Organismal fitness is influenced by a complex array of interacting and correlated traits that can be conceptualized as a causal sequence (Arnold, 1983; Walker, 2007; Langerhans, 2009c; Walker, 2010). Most closely related to fitness are aspects of performance, such as sprint speed, bite force, locomotor endurance, song production or foraging ability (Arnold, 1983; Koehl, 1996; Ghalambor et al., 2003; Walker, 2007; Irschick et al., 2008; Langerhans, 2009b). These aspects of performance are then influenced by a variety of interacting and correlated physical traits, including behaviour, physiology and morphology. Divergent selection between environments should lead to divergent selection on different aspects of performance, such as speed vs. endurance, with a by-product being divergence in the physical traits that influence performance. Most studies of adaptive divergence typically focus on these physical traits, most often morphology, but an increasing number are also examining adaptive divergence in aspects of performance (Ghalambor et al., 2003; Irschick et al., 2008; Langerhans, 2009b,c).
It is possible to consider adaptive divergence in an integrated fashion by comparing morphology–performance ‘mapping’ between populations from different environments. To be precise, the relationships between morphology and performance can be visualized as a surface with a multidimensional topology that reflects causal relationships (Koehl, 1996). Given the resulting complexity, the mapping between morphology and performance will likely change as populations diverge in morphological trait space (Emerson et al., 1990; Koehl, 1996), and the main cause of such divergence is expected to be differences in selection. The goals of this study are therefore to compare morphology, performance and morphology–performance mapping between conspecific populations adapting to different environments.
Comparisons of morphology–performance mapping between populations can be complicated by several factors. For instance, recent studies have shown that this mapping can differ between the sexes (McGuigan et al., 2003; Calsbeek, 2008; Herrel et al., 2008; Van Damme et al., 2008) and vary plastically between seasons (Irschick et al., 2006; Irschick & Meyers, 2007). Moreover, even within a given sex and season, variation among individuals in resource acquisition might cause traits and performance to show positive associations that do not have a genetic basis (van Noordwijk & de Jong, 1986). Studies of divergence in morphology–performance mapping would therefore benefit from controlling for such factors (e.g. sex) and from reducing the potential effects of plasticity. This last task can be accomplished by testing morphology–performance associations among individuals that have been reared for their entire lives in a common environment. Such analyses are rarely performed – but they make an important contribution to our understanding of adaptive divergence (Langerhans & Reznick, 2009; Langerhans, 2009b).
Our study employed a common-garden experiment to test for putative genetic divergence between two ecotypes (lake and stream) of threespine stickleback (Gasterosteus aculeatus) in morphology, performance and morphology–performance mapping for males and females. We focused on morphological traits and performance measures related to swimming, foraging and predation risk in the two environments. These are key targets of selection in fishes, and they can diverge between fish with different life styles and from different environments (Beamish, 1978; Webb, 1982, 1984; Domenici & Blake, 1997; Walker, 1997; Blake, 2004; Langerhans & Reznick, 2009; Langerhans, 2009a,b,c).
Stickleback, performance and morphology
Our work focused on stickleback populations in the inlet stream and lake of the Misty watershed on northern Vancouver Island (British Columbia, Canada). These two populations show very restricted gene flow (Thompson et al., 1997; Hendry et al., 2002; Moore et al., 2007) and strong divergence in several aspects of their morphology (Lavin & McPhail, 1993; Hendry et al., 2002; Moore et al., 2007), colour (Lavin & McPhail, 1993) and behaviour (Delcourt et al., 2008; Raeymaekers et al., 2009). Many of these differences are known to have a genetic basis, as revealed through common-garden rearing experiments (Lavin & McPhail, 1993; Hendry et al., 2002; Delcourt et al., 2008; Sharpe et al., 2009). No study, however, has yet examined differences between the two populations in performance, nor quantified how the divergent morphological traits relate to that performance.
One important aspect of performance in fish is the ability to swim for extended periods of time, variously called sustained (the term we will use), prolonged, endurance, steady or aerobic swimming ability (Beamish, 1978; Plaut, 2001; Blake, 2004). Adaptive divergence in sustained swimming ability is seen particularly clearly between populations that do or do not swim long distances (e.g. migration) or that do or do not hold their position in flowing water (Langerhans, 2009a). In stickleback, sustained swimming ability is greater in anadromous (migratory) populations than in resident freshwater populations (Taylor & McPhail, 1986; Tudorache et al., 2007; but see Schaarschmidt & Jürss, 2003) and is greater in limnetic than in benthic populations (Blake et al., 2005). For the Misty system, we expect greater sustained swimming ability in lake stickleback – because they are thought to swim long distances while searching for zooplankton in open water (Berner et al., 2008, 2009). Inlet stream stickleback, by contrast, range over shorter distances and avoid high-flow areas (J.S. Moore & A.P. Hendry, unpubl. data).
A second important aspect of performance in fish is the ability to rapidly accelerate, variously called burst (the term we will use), fast-start or C-start performance (Domenici & Blake, 1997; Blake, 2004; Walker et al., 2005). Adaptive divergence in burst swimming ability is seen particularly clearly between populations with or without exposure to predatory fishes (Langerhans et al., 2004; Ghalambor et al., 2004; Langerhans, 2009b,c). In stickleback, burst swimming ability is greater in freshwater resident populations than in anadromous populations (Taylor & McPhail, 1986), but does not differ between limnetic and benthic populations (Law & Blake, 1996). For the Misty system, we expect lake stickleback to be better burst swimmers, because fish predation rates are likely higher in the lake than in the inlet. Indirect evidence for this supposition is that Misty lake stickleback have longer spines than Misty inlet stickleback (Hendry et al., 2002), and longer spines generally indicate higher predation by fish (Hagen & Gilbertson, 1972; Gross, 1978; Reimchen, 1994).
A third important aspect of swimming performance in fish is the ability to turn at sharp angles. This ‘manoeuverability’ should be particularly important in complex environments, whether to avoid predators or to obtain prey or mates (Domenici & Blake, 1997; Domenici, 2003; Walker, 1997, 2004). Here, we estimate manoeuverability from the tightness of the turning radius when fish perform ‘C-starts’ during burst swimming trials. C-starts in stickleback are of the double-bend type (Law & Blake, 1996), involving the formation of a ‘C’ shape at the end of the first contraction of the lateral musculature, followed by a contralateral bend in the opposite direction (Domenici & Blake, 1997). No studies have formally examined manoeuverability in stickleback, but we expect it to be greater for inlet stickleback than for lake stickleback – owing to the greater habitat complexity in streams.
These expected performance differences might be influenced by several of the measured morphological traits (the assertions below are based on Webb, 1982, 1984; Domenici & Blake, 1997; Walker, 1997; Blake, 2004; Langerhans & Reznick, 2009; Langerhans, 2009a,b,c). First, sustained swimming ability should be greater for fish with shallower bodies (i.e. ‘fineness’ or ‘streamlining’) and larger pectoral fins, which are used for sustained swimming in stickleback (Taylor & McPhail, 1986; Law & Blake, 1996; Walker, 2004). Second, burst swimming ability should be greater for fish with deeper caudal regions and larger caudal fins. Third, manoeuverability should be greater for fish with deeper bodies towards their centre and for fish with fewer lateral plates. These associations were considered between the lake and inlet populations (based on population means) and among individuals within the populations.