SEARCH

SEARCH BY CITATION

Keywords:

  • antipredator behaviour;
  • Atlantic salmon;
  • diel activity rhythms;
  • growth;
  • Salmo salar

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

1. Much attention has been devoted to explaining the spatial distribution of foraging animals, but rather little to their temporal distribution (i.e. whether they are diurnal, nocturnal or crepuscular). Many animals face predictable diel cycles of food availability or predation risk, and so the approach of measuring the relative ratio of mortality risk to food gained (the ‘minimize μ/f’ rule) can be applied equally as well to different time periods of the day as to alternative food patches or habitats.

2. This method is used here to investigate the diel activity patterns of juvenile Atlantic salmon, which have previously been shown to become increasingly biased towards nocturnal activity in winter, hiding for much of the day in streambed refuges. Calculations based on published data show that nocturnal foraging in winter is far safer per unit of food obtained than is diurnal, despite greatly reduced food capture efficiency at night-time light levels.

3. Using an automated activity monitoring system based on passive integrated transponder (PIT) tags, this study shows that winter diel activity patterns in salmon are dependent on food availability. A change in food density led to a parallel change in time spent in the refuge, but (as predicted by the μ/f rule) the effect was greatest at the time of day with the least favourable ratio of predation cost to feeding benefit. Thus an experimental increase in food availability led to a 16% reduction in time spent in nocturnal foraging but a 98% reduction in time spent foraging by day, with fish spending only 0·6% of the daylight hours out of the refuge at the highest food density.

4. However, brief daytime foraging bouts had a major impact on growth rates (presumably because feeding efficiency was much greater in daylight), especially when food was scarce. Daytime feeding was thus profitable in terms of rapid food acquisition but normally suboptimal in terms of risk of predation.

5. Daily activity patterns are therefore suggested to be the result of a complex trade-off between growth and survival, which takes account of diel fluctuations in food availability, food capture efficiency and predation risk; individual variation in the extent of diurnal feeding in salmon may result from state-dependent differences in the benefits of rapid feeding and growth.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

It is now well established that the spatial distribution of foraging animals is strongly influenced by the abundance of potential predators as well as by food availability. The likelihood of being predated has been shown to affect a forager's choice of where to feed at a range of spatial scales from broad habitat types (e.g. Werner et al. 1983; Lindstro 1990) to adjacent food patches within a habitat (Grubb & Greenwald 1982; Bouskila 1995). The preferred foraging location should therefore be the one that provides the optimal balance between foraging returns and risk of mortality through predation. Several theoretical models have been developed to address this issue (Gilliam & Fraser 1987; Brown 1988; Ludwig & Rowe 1990; Abrams 1991; Newman 1991). Their predictions depend on the form of the fitness function chosen (e.g. whether it assumes the animal is attempting to maximize survival or reproductive output, and whether any reproduction is continuous or discreet; Brown 1992), but over a broad range of conditions the general finding is that foragers should be sensitive to the ratio of mortality (predation) rate to food intake rate in each potential food patch. More explicitly, Gilliam & Fraser (1987) show that nonreproducing animals with access to a refuge (which contains neither food nor predators) should favour the foraging patch that minimizes μ/f, where μ is instantaneous mortality rate and f is gross food intake rate. While the expected proportion of time spent in this patch (rather than the refuge) depends on the form of the fitness function (Kotler 1997), the general patch choice prediction has been upheld across a range of studies (Gilliam & Fraser 1987; Gotceitas 1990; Lindström 1990).

These theories have been developed as an aid in predicting the spatial distribution patterns of foraging animals. However, they can equally well be applied to the situation where there is temporal variation in foraging options. Many animals experience predictable diel fluctuations in food availability and/or predation risk, and are therefore potentially faced with decisions about when is the optimal time to feed during each 24 h cycle. If they can avoid predators when not feeding (by hiding in a refuge or by undertaking a diel migration; Zaret & Suffern 1976; Clark & Levy 1988), the optimal feeding period which maximizes survival should generally be the time of day at which μ/f is minimized, rather than that at which the feeding rate is maximized (Clark & Levy 1988). The daily period of foraging should therefore be that which minimizes the cumulative daily value of μ/f (subject to f exceeding a critical value H′ needed to offset death by starvation; Gilliam & Fraser 1987). Hence, if there is any reduction in the total time per day that must be spent foraging in order to reach H′ (as would happen if overall food availability were increased), foragers should show the greatest reduction in activity at those times of day when their risk per unit time spent foraging (μ/f) had been greatest.

In this paper these predictions are examined in juvenile Atlantic salmon (Salmo salar L.). Juvenile salmonids are highly unusual in showing a temperature-dependent shift in the balance of diel activity (Fraser, Metcalfe & Thorpe 1993; Fraser et al. 1995): at temperatures above ≈10 °C they are found in foraging locations throughout both day and night, but acquire most of their food by day (Higgins & Talbot 1985; Jørgensen & Jobling 1992; Fraser et al. 1993) because their feeding efficiency is much higher at day-time light levels (Fraser & Metcalfe 1997). They have traditionally therefore been considered to be diurnal/ crepuscular foragers (Hoar 1942; Higgins & Talbot 1985; Sagar & Glova 1988; Thorpe et al. 1988; Angradi & Griffith 1990). However, as the temperature drops they increasingly seek refuge during the day in crevices beneath stones in the streambed (Gardiner & Geddes 1980; Rimmer, Paim & Saunders 1983, 1984; Cunjak 1988) but emerge at night (Fraser et al. 1993, 1995; Heggenes et al. 1993; Riehle & Griffith 1993; Contor & Griffith 1995; Valdimarsson et al. 1997). At temperate latitudes they are thus predominantly diurnal (in terms of relative food intake obtained by day vs. by night) for most of the year but nocturnal in winter, while populations living in glacial rivers are nocturnal all year round (Fraser et al. 1995).

