ARTICLE

The effects of two fish predators on Wood Frog(Lithobates sylvaticus) tadpoles in a subarctic wetland: HudsonBay Lowlands, CanadaJ.M. Davenport, P.A. Seiwert, L.A. Fishback, and W.B. Cash

Abstract: Fish can have strong predatory impacts on aquatic food webs. Indeed, fish are known to have strong effects onamphibians, with some species being excluded from communities where fish are present. Most research with amphibians andfish has focused on lower latitudes and very little is known of amphibian–fish interactions at higher latitudes. Therefore, weconducted an enclosure experiment in a subarctic natural wetland to examine the predatory effects of two species of fish, brooksticklebacks (Culaea inconstans (Cuvier, 1829)) and ninespine sticklebacks (Pungitius pungitius (L., 1758)), on the survival and growthof Wood Frogs (Lithobates sylvaticus (LeConte, 1825)). We found no significant difference in survival and size at metamorphosisamong the two fish species treatments and fish-free treatments. We found that individuals from fish-free treatments metamor-phosed earlier than those from either fish species present treatment. Our work suggests that stickleback fish predation may nothave a major impact on Wood Frog tadpole survival and growth in a subarctic wetland. Sticklebacks may still have an impact onearlier developmental stages of Wood Frogs. This work begins to fill an important gap in potential factors that may impact larvalamphibian survival and growth at higher latitudes.

Key words: predation, subarctic, amphibian, temporary wetland, stickleback, metamorphosis.

Résumé : La prédation par les poissons peut avoir une importante incidence sur les réseaux trophiques aquatiques. Ainsi, il estétabli que les poissons peuvent avoir de forts effets sur les amphibiens, dont certaines espèces peuvent être exclues de commu-nautés où des poissons sont présents. La plupart des travaux sur les amphibiens et les poissons se sont intéressées aux basseslatitudes, de sorte que les connaissances sur les interactions entre amphibiens et poissons a plus hautes latitudes sont trèslimitées. Nous avons réalisé une expérience en enclos dans une zone humide subarctique naturelle afin d’examiner les effets dela prédation par deux espèces de poissons, l’épinoche a cinq épines (Culaea inconstans (Cuvier, 1829)) et l’épinoche a neuf épines(Pungitius pungitius (L., 1758)), sur la survie et la croissance de la grenouille des bois (Lithobates sylvaticus (LeConte, 1825)). Nousn’avons noté aucune différence significative en ce qui concerne la survie et la taille au moment de la métamorphose entreles grenouilles soumises a des traitements avec les deux espèces de poissons et les grenouilles soumises a des traitements sanspoisson. Nous avons observé que les individus soumis a des traitements sans poisson se métamorphosaient plus tôt que lesindividus soumis a des traitements avec l’une ou l’autre des espèces de poissons. Ces travaux semblent indiquer que la prédationpar les épinoches pourrait ne pas avoir un impact majeur sur la survie et la croissance des têtards de grenouille des bois dans unezone humide subarctique. Les épinoches pourraient toutefois quand même avoir un impact sur les stades de développementprécoces de la grenouille des bois. Ces travaux commencent a combler d’importantes lacunes en ce qui concerne la connaissancedes facteurs qui pourraient avoir une incidence sur la survie et la croissance des amphibiens larvaires a hautes latitudes. [Traduitpar la Rédaction]

Mots-clés : prédation, subarctique, amphibien, zone humide temporaire, épinoche, métamorphose.

