Population density and territory size in juvenile rainbow trout, Oncorhynchus mykiss: implications for population regulation
Jacquelyn L.A. Wood1, James W.A. Grant, and Moria H. Belanger
Department of Biology, Concordia University,
7141 Sherbrooke Street West, Montréal, Québec, H4B 1R6, Canada
1Phone: 514-848-2424 ext. 3439
Email: J.L.A. Wood: email@example.com (corresponding author)
J.W.A. Grant: firstname.lastname@example.org
M. H. Belanger: email@example.com
Abstract: We manipulated population density of juvenile rainbow trout (Oncorhynchus mykiss) across a range of realistic densities in artificial stream channels, while controlling food abundance in two different ways: in Experiment 1, the total amount of food was held constant over a three-fold increase in density, whereas in Experiment 2, the per capita amount of food was held constant over an eight-fold increase in density. We tested the contrasting predictions that territory size (1) is not affected by population density; (2) decreases with population density as 1/n, where n = the local population size; or (3) decreases with population density but towards an asymptotic minimum size. In Experiment 1, territory size decreased with increasing population density. With the broader range of densities used in Experiment 2, territory size initially decreased with density and then leveled off at a minimum territory radius of 20-30 cm. Our results suggest an asymptotic minimum size of about 0.2m2 for a 5-cm rainbow trout, similar to what is observed for high-density conditions in the wild. This minimum territory size could potentially set an upper limit on local population density and help regulate the population size of stream salmonids.
Key words: territoriality, population regulation, density dependence, trout
Whether or not territorial behaviour exerts strong population-regulating effects depends on how individual territory size is adjusted in response to changes in environmental conditions, such as food abundance and intruder pressure. At one extreme, an absolutely fixed territory size will set the upper limit of individuals that can settle in a particular habitat (Grant and Kramer 1992; Rodenhouse et al. 1997). At the opposite extreme, if there is no minimum acceptable territory size, then territory size will decrease as 1/n, wherein n is the local population size, with respect to the territory size at n=1. According to the latter hypothesis, territories serve only to space individuals within habitats (Fretwell and Lucas 1970; Maynard Smith 1974), and territoriality will play no role in population regulation. Intermediate between these two extremes, the elastic disc model (Huxley 1934) predicts that territories will initially decrease with increasing density until a minimum acceptable territory size is reached, which sets the maximum density for a particular habitat (Wilson 1975; Adams 2001). The minimum territory size in any given habitat may be affected by environmental conditions, such as food abundance, so that the maximum density is higher in optimal than in suboptimal habitats (Maynard Smith 1974). For example, warblers preferentially settle in flooded vs. dry habitats in response to prey abundance, leading to smaller minimum territory sizes in the preferred habitat (Petit and Petit 1996).
Stream dwelling salmonids have been a popular model system for investigating the population consequences of territoriality because juveniles defend feeding territories (Elliott 1990; Keeley and Grant 1995) and frequently exhibit density dependent growth, mortality, or emigration (Imre et al. 2005; Lobon-Cervia 2007; Utz and Hartman 2009). Hence, territoriality has frequently been implicated in the regulation of salmonid populations in the literature (e.g. Chapman 1966; Elliott 1994; Cattaneo et al. 2002). For example, Elliott (1993) demonstrated that mortality of brown trout was density dependent during early life stages as a result of competition for feeding territories among newly emerged fish, after which mortality was density independent (also see Milner et al. 2003; Einum and Nislow 2005). Indeed, studies of stream salmonids have often revealed regular, negative relationships between abundance and average body size (“thinning rules”, Grant and Kramer 1990; Elliott 1993; Bohlin et al. 1994) which have been taken as indirect evidence of density regulation (Jenkins et al. 1999), with territoriality as the mechanism.
Territory size data for stream salmonids have generally supported the two key predictions of optimal territory size models (e.g. Schoener 1983; Adams 2001): territory size decreases with increasing food abundance and population density (Slaney and Northcote 1974; Dill et al. 1981; Keeley 2000). However, Keeley’s (2000) elegant experiment used only three levels of population density, making it difficult to determine the actual shape of the territory size vs. density curve. Furthermore, most experimental studies have often used densities greater than 100 fish·m-2 (Slaney and Northcote 1974; Imre et al. 2004), much higher than found in the wild (e.g. Grant and Kramer 1992). Hence, the transferability of these results to natural situations is problematic.
