 1993). Laboratory exper- great northern diver [Gavia immer (BRr3NMCH)], goosander iments show that in total darkness Arctic char can grow well (Mergus merganser L.) and red-throated diver [Gavia stellata on food taken from the bottom (JORGENSEN & JOBLING 1990) (PONTOPPIDAN)]. and with continuous feeding provided throughout the 24-h The trout in Ellidavatn seem to take their prey, stickle- period, most of the food intake takes place at night (JOR- backs, mainly twice a day, in the evening and in the morning, GENSEN & JOBLING 1989). However, juvenile (20 cm) Arctic probably at dusk and dawn as many other piscivores do char kept in rearing tanks without feeding were more active (KEAsT & WELSH 1968; EMRY 1973; SWENSON 1977; KELSO during the day than night in all months of the year except in 1978; HELFMAN 1981, 1986). Direct observations by a diver October and November (LINNERet al. 1990). of 20 freshwater species in Ontario lakes suggested diurnal ALANARA& BRANNAS(1997) who studied the feeding be- feeding of 10 species, crepuscular feeding of 4 species and haviour of Arctic char and [ nocturnal feeding of 4 species (EMRY 1973). At dawn and mykiss (WALBAUM)] in rearing tanks with self-feeding sys- dusk, large changes in activity take place and large predators tems found that fish taking food at night were smaller than use this period of confusion to hunt. Predators swimming those taking food during the day. Radiotelemetry of adult near the bottom are hard to see but the prey above are more brook chat" in two Canadian lakes showed that they were easily seen against the lighter background (EMRY 1973; more active at night than in the day (BOURKE et al. 1996). In HELFMAN 1986). Sticklebacks which are diurnal feeders an experimental study, juvenile Arctic char were found to be (BEUKEMA 1968) are, thus, likely to be prone to predation more nocturnal than juvenile Atlantic salmon, and the juve- near dusk and dawn. Laboratory experiments have shown niles from a lake population of Arctic char were more noctur- nocturnal activity of brown trout (CHASTON1968). However, nal than the ones from a riverine population (VALDIMARSSON the main feeding of brown trout in an English river was near et al. 2000). These authors suggested that fish derived from a dawn and dusk in the spring and autumn but between lake population are better foragers in darkness because of 08.00-12.00 hours in the summer (CnASTON 1969). From adaptations to foraging in deeper (and hence darker) water October to May brown trout in a Norwegian stream showed than are riverine fish. nocturnal activity (HEGGENES et al. 1993). In a Norwegian In a Norwegian lake, Arctic char feeding almost exclu- lake during July - September, brown trout fed on zooplank- sively on zooplankton both day and night had higher stomach ton during the day and on surface insects and chironomid fullness during the day than at night in August and September pupae during the night (DERVO et al. 1991). (DERVO et al. 1991). The density of zooplankton prey was far The reduction in stomach content of char in Ellidavatn higher at night than during the day. Still the stomach fullness from 06.00-20.00 hours was similar to the decrease in stom- was highest in the evening and lowest in the morning indicat- ach content of brown trout which had been fed on amphipods ing that feeding efficiency decreased with reduced light or chironomid larvae at 6 °C (ELLIOTT1972). This indicates levels. that in autumn and spring the char ate virtually nothing dur- The nocturnal feeding of char in Ellidavatn does not indi- ing the day. The estimated food intake of the char in Ellida- cate that they rely primarily on vision for feeding as has been vatn during the 24-h period was 0.8% of body weight. In suggested for salmonids in general (NILSSON 1963; THORPE et comparison the maximum daily intake of chironomid larvae al. 1988; JORGENSEN ~; JOBLING 1990; MALMQUIST 1992, by a 100 g trout was 268 mg dry weight at 6 °C (ELLIOTT 1992b; MALMQUISTet al. 1992; GUDBERGSSON& ANTONSSON 1975a) which is equivalent to 2.2 g wet weight, assuming 1996). Instead the results are more consistent with the hy- 87.9% water content of the chironomid larvae (ELLIOTT pothesis that char in Ellidavatn rely on taste and/or tactile 1972). Arctic char has a better growth potential than brown stimulus to detect their prey at night. The high percentage of trout at lower temperatures than 12 °C (JENSEN 1985), which plant remains and gravel in the stomach content of char but suggests that maximum food intake is higher in char than not trout in Ellidavatn (BJORNSSON2001) suggests that char trout at low temperatures. Thus, apparently the daily food in- shift through the bottom substrate in search for prey. The gill- take of the char in Ellidavatn was less than half the maximum takers may aid in filtering particles of proper size from the food intake of char at 6 °C. This is consistent with the evi-