It has been suggested that this temperature-dependent shift in diel activity might be a result of changes in predation risk (Fraser et al. 1993), because daytime foraging might be so risky as to make it a suboptimal foraging period (despite the greater feeding success). Nocturnal foraging, while inefficient, would therefore be preferred because of the lower value of μ/f at night: feeding by day would maximize growth rate, but feeding by night maximizes survival. Fish should therefore feed for the majority of the night but only for as much of the day as is necessary to achieve the required food intake. As food requirements are greater at warmer temperatures, they should be forced to increase their daytime activity as temperature increases. In this paper this hypothesis is first evaluated by calculating approximate daytime and night-time values for μ/f for wild juvenile salmon. This is done for the winter period, because it is at this time that the fish do not need to be foraging 24 h per day and their digestion rates are so slow that they are unable to eat to satiation both by night and by day (Elliott 1975a); they therefore have the greatest scope for adjusting their activity in order to minimize μ/f. Experimental tests are then provided of the following novel predictions that arise out of the hypothesis:

1. The proportion of time that is spent foraging should contract as food availability increases, but this contraction in activity should be much greater during the day than at night.

2. At a given level of food availability, the growth rate of the fish should be proportional to the amount of daytime foraging activity that they exhibit.

Diel variation in winter predation and feeding rates

For the situation where the fish has a safe refuge and can either forage by night N or by day D, it should only prefer to forage during the night if μNDfN/fD (Gilliam & Fraser 1987). All four parameters will now be estimated from published literature. The main potential predators of juvenile salmonids in fresh water are piscivorous fish, birds such as herons, sawbill ducks and kingfishers, and certain species of mustelid. In combination these predators account for the majority of the mortality of juvenile salmonids after the fry stage (Alexander 1979), and all rely primarily on visual detection of the fish (Sjöberg 1988; Dunstone 1993; Kruuk 1995). Juvenile salmon have therefore evolved colour patterns that render them cryptic against the stony substrate of a river or stream (Donnelly & Dill 1984), but they become conspicuous when moving from their vantage position to intercept passing food items (Martel & Dill 1995). Feeding activity therefore increases their risk of being predated. The principal means by which they attempt to escape from a predatory attack is by a burst acceleration, usually towards a place of concealment under a stone in the bed of the stream. Both the time taken to react and the rate of burst acceleration are strongly temperature-dependent, so that salmonids are increasingly sluggish as temperatures drop below 10 °C (Webb 1978; Johnson, Bennett & McLister 1996). This makes them especially vulnerable to their endothermic predators at low temperatures. (Conversely the risk from piscivorous fish decreases because (i) temperature has an equivalent effect on the speed of attack of these fish, and (ii) their food demand is much reduced at low temperatures.)

Quantifying the rate (and therefore the risk) of predation by specific species of predator in wild populations is extremely difficult, but recent parallel and detailed studies of the impact of the three main predators on one species of salmon within one geographical area now allow more detailed examination of this hypothesis. For juvenile Atlantic salmon parr inhabiting rivers in central Scotland the main endothermic predators are the goosander (Mergus merganser L.), the otter (Lutra lutra L.) and the mink (Lutreola lutreola L.) [the only other common freshwater piscivore, the Grey Heron (Ardea cinerea L.), rarely eats juvenile salmon; Adams & Mitchell 1995). All three of the important predators are resident on the rivers in both summer and winter, and take salmon parr of a similar size range (5–12 cm length; Wise, Linn & Kennedy 1981; Marquiss & Feltham 1991; Kruuk et al. 1993). However, goosanders are exclusively diurnal (Nilsson 1970), while otter and mink are predominantly nocturnal (Dunstone 1993; Kruuk et al. 1993; Kruuk 1995).

Data on the density of these three main predators per km stretch of river, the percentage of salmon in their diet and their calculated impact on the fish are given in Table 1, which shows that goosanders account for ≈90% of the total predation rate. More predation occurs overall by day than by night. Moreover, in winter the relative difference in risk to the fish between day and night is even greater than is implied by these figures, because at typical winter water temperatures in central Scotland (<6 °C) the fish are actually hiding (and so unavailable to the predators) for ≈90% of the day but only 10% of the night (mean figures from (Fraser et al. 1993). For a winter photoperiod of 8 h L : 16 h D this translates into 48 min of foraging by day vs. 864 min by night; the relative predation risk per unit time foraging is therefore more than 150 times greater by day than by night [(89·5/48)/(10·5/864) = 153·4]. In the terms of the model, μND = 1/153·4 = 0·00652. While these calculations are necessarily somewhat approximate, the extent of the effect is unquestionable: it is clearly much riskier for juvenile salmon to forage by day than by night, and this is especially true in winter. The situation in summer is similar but less extreme, because the fish are almost equally available by day and by night (Valdimarsson et al. 1997) and the ratio of day to night length is reversed; this is not discussed further here because the tests of the model have been carried out on fish subjected to winter temperatures.

Table 1.  Potential impact of the major endothermic predators of juvenile (1+) salmon on Scottish rivers. Data are presented separately for those that are diurnal and those that are nocturnal. Densities are expressed per km of river length, and intake values are calculated from energetic requirements of adults of average weight; the consumption rate per km of river is derived from the predator density multiplied by the daily intake per adult
   Juvenile salmon consumed per day (g) 
SpeciesDensity (adults km−1)% of diet that is juvenile salmonper adult predatorper km of river% of total
  1. †Mean for Scottish rivers (Gibbons, Reid & Chapman 1993).

  2. ‡Median for salmonid rivers, Central Scotland (Kruuk et al. 1993).

  3. §Mean for Irish salmonid rivers (Smal 1991); no relevant data for Scotland.

  4. †† Wise et al. (1981) (with assumption that mink take the same ratio of juvenile salmon to trout as otters).

  5. ‡‡ Dunstone (1993).

