Journal of Fish Biology (2011) 79, 533–538 doi:10.1111/j.1095-8649.2011.03029.x, available online at wileyonlinelibrary.com
Behavioural responses of prey fishes to habitat complexity and predation risk induce bias in minnow trap catches A. Dupuch*†, Y. Paradis*‡ and P. Magnan* *Centre de recherche sur les interactions bassins versants, e´ cosyst`emes aquatiques (RIVE), Universit´e du Qu´ebec a` Trois-Rivi`eres, C.P. 500, Trois-Rivi`eres, QC, G9A 5H7 Canada (Received 10 February 2011, Accepted 6 May 2011) The effects of predation risk and habitat complexity on the efficiency of minnow traps to catch northern redbelly dace Chrosomus eos in laboratory experiments were investigated. Trap efficiency significantly decreased in the presence of vegetation and predators. These results suggest that the © 2011 The Authors various antipredator behaviours used by prey fishes can affect trap efficiency. Journal of Fish Biology © 2011 The Fisheries Society of the British Isles
Key words: antipredator behaviour; Chrosomus eos; fish trapping data; sampling bias; sheltering; shoaling.
Traps are among the oldest fishing techniques used to assess fish abundance and spatial distribution and are used in a variety of habitats (Jackson & Harvey, 1997; Robichaud et al., 2000; Layman & Smith, 2001). An underlying assumption when comparing trap catches either spatially or temporally is that their efficiency (i.e. proportion of individuals caught by the trap relative to population size) is constant under different environmental conditions (Rozas & Minello, 1997). Trap catches, however, have been reported to depend on several factors related to trap location, such as water depth (Blaustein, 1989; Magnan, 1991) and habitat complexity (Robichaud et al., 2000; Layman & Smith, 2001). Even though external factors can influence catches, traps might be the only method available to make effective studies of fish populations and understanding factors influencing their efficiency is essential. The range of behavioural decisions that could influence fishes to enter a trap is complex (Robichaud et al., 2000). For prey fishes that occur sympatrically with piscivorous predators, predation risk may be the major force governing behavioural decisions (Lima & Dill, 1990; Lima, 1998). For example, it is well known that prey fishes use structurally complex habitats to reduce both encounter rates with predators and capture efficiency of predators (Savino & Stein, 1989; Persson & Ekl¨ov, 1995). †Author to whom correspondence should be addressed at present address: NSERC-Laval University Industrial Research Chair in Silviculture and Wildlife, Department of Biology, Universit´e Laval, 1045 Avenue de la M´edecine, Pavillon Alexandre-Vachon, Qu´ebec, QC, G1V 0A6 Canada. Tel.: +1 418 656 2131 poste 6582; email:
[email protected] ‡Present address: Minist`ere des Ressources naturelles et de la Faune du Qu´ebec, Service de la faune aquatique, 880 Chemin Ste-Foy, Qu´ebec, QC, G1S 4X4 Canada
533 © 2011 The Authors Journal of Fish Biology © 2011 The Fisheries Society of the British Isles