286 Limnologica 31 (2001) 4 dence that the growth rate of the char in Ellidavatn was food - & PERSSON, L. (1978): The estimation of daily rates of food con- limited (BJORNSSON 2001). Feeding for 24 hours might in- sumption for fish. J. Anim. Ecol. 47: 977-991. crease the daily food intake and growth rate of char but most EMRY, A.R. (1973): Preliminary comparisons of day and night habits likely at the expense of increased predation risk (METCALFE of in Ontario lakes. J. Fish. Res. Board Can. 30: 761-774. et al. 1998, 1999). ERIKSSON, L.O. (1973): Spring inversion of the diel rhythm of loco- The resource partitioning and niche shift of char and trout motor activity in young sea-going brown trout, Salmo trutta trutta in Ellidavatn indicates that char are better adapted to preying L., and Atlantic salmon, Salmo salar L. Aquilo, Ser. Zool. 14: on small benthic invertebrates than trout, and trout are better 68-79. adapted to piscivory than char (BJORNSSON 2001). Thus, char FRASER, N.H.C., HEGOENES, J., METCALFE, N.B. & THORPE, J.E. and trout netted in the same location in Ellidavatn not only (1995): Low summer temperatures cause juvenile Atlantic salmon feed on different types of prey, they also feed on them at dif- to become nocturnal. Can. J. Zool. 73: 446-451. ferent times of day. This may be one more example of how - & METCALFE,N.B. (1997): The cost of becoming nocturnal: feed- char and trout can coexist in a small lake by avoiding direct ing efficiency in relation to light intensity in juvenile Atlantic confrontation while utilizing the available food resource in a salmon. Funct. Ecol. 11: 385-391. common habitat. GREGORY, R.S. & NORTHCOTE,T.G. (1993): Surface, planktonic, and benthic foraging by juvenile chinook salmon (Oncorhynchus tshawytscha) in turbid laboratory conditions. Can. J. Fish. Aquat. Acknowledgements: Mr. J0N KRISTJA,NSSON, Institute of Fresh- Sci. 50: 233-240. water , provided advice and nets. Mr. SMARI HARALDSSON GUDBERGSSO~,T, G. & ANTONSSON,TH. (1996): Freshwater fishes of and Ms. LARA HANSDOTTIR assisted with the data collection. Dr. Iceland. ReykjavN (in Icelandic). ARNTHOR GARDARSSON,University of Iceland, Dr. SIGURDURGUD- JONSSON, Institute of Freshwater Fisheries, and Dr. PI~TURM. JONAS- HEGGENES, J., I~oa, O.M.W., LINDXS, O.R., DOKK, J.G. & BREMNES, SON, University of Copenhagen, made helpful comments on the T. (1993): Homeostatic behavioural responses in a changing envi- manuscript. ronment: brown trout (Salmo trutta) become nocturnal during winter. J. Anim. Ecol. 62: 295-308. HELFMAN, G.S. (1981): Twilight activities and temporal structure in a References freshwater community. Can. J. Fish. Aquat. Sci. 38: 1405-1420. - (1986): Fish behaviour by day, night and twilight. In: T.J. P~TCHER (ed.), The behaviour of fishes, pp. 366-387. London. ALANARA,A. & B RANNAS,E. (1997): Diurnal and nocturnal feeding activity in Arctic char (Salvelinus alpinus) and rainbow trout (On- JENSEN, J.W. (1985): The potential growth of salmonids. 48:223-231. corhynchus mykiss). Can. J. Fish. Aquat. Sci. 54: 2894-2900. BEUKEMA, J.J. (1968): Predation by the three-spined stickleback JORGENSEN, E.H. & JOBLING, M. (1989): Patterns of food intake in (Gasterosteus aculeatus L.): the influence of hunger and experi- Arctic charr, Salvelinus aIpinus, monitored by radiography. Aqua- ence. Behaviour 31: 1-126. culture 81: 155-160. BJORNSSON,B. (2001): The trophic ecology of Arctic char (Salvelinus - & JOBLING, M. (1990): Feeding modes in Arctic charr, SalveIinus alpinus) and brown trout (Salrno trutta) in Ellidavatn, a small lake alpinus L.: the importance of bottom feeding for the maintenance in southwest Iceland. Limnologica 31: 199-207. of growth. Aquaculture 86: 379-385. BOURKE, P., MAGNAN,P. & RODR[GUEZ,M.A. (1996): Diel locomotor KEAST,A. & WELSH, L. (1968): Daily feeding periodicities, food up- activity of brook charr, as determined by radiotelemetry. J. Fish take rates, and dietary changes with hour of day in some lake fish- Biol. 49: 1174-1185. es. J. Fish. Res. Board Can. 25: 1133-1144. CAMPBELL, R. C. (1967): Statistics for biologists. Cambridge. KELSO, J.R.M. (1978): Diel rhythm in activity of walleye, Stizoste- CnASTON, I. (1968): Influence of light on activity of brown trout dion vitreum vitreum. J. Fish Biol. 12: 593-599. (Salmo trutta). J. Fish. Res. Board Can. 25: 1285-1289. LINNt~R, J., BR~,NN~S, E., WIKLUND,B.