Diurnal    89·5
Goosander0·36066·6334·0120·2 
Nocturnal    10·5
Otter0·06615·9135·08·9 
Mink0·490§3·9††10·6‡‡5·2 

The optimal diel activity period for the fish is dependent on the food supply as well as on the predation risk: the higher predation risk by day could, in theory, be offset by an even higher rate of feeding. Juvenile salmon are visual foragers, with the majority of the diet made up of drifting invertebrates (Power 1993). At typical night-time light intensities the fish are ≈ 1/10th as efficient at capturing drifting prey items as during the day (Fraser & Metcalfe 1997), but this is partially offset by a slight increase in the density of drifting prey at night: the mean ratio of night to day drift abundance from December to February ranges from 1·05 ± 0·34 (n = 3 sampling months; data from Heggenes et al. 1993, Table 4) to 2·21 ± 0·37 (n = 9 site-months; data from Elliott 1967, Figs 9–11). Taking the mean of these and incorporating the tenfold difference in feeding efficiency gives the night to day ratio of feeding success (units of food obtained per unit time foraging) fN/fD = 0·163. Because 0·00652 Š 0·163, the original hypothesis put forward by Fraser et al. (1993) appears to be supported by these data: in winter the predation risk per unit of accessible food is ≈25 times greater by day than it is by night. Therefore seeking refuge during the day rather than the night in winter is consistent with minimizing the mortality costs of foraging.

Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

The experimental test of the predictions arising from the foraging model was carried out on a group of 15 juvenile 1+ Atlantic salmon that were the offspring of sea-run, wild adults caught in the Loch Lomond catchment, west Scotland. The fish (mean forklength at the start of the experiment = 7·63 ± 0·18 cm, mean initial weight = 4·44 ± 0·34 g) had been reared under hatchery conditions and had come from the lower modal group of the size distribution so would remain in freshwater for at least another year (Thorpe 1977). They were individually tagged 5 weeks prior to the start of the experiment with passive integrated transponder (PIT) tags by subcutaneous insertion into the abdominal cavity. They were then placed in a 1 m diameter holding tank under conditions of constant photoperiod (12 h L : 12 h D, in order to create equal opportunities to feed by day or by night) and at a typical winter temperature (mean of 5·55 ± 0·08 °C throughout the 5-month experiment). An advantage of using this temperature was the slow speed of digestion: fish given food ad libitum could not feed to satiation twice in 24 h (Elliott 1975a), and so any difference between day and night in the percentage of time spent feeding could not simply be a result of fish feeding to satiation in both periods but taking longer to do so at night. They were fed to excess with commercial salmon pellets using an automatic feeder which delivered a trickle of pellets into the moving water column every 10 min throughout the day and night. The tank was modified by fitting a vertical transparent tube (30 cm long by 3·5 cm internal diameter) leading from the base of the tank down into a plastic lightproof chamber (10 cm diameter by 15 cm deep) (Fig. 1). This simulated the situation faced by fish in a natural stream, because the fish could choose to move downwards from an exposed foraging habitat (the open tank) into a dark refuge (the chamber), in a manner analogous to fish moving down into a crevice in the streambed. After a week of training they readily used the tube to move between the tank and the refuge.

image

Figure 1. Diagram of an experimental tank showing the PIT tag antennae that logged the movements of fish through the transparent tube linking the refuge and foraging habitat.

Download figure to PowerPoint

The fish were then left for a further month, to allow adjustment to the tank environment. They were then moved to an adjacent identical tank which had two PIT tag coil antennae wound (15 cm apart) around the vertical tube (Fig. 1). These were interrogated alternately every 200 ms by an automated system which logged the identity of any tagged fish passing through the coils onto a computer (Burns, Fraser & Metcalfe 1997; see also Prentice et al. 1990; Brännäs et al. 1994; Armstrong, Braithwaite & Rycroft 1996). Two antennae were used so that the direction of movement (in or out of the refuge) could be determined from the sequence in which the antennae were triggered. This system allowed monitoring of the precise timing and duration of every movement by each fish to and from the refuge. The vertical orientation of the tube coupled with its transparency deterred fish from resting within the detection range of the antennae (and thus triggering them continually), while its narrow diameter prevented the simultaneous movement of more than one fish. The photoperiod was kept constant (12 h L : 12 h D), with dawn at 07.00 h and dusk at 19.00 h. Daytime illumination was provided by a fluorescent strip light (giving 450 lx at the water surface), while a second filtered strip light gave a night-time surface illumination of 0·1 lx (equivalent to full moon and a clear sky; Fraser & Metcalfe 1997). Water velocities within the foraging habitat were less than 10 cm s−1, and so were within the preferred range for juvenile salmon at night-time light intensities (Metcalfe, Valdimarsson & Fraser 1997).

The experiment involved monitoring the responses of the fish to differing densities of drifting food, dispensed as before by the automatic feeder. Density was varied by altering the amount of pelleted food that was dispensed into the water current every 10 min when the feeder was activated. Four different feeding levels were used: 2·5% of total fish wet weight delivered per 24 h, equivalent to slightly excess rations and so expected to produce a high growth rate; 1·0%, equivalent to a low growth ration; 0·13%, calculated (from Elliott 1975b) to be the feeding level at which the fish would, on average, maintain weight but not grow); and 0·00%. Because the feeders operated 24 h per day, the density of drifting food was equal by day and by night, although the reduced ability of the fish to intercept food at low light levels (Fraser & Metcalfe 1997) would result in lower potential feeding rates at night. The feeder was positioned so that food items were carried away from the vicinity of the entrance to the refuge by the water circulation; a lip around the junction of the tube and the tank floor prevented any stray pellets from falling into the refuge, so that the fish could only feed if they emerged from the refuge. The tank was left undisturbed except for being cleaned once per day throughout the experiment. The same fish were exposed to the four levels of food availability, presented in a random sequence [with the proviso that the trial with no food (the only one in which fish were, on average, expected to lose weight) came last, so that fish would not be in a poor nutritional state at the beginning of any of the trials].

