Present address: Department of Limnology, Evolutionary Biology Center, Uppsala University, Norbyv. 20, SE-752 36 Uppsala, Sweden.
Morphology in perch affects habitat specific feeding efficiency
Article first published online: 20 JUL 2004
Volume 18, Issue 4, pages 503–510, August 2004
How to Cite
SVANBÄCK, R. and EKLÖV, P. (2004), Morphology in perch affects habitat specific feeding efficiency. Functional Ecology, 18: 503–510. doi: 10.1111/j.0269-8463.2004.00858.x
- Issue published online: 20 JUL 2004
- Article first published online: 20 JUL 2004
- Received 22 September 2003; revised 27 November 2003; accepted 19 December 2003
- Attack behaviour;
- foraging efficiency;
- search behaviour;
- trophic polymorphism
- 1Trophic polymorphism is a common phenomenon in many species. Trade-offs in foraging efficiency on different resources are thought to be a primary cause of such polymorphism.
- 2To test for a trade-off in foraging efficiency perch (Perca fluviatilis L.) were used from a population that differs in morphology between the littoral and pelagic habitat of a lake. Indoor aquarium experiments were performed with three different prey types in two different environments. It was predicted that the morphology of the individual would affect foraging efficiency in the different environments and on the different prey types through search and attack behaviour.
- 3Overall the foraging efficiency of perch was found to be related to individual morphology. A connection was also found between individual morphology and search and attack behaviour. Search behaviour but not attack behaviour was affected by the structure in the aquaria. Furthermore our results show that there are relations between search behaviour and detection rates and between attack behaviour and attack success.
- 4Our results give a mechanistic explanation for the differences in foraging efficiency between littoral and pelagic perch. These differences are probably driven by a functional trade-off between foraging performance and general body form.
Trade-offs in foraging on different food resources may lead to phenotypically divergent populations and are hypothesized to be a primary cause of trophic polymorphisms and adaptive radiations (e.g. Ehlinger & Wilson 1988; Schluter 1993; Skúlason & Smith 1995; Robinson, Wilson & Shea 1996; Svanbäck & Eklöv 2003). In fish it has been proposed that individuals that forage in the open water (pelagic zone) searching for widely distributed and conspicuous prey should use a high search rate for which a more streamlined body is better (Andersson 1981; Gendron & Staddon 1983, 1984; Webb 1984; Gendron 1986; Ehlinger 1990). On the other hand, foraging in structurally complex habitats such as the vegetated littoral zone and searching for more cryptic prey requires a lower search rate and higher manoeuvrability (Andersson 1981; Gendron & Staddon 1983, 1984; Gendron 1986; Ehlinger 1989). Deeper body morphology has been proposed to be better for high manoeuvrability (Webb 1984; Ehlinger & Wilson 1988; Ehlinger 1990). Thus a trade-off in morphology and search behaviour between the littoral and pelagic habitats is expected. Corresponding to this, a resource polymorphism has been observed with individual fish occupying the pelagic zone being more streamlined than individuals occupying the littoral zone (see Skúlason & Smith 1995 for review).
Differences in phenotype can also be related to variation in attack behaviour (Norton 1991, 1995). For example, foraging on attached prey requires more suction and manoeuvrability in the attack, for which a deeper body is optimal, whereas foraging on more elusive prey requires a higher attack speed, for which a more slender body has been hypothesized to be optimal (Norton 1991, 1995).
In previous studies of perch (Perca fluviatilis L.) we have found that individuals in the littoral zone of lakes are deeper bodied than individuals in the pelagic zone, though there is an overlap in morphology between the habitats (Svanbäck & Eklöv 2002, 2003). The morphology of individual perch has been found to be correlated with diet and growth of perch both within and between habitats (Svanbäck & Eklöv 2002, 2003). Svanbäck & Eklöv (2003) found that for individuals caught in the littoral zone, a deeper body was associated with higher growth whereas for the individuals in the pelagic zone, a more streamlined body was associated with higher growth (see also Hjelm et al. 2001). Furthermore, the foraging efficiency on littoral and pelagic prey types has been shown to differ between perch from the two habitats (Svanbäck & Eklöv 2003). In this study we used the results of Svanbäck & Eklöv (2003) and ask what mechanisms lead to differences in foraging efficiency between littoral and pelagic perch. In particular, we relate the individual foraging components search velocity, attack velocity, detection rate and attack success to morphology. Thus we investigate the whole foraging cycle from searching to capturing prey, i.e. (1) how morphology affects search behaviour and how that in turn affects detection rate, and (2) when the prey is detected how the morphology affects attack behaviour and how that in turn affects the attack success.
