Habitat-related specialization of lateral-line ... - Wiley Online Library

Journal of Fish Biology (2016) 88, 1631–1641 doi:10.1111/jfb.12912, available online at wileyonlinelibrary.com

Habitat-related specialization of lateral-line system morphology in a habitat-generalist and a habitat-specialist New Zealand eleotrid J. P. Vanderpham*†, S. Nakagawa‡§, A. M. Senior‖ and G. P. Closs§ *Vanderpham Consulting, 11027 50th Ave SE, Everett, WA 98208, U.S.A., ‡School of Biological, Earth and Environmental Sciences, The University of New South Wales, Sydney, NSW 2052, Australia, §Department of Zoology, University of Otago, P. O. Box 56, Dunedin 9054, New Zealand and ‖Charles Perkins Centre and School of Mathematics and Statistics, The University of Sydney, Sydney, NSW, 2006, Australia

(Received 28 May 2015, Accepted 27 December 2015) An investigation of intraspecific habitat-related patterns of variation in oculoscapular lateral-line superficial neuromasts (SN) identified a decrease in the ratio of total SNs to pores, and a trend towards decreased asymmetry in SNs in the habitat-generalist common bully Gobiomorphus cotidianus from fluvial habitats compared to lacustrine habitats, suggesting habitat-related phenotypic variability. A greater ratio of pores to SNs, as well as less variation in the total number and asymmetry of SNs observed in the fluvial habitat-specialist redfin bully Gobiomorphus huttoni may provide further evidence of variations in the oculoscapular lateral-line morphology of fluvial habitat G. cotidianus individuals serving as adaptations to more turbulent environments. © 2016 The Fisheries Society of the British Isles

Key words: neurology; phenotypic variation; plasticity; sensory system.

Some species of animals, known as habitat-generalists, exhibit, and are able to use, phenotypic variability to function effectively across a range of conditions and hence utilise a wide range of habitat-types (Imre et al., 2002; Bolnick et al., 2003; McCauley et al., 2008). There are many examples of generalist fishes displaying some degree of phenotypic adaptation to their individual habitats, such as modifications in fin size, body shape and mechanosensory lateral-line system (Ehlinger, 1990; Matthews et al., 2010; Wellenreuther et al., 2010; Gerry et al., 2011). The lateral-line system of fishes is a network of clustered mechanoreceptor organs (neuromasts), located externally on the epidermis of the head, trunk and tail of the fish (superficial neuromasts; SN), or within sub-epidermal fluid filled canals (canal neuromasts) connected by pores to the epidermal surface (Montgomery et al., 2000; Coombs & Braun, 2003; Janssen, 2004; van Netten & McHenry, 2014). Superficial †Author to whom correspondence should be addressed. Tel.: +1 425 971 9746; email: [email protected]

1631 © 2016 The Fisheries Society of the British Isles

1632

J . P. VA N D E R P H A M E T A L.