Each level of food density was maintained for 1 month, and data on the movements of the fish were recorded over the last 12 days of the month. All fish were anaesthetized and weighed at the start and end of each monthly period in order to calculate both specific growth rates (% change in weight per day; Ricker 1979) under each level of food density and the correct amount of food to be delivered for the next month. Fish were killed and dissected at the end of the experiment in order to determine their sex. However, the sex of the fish had no effect in any of the analyses presented here (P > 0·15 in all cases), and so this variable has been omitted for clarity of presentation. Because the experiment employed a repeated-measures design, the final analyses only included those individual fish for which data were obtained under all four feeding conditions (n = 12). All statistical analyses were performed using the SPSS for Windows statistical package; in the repeated-measures anovas the multivariate tests of significance were used whenever the Mauchly sphericity test was significant (P < 0·05), otherwise the univariate significance tests are given. All quoted probabilities are for two-tailed statistical tests.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

The effect of food density on growth

The four levels of food density had more or less the expected effect on the growth of the fish, with the highest drift density producing relatively fast rates of growth (given the low temperatures), the two intermediate levels producing negligible changes in weight and the month without food producing a significant weight loss (Fig. 2; repeated-measures anova comparing specific growth rates between food density levels: F3,9 = 29·44, P < 0·001).

image

Figure 2. Mean (+SE) specific growth rates (% weight change per day) of juvenile salmon (n = 12) under the four different levels of food availability (food provided expressed as percentage of the wet weight of the fish per 24 h). See text for statistical analysis.

Download figure to PowerPoint

Patterns of activity

There was marked temporal variation in the likelihood of fish emerging from the refuge. The biggest peak in emergence, regardless of food availability, was during the hour following dusk. Few fish emerged during the rest of the night, while the rate of emergence in daylight tended to increase as the day progressed: the Spearman's rank correlation between time since dawn and rate at which excursions were initiated (per fish per hour) was significant at two of the four levels of food availability (2·5% of bodyweight per 24 h: rsp = 0·300, n = 12 hourly periods, NS; 1·0%: rsp = 0·769, n = 12, P = 0·003; 0·13%: rsp = –0·303, n = 12, NS; 0·0%: rsp = 0·853, n = 12, P < 0·001). When fish emerged from the refuge they generally stayed out for either very short (typically less than 10 min) or very long periods (≈12 h), so that the frequency distribution of such ‘foraging excursions’ was markedly bimodal. However, the ratio of short to long excursions was markedly different between day and night, with a far greater proportion of longer excursions at night. By day 80·6 ± 9·0% (n = 4 food densities) of excursions lasted less than 10 min and 8·8 ± 5·4% lasted more than 11 h, whereas at night the proportions were more or less reversed (15·9 ± 6·0% and 66·3 ± 10·1%, respectively). The same pattern was broadly true for all levels of food density, but the relative frequency of longer daytime excursions decreased as the food availability increased (25·0% were longer than 11 h in the no food condition, but only 3·8% at a ration of 2·5%).

Food availability also affected the number of such excursions per day and night. The number of daytime foraging bouts peaked at intermediate food densities Fig. 3a,b; repeated-measures anova, F3,33 = 3·96, P = 0·016; linear term of orthogonal polynomial contrasts t = 0·11, P = 0·91, quadratic term t = −2·63, P = 0·02), whereas at night the number increased directly with food availability (F3,9 = 8·67, P < 0·01; linear term t = −3·59, P < 0·01, quadratic term t = 0·08, P = 0·94). However, while fish tended to make more excursions when food was more plentiful, this was more than offset by the shorter duration of these feeding bouts, so that the overall percentage of time spent out of the refuge decreased with increasing food levels. As predicted, this effect was strongest during the day: fish spent an average of 28·1% of the day out of the refuge when no food was provided, but reduced this to only 0·6% (a decrease of 98% from the no-food condition) at the highest food density (Fig. 3c,d; repeated-measures anova on arcsin-transformed data, F3,33 = 8·24, P < 0·001). There was thus a significant trend for the fish to spend increasingly more time out as the level of food decreased (linear term of orthogonal polynomial contrast, t = 4·44, P = 0·001; higher order terms nonsignificant). The percentage of the night spent out of the refuge did not vary significantly with food availability, decreasing from an average of 91·9% when there was no food to 77·9% at the highest food density (a drop of only 16% repeated-measures anova on arcsin-transformed data: F3,33 = 2·58, P = 0·07).

Relationship between activity and growth

As predicted, the growth rate of a fish was strongly correlated with its daytime foraging activity, but only at the two highest levels of food availability. Specific growth rates correlated most strongly with the number of foraging excursions made per day, but were unrelated to either the number of night-time excursions or the percentage of the night spent out of the refuge (Table 2). These relationships were quantified in a stepwise multiple regression of food density and daytime activity on growth rate. A total of 73·2% of the individual variation in specific growth rate was explained by the regression (F2,45 = 61·37, P < 0·0001). Both food density (P < 0·0001) and the interaction between food density and the number of foraging excursions per day (P = 0·0015) were significant. Thus the frequency with which fish emerged to forage during daylight hours had little effect on growth rates when the food density was low, but had an increasingly important effect as food availability increased (Fig. 4).

Table 2.  Correlation coefficients for the relationship between day or night activity and specific growth rate in individual juvenile salmon, at four different levels of food density. Activity is measured as (a) the percentage of the day or night spent out of the refuge, and (b) the number of excursions from the refuge by day or night. Coefficients are those of (a) Spearman's Rank and (b) Pearson Correlations; n = 12 fish for all correlations, asterisks indicate significant correlations after sequential Bonferroni correction (Rice 1989). Note that the strongest relationships are for the higher levels of food and for daytime activity
 Food density (% bodyweight per 24 h) 0·00%0·13%1·0%2·5%
  • **

    P < 0·01,

  • ***

    P < 0·001

(a) Time out of refuge (%)
Day−0·380 0·238 0·6060·235
Night−0·129 0·042 0·5880·406
(b) Excursions (n)
Day 0·231−0·252 0·933***0·747**
Night 0·062−0·362−0·0600·005
image

Figure 4. Effect of food availability (food provided expressed as percentage of the wet weight of the fish per 24 h) and number of daytime foraging bouts initiated per day on growth rate (% per day). Graph illustrates the three-dimensional plot of the multiple regression analysis described in the text.