IntroductionPredation can act as a selective force to structure ecological

communities. Within freshwater communities, fish are often thetop predators exerting strong top–down effects on other speciesembedded in food webs (Brooks and Dodson 1965). These strongeffects span from altering community composition (Morin 1984;Werner and McPeek 1994), oviposition site selection (Angelon andPetranka 2002; Binckley and Resetarits 2005), life history (Reznicket al. 1990), and trait variability (Tollrian 1995) across many differ-ent taxa (Wellborn et al. 1996). One particular group known to bestrongly impacted by fish in freshwater communities is amphibi-ans (Wells 2007). The strong impacts have been confirmed over

small and large spatial scales with surveys and experiments(Wellborn et al. 1996; Hecnar and M’Closkey 1997; Smith et al.1999; Rieger et al. 2004). Amphibians are known to shift foraginghabits (Kats et al. 1988; Lawler et al. 1999), express morphologicaldefenses (Relyea 2001; Teplitsky et al. 2003) and cryptic coloration(Wassersug 1971), and use chemical cues to detect and avoid fish(Petranka et al. 1987; Resetarits and Wilbur 1989).

With amphibians declining worldwide, it is important to knowwhat factors are regulating amphibian populations. One poten-tially strong regulating factor is fish predation (Wells 2007). Themajority of the research focusing on the effects of fish on amphib-ians, however, has been in temperate regions (Wells 2007). Am-phibian populations in other regions (e.g., tropical and subarctic)

Received 22 April 2013. Accepted 9 October 2013.

J.M. Davenport. Division of Biological Sciences, University of Montana, Missoula, MT 59812, USA.P.A. Seiwert and W.B. Cash. Department of Biology, University of Central Arkansas, Conway, AR 72035, USA.L.A. Fishback. Churchill Northern Studies Centre, Churchill, MB R0B 0E0, Canada.Corresponding author: J.M. Davenport (e-mail: jon.davenport@mso.umt.edu).

866

Can. J. Zool. 91: 866–871 (2013) dx.doi.org/10.1139/cjz-2013-0091 Published at www.nrcresearchpress.com/cjz on 15 October 2013.

Can

. J. Z

ool.

Dow

nloa

ded

from

ww

w.n

rcre

sear

chpr

ess.

com

by

SUN

Y A

T S

TO

NY

BR

OO

K o

n 11

/11/

14Fo

r pe

rson

al u

se o

nly.

may also be influenced by biotic factors, such as fish predation. Atthis point, little information is known about factors regulatingamphibian populations at higher latitudes (but see Laurila et al.2008; Lindgren and Laurila 2010; Hjernquist et al. 2012 for Euro-pean examples).

Organisms with wide distributions are in general probably atlower risk of extirpation; however, edge populations may be moresusceptible to higher risk (Gaston 2003). Amphibians with largedistributions are ideal for comparative work across large spatialscales that can help to elucidate what factors are important ininfluencing populations along range edges. Wood Frogs (Lithobatessylvaticus (LeConte, 1825)) are an ideal candidate species for com-parative work on factors regulating populations for two reasons.First, Wood Frogs are one of the most widespread amphibians inthe world with a distribution ranging from the southeasternUnited States to the northern tree line at the edge of the Arctic(Redmer and Trauth 2005). The Wood Frog populations at higherlatitudes are edge populations and could be more susceptible toextirpation than core populations (Hardie and Hutchings 2010).Second, several studies in temperate regions have been conductedwith Wood Frogs and fish. Adult Wood Frogs in temperate regionsuse chemical cues to avoid egg deposition in wetlands with pred-atory sunfish (Hopey and Petranka 1994). Wood Frog tadpoles ex-perience low survival in wetlands with fish and can alter traits(behavioral and morphological) to increase survival with fish(Chivers and Mirza 2001; Relyea 2001). These behavioral and mor-phological responses have been hypothesized to result in a trade-off with larval period and size at metamorphosis (Werner andAnholt 1993; Relyea and Werner 1999). Surprisingly, few experi-mental studies have addressed the lethal effects of a predator onWood Frog life-history responses (Relyea 2007, Appendix 2). Inone of the few studies to document the effects of fish predationon Wood Frogs at higher latitudes, Eaton et al. (2005) foundthat Wood Frog young-of-the-year abundance increased afterwinter fish kills in boreal lakes of northern Alberta. WhileEaton et al. (2005) suggest fish predation may play a major role,they did not explicitly test the effects of fish predators on WoodFrogs.