neuromasts function as water velocity detectors and are important in rheotaxis, or orientation in water currents, while the canal neuromasts detect pressure gradients and are particularly important in feeding and predator avoidance (Montgomery et al., 2000; Coombs & Braun, 2003; Janssen, 2004). Higher water velocities may reduce the effectiveness of SNs, and canal neuromasts, sheltered from direct flow, may be of greater use (Engelmann et al., 2002; Kanter & Coombs, 2003; Bleckmann & Zelick, 2009). Both inter- and intra-specific habitat-related variation has been observed in lateral-line systems (Wark & Peichel, 2010; Trokovic et al., 2011), and several studies have associated systems comprising more canal neuromasts and fewer SNs with habitats subject to high water velocities (Dijkgraaf, 1963; Coombs et al., 1988; Wellenreuther et al., 2010). In New Zealand, the common bully Gobiomorphus cotidianus McDowall 1975, a small, benthic eleotrid, is a habitat-generalist widespread in both rivers and lakes (McDowall, 2000). Recent studies have identified a pattern of more oculoscapular canal pores and decreased pore asymmetry (difference between the number of pores on the right and left sides of the head) in fluvial compared to lacustrine individuals (Michel et al., 2008; Vanderpham et al., 2013a). This suggests some level of phenotypic specialization of individuals to fluvial habitats. Bassett et al. (2006) showed that G. cotidianus lateral-line systems of only SNs were compromised in flowing water, which may reflect habitat-related variation in pores as an adaptation to higher water velocities. If SNs, like lateral-line pores, are a variable phenotypic feature, certain variations might also serve as adaptations to particular hydrodynamic conditions, facilitating the ability of G. cotidianus to survive in a wide range of habitats. Recent research (Vanderpham et al., 2013a) has also identified significantly more variation in the oculoscapular canal pore morphology of G. cotidianus than in the closely related, often co-existing, fluvial-specialist redfin bully Gobiomorphus huttoni (Ogilby 1894) (McDowall, 2000; Jowett & Richardson, 2003; NIWA, 2015). Gobiomorphus huttoni tends to have more total oculoscapular canal pores with almost no variation in total pores or pore symmetry (Vanderpham et al., 2013a, b). The apparently more refined lateral-line morphology of G. huttoni and greater number of oculoscapular canal pores may indicate their importance as a specialized adaptation to swift flowing habitats. Likewise, adaptations in SNs may be present, but the SN systems of G. huttoni or G. cotidianus have not been studied in detail. The primary aim of this study was to test for intraspecific habitat-related patterns of variation in SNs within a generalist, specifically G. cotidianus, as an indication of habitat-specific adaptations at the individual or population level. The prediction was that G. cotidianus inhabiting fluvial habitats would have fewer oculoscapular SNs and decreased asymmetry, as well as a decrease in the ratio of oculoscapular SNs to oculoscapular lateral-line pores compared to G. cotidianus from lacustrine habitats. This prediction was based on reduced effectiveness of SNs in higher velocity and flowing habitats and the inferred greater importance of canal neuromasts, which are sheltered from the direct disturbance (i.e. over-stimulation) of flow (Engelmann et al., 2002; Kanter & Coombs, 2003; Bassett et al., 2006; Bleckmann & Zelick, 2009). This study also compared the lateral line morphology of G. cotidianus to G. huttoni, predicting that the fluvial-specialist G. huttoni would have a lateral-line system with less variation and a lower ratio of SNs to pores, further suggesting habitat-related morphological specialization within the genus.

© 2016 The Fisheries Society of the British Isles, Journal of Fish Biology 2016, 88, 1631–1641

1633

H A B I TAT- R E L AT E D L AT E R A L- L I N E S P E C I A L I Z AT I O N

Table I. Summary of total Gobiomorphus huttoni and Gobiomorphus cotidianus collected from each water-body, mean ± s.d. standard length (LS ) n

Mean ± s.d. LS (mm)

Inland Canterbury Coastal Otago Inland Canterbury

9 22 11

42⋅2 ± 5⋅8 37⋅9 ± 3⋅5 52⋅2 ± 10⋅8

Coastal Otago

20

48⋅5 ± 8⋅5

Coastal Otago Coastal Otago Coastal Otago

12 15 20

46⋅1 ± 10⋅7 37⋅7 ± 6⋅4 45⋅4 ± 8⋅5

Coastal Otago Coastal Westland Coastal Otago Coastal Southland

20 8 1 9

59⋅1 ± 8⋅1 61⋅8 ± 16⋅0 31⋅0 33⋅1 ± 4⋅7

Species

Habitat

Water-body

Region

G. cotidianus

Lake

Lake Benmore Lake Waihola Stoney Creek (tributary of Lake Benmore) Alex Creek (tributary of Lake Waihola) Waitaki River Tokomairiro River Sawmill Creek Alex Creek Waihuka Stream Tokomairiro River Waiau River

Upstream river

Coastal river

G. huttoni

Coastal river (pool) Upstream river Coastal river

Gobiomorphus cotidianus and G. huttoni were collected from the southern half of the South Island of New Zealand (Table I) by electrofishing (fluvial habitats) and seining (lacustrine habitats). Sampling occurred across a range of habitats in which both species occur, including differences in life histories. While coastal river populations of both species are considered amphidromous, some lake populations and populations in tributary streams upstream of lakes are known to display facultative amphidromy, spawning in lakes rather than marine environments (Closs et al., 2003; Vanderpham et al., 2013a). The sampling locations for G. cotidianus included three coastal rivers, two lakes and two tributary rivers upstream of the lakes. Gobiomorphus huttoni were collected from two coastal rivers and two rivers upstream of lakes. One of the coastal rivers, Sawmill Creek, was sampled for G. cotidianus because of the unique conditions of the collection site, a large main-channel pond-like pool with no visible surface turbulence. Flow is minimal in this downstream reach of the creek as the creek’s outlet to the sea is typically blocked by a naturally formed sand-dune berm (Lill et al., 2012), allowing for further investigation of habitat-specific adaptations of G. cotidianus in a still habitat within a river. Only fishes >30 mm standard length (LS ) were collected to ensure only adult morphology was assessed (Stephens, 1982; McDowall, 1990; Kattel & Closs, 2007), as adult fishes are expected to display fully developed morphological characteristics (Faustino & Power, 1999; Fischer-Rousseau et al., 2009; Jafari et al., 2009). Fishes were fixed in 10% formalin, and preserved in 70% ethanol. Digital photographs were taken of each fish using a 16⋅2 mega-pixel digital SLR camera (www.nikon.com), with an 18–55 mm, 1:3⋅5–5⋅6G lens zoomed to 35 mm, and a 12 mm extension tube, with a 1/15 exposure value, F5 and ISO 320 settings and visually analysed. Oculoscapular SNs and oculoscapular canal pores (those from the anterior tip of the snout to just posterior to the eyes, up to, but not including the medial lateral-line running along the body of the fish from the posterior side of the eye towards the tail) were counted. Oculoscapular SNs and oculoscapular pores were visually observable in