Our goal was to determine how population density affects territory size while holding food density constant in two different ways. In Experiment 1, we tested whether territory size decreased with increasing population density when the total food supply was held constant, with two advances over previous studies: densities typically found in nature were used and fish were not allowed to emigrate so the densities were held constant in a trial. Because total food abundance was kept constant across a three-fold increase in density in Experiment 1, the per capita food abundance was actually decreasing as density increased. In nature, this experiment might be analogous to a situation where the initial density of young-of-the-year salmonids is set by the proximity to a redd, and is independent of food abundance (e.g. Teichart et al. 2011). In Experiment 2, we modified the experiment in two important ways: the per capita food supply was maintained at a constant level among density treatments and the range of density treatments used was increased to an eight-fold range, in an attempt to determine the shape of the territory size vs. density curve. Under natural conditions, a situation where the per capita food abundance remains relatively constant among different local population densities might be equivalent to an ideal free distribution between adjacent sites within a stream that differ in food abundance (i.e. an adjacent pool and riffle). If fish density is directly proportional to food abundance, then population density will vary while per capita food abundance is held constant. In summary, we conducted two different experiments by holding food constant in two different ways to test the contrasting predictions that territory size (1) is not affected by population density; (2) decreases with population density as 1/n; or (3) decreases with population density but towards an asymptotic minimum size.
Material and Methods
Young-of-the-year rainbow trout were obtained from Pisciculture des Arpents Verts, Ste-Edwidge-de-Clifton, Quebec, Canada. The trout were maintained in 133 L holding tanks on a 12 hour light: 12 hour dark cycle (08:00-20:00). Water temperature varied over the course of the experiments (from 11oC in winter to 21oC in summer), but was still within the range of preferred temperatures for rainbow trout (Kerr and Lasenby 2000). The fish were fed ad libitum daily with ground trout chow pellets (Vigor #4, Corey Feed Mills).
Artificial Stream Channels
In order to simulate natural conditions, all trials were conducted in 1.92 m x 0.76 m (l × w; 1.47 m2) artificial stream channels. However, only 1.08 m2 of space downstream from the automatic feeder was suitable space for foraging. Water temperature in the stream channels varied with the outdoor temperature, and was recorded daily for each trial (mean ± SD temperature = 12.1 ± 3.1oC). Water depth, measured at the centre of each channel, was 24.5 ± 1.4 cm (mean ± SD), whereas current velocity, measured at 40% of the water depth at 20 positions, was 0.056 ± 0.033 m·s-1 (mean ± SD). The substrate of each stream channel consisted of a layer of light colored aquarium gravel overlaid by a four by eight grid of river stones (mean max. diameter = 7.84 cm; range = 5.7-10.5 cm). The stones were spaced at equal distances (15.3 cm apart horizontally, and 21.3 cm apart vertically) from each other, acting as visual markers to aid fish in establishing territories, and to provide a frame of reference for recording the positions of focal fish during observations.
Since stream-dwelling salmonids feed primarily on drifting aquatic invertebrates (e.g. Keeley and Grant 1995), food was presented in a manner simulating stream drift. The daily ration of food was spread evenly over the belt of an automatic belt feeder, which delivered the food at a constant rate over a period of 12 hours (08:00-20:00). To promote growth over the course of the experiment, fish received a daily ration of food (Optimum 0.7 granulated fish feed, Corey Feed Mills) that was equivalent to 5% of the biomass of four fish for each treatment in Experiment 1 (i.e. 5% × 4 × mean mass of a fish at the beginning of the trial), and approximately 5% of the total fish biomass in the stream channel in Experiment 2. Each morning the feeder belts were checked for any food that had not fallen into the stream channel during the previous day (mean ± SD % food = 25.23 ± 6.48 %); this amount was weighed and subtracted from the original planned ration. Planned and actual ration of food were highly correlated across trials (linear regression: actual ration (g) = 0.743 planned ration (g) + 0.001, r2 = 0.971, n = 191, P < 0.001).
Each stream channel was stocked with juvenile rainbow trout (Experiment 1; mean ± SD mass = 1.01 ± 0.38 g, mean ± SD fork length = 3.55 ± 0.52 cm, and Experiment 2; mean ± SD mass = 1.31 ± 0.38 g; mean ± SD fork length = 4.95 ± 0.52 cm) in one of three different density treatments (4, 8, or 12 fish) in Experiment 1, and in one of five different density treatments (2, 4, 8, 12, or 16 fish) in Experiment 2. Five replicates of each density treatment were rotated successively through each of three artificial stream channels to control for any differences between channels.