We carried out aquarium experiments during late summer 1999. We caught the fish in Lake Trehörningen (situated in central Sweden, latitude: 64°00′50″ N, longitude: 20°08′00″ E) by angling with barb-free hooks (pelagic and littoral fish) and by seine netting (littoral fish). We kept the fish from the two habitats separate in 500-l holding tanks with flow-through water before and between trials. Before experiments, we allowed the fish to acclimatize to indoor conditions for at least 2 weeks. We used two size classes of perch in the experiments: small (128·2 ± 0·7 mm, N = 48, mean ± SE) and large (171·7 ± 1·0 mm, N = 48, mean ± SE). There was no difference in size between the littoral and the pelagic perch within each size class (t-test: small; t46 = 0·107, P = 0·916, large; t46 = 0·753, P = 0·455).
We carried out the experiments in four 600-l glass aquaria (200 × 60 × 50 cm) with a water depth of 45 cm. We placed fluorescent tubes (2 × 30 W) 40 cm above the water surface. The background of the aquaria was pale green, the bottom was covered with fine sand and the sides were covered with black plastic sheets. We divided the front of the aquarium with vertical (5 cm apart) and horizontal lines (10 cm apart) and used the grid to estimate swimming velocity and location of the predators. We divided each aquarium into two smaller and one larger compartment by grey PVC-sheets that could slide vertically. We assigned the smaller compartments (1/6 of the aquarium length, on both ends of the aquarium) as holding chambers for the predators and the larger compartment as the experimental chamber. We carried out trials with (vegetation trials) or without (open water trials) artificial vegetation. The artificial vegetation consisted of white polypropylene strings attached to a plastic net (mesh size 10 mm, 200 strings m−2), which we buried in the sand.
We carried out trials with Daphnia sp. (density 4 l−1, length 1·13 ± 0·04 mm, N = 38, mean ± SE), Chaoborus sp. (density 0·2 l−1, length 9·18 ± 0·08 mm, mean ± SE) or mayfly larvae (Cloeon sp.) (density 100 m−2, mean length 4·39 ± 0·09 mm, mean ± SE) as prey. We chose Daphnia sp. and Chaoborus sp. as pelagic prey and Cloeon sp. as littoral prey because they are common prey types of pelagic and littoral perch and they represents distinct resource groups from both habitats (Svanbäck & Eklöv 2002, 2003). We conditioned the perch on the prey type used in a trial for at least a week in advance and allowed them to acclimatize in the aquarium for at least 24 h before a trial. One hour before each trial we transferred the predators to the holding chamber. Because perch are social foragers (Eklöv 1992), we used two individuals in each trial. Each combination of predator, prey and presence or absence of vegetation was replicated at least four times. No fish was used more than once. At the end of the experiment we anaesthetized the perch and deep-froze them for later analysis of their morphology (see below). We followed Swedish guidelines concerning the care and welfare of the experimental fish. Because many of the small perch got infections in the middle of the trial period we terminated the trials on the small perch to avoid changes in morphology due to the infection. Since the infection on the small perch occurred between experimental periods, we are confident that it did not affect the results from the previous trials (mayfly larvae). Therefore, we only have data on all prey types for the larger perch.
We added the prey to the experimental chamber 10–20 min before the predators were released. Each trial was 3 min long and started at the first attack by a predator. We videotaped each trial with an S-VHS video camera and the outcome of attacks (success or failure) was orally recorded by an observer. The videotapes were later analysed for fish behaviour with frame-by-frame analysis of the video recordings (50 frames/s). We quantified attacks (failure or success), capture rate and detection rate.
We calculated the proportions of the attacks that were successful (capture success) from the 10 first strikes made by a predator. We quantified capture rate as the number of prey captured per second in a trial. Detection rate was determined as the number of successful and unsuccessful strikes per second.
Search velocity, measured as swimming velocity while the fish were searching for prey, was determined from the position of each fish every other second, using the fish's eye as the focal point.
The attack velocity was measured from the film frame just prior to the opening of the mouth to the frame when the mouth was closed. We only used attacks that were perpendicular to the camera in the attack velocity measures.