1634


(a)

(b)

Eye

Eye

0

(c)

5 mm

(d)

Eye

0

0

5 mm

5 mm

Eye

0

5 mm

Fig. 1. Examples of cephalic dorsal lateral-line system photographs, indicating, on one side of the head only, superficial neuromasts (SN; ) and pores ( ) included in the analyses. (a) Gobiomorphus huttoni (65 mm standard length, LS ) with typical oculoscapular canal pore and SN morphology and Gobiomorphus cotidianus of (b) 43, (c) 44 and (d) 56 mm LS , demonstrating some of the variation in pore and SN numbers and arrangements observed. Note that only SNs within the defined oculoscapular lateral-line were included in the analyses, and those just adjacent were not.

digital images without staining. Only neuromasts within the oculoscapular lateral-line on the dorsal surface of the head, and not adjacent to it, were included in the analysis as additional neuromasts were less distinct, and thus difficult to count accurately (Fig. 1). For consistency, all neuromasts within the defined observable area viewing the surface of the fish perpendicularly were counted, including those at, or just within a pore opening. As only oculoscapular SNs and oculoscapular canal pores were recorded in this study (SNs and pores). Statistical analyses were conducted in the statistical environment R (Version 2.13.0; R Development Core Team; www.r-project.org). To explore intraspecific and interspecific variation in the morphology of Gobiomorphus spp., generalized linear mixed models (GLMM) were used. GLMMs were implemented with the glmmPQL function



1635

in the package MASS (version 7.3-16, Venables & Ripley, 2002). GLMMs were used as the morphological measures taken conform to non-Gaussian distributions, which can be modelled with appropriate link functions. Count variables (e.g. total SNs) were modelled with quasi-Poisson GLMMs (i.e. log-link), and ratios (e.g. SNs to pores) and asymmetry with quasi-binomial (i.e. logit-link) GLMMs. Additionally, because multiple data have been sampled from the same populations (Table I) correlated structures (i.e. non-independence) may be present, which were taken into account by fitting the sampling population as a random factor. Finally, morphological responses, such as total pore number, may increase with individual size. Given there may be size differences between fish populations and species (Table I), all GLMMs controlled for individual sizes by also fitting LS of the individual (Z transformed over the dataset of interest) as a numeric fixed-effect (as opposed to standardizing total pore number to LS , giving pores per mm; Kronmal, 1993). In all intra-specific models, the habitat-type of origin for individual G. cotidianus was fitted as a four level categorical fixed effect (i.e. coastal river, coastal river pool, lake and upstream river) along with the factors specified above. To investigate habitat-related patterns in total superficial neuromast variation within G. cotidianus, a GLMM was used, with the counts of total SNs on an individual as the response. To compare the ratio of SNs to pores between habitat-types, a GLMM was specified with the relative counts of the two traits as the response. Finally, a GLMM was used to assess the level of asymmetry in SNs, with the relative counts of SNs from one side of the head (with most) and the other side (with the fewest) as the response. Interspecific differences in the level of variation in SNs were assessed with a variance test using F statistics, which compares heterogeneity in variance. To account for differences in LS between individuals of each species, the numbers of total SNs were first standardized to LS and ln transformed before the F-test of analysis of variation. To assess inter-specific differences in morphology GLMMs were used with species fitted as a two level categorical fixed-effect (along with the predictors described above). Differences in total counts of SNs were assessed by fitting counts thereof as a response in a GLMM. Inter-specific comparisons in the ratio of total SNs to total pores were modelled in a GLMM with the relative counts of the two traits as the response variable. Finally, species differences in asymmetry of SNs were assessed by GLMM with the response as the relative counts of SNs from one side of a fish’s head (with the highest count) to the other (the lowest count). Gobiomorphus cotidianus from the four habitat-types had similar total numbers of SNs, with no significant difference between them estimated by GLMM (Table II). Coastal river pool G. cotidianus had the greatest mean ratios of SNs to pores, followed by lake, upstream river and lastly coastal river G. cotidianus. Lake G. cotidianus had a significantly higher ratio of SNs to pores than coastal river G. cotidianus (Table II). Coastal river pool G. cotidianus were also estimated to have a similarly larger ratio than coastal river G. cotidianus, although this difference was not statistically significant (Table II). Asymmetry in SNs was present in 56–77% of G. cotidianus in each of the habitat types, and the level of asymmetry was greatest in lake and coastal river G. cotidianus. Differences between habitat-types in asymmetry of SNs, however, were not found to be significant by GLMM (Table II). Little variation in SNs and pores was observed in G. huttoni, but was widespread within G. cotidianus. Variation in total SNs, standardized to LS , was significantly greater in G. cotidianus (mean = 0⋅70, s.d. = 0⋅15, s.e. = 0⋅01, var. = 0⋅02) than