In Experiment 1, 15 trials were carried out between 1 May and 21 July, 2006, with each trial lasting 7 days. In Experiment 2, 25 trials were carried out between 21 January and 21 June, 2007, with each trial lasting for 8 days. On the morning prior to the first day of each trial, fish were chosen and randomly assigned to a density treatment, anaesthetized using clove oil (Keene et al. 1998), and the initial fork length (± 0.1 cm) and mass (± 0.0001 g) were recorded. Each fish received a unique mark by subcutaneous injection of a small amount of fluorescent red, green, or yellow elastomer in either the dorsal or caudal fin rays (Dewey and Zigler 1996), which allowed us to monitor the movements of focal animals in the stream channel. The fish were allowed 15-20 minutes to recover from anaesthesia before being released into a small mesh enclosure (diameter = 18.0 cm) within the stream channel. After 2 hours, the enclosure was removed, and the fish were able to enter the main body of the stream channel. A 6-hour conditioning period then followed, during which a small amount of food (approx. 1% of fish biomass) was spread on the belt feeders to allow fish to acclimate to the method of food delivery. New test fish were selected for each trial and there was no mortality during any trial for either of the experiments in this study.
Beginning the next day, Day 1 of each trial, scan samples were conducted 3 times per day at approximately 10:00, 13:00, and 16:00 h; the position of each fish in each stream channel was recorded on a schematic map. On Days 6 or 7 of each trial in Experiment 1, we conducted 20-minute observations of territory size and space-use of 4 focal fish for each density treatment. Because the rates of aggression were not extremely high in Experiment 1, we collected more information about space use for each focal fish in Experiment 2. Hence, one 15-minute observation was conducted on day 4 of each trial, and one 30-minute observation was conducted on Days 7 or 8. At the conclusion of each trial, all fish were captured, and the final fork length and mass were recorded to determine % specific growth rate (100 × [loge final weight of fish – loge initial weight of fish] · trial length in days-1).
During observations, the grid of river stones was used to estimate the relative position of fish within the channel (± 5.1 cm horizontally, and ± 7.1 cm vertically). The location of each foraging station, and the direction (1-12 o’clock, with 12 o’clock as directly upstream) and distance (in body lengths) of aggressive acts initiated from each foraging position (see Steingrímsson and Grant 2008) were recorded on a digital voice recorder. We defined foraging stations as locations where the fish held its position against the current for at least 5 seconds. All of the aggression recorded in this study was either nipping, when one fish bites another, charges, when a fish swims quickly and directly at an intruder usually ending with a nip, or chases, when a fish repeatedly charges while attempting to nip at the tail of the retreating fish (Keenleyside and Yamamoto 1962).
Quantifying Space Use
In Experiment 1, we focused primarily on the dominant fish in each trial because too few subordinates engaged in aggressive behaviour. To estimate the territory size of each focal fish, we first calculated the mean aggressive radius for each dominant fish in Experiment 1, and for each focal fish in Experiment 2 that exhibited agonistic behaviour. Aggressive radius was the distance between the defender and the intruder when the former initiated the attack. These defended areas overlapped little and were stable over time.
To provide an independent estimate of the tank area successfully monopolized by the dominant fish in Experiment 2, we applied the minimum convex polygon (MCP) method to the coordinates of the positions of all subordinate fish obtained from the daily scan samples (i.e. 24 for each fish). Digital maps of the stream channels and territories of each focal fish were created using ArcView GIS 3.2, in conjunction with the Animal Movement Extension (Hooge and Eichenlaub 2000). Each stream channel was treated as a simple x-y coordinate system with the 0, 0 (x, y) position in the downstream, left corner of the stream channel.
Prior to analysis, all continuous variables were subjected to one-sample Kolmogorov-Smirnov tests for normality; all variables met the assumptions for parametric tests. Because each tank contained one dominant individual that defended a relatively large portion of the stream channel compared to other fish, two-way ANOVAs were used to determine whether there was a main effect of either dominance status (dominant versus subordinate) or density treatment on patterns of space use and growth rate among test fish. Average water temperature was also added as a covariate in all analyses. However, there was no significant effect of water temperature on any variable except specific growth rate in Experiment 2, so we reported the results of the two-way ANOVAs only, except in the section on growth rate where we reported the results of the two-way ANCOVA. In instances where aspects of space use for dominant or subordinate fish were considered separately, one-way ANOVAs were used. Finally, we used quadratic and piecewise regression models to investigate whether the territory size of focal fish leveled off with increasing population density, and used Akaike’s Information Criterion (AIC) to determine the best-fit model. All statistical analyses were performed using SPSS v.19 statistical software with the exception of AIC model selection which was carried out using R v.2.14.0. (R Development Core Team 2011).
At the beginning of all trials in both experiments, fish initially congregated at the downstream end of the stream channel. Within 24 hours, a dominant fish moved forward in the stream channel and adopted a central upstream station, directly downstream of the feeder, and initiated the majority of forages and chases from this position. Depending on fish density, subordinates began adopting positions further upstream soon after. In general, it took 1-3 days for fish behaviour to stabilize within the stream channels. On average, dominant fish were 25.0% heavier than the average subordinate in the tank at the beginning of trials in Experiment 1 (two-way ANOVA: F1, 24 = 5.78, P = 0.024), and 17.5% heavier than the average subordinate in the tank in Experiment 2 (two-way ANOVA: F1, 38 = 1.74, P = 0.20).