We designed an enclosure experiment in a subarctic naturalwetland to investigate the predatory effects of two fish species,brook sticklebacks (Culaea inconstans (Cuvier, 1829)) and ninespinesticklebacks (Pungitius pungitius (L., 1758)), on survival and life his-tory of the Wood Frog. Brook and ninespine sticklebacks areclosely related (Kawahara et al. 2009) and are common in ourstudy area in the Hudson Bay Lowlands near Churchill, Manitoba.Previous work has shown that these species may be ecologicallysimilar and compete for resources, but can undergo ecologicalcharacter displacement when co-occurring (Gray and Robinson2002). This displacement suggests that although these speciesmay be closely related, each species may have functionally differ-ent consumptive effects on a shared prey resource (Gray et al.2005). Our experiment drew from 3 years of landscape-level sur-vey data depicting limited spatial overlap between sticklebacksand Wood Frogs (W.B. Cash, L.A. Fishback, and J.M. Davenport,unpublished data). Physical factors (such as canopy cover, percentsedge cover, etc.) may also explain the limited spatial overlap be-tween sticklebacks and Wood Frogs; however, we also have prelim-inary data from gut content analyses confirming that brooksticklebacks do consume Wood Frog tadpoles. Therefore, our a priorihypothesis for this experiment was that sticklebacks would have aconsumptive effect on Wood Frogs, but that the magnitude of thiseffect may differ between the two species (per character displace-ment; Gray and Robinson 2002). We hypothesize that both fish spe-cies would reduce Wood Frog survival. We also expect that larvalperiod would be shorter and size at metamorphosis would besmaller for surviving Wood Frogs with fish relative to Wood Frogs infish-free controls. The rationale for these expectations is based ontheoretical models predicting time to and size at metamorphosis for

complex-lived organisms (Werner 1986; Ludwig and Rowe 1990;Rowe and Ludwig 1991). In larval environments with risky predators,prey should metamorphose from the larval environment as early aspossible, which often results in the trade-off of a smaller size atmetamorphosis (Werner 1986; Ludwig and Rowe 1990; Rowe andLudwig 1991).

Materials and methods

Study systemThe Hudson Bay Lowlands (HBL) is one of the most extensive

subarctic wetland landscapes in the world. The low topographicrelief (<5 m) of the northern HBL region in the Churchill area isunderlain by continuous permafrost, poorly drained, and domi-nated by wetlands. In the HBL ecoregion, 25%–40% of the land iscovered by these shallow water bodies (Duguay and Lafleur 2003;Macrae et al. 2004). Wetlands in the HBL typically range between400 and 50 000 m2 in surface area, with mean water depth rangesapproximately 0.1–0.6 m (Macrae et al. 2004). The abundant shal-low wetlands of the HBL play several important roles in the ecore-gion (e.g., Symons et al. 2012), including habitat for amphibiansand fish.

The Wood Frog is the predominant amphibian species thatbreeds in wetlands in the Churchill area. Wood Frogs breed pri-marily in temporary wetlands, but can be found in more perma-nent bodies of water (Redmer and Trauth 2005). Wood Frogtadpoles are omnivorous, with some studies demonstrating thattadpoles can be cannibalistic on conspecifics (Bleakney 1958) andwill eat eggs and embryos of other amphibians (Petranka et al.1994, 1998). Wood Frog tadpoles are highly palatable by many fishspecies and are often excluded from communities with fish (Katset al. 1988).