Coastal river Coastal river pool Lake Upstream


SN asymmetry


Habitat

SNs:pores

SNs

Trait

1⋅26 ± 0⋅32 1⋅70 ± 0⋅32 1⋅65 ± 0⋅24 1⋅45 ± 0⋅26

16⋅85 ± 4⋅28 39⋅55 ± 6⋅94 34⋅95 ± 4⋅80 26⋅39 ± 5⋅74

29⋅26 ± 0⋅82 31⋅80 ± 1⋅34 28⋅81 ± 0⋅96 31⋅25 ± 1⋅40

Mean ± s.e.

27 20 31 31

27 20 31 31

27 20 31 31

n Coastal river to lake Coastal river to coastal river pool Coastal river to upstream Lake to upstream Lake to coastal river pool Coastal river pool to upstream Coastal river to lake Coastal river to coastal river pool Coastal river to upstream Lake to upstream Lake to coastal river pool Coastal river pool to upstream Coastal river to lake Coastal river to coastal river pool Coastal river to upstream Lake to upstream Lake to coastal river pool Coastal river pool to upstream

Habitat comparison −0⋅01 0⋅05 −0⋅03 −0⋅03 0⋅06 −0⋅09 0⋅84 0⋅81 0⋅42 −0⋅42 −0⋅03 −0⋅38 0⋅03 0⋅02 0⋅10 >0⋅10 >0⋅10 >0⋅10 >0⋅10 >0⋅10 0⋅05 >0⋅10 >0⋅10 >0⋅10 >0⋅10 >0⋅10 >0⋅10 >0⋅10 >0⋅10 >0⋅10 >0⋅10

P

Table II. Comparisons of total Gobiomorphus cotidianus oculoscapular superficial neuromasts (SN), the ratio of total oculoscapular SNs to total oculoscapular canal pores and the level of SN asymmetry between habitat-types, showing means, s.e. and results of GLMM analyses. b estimates for total SNs are on ln scales and all other estimates are on the logit scale