Foraging by the dominant fish was concentrated in a forward direction, in the immediate area where food dropped from the feeders. However, the dominant fish defended much of the stream channel, particularly in the low-density treatments, and rarely tolerated subordinate individuals in the extreme upstream end of the stream channel. In the 2 and 4-fish treatments, few chases were observed by the dominant individual, presumably because of low encounter rates with the subordinate fish. However, subordinate fish were still effectively confined to the downstream end of the stream channel by the dominant, which frequently patrolled the tank, periodically chasing subordinates that attempted to move farther upstream. At these densities, dominant individuals easily monopolized available food and space. In the higher-density treatments, dominant fish were initially able to defend a large area of the stream channel as at lower densities. As the trial progressed, however, subordinate individuals gradually spread out and moved upstream, encroaching on the dominant’s foraging area. Dominant individuals still effectively defended the extreme upstream end of the stream channel, but subordinate fish were able to dart in from outside the central feeding area of the dominant to intercept food items as they fell from the feeder.
Experiment 1: total food held constant
Dominant fish were more frequently aggressive than the average subordinate fish (Fig. 1a; two-way ANOVA: F1, 24 = 96.11, P = 0.001). Subordinates did not initiate any aggression in the 4-fish treatment, and for the remaining treatments dominant fish initiated, on average, 6.7 times more chases than the average subordinate. Across density treatments, rates of aggression increased significantly with increasing density (Fig. 1a; two-way ANOVA: F2, 24 = 8.96, P = 0.001). The average aggressive radius of focal fish decreased with increasing density (Fig. 1b; one-way ANOVA: F1, 24 = 47.91, P = 0.001; Linear regression: t1 = -6.92, P = .001). For all density treatments, dominant fish had a higher specific growth rate than subordinate fish over the course of the seven-day trials (Fig. 1c; two-way ANOVA: F 1, 24 = 5.26, P = 0.031). Specific growth rate tended to decrease with increasing density, but the difference was not significant (two-way ANOVA: F2, 24 = 1.98, P = 0.161).
Experiment 2: per capita food held constant
The collective space used by the subordinate fish in Experiment 2 increased with increasing density (Fig. 2a; one-way ANOVA: F4, 19 = 7.89, P = 0.001). A polynomial regression yielded a marginally non-significant quadratic term (Fig. 2a; t1 = -1.98, P = 0.06) after the linear term was included in the model, suggesting that subordinate fish collectively used more space with initial increases in density, but this space use leveled off at about 0.85 m2. We confirmed this result using a piecewise regression model which revealed a significant increase in subordinate space use with increasing density up to the 8-fish treatment (t1 = 5.11, P = 0.001), after which space use remained constant (t1 = -0.27, P = 0.79). The piecewise regression was the best fitting model with an AIC value of -1.88, lower than either the linear (AIC = 2.20) or the quadratic (AIC = 0.080) models.
Dominant fish were more frequently aggressive than the average subordinate fish (two-way ANOVA: F1, 38 = 8.21, P = 0.007), even when one extremely aggressive fish in the 12-fish treatment was not included (Fig. 2b; two-way ANOVA: F1, 37 = 11.39, P = 0.002). Subordinates did not initiate any aggression in the 2-fish treatment, and for the remaining treatments dominant fish initiated, on average, 3.2 times more chases than the average subordinate. Across density treatments, rates of aggression increased significantly with increasing density (Fig. 2b; two-way ANOVA: F4, 37 = 7.21, P = 0.001). The proportion of aggression initiated by the dominant fish, however, decreased with increasing density (Fig. 2c; one-way ANOVA: F3, 18 = 12.83, P = 0.001) before leveling off at 8-fish per tank (Piecewise regression: ≤ 8 fish per tank: t = -3.85, P = 0.003; ≥ 8 fish per tank: t = -0.78, P = 0.452). We confirmed that the piecewise regression was the best fit model with AIC (linear: 5.05; quadratic: 0.90; piecewise: -2.59).
Dominant fish had greater chase distances than subordinate fish (data not shown; two-way ANOVA: F1, 32 = 17.78, P = 0.001) across all trials. The average aggressive radius of focal fish defending territories decreased with increasing density (Fig. 3a; two-way ANOVA: F4, 32 = 8.98, P = 0.001) and appeared to level off at high population densities, as indicated by a significant quadratic term in a polynomial regression ( t1 = 2.16, P = 0.037) and AIC (linear: 361.55; piecewise: 361.22; quadratic: 358.81) Tukey’s post-hoc tests confirmed that the aggressive radius in the 16-fish treatment was not significantly different from either the 12-fish (P = 0.876) or 8-fish treatments (P = 0.389).