We analysed the morphology of the experimental fish used in this study together with littoral and pelagic perch caught in a field survey from the same population (Svanbäck & Eklöv 2003) to be able directly to compare the results from the field survey with the results from this laboratory study. The morphology was analysed using landmarks captured by a Polhemus 3D-digitizing tablet (Colchester, VT) and Ds-digit (Slice 1994). We digitized 24 landmarks on the left side of each specimen. We used the digitized landmarks to analyse the relative position of each landmark and variation in body shape using TPSRW (Thin-Plate Spline Relative Warp) (Rohlf 1993a). This analysis is a flexible, powerful and interpretable multivariate technique for analysis of morphometric shape variation (Bookstein 1991; Rohlf 1993b; Marcus et al. 1996). We used TPSRW to calculate partial warp and uniform scores of the individuals. The partial warp and uniform scores were analysed with a multivariate discriminant function analysis of all individuals pooled and classified by the habitat in which they were captured. This technique combines all partial warp and uniform scores into a single morphological index (MI) for each fish that maximally discriminates between the two habitats. Shape changes associated with the morphological index were visualized as deformations by using the TPSREGR program (see Rohlf 1993b) to display the regression of the original coordinates on the morphological index. The shape changes associated with the individual morphological index were then used as covariates in ancovas to analyse the influence of perch morphology on the dependent variables.
In trials with Daphnia sp. and Chaoborus sp., more streamlined perch had a higher capture rate than deep-bodied perch in open water trials, whereas the opposite pattern was found in vegetation trials (Fig. 1a,b; Table 1a,b). In trials with mayfly larvae as prey, deeper-bodied perch had higher capture rates than streamlined perch in both open water and vegetation trials. Capture rates of both deep-bodied and streamlined perch were lower in the vegetation trials (Fig. 1c,d; Table 1c,d).
|Capture rate||Search velocity||Attack velocity|
|(a) Daphnia sp.|
|Predator-type * Environment||0·70||3·54||0·31|
|Predator-type * MI||1·31||0·07||0·05|
|Environment * MI||4·90*||3·26||1·04|
|Predator-type * Environment * MI||0·20||0·47||<0·001|
|(b) Chaoborus sp.|
|Predator-type * Environment||2·21||3·61||5·44*|
|Predator-type * MI||0·14||10·38**||1·00|
|Environment * MI||3·41||18·10***||0·26|
|Predator-type * Environment * MI||0·01||5·35*||0·03|
|(c) Small perch on mayfly larvae|
|Predator-type * Environment||0·19||9·67*||0·04|
|Predator-type * MI||0·58||1·66||3·32|
|Environment * MI||0·45||8·27**||0·74|
|Predator-type * Environment * MI||1·23||3·21||1·31|
|(d) Large perch on mayfly larvae|
|Predator-type * Environment||0·38||3·73||0·11|
|Predator-type * MI||0·23||0·10||0·55|
|Environment * MI||0·05||4·27*||0·20|
|Predator-type * Environment * MI||0·001||0·01||0·67|
The search velocity of perch was affected by the habitat the perch was caught in, the structure in the aquarium, and the morphology of the fish (Fig. 2; Table 1). Overall, pelagic perch had a higher search velocity in open water trials than littoral perch whereas the opposite was true for vegetation trials. Independent of which habitat the perch originated from, more streamlined individuals used higher search velocity in open water trials, whereas deeper bodied individuals used higher search velocity in vegetation trials, indicated by the interaction between MI and treatment (Table 1).
Detection rates on different prey were correlated with search velocity. Using Daphnia sp. and Chaoborus sp. as prey, search velocity was positively correlated with detection rate in open water trials (Fig. 3a,b, Daphnia sp. r = 0·599, n = 12, P = 0·040; Chaoborus sp. r = 0·819, n = 11, P = 0·002), whereas there was no correlation between search velocity and detection rate in vegetation trials (Daphnia sp. r = 0·341, n = 12, P = 0·278; Chaoborus sp. r = 0·305, n = 11, P = 0·361). In trials with mayfly larvae as prey there was a negative relationship between search velocity and detection rate in open water trials for both perch sizes (Fig. 3c,d, small perch: r = −0·561, n = 17, P = 0·019; large perch: r =−0·659, n = 16, P = 0·006), whereas there was no correlation between search velocity and detection rate in vegetation trials (small perch: r = 0·295, n = 18, P = 0·234; large perch: r = 0·241, n = 12, P = 0·449).
Littoral and pelagic perch used different attack velocities. Pelagic perch had higher attack velocities on all prey types than littoral perch and attack velocity was not affected by vegetation (Table 1). Independent of which habitat the perch originated from, attack velocity was greater for more streamlined fish on all prey types except on Daphnia sp. (Fig. 4, Table 1).