1636 J . P. VA N D E R P H A M E T A L.



1637

G. huttoni (mean = 0⋅25, s.d. = 0⋅08, s.e. = 0⋅01, var. = 0⋅01; F-test of variance F 108,37 = 0⋅47, 95% c.i. = 0⋅27–0⋅78, P < 0⋅01). The total number of SNs was significantly greater in G. cotidianus (mean = 30⋅17, s.e. = 0⋅58) than in G. huttoni (mean = 11⋅84, s.e. = 0⋅17; GLMM; b = −1⋅02, s.e. = 0⋅05, d.f. = 135, t = −20⋅0, P < 0⋅01). The ratio of total SNs to total pores was also significantly greater (i.e. more SNs to every pore) in G. cotidianus (mean = 28⋅88, s.e. = 2⋅78) than G. huttoni (mean = 2⋅12, s.e. = 0⋅04; GLMM; b = −1⋅68, s.e. = 0⋅19, d.f. = 135, t = −9⋅10, P < 0⋅01). The degree of asymmetry in the number of SNs was significantly greater in G. cotidianus than in G. huttoni (GLMM; b = −0⋅08; s.e. = 0⋅03, t = −3⋅05, P < 0⋅05). This study identified habitat-related patterns of lateral-line system variation within G. cotidianus. No clear habitat-related pattern of greater or fewer oculoscapular SNs was found. Although relationships between flow and turbulence and the total number of SNs and canals and pores in fishes have been identified in some species (Dijkgraaf, 1963; Coombs et al., 1988; Engelmann et al., 2002; Wellenreuther et al., 2010), several studies have also found no relationships. For example, in comparing two species in the family Pinguipedidae, Carton & Montgomery (2004) found that Cheimarrichthys fosteri Haast 1874 has a high number of SNs despite living in a turbulent environment, while Parapercis colias (Forster 1801), which prefers quiet water habitats, has a low number of SNs. Beckmann et al. (2010) also found no relationship between the habitat occurrence (reophilic or limnophilic) and the total number of SNs in 12 species of European Cypriniformes. Although the functionality of superficial neuromasts may be reduced under some conditions (e.g. turbulence), they remain important for fishes living in fluvial habitats as they provide information on objects in the water and are used in rheotaxis (Beckmann et al., 2010), and therefore, total CNs may not be reduced in fluvial G. cotidianus. In addition, micro-habitat selection (e.g. areas of laminar flow) by G. cotidianus in fluvial habitats may limit the magnitude of disturbance associated with water flow to lateral-line systems. A habitat-related pattern in the ratio of total oculoscapular SNs to total oculoscapular canal pores was, however, observed, and is concordant with the results of previous research of oculoscapular pores alone (Vanderpham et al., 2013a). This finding suggests patterns of variation in total pores are more closely associated with habitat-type than are total SNs. The greatest ratio of SNs to pores was found in G. cotidianus from the coastal river pool habitat, followed by lake, upstream river and lastly coastal river. Only the difference between ratios of SNs to pores for coastal river to coastal river pool G. cotidianus was statistically significant, although other model estimates were also consistent (if not significant) with the prediction of a decreased ratio as an adaptation to more turbulent habitats (e.g. coastal river pool > coastal river). In turbulent conditions, the effectiveness of SNs is probably reduced in G. cotidianus (Bassett et al., 2006), and pores and associated canal neuromasts are of greater importance in sensing the surrounding environment (Engelmann et al., 2002; Kanter & Coombs, 2003; Bleckmann & Zelick, 2009). Furthermore, this pattern also matches the lower ratio of SNs to pores identified in the G. huttoni which occurs in turbulent rivers. Although not statistically significant, the trend of greater asymmetry in oculoscapular SNs in G. cotidianus from the coastal river pool and lake habitat compared to coastal rivers and rivers upstream of lakes supports the present prediction. Previous research identified significantly more asymmetry in oculoscapular canal pores of G. cotidianus in lacustrine than fluvial habitats (Vanderpham et al., 2013a). Reduced asymmetry may suggest a selective pressure for more refined lateral-line system