Dominant fish had a higher specific growth rate than subordinate fish over the course of the eight-day trials (Fig. 3b; two-way ANOVA: F 1, 38 = 27.84, P = 0.001). Water temperature had a significant positive effect on growth rate (two-way ANCOVA: F1, 37 = 4.64, P = 0.038), but there was still a significant main effect of dominance status after statistically removing the effect of water temperature (two-way ANCOVA: F1, 37 = 30.50, P = 0.001). Growth rate also tended to increase with density (two-way ANCOVA: F4, 37 = 2.87, P = 0.036).
To examine the effect of food abundance on territory size, we compared the aggressive radius for focal fish in the 4, 8, and 12-fish treatments in both experiments. Territory size was greater in Experiment 1 than in Experiment 2, but the result was not significant (Fig. 3c; two-way ANOVA: F1, 48 = 2.69, P = 0.107). As already reported, territory size decreased with increasing density over the range from 4 to 12 fish per tank (Fig. 3c; two-way ANOVA: F2, 48 = 10.90, P = 0.001).
In our study, territory sizes were neither incompressible nor followed the 1/n curve. Rather, territory size decreased with increasing population density, when food density was held constant in two different ways. The lack of a significant difference in territory size between the two experiments was surprising given that food was more abundant in Experiment 2 than in 1. While previous studies indicated a negative relationship between territory size and food availability (Slaney and Northcote 1974; Keeley 2000), our data were consistent with the general finding that the magnitude of decrease in territory size is small, even when food abundance is doubled (Grant et al. 1998; Imre et al. 2004). In Experiment 2, two lines of evidence were consistent with the prediction that territory size approached an asymptotic minimum size of about 0.2m2 as density increased. If salmonid territories are approximately circular (see Keeley and Grant 1995), then the minimum territory radius of 20-30 cm translated into a minimum territory size of 0.13-0.28m2. In addition, the collective space used by subordinates leveled off at about 0.85m2, implying that the dominant monopolized 0.23m2 (i.e. 1.08-0.85). Both curves leveled off between densities of 8 to 16 fish, or 7.4-14.8 fish· m-2.
At low densities in both experiments, dominant fish were despotic, monopolizing most of the usable area of the stream channel, and restricting subordinates to the downstream corners of the channel. Because food items arrived asynchronously, dominant fish were able to use the “free” time available between the arrival of food items to patrol its territory, intimidate subordinates, and discourage them from stealing food (Grant and Kramer 1992; Praw and Grant 1999). As the density of individuals increased, the subordinate fish shared the costs of the dominant’s aggression, were able to more effectively intercept food items, and eventually established a foraging space or even a territory in the stream channel. In Experiment 2, the total food abundance increased with density, so there was less free time available between the arrival of food items for the dominant to spend on aggressive behaviour. The per capita rate of aggression did not peak at intermediate densities as expected (Grant 1993), but increased with increasing population density. This result was likely because the social system shifted from a despotic society, with the dominant individual defending the entire tank, to a territorial mosaic (sensu Keenleyside 1979), where the dominant and several subordinates each defended individual territories. Consistent with this interpretation, the dominant’s share of the aggressive behaviour decreased from 100% in the 2-fish treatment to about 15% in the 16-fish treatment in Experiment 2.
In stream salmonids under experimental conditions, relative competitive ability of individuals is usually determined by body size (Chapman 1962; Noakes 1980), and the most profitable foraging sites are often occupied by dominant individuals that grow faster than subordinates (Chapman 1962; Fausch 1984; Metcalfe et al. 1989). In our experiments, dominant individuals adopted a central upstream station, directly beneath the feeder, where they had priority of access to falling food items. While dominant fish were not always the largest fish initially, they were more aggressive, had a larger chase radius, and a higher growth rate than subordinates. One upstream source of food in this study may have made the longitudinal hierarchy more pronounced than it would be in nature. However, the situation is not entirely un-natural because food primarily originates in the fast upstream riffles for fish in slower raceways and pools (Elliott 1967; Allan and Russek 1985; Nakano 1995).
Even though our experiments were conducted in laboratory stream channels, our results likely have implications for field conditions. The minimum territory size of about 0.2m2 in our study was comparable to those predicted by allometric regressions for similar sized fish in wild populations: 0.08m2 for brown trout (Elliott 1990); 0.28m2 for steelhead trout (Keeley and McPhail 1998); and 0.31m2 for Atlantic salmon (Keeley and Grant 1995). Before the advent of non-invasive tagging techniques, most measurements of territory size in the wild were of sedentary individuals in high density populations. The similarity between these field data and our results suggest that the contiguous territories observed in territorial mosaics in the wild (sensu Keenleyside 1979) are at or near the asymptotic territory size. Using elastomer tags, Steingrímsson and Grant (2008) quantified the territory size of more mobile fish at lower densities; their YOY Atlantic salmon defended the largest territories yet documented (0.9m2), equivalent to a 50-cm territory radius in our study. Hence, the range in territory radii observed in our study encompass the range of territory sizes observed in the wild for juvenile salmonids of about 5cm in length.