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

The majority of animal species have evolved to be active either by day or by night, and it is now difficult to establish what is the primary cause of their diel activity pattern. However, those species that can be either nocturnal or diurnal can be used to test whether diel activity patterns are responsive to variations in predation risk and food availability. There is increasing evidence that freshwater herbivorous insects can shift their activity pattern in relation to predation risk, either over an evolutionary time scale (e.g. Flecker 1992) or as an immediate facultative response to the presence of predators (Cowan & Peckarsky 1994; Douglas, Forrester & Cooper 1994). Diel variation in potential feeding rates is less likely to be important for such species, because both the food supply and their ability to harvest it may be independent of light intensity. However, in the case of visually feeding freshwater salmonids their reduced ability to catch prey items at low light levels (Henderson & Northcote 1985; Fraser & Metcalfe 1997) makes the situation more complicated: their potential feeding rate is higher by day, but so is the mortality rate. The values for μN, μD, fN and fD presented here are only estimates, and there is bound to be regional variation, but the magnitude of the calculated difference in risk between night and day activity in winter is so great that nocturnal foraging is clearly optimal at this time of year, despite the impairment in feeding efficiency. There are few comparative data available with which to compare and validate the impacts of the different predators. However, Alexander (1979) carried out a detailed study of the relative impact of different predators (including goosander, otter and mink) on salmonid [brown trout (Salmo trutta L.) and brook trout (Salvelinus fontinalis L.)] populations on an equivalent small river in Michigan. Given the differences in location and sampling methods, he recorded densities of these predators in winter (November–April) that were remarkably similar (both in absolute and relative terms) to those quoted in the present study (0·23 goosanders per km river vs. 0·36 quoted here, 0·087 otters vs. 0·066, and 1·18 mink vs. 0·49). His figures for relative predation rates also show the same basic result as the present study: goosanders were the most important predators of 1 + fish in winter, consuming an estimated 59·7% of all brown trout and 52·5% of all brook trout that died during the winter period (data from Tables 2 and 4, Alexander 1979); diurnal predators therefore had a greater impact than did nocturnal ones. These values also indicate the importance of predation as an agent of mortality, demonstrating that there will be a substantial selection pressure for appropriate predator-sensitive foraging tactics.

Other explanations have been put forward for the shift towards predominantly nocturnal activity in wintering salmonids. The tendency to seek shelter at low temperatures was interpreted by Rimmer et al. (1984) as a need to conserve energy, but fish prefer shelters that offer concealment over those that only provide a screen against the water flow (Gregory & Griffith 1996; Valdimarsson & Metcalfe 1998). Heggenes et al. (1993) suggested that the fish came out of their streambed refuges at night in order to avoid becoming trapped in anchor ice, which tends to form at that time. However, this cannot account for the adoption of pronounced nocturnal behaviour in environments where anchor ice virtually never forms (e.g. in most UK rivers, where winter temperatures fluctuate between 1 and 7 °C). Recent work has shown that the fish become increasingly diurnal as spring progresses, a trend which is independent of temperature (Valdimarsson et al. 1997). This may be in response to the change in photoperiod, which reduces the amount of time available for safer nocturnal feeding. As temperatures increase in the spring the fish will also require more food, thus forcing them to increase their period of activity further and further into the day. The fact that fish in the present experiments were increasingly likely to initiate a foraging excursion the longer the time elapsed since dawn (which roughly equates with time since the last foraging bout), further implies that their daily activity patterns are partially driven by nutritional need. This is complementary to the manner in which longer-term energetic requirements have been shown to determine the overall daily foraging intensity of these wintering fish (Metcalfe & Thorpe 1992; Bull, Metcalfe & Mangel 1996).

It might be argued that the design of the experiment was inappropriate because it did not incorporate actual or simulated predators. However, it is well known that animals have evolved inherent behaviour patterns that take account of anticipated (rather than observed) predation risks. While these behaviour patterns may be fine-tuned through experience, they are sensitive to the perceived riskiness of the environment, even in the absence of any actual predators. Thus stream invertebrates that have evolved nocturnal activity patterns in response to diurnal predators will remain nocturnal when placed in a predator-free environment (Flecker 1992; Cowan & Peckarsky 1994). Moreover, animals with imperfect knowledge should always tend towards overestimating predation hazard, because this is less risky than the reverse (Bouskila & Blumstein 1992). That the fish in the present study used the refuge extensively (despite it being less than 0·5% of the volume of the foraging area) and used it increasingly as their time required for foraging reduced (as a result of increased food availability) further indicates that they regarded the foraging area as a risky habitat. The effect on time budgets of manipulating the perceived riskiness of this habitat will be presented in a future paper.

It has previously been shown that juvenile salmon make fine-tuned adjustments to predation risk, responding to an increased risk by reducing their level of territorial defence (Martel 1996), feeding activity (Dill & Fraser 1984; Metcalfe, Huntingford & Thorpe 1987a; Angradi 1992) and attention to prey items (Metcalfe, Huntingford & Thorpe 1987b). The results presented here indicate that they also adjust their patterns of feeding activity to minimize their exposure to predators, by preferentially reducing daytime (in contrast to night-time) activity when faced with an increased food supply. The first prediction was thus upheld. However, the fish also grew faster when experiencing the highest food density, indicating that they were not simply reducing their activity to the level that produced the same food intake as in the reduced ration trials. Instead they adopted an intermediate level of foraging activity, so were able to both reduce activity and obtain more food. This pattern of response is predicted when there is a fitness benefit to acquiring additional food above the level H′ needed for survival (Brown 1988). It has been observed in those species that are able to put excess resources into growth (e.g. fish; Gilliam & Fraser 1987) or that can cache surplus food as an insurance against later food shortage (e.g. gerbils Gerbillus spp.; Kotler 1997). The typical feeding behaviour of over-wintering juvenile salmon indicates that their primary objective is to survive rather than to grow (Bull et al. 1996), and indeed the food supply in the wild may be so low that underyearling fish may have difficulty avoiding starvation (Gardiner & Geddes 1980). However, in conditions where food availability is high and the cost of attempting to grow is therefore reduced, growth would be beneficial because larger fish are able to store proportionally more fat reserves (Elliott 1976; Kane 1988; Simpson et al. 1997) and (in the case of males) are more likely to be able to mature early (Berglund 1992; Simpson 1992).