534
A. DUPUCH ET AL.
This suggests that prey fishes might be more attracted to the trap when predation risk is high because it provides a physical structure in the environment (Layman & Smith, 2001). Furthermore, traps placed in less complex habitats may be more attractive than those placed in highly complex habitats (Robichaud et al., 2000). Thus, it is possible that local habitat complexity combined with the level of predation risk will influence trap attractiveness (and so trap efficiency) through sheltering behaviour of prey fishes. When under risk of predation, however, prey fishes can adopt various antipredator tactics, not only sheltering (Brown & Warburton, 1997). Shoaling and reducing activity levels are antipredator behaviours that are widely observed in prey fishes (Godin, 1997). Consequently, both habitat complexity and predation risk may induce various antipredator behaviours in prey fishes with consequences for trap efficiency. In this study, an experimental approach was used to determine whether predation risk and the structural complexity of the habitat would affect minnow trap efficiency. Northern redbelly dace Chrosomus eos (Cope 1861) [synonym of Phoxinus eos (Cope 1861); Strange et al., 2009] was used as a model species for three reasons. First, this species is prey for several piscivorous predators such as brook trout Salvelinus fontinalis (Mitchill 1814) and adult creek chub Semotilus atromaculatus (Mitchill 1818) in Canadian Shield lakes (Lacasse & Magnan, 1992; Dupuch et al., 2009a). Second, field and laboratory studies have shown that C. eos avoid predator-rich habitats and preferentially use structured habitats to reduce their predation risk (Naud & Magnan, 1988; Dupuch et al., 2009a, b). Third, C. eos are capable of adapting the intensity of their antipredator response to the level of predation risk (Dupuch et al., 2004). Both C. eos and adult S. atromaculatus used in the experiments were collected from Lake de la Grosse in the Mastigouche Reserve (Qu´ebec, Canada; 46◦ 40 N; 73◦ 20 W) with baited minnow traps and fyke nets and transferred to the laboratory of the Research Group on Aquatic Ecosystems (located at the Universit´e du Qu´ebec a` Trois-Rivi`eres). Salvelinus fontinalis were raised from eggs to adults in the laboratory. One thousand C. eos (mean ± s.d. total length, LT , = 6·4 ± 0·5 cm), 20 S. atromaculatus and 60 S. fontinalis were used in the experiments. For each species (C. eos, S. atromaculatus and S. fontinalis), fish were kept in two holding tanks (each containing about half of the individuals) at 14◦ C, range ± 1◦ C and under a 12L:12D regime. Fishes were fed ad libitum once a day with commercial trout pellets (a mixture of Corey Aquaculture 0.5 GR and 1.0 GR for C. eos; Corey Aquaculture 1.5 GR for S. atromaculatus and S. fontinalis www.coreyaqua.ca). In the laboratory, minnow trap efficiency was estimated for different conditions of habitat structural complexity (presence or absence of artificial vegetation) and predation risk (no predators, with S. atromaculatus or with S. fontinalis), resulting in a 2 × 3 factorial design. Fifteen replicates were conducted for each treatment (a total of 90 trials), and trials were randomly performed in two similar experimental tanks (300 cm × 135 cm × 40 cm). An opaque curtain was placed around each experimental tank to reduce disturbance. Experimental fish of each species were randomly taken from one of the two holding tanks and assigned to a treatment level and experimental tank. After each trial, experimental fishes were returned to their original holding tank. No individual could be subjected to more than one trial in 2 days because the holding tanks from which the experimental fishes were selected were alternated. A group of 30 C. eos was transferred to the experimental tank 1 h © 2011 The Authors Journal of Fish Biology © 2011 The Fisheries Society of the British Isles, Journal of Fish Biology 2011, 79, 533–538
S A M P L I N G B I A S I N D U C E D B Y B E H AV I O U R A L R E S P O N S E S
535