1638


morphology in turbulent habitats. As water velocity and turbulence increases, the challenges of swimming, orientating and maintaining position increase (Liao, 2007; Tritico & Cotel, 2010), all of which are linked to lateral-line system function (Coombs & Braun, 2003; Bleckmann & Zelick, 2009). Detailed investigations of the effects of lateral-line asymmetry on abilities or behaviours of the fishes were not conducted, however, the low asymmetry in both oculoscapular canal pores and oculoscapular SNs in G. huttoni observed in this and previous research (Vanderpham et al., 2013a, b) indicate the importance of refinement of these sensory systems in fluvial habitats. This comparison between G. cotidianus and G. huttoni may provide further evidence that variations in the oculoscapular lateral-line morphology of fluvial habitat G. cotidianus individuals serve as adaptations to more turbulent environments. The significantly lower ratio of SNs to canal pores observed in G. huttoni than G. cotidianus may suggest specialization of the G. huttoni lateral-line system to more turbulent (i.e. fluvial) environments (assuming that fluvial habitats favour a lower ratio of superficial neuromast to canal pores as the present study suggest). This is consistent with the habitat-related variation observed in G. cotidianus, with a lower ratio of SNs to pores in G. cotidianus from more turbulent habitats (coastal and upstream rivers). Significantly less variation in total numbers and asymmetry of oculoscapular SNs identified in G. huttoni compared to G. cotidianus may also suggest a more refined system of oculoscapular SNs. This finding follows the pattern of less variation and asymmetry in G. huttoni oculoscapular canal pores compared to G. cotidianus observed in previous research (Vanderpham et al., 2013a, b). With habitat-specialization of G. huttoni, genotypes displaying particularly beneficial traits may be selected for, resulting in a reduction in variation (Futuyma & Moreno, 1988; Debat & David, 2001). For G. cotidianus, occurring as they do across a much wider range of hydrodynamic conditions, the large amount of phenotypic variability observed within their lateral-line system may be a result of less consistent specific habitat-selective pressures as opposed to that experienced by G. huttoni (Kassen, 2002; Kawecki & Ebert, 2004). It is important, however, to recognize the possible influence of microhabitat selection on flow and turbulence conditions experienced by both species. For example in a river, a fish may select microhabitat with more laminar flow, or slower flow (e.g. pool habitat), therefore experiencing conditions more similar to a lentic environment, with similar levels of disturbance (e.g. caused by turbulence) on lateral-line system functionality. Although inferences of adaptation are limited when making a comparison of only two species (Garland & Adolph, 1994), the findings of the present study are consistent with morphological specialization in G. huttoni, and phenotypic adaptability in G. cotidianus. Like many other habitat-generalists, G. cotidianus may be a less competitive species in any one particular habitat-type compared to a specialist (such as G. huttoni) adapted to that habitat (Partridge & Green, 1987; Wilson & Yoshimura, 1994; Caley & Munday, 2003), but lateral-line system variability may be vital for their ability to utilize a range of hydrodynamic conditions. The pattern of a decreased ratio of oculoscapular SNs to oculoscapular canal pores in G. cotidianus from more turbulent fluvial habitats compared to still water habitats may have important implications for fish survival strategies (Coombs & Braun, 2003; Janssen, 2004). Canal neuromasts have been shown to be particularly important in locating prey and avoiding predators, while SNs are particularly important for rheotaxis (Montgomery et al., 1997, 2000; Coombs & Braun, 2003; Janssen, 2004). The interaction of canal and SNs plays an important role in feeding (Montgomery



1639

et al., 2002). Therefore, the differences in ratios between G. cotidianus and G. huttoni, and between the different habitats of G. cotidianus could be expected to influence their behaviour, abilities and habitat selection. The physiology of G. cotidianus and G. huttoni lateral-line systems are poorly understood, however, and the implications of the variation cannot be accurately predicted. Although Bassett et al. (2006) showed impaired detection capabilities of G. cotidianus with a lateral-line system lacking canal neuromasts, the number of canal neuromasts within each canal and associated with each pore remains unclear (Webb & Shirey, 2003). Furthermore, the function of individual oculoscapular neuromasts, as well as neuromasts located along other regions of the bodies of G. cotidianus and G. huttoni are not known. Future behavioural experiments with detailed morphological analyses may provide a better understanding of lateral-line system function and ecological implications of phenotypic variation in habitat-generalist and specialist species. The Department of Zoology at the University of Otago provided laboratory and field equipment, and all fish collections and experimental procedures were approved by the University of Otago Animal Ethics Committee. Golder Associates, Ltd (www.golder.com) is also acknowledged for their support through the allocation of time required for the preparation of this manuscript. We thank A. Hicks, K. Garret, E. dos Santos, R. Patterson and M. Warburton for assistance with fish collections and statistical analyses. We also thank all members of the Closs and Nakagawa research groups at the University of Otago for their feedback and support, and N. McHugh and M. Downes for assistance with laboratory analyses.