Not all individual salmon adopted the same level of day or night activity, and this had pronounced effects on their growth rates: those salmon that made most daytime foraging excursions also grew fastest, provided that the drift density was high. Therefore at high levels of food availability those fish that showed minimal diurnal activity incurred a significant cost in terms of lost growth, but presumably these fish would normally incur a reduced risk of predation. While it is not fully understood what determines the level of daytime vs. night-time activity expressed by an individual, there is some indication that this may be linked to their social status (Alanärä & Brännäs 1997). Moreover, diurnal feeding activity is greatest in those salmon that have lost most weight during early winter, and in the smallest of the fish that are about to undertake the arduous seaward migration (Metcalfe, Fraser & Burns 1998). It is therefore evident that diel activity patterns can be dynamic, involving complex state-dependent trade-offs between competing demands.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

We thank the Loch Lomond Angling Improvement Association for providing salmon eggs, John Laurie and Roddy MacDonald for looking after the fish, BOCM Pauls for provision of fish food and Malcolm Elliott, Chris Luecke, Marc Mangel and Sveinn K. Valdimarsson for comments on the manuscript. This project was funded by a grant from the Natural Environment Research Council to NBM.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  • Abrams, P.A. 1991 Life history and the relationship between food availability and foraging effort. Ecology, 72, 12421252.
  • Adams, C.E. & Mitchell, J. 1995 The response of a Grey Heron Ardea cinerea breeding colony to rapid change in prey species. Bird Study, 42, 4449.
  • Alanärä, A., Brännä S & E. 1997 Diurnal and nocturnal feeding activity in arctic char and rainbow trout. Canadian Journal of Fisheries and Aquatic Sciences, 54, 28942900.
  • Alexander, G.R. 1979 Predators of fish in coldwater streams. Predator–Prey Systems in Fisheries Management (eds R. H.Stroud & H.Clepper), pp. 153170. Sport Fishing Institute, Washington, DC.
  • Angradi, T.R. 1992 Effects of predation risk on foraging behavior of juvenile rainbow trout (Oncorhynchus mykiss). Canadian Journal of Zoology, 70, 355360.
  • Angradi, T.R. & Griffith, J.S. 1990 Diel feeding chronology and diet selection of rainbow trout (Oncorhynchus mykiss) in the Henry's Fork of the Snake River, Idaho. Canadian Journal of Fisheries and Aquatic Sciences, 47, 199209.
  • Armstrong, J.D., Braithwaite, V.A. & Rycroft, P. 1996 A flat-bed passive integrated transponder antenna array for monitoring behaviour of Atlantic salmon parr and other fish. Journal of Fish Biology, 48, 539541.
  • Berglund, I. 1992 Growth and early sexual maturation in Baltic salmon (Salmo salar) parr. Canadian Journal of Zoology, 70, 205211.
  • Bouskila, A. 1995 Interactions between predation risk and competition—a field-study of kangaroo rats and snakes. Ecology, 76, 165178.
  • Bouskila, A. & Blumstein, D.T. 1992 Rules of thumb for predation hazard assessment—predictions from a dynamic model. American Naturalist, 139, 161176.
  • Brännä S. E., Lundqvist, H., Prentice, E., Schmitz, M., Brännä S, K. & Wiklund, B.S. 1994 Use of the passive integrated transponder (PIT) in a fish identification and monitoring system for fish behavioral studies. Transactions of the American Fisheries Society, 123, 395401.
  • Brown, J.S. 1988 Patch use as an indicator of habitat preference, predation risk, competition. Behavioral Ecology and Sociobiology, 22, 3747.
  • Brown, J.S. 1992 Patch use under predation risk: I. Models and predictions. Annales Zoologici Fennici, 29, 301309.
  • Bull, C.D., Metcalfe, N.B. & Mangel, M. 1996 Seasonal matching of foraging to anticipated energy requirements in anorexic juvenile salmon. Proceedings of the Royal Society of London B, 263, 1318.
  • Burns, M.D., Fraser, N.H.C. & Metcalfe, N.B. 1997 An automated system for monitoring fish activity patterns. Transactions of the American Fisheries Society, 126, 10361040.
  • Clark, C.W. & Levy, D.A. 1988 Diel vertical migrations by juvenile sockeye salmon and the anti-predation window. American Naturalist, 131, 271290.
  • Contor, C.R. & Griffith, J.S. 1995 Nocturnal emergence of juvenile rainbow trout from winter concealment relative to light intensity. Hydrobiologia, 299, 179183.
  • Cowan, C.A. & Peckarsky, B.L. 1994 Diel feeding and positioning periodicity of a grazing mayfly in a trout stream and fishless stream. Canadian Journal of Fisheries and Aquatic Sciences, 51, 450459.
  • Cunjak, R.A. 1988 Behaviour and microhabitat of young Atlantic salmon (Salmo salar) during winter. Canadian Journal of Fisheries and Aquatic Sciences, 45, 21562160.
  • Dill, L.M. & Fraser, A.H.G. 1984 Risk of predation and the feeding behaviour of juvenile coho salmon (Oncorhynchus kisutch). Behavioral Ecology and Sociobiology, 16, 6571.
  • Donnelly, W.A. & Dill, L.M. 1984 Evidence for crypsis in coho salmon, Oncorhynchus kisutch (Walbaum) parr: substrate colour preference and achromatic reflectance. Journal of Fish Biology, 25, 183195.