Two of the most common fish species encountered in theHBL wetlands are brook and ninespine sticklebacks (W.B. Cash,L.A. Fishback, and J.M. Davenport, unpublished data). Both brookand ninespine sticklebacks are small generalist predators in streams,wetlands, and lakes of Canada (Scott and Crossman 1998). Stickle-backs can be voracious predators of ranid tadpoles of Europe(Teplitsky et al. 2005). In the Churchill, Manitoba, region, spatialoverlap in the distribution of Wood Frogs and sticklebacks is patchy.Preliminary data from 3 years of surveys across 232 ephemeral wet-lands suggest that Wood Frogs co-occur with sticklebacks inless than 20% of wetlands across the landscape. Briefly, dip-netand trap surveys were conducted of ephemeral wetlands within40 km radius of the Churchill Northern Studies Centre (CNSC;58.737746°N, 93.847763°W). During the course of the surveys, whenwe encountered Wood Frogs in wetlands with fish, tadpoles oftenremain close to the shore along with emergent vegetation. Bothspecies of sticklebacks also favor inshore littoral habitats whenforaging (Wooton 1976; Scott and Crossman 1998). Preliminarydietary data from Churchill, Manitoba, also demonstrates thatsticklebacks do consume tadpoles as food items in northern lati-tudes (W.B. Cash, L.A. Fishback, and J.M. Davenport, unpublisheddata).

Experimental designOur experiment was conducted with mesh enclosures deployed

in a natural subarctic wetland (unofficial name, Rocket Pond;58.287777°N, 93.817683°W). Rocket Pond is a medium-sized wet-land (125 000 m2) with a mean water depth of approximately 0.6 mand was chosen because it does not completely dry every year andcontains populations of Wood Frogs and both stickleback species.For our experimental design, we had three food web treatments:fish-free and only tadpoles, one brook stickleback with tadpoles,and one ninespine stickleback with tadpoles. All enclosures re-ceived 100 newly hatched Wood Frog tadpoles, evenly distributedfrom 12 clutches. Each of the 12 clutches came from wetlandswithout recent contact (last 6 years: W.B. Cash, L.A. Fishback, and

Davenport et al. 867

Published by NRC Research Press

Can

. J. Z

ool.

Dow

nloa

ded

from

ww

w.n

rcre

sear

chpr

ess.

com

by

SUN

Y A

T S

TO

NY

BR

OO

K o

n 11

/11/

14Fo

r pe

rson

al u

se o

nly.

J.M. Davenport, unpublished data) with either stickleback species.Sticklebacks were weighed in the laboratory and matched for size(approximately 1.9 g) across all replicates. Each treatment wasreplicated four times yielding 12 total experimental units. Repli-cates were arranged in spatial blocks along the periphery of thewetland. Each of the three treatments was randomly assigned toone enclosure within each spatial block. Densities of organismsstocked to each enclosure are reflective of the range of densitiesobserved in natural wetland communities surrounding Churchill(W.B. Cash, L.A. Fishback, and J.M. Davenport, unpublisheddata). Based on estimated densities of sticklebacks in RocketPond, one individual per enclosure is on the high end of stick-leback densities.

All field procedures described below were performed on a blockby block basis. Enclosures were 0.91 m × 0.91 m × 0.51 m and madeof wood framing and fine screen window mesh (1.5 mm2). Meshcompletely enclosed the wooden frame of the enclosure to pre-vent colonization of unwanted organisms (predatory inverte-brates, fish, etc.) and escape of study organisms. Primary fooditems, zooplankton and phytoplankton, were small enough tomove across the mesh screening. Enclosures were deployed intoRocket Pond on 2 June 2012. Each enclosure received 100 g of driedsedge on 2 June 2012 to provide natural refuge.

Wood Frog egg masses and sticklebacks were collected fromwetlands near the CNSC between 29 May and 1 June 2012. Twentyegg masses were brought back to the laboratory on 29–30 May2012 and maintained until hatching of embryos occurred on3 June 2012. After hatching, tadpoles were counted from 12 eggmasses and randomly assigned to one of the three experimentaltreatments. Wood Frog tadpoles were added to enclosures on6 June 2012. Hatchling tadpoles added to experimental enclosureswere all within the gape limits of both stickleback species. Bothstickleback species were collected by deploying minnow traps inwetlands near the CNSC. Brook sticklebacks added to enclosureshad a mean mass of 1.914 g (SE ±0.102 g). Ninespine sticklebacksadded to enclosures had a mean mass of 1.923 g (SE ±0.042 g). Allsticklebacks used in the experiment were mature adults. Stickle-backs were randomly assigned to experimental treatments and in-troduced into enclosures on 7 June 2012 to begin the experiment.