There was no correlation between attack velocity and capture success on Daphnia sp. since all but two attacks were successful. In the open water trials with Chaoborus sp. there was a positive relation between attack velocity and capture success (correlation; r = 0·889, N = 11, P < 0·001) (Fig. 5a) whereas in vegetation trials the relationship was negative (correlation; r = −0·944, N = 9, P < 0·001). Capture success on mayfly larvae decreased with increasing attack velocity in both open water and vegetation trials for both predator size classes (Fig. 5b,c, correlation; open water small predators; r = −0·681, N = 17, P = 0·003, open water large predators; r = −0·768, N = 16, P < 0·001, vegetation small predators; r = −0·581, N = 18, P = 0·011, vegetation large predators; r = −0·808, N = 10, P = 0·005).
This study showed that deeper-bodied perch were more efficient than more streamlined individuals on all resources in the vegetation. Conversely, the more streamlined individuals, independent of origin, were superior foragers in the open water trials on Daphnia and Chaoborus, suggesting a trade-off between foraging on littoral and pelagic prey types. These results are in agreement with the field patterns of the same population of perch (Svanbäck & Eklöv 2003), suggesting a mechanistic explanation for perch morphology found in the field and habitat specific foraging efficiency.
Similarly, Schluter (1993, 1995) showed that in a species pair complex of sticklebacks (Gasterosteus spp.) the benthic form with a deeper body had higher capture rate and growth on littoral prey compared with the streamlined limnetic form. The limnetic form on the other hand had higher foraging efficiencies and growth on pelagic prey (Schluter 1993, 1995).
Our results suggest that there exists a functional trade-off between general body form and ecological performance in a single population of perch. The trade-off comes from different optimal morphologies for searching and attacking prey in the littoral and pelagic habitats of a lake. In a previous study we found that the more streamlined individuals from the pelagic zone grew faster than deeper-bodied individuals, whereas the opposite pattern was found in the perch from the littoral zone (Svanbäck & Eklöv 2003). Trade-offs in foraging efficiency have been shown to be important in creating intraspecific differences in morphology between resources and habitats (e.g. Smith 1987; Ehlinger 1990; Schluter 1993, 1995). These trade-offs in foraging on littoral and pelagic prey have been hypothesized to depend both on search and attack behaviour (e.g. Webb 1984; Ehlinger 1990; Norton 1991, 1995; Schluter 1993).
morphology and foraging behaviour
A deep body is thought to be connected to high manoeuvrability, a trait that is important when foraging in the vegetated littoral zone for cryptic prey while a more streamlined body is better for foraging for widely dispersed prey in the pelagic zone (Webb 1984; Ehlinger 1990; Schluter 1993). Correspondingly, perch caught in the littoral zone feed more on littoral prey than the pelagic perch, which feed more on pelagic prey types such as Daphnia and copepods (Svanbäck & Eklöv 2002, 2003).
It has been proposed that the detection rate of a predator reflects a trade-off between the rate at which prey are encountered (which increases with search rate) and the probability of detecting an encountered prey item (which decreases with search rate). Given a certain search rate, a conspicuous prey type is easier to detect than a more cryptic prey type. Thus, a lower search rate results in a higher detection rate on cryptic prey, and a higher search rate is connected to a higher detection rate on conspicuous prey (Andersson 1981; Gendron & Staddon 1983, 1984; Gendron 1986; Ehlinger 1989). Correspondingly, we found for perch feeding on Daphnia and Chaoborus that detection rate increased with search rate (velocity) whereas the detection rate declined with search rate when feeding on mayfly larvae. Similarly, Ehlinger (1990) found that Bluegills using shorter hover duration (higher search rate) had higher detection rate on Daphnia compared with Bluegills that used longer hover duration (lower search rate). On damselfly nymphs (Enallagma sp.), on the other hand, the Bluegills that used longer hover duration had a higher detection rate than Bluegills using shorter hover duration. More streamlined perch had higher search velocities in open water and thus had higher detection rates on Daphnia and Chaoborus but a lower detection rate on mayfly larvae. In vegetation trials the pattern was reversed although not significant. These results suggest that the relationship between search velocity and detection rate is also influenced by the complexity of the habitat. Hence, the different search behaviours of the littoral and pelagic perch clearly represent adaptations to searching for conspicuous prey in the pelagic zone and cryptic prey in the littoral zone (see also Svanbäck & Eklöv 2003).
Besides the trade-off in search behaviour on cryptic and conspicuous prey, there is also a trade-off in attack behaviour between elusive and attached prey. Foraging on elusive prey has been hypothesized to require higher attack velocities for which a more slender morphology is hypothesized to be optimal (Webb 1984; Norton 1991, 1995). On the other hand foraging on attached prey requires manoeuvring and more suction (slower attack velocity) in the attack and for such behaviour a deeper body is thought to be optimal (Webb 1984; Norton 1991, 1995). In accordance with these hypotheses we found that the more slender perch used a higher attack velocity than deeper-bodied perch. A higher attack velocity was also correlated with decreased attack success on mayfly larvae and on Chaoborus in trials with vegetation. On the other hand, more successful attacks on Chaoborus in trials without vegetation were correlated with a higher attack velocity. These results suggest that the relationship between attack velocity and attack success can be influenced by the structural complexity of the habitat (see also Svanbäck & Eklöv 2003).