before each trial to allow fish to acclimate to the system. The probability of the 90 groups of C. eos being composed of exactly the same individuals was negligible because of the randomized selection of fish (i.e. the group of fish is the experimental unit). A Gee minnow trap (42 cm long by 22·5 cm in diameter, with 10 mm mesh and a 25 mm diameter opening; Dynamic Aqua-Supply Ltd, model MT14-BV; http://www.dynamicaqua.com/streamsampling.html/) was placed in the centre of the tank in which the artificial vegetation and predators were present or absent. Vegetation, when present, was made up of 35 plastic plants that mimic Myriophyllum spp. (mean height = 40 cm) that were evenly placed in the experimental tank. Different conditions of predation risk were created by introducing two S. atromaculatus (LT range = 16–18 cm) or two S. fontinalis (LT range = 22–25 cm) into the experimental tank after the C. eos acclimation period and just before introduction of the trap to the tank. Predation risk on C. eos was considered as different in the presence of S. atromaculatus than in the presence of S. fontinalis due to (1) their differential body size and potentially predatory behaviour and (2) the fact that prey fishes are rarely found in the stomach contents of wild adult S. atromaculatus (Magnan & Fitzgerald, 1984) but can represent >40% of the diet of S. fontinalis (Lacasse & Magnan, 1992). In experimental tanks, C. eos were exposed to the following conditions of predation risk: no predators, presence of S. atromaculatus and presence of S. fontinalis. The minnow trap was set at the water surface during the C. eos acclimation period; the bottom of the trap was submerged, but both openings were kept out of the water to prevent C. eos from entering. After the acclimation period, the trap was gently lowered into the water in the centre of the tank. The submersion of the trap caused little or no disturbance to the fish. The number of C. eos in the trap was counted 1 h after its introduction to the experimental tank. In experiments with predators, C. eos outside the trap were also counted to determine the number (if any) eaten by S. fontinalis or S. atromaculatus. Trap efficiency was expressed as the proportion of C. eos in the trap relative to C. eos abundance in the experimental tank (between 26 and 30 individuals depending on the number of predation events during the trial). Trials in which S. fontinalis or S. atromaculatus stayed motionless in the experimental tank (three of 60 trials) or trials where the number of predation events was higher than four (to keep C. eos density relatively similar among trials; two of 60 trials) were not taken into account and replaced by successful ones. A two-way ANOVA was performed to determine the effects of predation risk (0 = no predators; 1 = presence of S. atromaculatus; 2 = presence of S. fontinalis), habitat structural complexity (0 = open-water habitat; 1 = structured habitat), and their interaction on minnow trap efficiency. Mean trap efficiencies were normalized using arcsine square-root transformations and then compared with a Bonferroni post hoc multiple comparisons test. Preliminary experiments allowing only visual contact between prey and predator showed that C. eos did not react to predator presence. It was thus necessary that C. eos and predators physically encounter each other to test the hypotheses. To reduce the number of predations during experiments (Huntingford, 1984), C. eos exposure to predators was limited to 1 h. This research received prior approval from the Animal Care Committee of the Universit´e du Qu´ebec a` Trois-Rivi`eres (protocol approval number #2006-05-24-026-04-S-P). © 2011 The Authors Journal of Fish Biology © 2011 The Fisheries Society of the British Isles, Journal of Fish Biology 2011, 79, 533–538
536
A. DUPUCH ET AL.
Minnow trap efficiency
1·2
c
d
c
ab
ad
b
1·0 0·8 0·6 0·4 0·2 0·0 No predators
With S. atromaculatus
With S. fontinalis
Fig. 1. Minnow trap efficiency (mean ± s.d. proportion of Chrosomus eos in the trap) under different conditions of predation risk (no predators, presence of Semotilus atromaculatus, or presence of Salvelinus fontinalis) and habitat structural complexity ( , without artificial vegetation; , with artificial vegetation). Means (calculated from 15 data) with different lower case letters are significantly different (Bonferroni post hoc multiple comparisons test; P < 0·05).