-S. & LUNDQUIST,H. (1990): - (1969): Seasonal activity and feeding pattern of brown trout Diel and seasonal locomotor activity patterns in Arctic charr, (Salmo trutta) in a Stream in relation to availability of Salvelinus alpinus (L.). J. Fish Biol. 37: 675-685. food. J. Fish. Res. Board Can. 26: 2165-2171. MALMQUISr, H.J. (1992a): Trophic ecology of Arctic charr (Salveli- DERVO, B.K., HEGGE, O., HESSEN, D.O. & SKURDAL,J. (1991): Diel nus alpinus) in Thingvallavatn, Iceland: eco-morphological spe- food selection of pelagic Arctic chart, Salvelinus alpinus (L.), and cializations in a polymorphic fish species. Doctoral Thesis, Uni- brown trout, Sabno trutta L., in lake Atnsj¢, SE . J. Fish versity of Copenhagen, Denmark. Biol. 38: 199-209. - (1992b): Phenotype-specific feeding behaviour of two Arctic charr DONNELLY, W.A. & DILL, L.M. (1984): Evidence for crypsis in coho Salvelinus alpinus morphs. Oecologia 92:354-361. salmon, Oncorhynchus kisutch (WALBAUM),parr: substrate colour -- SNORRASON,S.S., SKI)LASON,S., JONSSON,B., SANDLUND,O.T. & preference and achromatic reflectance. J. Fish Biol. 25: 183-195. JONASSON, RM. (1992): Diet differentiation in polymorphic Arctic ELLIOTT, J.M. (1972): Rates of gastric evacuation in brown trout, chart in Thingvallavatn, Iceland. J. Anita. Ecol. 61: 21-35. Sahno trutta L. Freshw. Biol. 2: 1-18. METCALFE, N.B., FRASER, N.H.C. & BURNS, M.D. (1998): State-de- - (1975a): Number of meals in a day, maximum weight of food con- pendent shifts between nocturnal and diurnal activity in salmon. sumed in a day and maximum rate of feeding for brown trout, Proc. R. Soc. Lond., Ser. B: Biol. Sci. 265: 1503-1507. Salmo trutta L. Freshw. Biol. 5: 287-303. -- FRASER,N.H.C. & BURNS, M.D. (1999): Food availability and the - (1975b): The growth rate of brown trout, Salmo trutta L., fed on nocturnal vs. diurnal foraging trade-off in juvenile salmon. J. maximum rations. J. Anim. Ecol. 44: 805-821. Anita. Ecol. 68: 371-381.

Limnologica 31 (2001) 4 287 -- VALDIMARSSON.S.K. &FRASER, N.H.C. (1997): Habitat profitabil- populations of Atlantic salmon (Salmo salar) and Arctic char ity and choice in a sit-and-wait predator: juvenile salmon prefer (Salvelinus alpinus). Can. J. Fish. Aquat. Sci. 57: 719-724. slower currents on darker nights. J. Anita. Ecol. 66: 866-875. - METCALFE, N.B., THORPE, J.E. & HUNTINGFORD, F.A. (1997): Sea- NmSSON, N.-A. (1963): Interaction between trout and char in Scandi- sonal changes in sheltering: effect of light and temperature on diel navia. Trans. Am. Fish. Soc. 92: 276-285. activity in juvenile salmon. Anim. Behav. 54: 1405-1412. SWENSON, W.A. (1977): Food consumption of walleye (Stizostedion WARE, D.M. (1972): Predation by rainbow trout (Salmo gairdneri): vitreum vitreum) and sauger (S. canadense) in relation to food the influence of hunger, prey density and prey size. J. Fish. Res. availability and physical environmental conditions in Lake of the Board Can. 29: 1193-1201. Woods, , Shagawa Lake, and western . J. Fish. Res. Board Can. 34: 1643-1654. Received: February 1, 2001 THORPE, J.E., MORGAN, R.I.G., PRETSWELL, D. & HIGGINS, RJ. Accepted: March 15,2001 (1988): Movement rhythms in juvenile Atlantic salmon, Salmo salar L. J. Fish Biol. 33:931-940. VALD~MARSSON,S.K. & METCALFE,N.B. (1998): Shelter selection in juvenile Atlantic salmon, or why do salmon seek shelter in winter? J. Fish Biol. 52: 42-49. Author's address: BJORN BJORNSSON, Marine Research Institute, - METCALFE, N.B. & SKt~LASON, S. (2000): Experimental demon- Skfilagata 4, P.O. Box 1390, 121 ReykjavN, Iceland; stration of differences in sheltering behaviour between Icelandic Tel.: +354 552 0240; Fax: +354 562 3790; e-mail: [email protected]

288 Limnologica 31 (2001) 4