In conclusion, our results give a mechanistic explanation to the littoral and pelagic population differences of perch (Svanbäck & Eklöv 2003) and these differences are probably driven by a functional trade-off between foraging performance and general body form. Such trade-offs in foraging performance are probably the cause of most cases of resource polymorphism.
A special thanks to the Lake Trehörningen's Fish Co-operative for the right to use this lake. Thanks also to Pär Byström, Felipe Gonzalez and Sofia Nilsson for assistance in the field and laboratory and to Göran Arnqvist for helpful advice on morphometrics and statistics. Valuable comments on earlier drafts by Mario Quevedo and two anonymous reviewers greatly improved the manuscript. This research was supported by grants from JA Ahlstrands Fond and Hjerta-Rezius Stipendiefond to Richard Svanbäck and by Swedish Council for Forestry and Agricultural Research to Peter Eklöv.
- 1981) On optimal predator search. Theoretical Population Biology 19, 58–86. (
- 1991) Morphometric Tools for Landmark Data: Geometry and Biology. Cambridge University Press, Cambridge. (
- 1989) Learning and individual variation in bluegill foraging: habitat-specific techniques. Animal Behaviour 38, 643–658. (
- 1990) Habitat choice and phenotype-limited feeding efficiency in bluegill: individual differences and trophic polymorphism. Ecology 71, 886–896. (
- 1988) Complex foraging polymorphism in bluegill sunfish. Proceedings of the National Academy of Science USA 85, 1878–1882. & (
- 1992) Group foraging versus solitary foraging efficiency in piscivorous predators: the perch, Perca fluviatilis, and pike, Esox lucius, patterns. Animal Behaviour 44, 313–326. (
- 1986) Searching for cryptic prey: evidence for optimal search rates and the formation of search images in quail. Animal Behaviour 34, 898–912. (
- 1983) Searching for cryptic prey: the effect of search rate. American Naturalist 121, 172–186. & (
- 1984) A laboratory simulation of foraging behavior: the effect of search rate on the probability of detecting prey. American Naturalist 124, 407–415. & (
- 2001) Diet-dependent body morphology and ontogenetic reaction norms in Eurasian perch. Oikos 95, 311–323. , , , & (
- 1996) Advances in Morphometrics. Plenum Press, New York. , , , & (
- 1991) Capture success and diet of cottid fishes: the role of predator morphology and attack kinematics. Ecology 72, 1807–1819. (
- 1995) A functional approach to ecomorphological patterns of feeding in cottid fishes. Environmental Biology of Fishes 44, 61–78. (
- 1996) Trade-offs of ecological specialization: an intraspecific comparison of pumpkinseed sunfish phenotypes. Ecology 77, 170–178. , & (
- 1993a) TPSRW: Thin-Plate Spline Relative Warp Analysis. Department of Ecology and Evolution. State University of New York, Stony Brook, New York. Free computer software available online at: http://life.bio.sunysb.edu/morph/. (
- 1993b) Relative warp analysis and an example of its application to mosquito wings. Contributions to Morphometrics (eds L.F.Marcus, E.Bello & A.Garcia-Valdescasas), pp. 131–159. Monografias del Museo Nacional de Ciencias Naturales, Madrid. (
- 1993) Adaptive radiation in sticklebacks: size, shape, and habitat use efficiency. Ecology 74, 699–709. (
- 1995) Adaptive radiation in sticklebacks: trade-offs in feeding performance and growth. Ecology 76, 82–90. (
- 1995) Resource polymorphism in vertebrates. Trends in Ecology and Evolution 10, 366–370. & (
- 1994) DS-DIGIT: Basic Digitizing Software. Department of Ecology and Evolution. State University of New York, Stony Brook, New York. Free computer software available online at: http://life.bio.sunysb.edu/morph/. (
- 1987) Bill size polymorphism and intraspecific niche utilization in an African finch. Nature 329, 717–719. (
- 2002) Effects of habitat and food resources on morphology and ontogenetic trajectories in perch. Oecologia 131, 61–70. & (
- 2003) Morphology dependent foraging efficiency in perch: a trade-off for ecological specialization. Oikos 102, 273–284. & , (
- 1984) Body form, locomotion and foraging in aquatic vertebrates. American Zoologist 24, 107–120. (