Our results suggest that an asymptotic minimum territory size can limit the maximum local density of individuals, independent of food supply. The individuals who were unable to establish territories in our experiments would presumably have lost weight and died in longer-term experiments (e.g. Imre et al. 2004). In laboratory experiments with a small number of confined fish, differences between dominant and subordinate fish are often magnified compared to field conditions (reviewed in Sloman et al. 2008). On the other hand, when fish are free to disperse in the wild, local competition is often manifested in terms of density dependent emigration rather than by mortality (Einum et al. 2006). Hence, it is conceivable that a relatively inelastic minimum territory size might trigger density-dependent dispersal and have profound effects on the population regulation of wild populations.
Our results contribute to the growing literature suggesting that territorial behaviour can limit the density of individuals settling in a habitat in a wide variety of taxa (Adams 2001; Lopez-Sepulcre and Kokko 2005). In most other taxa, however, adult individuals defend relatively permanent all-purpose territories that are insensitive to changes in population density, so that excess non-breeding individuals exist as floaters in the population (Tricas 1989; Newton 1998; Both and Visser 2003; Mougeot et al. 2003). By contrast, data for stream-dwelling salmonids suggest that similar dynamics can occur, even when juveniles defend feeding territories that shift in size and location as the individuals grow (Keeley and Grant 1995). Because the population regulation of juvenile salmonids appears to be primarily density-dependent in freshwater habitats (Elliott 1994; Jonsson et al. 1998), an asymptotic minimum territory size during critical periods of habitat limitation (Elliott 1990) can play an important role in population regulation.
We thank Grant Brown and Jae Woo Kim for logistical support and the Associate Editor and anonymous referees for helpful comments on the manuscript. This study was financially supported by a Discovery Grant to J.W.A.G, and an Undergraduate Student Research Award to M.H.B from the Natural Sciences and Engineering Research Council of Canada.
Adams, E.S. 2001. Approaches to the study of territory size and shape. Annu. Rev. Ecol. Syst. 32: 277-303.
Allan, J.D., and Russek, E. 1985. The quantification of stream drift. Can. J. Fish. Aquat. Sci. 42: 210-215.
Allen, K.R. 1969. Limitations on production in salmonid populations in streams. In Symposium on salmon and trout in streams. Edited by T.G. Northcote. University of British Columbia, Vancouver, British Columbia, Canada. pp. 3-18.
Bohlin, T., Dellefors, C., Faremo, U., and Johlander, A. 1994. The energetic equivalence hypothesis and the relation between population density and body size in stream-living salmonids. Am. Nat. 143: 478-493.
Both, C. and Visser, M.E. 2003. Density dependence, territoriality, and divisibility of resources: from optimality models to population processes. Am. Nat. 161: 326-336.
Cattaneo, F., Lamouroux, N., Breil, P., and Capra, H. 2002. The influence of hydrological and biotic process on brown trout (Salmo trutta) population dynamics. Can. J. Fish. Aquat. Sci. 59: 12-22.
Chapman, D.W. 1962. Aggressive behavior in juvenile coho salmon as a cause of emigration. J. Fish. Res. Board. Can. 19: 1047-1080.
Chapman, D.W. 1966. Food and space as regulators of salmonid populations in streams. Am. Nat. 100: 345-357.
Dewey, M.R., and Zigler, S.J. 1996. An evaluation fluorescent elastomer for marking bluegills in experimental studies. Prog. Fish. Cult. 58: 219-220.
Dill, L.M., Ydenberg, R.C., and Fraser, A.H.G. 1981. Food abundance and territory size in juvenile coho salmon (Oncorhynchus kisutch). Can. J. Zool. 59: 1801-1809.
Einum, S., and Nislow, K.H. 2005. Local-scale, density-dependent survival of mobile organisms in continuous habitats: an experimental test using Atlantic salmon. Oecologia 143: 203-210.
Einum, S., Sundt-Hansen, L., and Nislow, K.H. 2006. The partitioning of density-dependent dispersal, growth and survival throughout ontogeny in a highly fecund organism. Oikos 113: 489-496.
Elliott, J.M. 1967. The food of trout (Salmo trutta) in a Dartmoor stream. J. Appl. Ecol. 4: 59-71.
Elliott, J.M. 1990. Mechanisms responsible for population regulation in young migratory trout, Salmo trutta. III. The role of territorial behaviour. J. Anim. Ecol. 59: 803-818.