On average, predation events occurred in 37% of the trials with predators (i.e. 22 of 60 trials). The occurrence of a predation event and the total number of predation events were lower in the presence of S. atromaculatus (trials with predation events: 26%; mean ± s.d. number of predation events per trial = 1·2 ± 0·4) than in the presence of S. fontinalis (trials with predation events: 46%; mean ± s.d. number of predation events per trial = 1·8 ± 1·0), suggesting that predation risk was higher with S. fontinalis than with S. atromaculatus. Minnow trap efficiency was significantly higher in the absence than in the presence of vegetation (partial r 2 = 0·35, F1,84 = 88·33, P < 0·001; Fig. 1). On average, 76% of C. eos entered the trap when vegetation was absent compared to 27% in the presence of vegetation. Minnow trap efficiency was also significantly higher without predators and in the presence of S. atromaculatus than with S. fontinalis (partial r 2 = 0·29, F2,84 = 38·92, P < 0·001; Fig. 1). Sixty seven per cent of C. eos entered the trap in the absence of predators and in the presence of S. atromaculatus compared to 21% in the presence of S. fontinalis. Furthermore, the decrease in trap efficiency under predation risk was higher in the presence than in the absence of vegetation (the predation risk × vegetation term was significant; partial r 2 = 0·03, F2,84 = 3·90, P < 0·05; Fig. 1). In the absence of vegetation, minnow trap efficiency did not decrease in the presence of S. atromaculatus but was 2·5 times lower in the presence of S. fontinalis than in the absence of predators (Fig. 1). In the presence of vegetation, trap efficiency was three times and 19·6 times lower in the presence of S. atromaculatus and S. fontinalis, respectively, than in the absence of predators (Fig. 1). This study shows that both habitat complexity and predation risk can affect minnow trap efficiency. Trap efficiency was lower when vegetation was present than absent, supporting the hypothesis that prey fishes were less attracted to the trap in structured than in unstructured habitats. Minnow trap efficiency also decreased in the presence of predators, suggesting a negative correlation between trap attractiveness (i.e. refuge use) and predation risk in C. eos. This behaviour, also observed in the tethering field experiment by Dupuch et al. (2009b) in seven Canadian Shield lakes, © 2011 The Authors Journal of Fish Biology © 2011 The Fisheries Society of the British Isles, Journal of Fish Biology 2011, 79, 533–538
S A M P L I N G B I A S I N D U C E D B Y B E H AV I O U R A L R E S P O N S E S
537
is surprising compared to the general observation that refuge use by prey increases with predation risk (Lima & Dill, 1990; Lima, 1998). This could be explained by the flexibility in antipredator behaviour of C. eos. Dupuch et al. (2004) showed that both shoaling and freezing behaviours increased with the level of predation risk. Furthermore, Pink et al. (2007) showed that C. eos did not increase their use of vegetation in the littoral zone after the introduction of S. fontinalis in lakes but significantly increased their shoal size. The preference for shoaling over sheltering in response to increasing predation risk was also observed in other fish species (Brown & Warburton, 1997; Jacobsen & Berg, 1998) and could explain the negative relationship observed between refuge (i.e. submerged vegetation or trap) use and the level of predation risk in the field (Dupuch et al., 2009b) and in the laboratory experiments. Although not directly measured during the experiment, results suggest that when predation risk was high (i.e. in the presence of S. fontinalis), C. eos preferred to use an alternative tactic (shoaling) instead of hiding in a trap (A. Dupuch & Y. Paradis; pers. obs.). Further experiments need to be conducted in order to ascertain behavioural mechanisms producing the observed patterns. These results also showed that minnow trap efficiency decreased in the presence of predators in both habitat types, but the decrease was significantly more pronounced in structured than in unstructured habitats. This effect, while interesting and significant, was weak compared to the main effects of predation risk and presence of vegetation. This study, showed that the effects of habitat complexity and predation risk on trap efficiency may result in biased estimates of fish abundance. Consequently, comparisons of minnow trap catches between different sampling sites that vary in habitat complexity and predation risk could lead to a biased picture of the spatial distribution of fishes. The results of this study stress the importance of assessing trap efficiency under various environmental conditions before making inferences from direct trap counts. We thank V. Boily, C. Duchesne, C. Fournier, S. Fradette and B. Jacob for their assistance at various stages of this research and A. Bertolo, M. Bertrand, K. Turgeon, D. Kramer and L. Devine for their comments on earlier versions of this paper. This work was supported by grants from the Natural Sciences and Engineering Research Council of Canada and the Canada Research Chair Program to P.M.