  • Douglas, P.L., Forrester, G.E. & Cooper, S.D. 1994 Effects of trout on the diel periodicity of drifting in Baetid mayflies. Oecologia, 98, 4856.
  • Dunstone, N. 1993The Mink. T. & A.D. Poyser, London.
  • Elliott, J.M. 1967 Invertebrate drift in a Dartmoor stream. Archiv Für Hydrobiologie, 63, 202237.
  • Elliott, J.M. 1975a Number of meals in a day, maximum weight of food consumed in a day and maximum rate of feeding for brown trout, Salmo trutta L. Freshwater Biology, 5, 287303.
  • Elliott, J.M. 1975b The growth rate of brown trout (Salmo trutta L.) fed on reduced rations. Journal of Animal Ecology, 44, 823842.
  • Elliott, J.M. 1976 Body composition of brown trout (Salmo trutta L.) in relation to temperature and ration size. Journal of Animal Ecology, 45, 273289.
  • Feltham, M.J. 1995 Consumption of Atlantic salmon smolts and parr by goosanders; estimates from doubly-labelled water measurements of captive birds released on two Scottish rivers. Journal of Fish Biology, 46, 273281.
  • Flecker, A.S. 1992 Fish predation and the evolution of invertebrate drift periodicity—evidence from neotropical streams. Ecology, 73, 438448.
  • Fraser, N.H.C., Heggenes, J., Metcalfe, N.B. & Thorpe, J.E. 1995 Low summer temperatures cause juvenile Atlantic salmon to become nocturnal. Canadian Journal of Zoology, 73, 446451.
  • Fraser, N.H.C. & Metcalfe, N.B. 1997 The costs of becoming nocturnal: feeding efficiency in relation to light intensity in juvenile Atlantic salmon. Functional Ecology, 11, 385391.
  • Fraser, N.H.C., Metcalfe, N.B. & Thorpe, J.E. 1993 Temperature-dependent switch between diurnal and nocturnal foraging in salmon. Proceedings of the Royal Society of London B, 252, 135139.
  • Gardiner, W.R. & Geddes, P. 1980 The influence of body composition on the survival of juvenile salmon. Hydrobiologia, 69, 6772.
  • Gibbons, D.W., Reid, J.B. & Chapman, R.A. 1993 The New Atlas of Breeding Birds in Britain and Ireland1988–91. T. & A.D. Poyser Ltd, London.
  • Gilliam, J.F. & Fraser, D.F. 1987 Habitat selection under predation hazard-test of a model with foraging minnows. Ecology, 68, 18561862.
  • Gotceitas, V. 1990 Foraging and predator avoidance: a test of a patch choice model with juvenile bluegill sunfish. Oecologia, 83, 346351.
  • Gregory, R.S. & Griffith, J.S. 1996 Winter concealment by subyearling rainbow trout—space size selection and reduced concealment under surface ice and in turbid water conditions. Canadian Journal of Zoology, 74, 451455.
  • Grubb, T.C. & Greenwald, L. 1982 Sparrows and a brushpile: foraging responses to different combinations of predation risk and energy cost. Animal Behaviour, 30, 637640.
  • Heggenes, J., Krog, O.M.W., LindaÅ S. O.R., Dokk, J.G. & Bremnes, T. 1993 Homeostatic behavioural responses in a changing environment: brown trout (Salmo trutta) become nocturnal during winter. Journal of Animal Ecology, 62, 295308.
  • Henderson, M.A. & Northcote, T.G. 1985 Visual prey detection and foraging in sympatric cutthroat trout (Salmo clarkii clarkii) and Dolly Varden (Salvelinus malma). Canadian Journal of Fisheries and Aquatic Sciences, 42, 785790.
  • Higgins, P.J. & Talbot, C. 1985 Growth and feeding in juvenile Atlantic salmon. Nutrition and Feeding in Fish (eds C.B.Cowey, A.M.Mackie & J.G.Bell), pp. 243263. Academic Press, London.
  • Hoar, W.S. 1942 Diurnal variations in feeding activity of young salmon and trout. Journal of the Fisheries Research Board of Canada, 6, 90101.
  • Johnson, T.P., Bennett, A.F. & McLister, J.D. 1996 Thermal dependence and acclimation of fast start locomotion and its physiological basis in rainbow trout (Oncorhynchus mykiss). Physiological Zoology, 69, 276292.
  • Jørgensen, E.H. & Jobling, M. 1992 Feeding behaviour and effect of feeding regime on growth of Atlantic salmon, Salmo salar. Aquaculture, 101, 135146.
  • Kane, T.R. 1988 Length–weight relationship of hatchery-reared Atlantic salmon. Progressive Fish-Culturist, 50, 2830.
  • Kotler, B.P. 1997 Patch use by gerbils in a risky environment: manipulating food and safety to test four models. Oikos, 78, 274282.
  • Kruuk, H. 1995Wild Otters: Predation and Populations. Oxford University Press, Oxford.
  • Kruuk, H., Carss, D.N., Conroy, J.W.H. & Durbin, L. 1993 Otter (Lutra lutra L.) numbers and fish productivity in rivers in north-east Scotland. Symposium of the Zoological Society of London, 65, 171191.
  • Lindströ M & A. 1990 ³. The role of predation risk in stopover habitat selection in migrating bramblings, Fringilla montifringilla. Behavioral Ecology, 1, 102106.
  • Ludwig, D. & Rowe, L. 1990 Life-history strategies for energy gain and predator avoidance under time constraints. American Naturalist, 135, 686707.
  • Marquiss, M. & Feltham, M.J. 1991 Sawbills (goosanders and red-breasted mergansers) as salmon predators. Annual Review of the Freshwater Fisheries Laboratory Pitlochry, 1990–91, 3035.
  • Martel, G. 1996 Growth rate and influence of predation risk on territoriality in juvenile coho salmon (Oncorhynchus kisutch). Canadian Journal of Fisheries and Aquatic Sciences, 53, 660669.