Elliott, J.M. 1993. The self-thinning rule applied to juvenile sea trout, Salmo trutta. J. Anim. Ecol. 62: 371-379.
Elliott, J.M. 1994. Quantitative ecology and the brown trout. Oxford University Press, Oxford.
Fausch, K.D. 1984. Profitable stream positions for salmonids: relating specific growth rate to net energy gain. Can. J. Zool. 62: 441-451.
Fretwell, S. D., and Lucas, H. L. 1970. On territorial behavior and other factors influencing habitat distribution in birds. I. Theoretical development. Acta Biotheor. 19:16-36.
Grant, J.W.A. 1993. Whether or not to defend? The influence of resource distribution. Mar. Freshw. Behav. Physiol. 23: 137-153.
Grant, J.W.A. 1997. Territoriality. In Behavioural ecology of teleost fishes. Edited by J.-G.J. Godin. Oxford University Press, New York. pp. 81-103.
Grant, J.W.A., and Kramer, D.L. 1990. Territory size as a predictor of the upper limit to population density of juvenile salmonids in streams. Can. J. Fish. Aquat. Sci. 47: 1724-1737.
Grant, J.W.A., and Kramer, D.L. 1992. Temporal clumping of food arrival reduces its monopolization and defence by zebrafish, Brachydanio rerio. Anim. Behav. 44: 101-110.
Grant, J.W.A., Steingrímsson, S.ó., Keeley, E.R., and Cunjak, R.A. 1998. Implications of territory size for the measurement and prediction of salmonid abundance in streams. Can. J. Fish. Aquat. Sci. 55(S1): 181-190.
Hooge, P.N., and Eichenlaub, B. 2000. Animal movement extension to Arcview, version 2.0. Alaska. Science Center – Biological Science office, U.S. Geological Survey, Anchorage, AK, USA.
Huxley, J.S. 1934. A natural experiment on the territorial instinct. Brit. Birds. 27: 270-277.
Imre, I., Grant, J.W.A., and Keeley, E.R. 2004. The effect of food abundance on territory size and population density of juvenile steelhead trout (Oncorhynchus mykiss). Oecologia 138: 371-378.
Imre, I., Grant, J.W.A., and Cunjak, R.A. 2005. Density-dependent growth of young-of-the-year Atlantic salmon Salmo salar in Catamaran Brook, New Brunswick. J. Anim. Ecol. 74: 508-516.
Jenkins, T.M., Diehl, S., Kratz, K.W., and Cooper, S.D. 1999. Effects of population density on individual growth of brown trout in streams. Ecology 80: 941-956.
Jonsson, N., Jonsson, B., and Hansen, L.P. 1998. The relative role of density-dependent and density-independent survival in the life cycle of Atlantic salmon Salmo salar. J. Anim. Ecol. 67: 751-762.
Keeley, E.R. 2000. An experimental analysis of territory size in juvenile steelhead trout. Anim. Behav. 59: 477-490.
Keeley, E.R., and Grant, J.W.A. 1995. Allometric and environmental correlates of territory size in juvenile Atlantic salmon (Salmo salar). Can. J. Fish. Aquat. Sci. 52: 186-196.
Keeley, E.R., and McPhail, J.D. 1998. Food abundance, intruder pressure, and body size as determinants of territory size in juvenile steelhead trout (Oncorynchus mykiss). Behaviour 135: 65-82.
Keene, J.L., Noakes, D.L.G., Moccia, R.D., and Soto, C.G. 1998. The efficacy of clove oil as an anaesthetic for rainbow trout, Oncorhynchus mykiss (Walbaum). Aquat. Res. 29: 89-101.
Keenleyside, M. H. A. 1979. Diversity and adaptation in fish behaviour. In Zoophysiology, Vol. 11. Edited by M.H.A. Keenleyside. Springer-Verlag, Berlin. pp. 150–174.
Keenleyside, M.H.A., and Yamamoto, F.T. 1962. Territorial behaviour of juvenile Atlantic salmon (Salmo salar L.). Behaviour 19: 139-169.
Kerr, S.J., and Lasenby, T.A. 2000. Rainbow stocking in inland lakes and streams: An annotated bibliography and literature review. Fish and Wildlife Branch, Ontario Ministry of Natural Resources, Peterborough, Ontario.
Lobon-Cervia, J. 2007. Density-dependent growth in stream living brown trout, Salmo trutta L. Funct. Ecol. 21: 117-124.
Lopez-Sepulcre, A. and Kokko, H. 2005. Territorial defense, territory size, and population regulation. Am. Nat. 166: 317-329.
Maynard-Smith, J. 1974. Models in Ecology. Cambridge University Press, Cambridge.