References Bassett, D. K., Carton, A. G. & Montgomery, J. C. (2006). Flowing water decreases hydrodynamic signal detection in fish with an epidermal lateral-line system. Marine and Freshwater Research 5, 611–617. Beckmann, M., Erös, T., Schmitz, A. & Bleckmann, H. (2010). Number and distribution of superficial neuromasts in twelve common European Cypriniform fishes and their relationship to habitat occurrence. International Review of Hydrobiology 95, 273–284. Bleckmann, H. & Zelick, R. (2009). Lateral line system of fish. Integrative Zoology 4, 13–25. Bolnick, D. I., Svanbäck, R., Fordyce, J. A., Yang, L. H., Davis, J. M., Hulsey, C. D. & Forister, M. L. (2003). The ecology of individuals: incidence and implications of individual specialization. American Naturalist 161, 1–28. Caley, M. J. & Munday, P. L. (2003). Growth trades off with habitat specialization. Proceedings of the Royal Society B 270(Suppl. 2), S175–S177. Carton, A. G. & Montgomery, J. C. (2004). A comparison of lateral line morphology of blue cod and torrentfish; two sandperches of the family Pinguipedidae. Environmental Biology of Fishes 70, 123–131. Closs, G. P., Smith, M., Barry, B. & Markwitz, A. (2003). Nondiadromous recruitment in coastal populations of common bully (Gobiomorphus cotidianus). New Zealand Journal of Marine and Freshwater Research 37, 301–313. Coombs, S. & Braun, C. B. (2003). Information processing by the lateral line. In Sensory Processing of the Aquatic Environment (Collin, S. P. & Marshall, N. J., eds), pp. 122–138. New York, NY: Springer-Verlag. Coombs, S., Janssen, J. & Webb, J. (1988). Diversity of lateral line systems: evolutionary and functional considerations. In Sensory Biology of Aquatic Animals (Atema, J., Fay, R. R., Popper, A. N. & Tavolga, W. N., eds), pp. 553–594. New York, NY: Springer-Verlag. Debat, V. & David, P. (2001). Mapping phenotypes: canalization, plasticity and developmental stability. Trends in Ecology and Evolution 16, 555–561. Dijkgraaf, S. (1963). The functioning and significance of the lateral-line organs. Biological Reviews 38, 51–105. Ehlinger, T. J. (1990). Habitat choice and phenotype-limited feeding efficiency in bluegill: individual differences and trophic polymorphism. Ecology 71, 886–896.


1640


Engelmann, J., Hanke, W. & Bleckmann, H. (2002). Lateral line reception in still- and running water. Journal of Comparative Physiology A 188, 513–526. Faustino, M. & Power, D. M. (1999). Development of the pectoral, pelvic, dorsal and anal fins in cultured sea bream. Journal of Fish Biology 54, 1094–1110. Fischer-Rousseau, L., Cloutier, R. & Zelditch, M. L. (2009). Morphological integration and developmental progress during fish ontogeny in two contrasting habitats. Evolution and Development 11, 740–753. Futuyma, D. J. & Moreno, G. (1988). The evolution of ecological specialization. Annual Review of Ecology and Systematics 19, 207–233. Garland, T. Jr. & Adolph, S. C. (1994). Why not to do two-species comparative studies: limitations on inferring adaptation. Physiological Zoology 67, 797–828. Gerry, S. P., Wang, J. & Ellerby, D. J. (2011). A new approach to quantifying morphological variation in bluegill Lepomis macrochirus. Journal of Fish Biology 78, 1023–1034. Imre, I., McLaughlin, R. L. & Noakes, L. G. (2002). Phenotypic plasticity in brook charr: changes in caudal fin induced by water flow. Journal of Fish Biology 61, 1171–1181. Jafari, M., Kamarudin, M. S., Saad, C. H., Arshad, A., Oryan, S. & Bahmani, M. (2009). Development of morphology in hatchery-reared Rutilus frisii kutum larvae. European Journal of Scientific Research 38, 296–305. Janssen, J. (2004). Lateral line sensory ecology. In The Senses of Fish: Adaptations for the Reception of Natural Stimuli (Emde, G. V. D., Mogdans, J. & Kapoor, B. G., eds), pp. 231–264. Boston, MA: Kluwer. Jowett, I. G. & Richardson, J. (2003). Fish communities in New Zealand rivers and their relationship to environmental variables. New Zealand Journal of Marine and Freshwater Research 37, 347–366. Kanter, M. J. & Coombs, S. (2003). Rheotaxis and prey detection in uniform currents by Lake Michigan mottled sculpin (Cottus bairdi). Journal of Experimental Biology 206, 59–70. Kassen, R. (2002). The experimental evolution of specialists, generalists, and the maintenance of diversity. Journal of Evolutionary Biology 15, 173–190. Kattel, G. R. & Closs, G. P. (2007). Spatial and temporal variation in the fish community of a South Island, New Zealand coastal lake. New Zealand Journal of Marine and Freshwater Research 41, 1–11. Kawecki, T. J. & Ebert, D. (2004). Conceptual issues in local adaptation. Ecology Letters 7, 1225–1241. Kronmal, R. A. (1993). Spurious correlation and the fallacy of the ratio standard revisited. Journal of the Royal Statistical Society A 156, 379–392. Liao, J. C. (2007). A review of fish swimming mechanics and behaviour in altered flows. Philosophical Transactions of the Royal Society B 362, 1973–1993. Lill, A., Closs, G., Schallenberg, M. & Savage, C. (2012). Impact of berm breaching on hyperbenthic macroinvertebrate communities in intermittently closed estuaries. Estuaries and Coasts 35, 155–168. Matthews, B., Marchinko, K. B., Bolnick, D. I. & Mazumder, A. (2010). Specialization of trophic position and habitat use by sticklebacks in an adaptive radiation. Ecology 91, 1025–1034. McCauley, S. J., Davis, C. J. & Werner, E. E. (2008). Predator induction of spine length in larval Leucorrhinia intacta (Odonata). Evolutionary Ecology Research 10, 435–447. McDowall, R. M. (1990). New Zealand Freshwater Fishes: A Natural History and Guide. Auckland: Heinemann Reed. McDowall, R. M. (2000). The Reed Field Guide to New Zealand Freshwater Fishes. Auckland: Reed Books. Michel, C., Hicks, B. J., Stolting, K. N., Clarke, A. C., Stevens, M. I., Tana, R., Meyer, A. & van den Heuvel, M. R. (2008). Distinct migratory and non-migratory ecotypes of an endemic New Zealand eleotrid (Gobiomorphus cotidianus) - implications for incipient speciation in island freshwater fish species. BMC Evolutionary Biology 8, 49. Montgomery, J. C., Baker, C. F. & Carton, A. G. (1997). The lateral line can mediate rheotaxis in fish. Nature 389, 960–963. Montgomery, J., Carton, G., Voigt, R., Baker, C. & Diebel, C. (2000). Sensory processing of water currents by fishes. Philosophical Transactions of the Royal Society B 355, 1325–1327.