  • Martel, G. & Dill, L.M. 1995 Influence of movement by coho salmon (Oncorhynchus kisutch) parr on their detection by common mergansers (Mergus merganser). Ethology, 99, 139149.
  • Metcalfe, N.B., Fraser, N.H.C. & Burns, M.D. 1998 State-dependent shifts between nocturnal and diurnal activity in salmon. Proceedings of the Royal Society of London B, 265, 15031507.
  • Metcalfe, N.B., Huntingford, F.A. & Thorpe, J.E. 1987a The influence of predation risk on the feeding motivation and foraging strategy of juvenile Atlantic salmon. Animal Behaviour, 35, 901911.
  • Metcalfe, N.B., Huntingford, F.A. & Thorpe, J.E. 1987b Predation risk impairs diet selection in juvenile salmon. Animal Behaviour, 35, 931933.
  • Metcalfe, N.B. & Thorpe, J.E. 1992 Anorexia and defended energy levels in over-wintering juvenile salmon. Journal of Animal Ecology, 61, 175181.
  • Metcalfe, N.B., Valdimarsson, S.K. & Fraser, N.H.C. 1997 Habitat profitability and choice in a sit-and-wait predator: juvenile salmon prefer slower currents on darker nights. Journal of Animal Ecology, 66, 866875.
  • Newman, J.A. 1991 Patch use under predation hazard: foraging behavior in a simple stochastic environment. Oikos, 61, 2944.
  • Nilsson, L. 1970 Food-seeking activity of south Swedish diving ducks in the non-breeding season. Oikos, 21, 145154.
  • Power, G. 1993 Estimating production, food supplies and consumption by juvenile Atlantic salmon (Salmo salar). Production of Juvenile Atlantic Salmon, Salmo Salar, in Natural Waters, Vol. 118 (eds R.J.Gibson & R.E.Cutting), pp. 163174. Canadian Special Publications in Fisheries and Aquatic Sciences.
  • Prentice, E., Flagg, T.A., McCutcheon, C.S. & Brastow, D.F. 1990 PIT-tag monitoring systems for hydroelectric dams and fish hatcheries. American Fisheries Society Symposium, 7, 323334.
  • Rice, W.R. 1989 Analyzing tables of statistical tests. Evolution, 43, 223225.
  • Ricker, W.E. 1979 Growth rates and models. Fish Physiology, Vol. 8 (eds W. S.Hoar, D. J.Randall & J. R.Brett), pp. 677743. Academic Press, New York.
  • Riehle, M.D. & Griffith, J.S. 1993 Changes in habitat utilization and feeding chronology of juvenile rainbow trout (Oncorhynchus mykiss) in fall and the onset of winter in Silver Creek, Idaho. Canadian Journal of Fisheries and Aquatic Sciences, 50, 21192128.
  • Rimmer, D.M., Paim, U. & Saunders, R.L. 1983 Autumnal habitat shift of juvenile Atlantic salmon (Salmo salar) in a small river. Canadian Journal of Fisheries and Aquatic Sciences, 40, 671680.
  • Rimmer, D.M., Paim, U. & Saunders, R.L. 1984 Changes in the selection of micro-habitat by juvenile Atlantic salmon (Salmo salar) at the summer–autumn transition in a small river. Canadian Journal of Fisheries and Aquatic Sciences, 41, 469475.
  • Sagar, P.M. & Glova, G.J. 1988 Diel feeding periodicity, daily ration and prey selection of a riverine population of juvenile chinook salmon, Oncorhynchus tshawytscha (Walbaum). Journal of Fish Biology, 33, 643653.
  • Simpson, A.L. 1992 Differences in body size and lipid reserves between maturing and nonmaturing Atlantic salmon parr, Salmo salar L. Canadian Journal of Zoology, 70, 17371742.
  • Simpson, A.L., Metcalfe, N.B., Huntingford, F.A. & Thorpe, J.E. 1997 Pronounced seasonal differences in appetite of Atlantic salmon parr, Salmo salar: effects of nutritional state and life-history strategy. Functional Ecology, 10, 760767.
  • Sjöberg, K. 1988 Food selection, food-seeking patterns and hunting success of captive Goosanders Mergus merganser and Red-breasted Mergansers M. serrator in relation to the behaviour of their prey. Ibis, 130, 7993.
  • Smal, C.M. 1991 Population studies on feral American mink Mustela vison in Ireland. Journal of Zoology, London, 224, 233249.
  • Thorpe, J.E. 1977 Bimodal distribution of length of juvenile Atlantic salmon (Salmo salar L.) under artificial rearing conditions. Journal of Fish Biology, 11, 175184.
  • Thorpe, J.E., Morgan, R.I.G., Pretswell, D. & Higgins, P.J. 1988 Movement rhythms in juvenile Atlantic salmon, Salmo salar L. Journal of Fish Biology, 33, 931940.
  • Valdimarsson, S.K. & Metcalfe, N.B. 1998 Shelter selection in juvenile Atlantic salmon; or why do salmon seek shelter in winter? Journal of Fish Biology, 52, 4249.
  • Valdimarsson, S.K., Metcalfe, N.B., Thorpe, J.E. & Huntingford, F.A. 1997 Seasonal changes in sheltering: effect of light and temperature on diel activity in juvenile salmon. Animal Behaviour, 54, 14051412.
  • Webb, P.W. 1978 Temperature effects on acceleration in rainbow trout (Salmo gairdneri). Journal of the Fisheries Research Board of Canada, 35, 14171422.
  • Werner, E.E., Gilliam, J.F., Hall, D.J. & Mittelbach, G.G. 1983 An experimental test of the effects of predation risk on habitat use in fish. Ecology, 64, 15401548.
  • Wise, M.H., Linn, I.J. & Kennedy, C.R. 1981 A comparison of the feeding biology of Mink Mustela vison and otter Lutra lutra. Journal of Zoology, London, 195, 181213.
  • Zaret, T.M. & Suffern, J.S. 1976 Vertical migration in zooplankton as a predator avoidance mechanism. Limnology and Oceanography, 21, 804813.

Received 13 February 1998; revision received 3 July 1998