McNicol, R.E., and Noakes, D.L.G. 1984. Environmental influences on territoriality of juvenile brook charr, Salvelinus fontinalis, in a stream environment. Environ. Biol. Fishes 10: 29-42.
Metcalfe, N.B., Huntingford, F.A., Graham, W.D., and Thorpe, J.E. 1989. Early social status and the development of life-history strategies in Atlantic salmon. Proc. R. Soc. Lond. B Biol. Sci. 236: 7-19.
Milner, N.J., Elliott, J.M., Armstrong, J.D., Gardiner, R., Welton, J.S., and Ladle, M. 2003. The natural control of salmon and trout populations in streams. Fish. Res. 62: 111-125.
Mougeot, F. Redpath, S.M., Leckie, F., and Hudson, P.J. 2003. The effect of aggressiveness on the population dynamics of a territorial bird. Nature 421: 737-739.
Nakano, S. 1995. Competitive interactions for foraging microhabitats in a size-structured interspecific dominance hierarchy of two sympatric stream salmonids in a natural habitat. Can. J. Zool. 73: 1845-1854.
Newton, I. 1998. Population limitation in birds. Academic Press, London.
Noakes, D.L.G. 1980. Social behavior in young chars. In Charrs, salmonid fishes of the genus salvelinus. Edited by E.K. Balon. Dr. W. Junk Publishers, The Hague. pp. 683-701.
Petit, L.J., and Petit, D.R. 1996. Factors governing habitat selection in prothonotary warblers: field tests of the Fretwell-Lucas models. Ecol. Monogr. 66: 367-387.
Praw, J.C., and Grant, J.W.A. 1999. Optimal territory size in the convict cichlid. Behaviour 136: 1347-1363.
R Development Core Team. 2011. R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. Available from http://www.R-project.org.
Rodenhouse, N.L., Sherry, T.W., and Holmes, R.T. 1997. Site-dependent regulation of population size: a new synthesis. Ecology 78: 2025-2042.
Schoener, T.W. 1983. Simple models of optimal territory size: a reconciliation. Am. Nat. 121: 608-629.
Slaney, P.A., and Northcote, T.G. 1974. Effects of prey abundance on density and territorial behaviour of young rainbow trout (Salmo gairdneri) in a laboratory stream channel. J. Fish. Res. Board Can. 31: 1201-1209.
Sloman, K.A., Baker, D., Winberg, S., and Wilson, R.W. 2008. Are there physiological correlates of dominance in natural trout populations? Anim. Behav. 76: 1279-1287.
Steingrímsson, S.ó. and Grant, J.W.A. 2008. Multiple central-place territories in wild young-of-the-year Atlantic salmon (Salmo salar). J. Anim. Ecol. 77: 448-457.
Teichart, M.A.K., Foldvik, A., Forseth, T., Ugedal, O., Einum, S., Finstad, A.G., Hedger, R.D., and Bellier, E. 2011. Effects of spawning distribution on juvenile Atlantic salmon (Salmo salar) density and growth. Can. J. Fish. Aquat. Sci. 68: 43-50.
Tricas, T.C. 1989. Determinants of feeding territory size in the corallivorous butterfly fish, Chaetodon multicinctus. Anim. Behav. 37: 830-841.
Utz, R.M., and Hartman, K.J. 2009. Density-dependent individual growth and size dynamics of central Appalachian brook trout (Salvelinus fontinalis). Can. J. Fish. Aquat. Sci. 66: 1072–1080.
Wilson, E.O. 1975. Sociobiology: The new synthesis. Harvard University Press.
Figure 1. Mean (± SE, N=5) a) aggressive rate of dominant fish (●) and the average subordinate fish (■), b) aggressive radius of all focal fish, and c) specific growth rate of dominant (●) and the average subordinate fish (■) in Experiment 1.
Figure 2. Mean (± SE, N=5) a) 95% minimum convex polygon of area used by all subordinate fish in each trial, b) aggressive rate of dominant fish (●) and the average subordinate fish (■), and c) proportion of aggression initiated by dominant fish in Experiment 2. The dotted lines in a) and c) are the piecewise regression models (Y = 0.097X + 0.12 (X≤8) - 0.0046X + 0.85 (X ≥8), F3,20 = 7.80, r2 = 0.47, P = 0.001) and (Y = -0.11X + 1.07 (X≤8) - 0.011X + 0.36 (X ≥8), F3,18 = 12.83, r2 = 0.63, P = 0.001), respectively.
Figure 3. Mean (± SE, N=5) a) aggressive radius of all focal fish in Experiment 2, b) specific growth rate of dominant (●) and the average subordinate fish (■) in Experiment 2 and c) aggressive radius of focal fish in Experiment 1 (■) and for the 4, 8, and 12 fish treatments in Experiment 2 (●).