1641

Montgomery, J. C., Macdonald, F. F., Baker, C. F. & Carton, A. G. (2002). Hydrodynamic contributions to multimodal guidance of prey capture behavior in fish. Brain, Behavior and Evolution 59, 190–198. van Netten, S. M. & McHenry, M. J. (2014). The biophysics of the fish lateral line. In The Lateral Line System (Coombs, S., Bleckmann, H., Fay, R. R. & Popper, A. N., eds), pp. 99–119. New York, NY: Springer. Partridge, L. & Green, P. (1987). An advantage for specialist feeding in jackdaws, Corvus monedula. Animal Behaviour 35, 982–990. Stephens, R. T. T. (1982). Reproduction, growth and mortality of the common bully, Gobiomorphus cotidianus McDowall, in a eutrophic New Zealand lake. Journal of Fish Biology 20, 259–270. Tritico, H. M. & Cotel, A. J. (2010). The effects of turbulent eddies on the stability and critical swimming speed of creek chub (Semotilus atromaculatus). Journal of Experimental Biology 213, 2284–2293. Trokovic, N., Herczeg, G., McCairns, R. J., Ab Ghani, N. I. & Merila, J. (2011). Intraspecific divergence in the lateral line system in the nine-spined stickleback (Pungitius pungitius). Journal of Evolutionary Biology 24, 1546–1558. Vanderpham, J. P., Nakagawa, S. & Closs, G. P. (2013a). Habitat-related patterns in phenotypic variation in a New Zealand freshwater generalist fish, and comparisons to a closely related specialist. Freshwater Biology 58, 396–408. Vanderpham, J. P., Nakagawa, S. & Closs, G. P. (2013b). Feeding ability of a fluvial habitat-specialist and habitat-generalist fish in turbulent and still conditions. Ecology of Freshwater Fish 22, 596–606. Venables, W. N. & Ripley, B. D. (2002). Modern Applied Statistics with S, 4th edn. New York, NY: Springer. Wark, A. R. & Peichel, C. L. (2010). Lateral line diversity among ecologically divergent threespine stickleback populations. Journal of Experimental Biology 213, 108–117. Webb, J. F. & Shirey, J. E. (2003). Postembryonic development of the cranial lateral line canals and neuromasts in zebrafish. Developmental Dynamics 228, 370–385. Wellenreuther, M., Brock, M., Montgomery, J. & Clements, K. D. (2010). Comparative morphology of the mechanosensory lateral line system in a clade of New Zealand triplefin fishes. Brain, Behavior and Evolution 75, 292–308. Wilson, D. S. & Yoshimura, J. (1994). On the coexistence of specialists and generalists. American Naturalist 144, 692–707.

Electronic Reference NIWA (National Institute of Water and Atmospheric Research) (2015). New Zealand Freshwater Fish Database. Available at http://www.niwa.co.nz/our-services/online-services/fresh water-fish-database/ (last accessed 1 May 2015).