References Blaustein, L. (1989). Effects of various factors on the efficiency of minnow traps to sample mosquitofish (Gambusia affinis) and green sunfish (Lepomis cyanellus) populations. Journal of the American Mosquito Control Association 5, 29–35. Brown, C. & Warburton, K. (1997). Predator recognition and anti-predator responses in the rainbowfish Melanotaenia eachamensis. Behavioral Ecology and Sociobiology 41, 61–68. Dupuch, A., Magnan, P. & Dill, L. M. (2004). Sensitivity of northern redbelly dace, Phoxinus eos, to chemical alarm cues. Canadian Journal of Zoology 82, 407–415. Dupuch, A., Dill, L. M. & Magnan, P. (2009a). Testing the effects of inherent habitat riskiness and resource distribution on the simultaneous habitat selection of predators and prey. Animal Behaviour 78, 705–713. Dupuch, A., Magnan, P., Bertolo, A., Dill, L. M. & Proulx, M. (2009b). Does predation risk influence habitat use by northern redbelly dace Phoxinus eos at different spatial scales? Journal of Fish Biology 74, 1371–1382. Godin, J.-G. J. (1997). Evading predators. In Behavioural Ecology of Teleost Fishes (Godin, J.-G. J., ed.), pp. 191–236. Oxford: Oxford University Press. © 2011 The Authors Journal of Fish Biology © 2011 The Fisheries Society of the British Isles, Journal of Fish Biology 2011, 79, 533–538
538
A. DUPUCH ET AL.
Huntingford, F. A. (1984). Some ethical issues raised by studies of predation and aggression. Animal Behavior 32, 210–215. Jackson, D. A. & Harvey, H. H. (1997). Qualitative and quantitative sampling of lake fish communities. Canadian Journal of Fisheries and Aquatic Sciences 54, 2807–2813. Jacobsen, L. & Berg, S. (1998). Diel variation in habitat use by planktivores in field enclosure experiments: the effect of submerged macrophytes and predation. Journal of Fish Biology 53, 1207–1219. Lacasse, S. & Magnan, P. (1992). Biotic and abiotic determinants of the diet of brook trout, Salvelinus fontinalis, in lakes of the Laurentian Shield. Canadian Journal of Fisheries and Aquatic Sciences 49, 1001–1009. Layman, C. A. & Smith, D. E. (2001). Sampling bias of minnow traps in shallow aquatic habitats on the Eastern Shore of Virginia. Wetlands 21, 145–154. Lima, S. L. (1998). Stress and decision making under the risk of predation: recent developments from behavioral, reproductive, and ecological perspectives. Stress and Behavior 27, 215–290. Lima, S. L. & Dill, L. M. (1990). Behavioral decisions made under the risk of predation: a review and prospectus. Canadian Journal of Zoology 68, 619–640. Magnan, P. (1991). Unrecognized behavior and sampling limitations can bias field data. Environmental Biology of Fishes 31, 403–406. Magnan, P. & Fitzgerald, G. J. (1984). Ontogenetic changes in diel activity, food habits and spatial distribution of juvenile and adult creek chub, Semotilus atromaculatus. Environmental Biology of Fishes 11, 301–307. Naud, M. & Magnan, P. (1988). Diel onshore offshore migrations in northern redbelly dace, Phoxinus eos (Cope), in relation to prey distribution in a small oligotrophic lake. Canadian Journal of Zoology 66, 1249–1253. Persson, L. & Ekl¨ov, P. (1995). Prey refuges affecting interactions between piscivorous perch and juvenile perch and roach. Ecology 76, 70–81. Pink, M., Fox, M. G. & Pratt, T. C. (2007). Numerical and behavioural response of cyprinids to the introduction of predatory brook trout in two oligotrophic lakes in northern Ontario. Ecology of Freshwater Fish 16, 238–249. Robichaud, D., Hunte, W. & Chapman, M. R. (2000). Factors affecting the catchability of reef fishes in antillean fish traps. Bulletin of Marine Science 67, 831–844. Rozas, L. P. & Minello, T. J. (1997). Estimating densities of small fishes and decapod crustaceans in shallow estuarine habitats: a review of sampling design with focus on gear selection. Estuaries 20, 199–213. Savino, J. F. & Stein, R. A. (1989). Behavioural interactions between fish predators and their prey: effects of plant density. Animal Behaviour 37, 311–321. Strange, R. M. & Mayden, R. L. (2009). Phylogenetic relationships and a revised taxonomy for North American cyprinids currently assigned to Phoxinus (Actinopterygii: Cyprinidae). Copeia 2009, 494–501.
© 2011 The Authors Journal of Fish Biology © 2011 The Fisheries Society of the British Isles, Journal of Fish Biology 2011, 79, 533–538