Enclosures were monitored daily and metamorphosed frogs (in-dividuals with at least one forelimb emergence) were capturedand returned to the laboratory. All individuals were not consid-ered fully metamorphosed until after tail absorption was com-pleted. We recorded wet mass (g) of all metamorphs and date ofcollection of each individual. Wood Frog mass at metamorphosisis represented by the mean mass of all frogs that successfullymetamorphosed from a particular enclosure. Wood Frog larvalperiod was calculated as the difference in date of metamorphosisminus the date of experiment initiation. We focus on larval periodand size at metamorphosis because these variables have previ-ously been found to play an important role in adult fitness,demography, and regulation of Wood Frog populations (Berven1990). Wood Frog survival was measured as the natural logarithm(ln) of the proportion of individuals that survived to metamorpho-sis. The ln transformation of proportion of Wood Frogs survivingnot only assured statistical assumptions were met (Gotelli 2008),but also has biological meaning because it is a measure of instan-taneous per capita mortality rates (Davenport and Chalcraft 2012).No fish died during the experiments. Enclosures were pulled fromthe wetland on 14 August 2012 and remaining vegetation wassearched meticulously for any surviving organisms. The wet mass(g) of all remaining larval Wood Frogs and fish was recorded. Thefew surviving larval Wood Frogs (14 individuals across all repli-cates) were included in the survival analysis but excluded fromlife-history statistical analyses. All surviving organisms were re-leased back at original capture sites. It was discovered that mesh

from two replicates (one control and one ninespine) had becomecompromised and tadpoles had escaped. These two replicateswere excluded from all statistical analyses.

Statistical analysesWe evaluated the effect of fish predation on Wood Frog survival

and life history by conducting a one-way ANOVA with block and thethree fish treatment (absent (control), brook present, ninespine pres-ent) terms. A post hoc contrast with Ryan–Einot–Gabriel–Welsch testwas used to detect differences among treatment means. All statisti-cal analyses were considered statistically significant when P < 0.05.All statistical analyses were conducted with SAS version 9.0 software(SAS Institute Inc. 2010).

ResultsBlock effects were not significant for any of our response vari-

ables (P > 0.07), but were retained in all statistical analyses. Wefound no statistically significant difference among treatments forWood Frog survival (F[2,4] = 3.17, P = 0.149; Fig. 1A) or Wood Frogmass at metamorphosis (F[2,4] = 2.25, P = 0.221; Fig. 1B). There werestatistically significant differences among treatments on WoodFrog larval period (F[2,4] = 7.81, P = 0.042). Wood Frog time tometamorphosis was shorter in the fish-free treatment in compar-ison with either two of the fish present treatments (Fig. 1C).

DiscussionThe novel contribution of our study is that fish predation

does not appear to be a major factor regulating Wood Frogsurvival and growth in a subarctic wetland. We found that bothspecies of sticklebacks had no significant consumptive effectson Wood Frog survival and growth (Figs. 1A–1C). This is surpris-ing given the documented strong effects of fish on amphibians intemperate regions. The majority of that work, however, has not beenconducted at higher latitudes of the world. There was a nonsignifi-cant trend for brook sticklebacks to consume more tadpoles (meansurvival of Wood Frogs was 65%) than ninespine sticklebacks. Thenonsignificant trend in survival may have been due to a loss of sta-tistical power (one control replicate and one ninespine treatment).However, we do believe that replication was sufficient enough. Thisis based on the fact that our replication (3–4 replicates/treatment)was similar to (Skelly 1995; Davenport and Chalcraft 2012) or greaterthan (Smith et al. 1999) other experimental work with amphibians(Wilbur 1997). The consumptive effect of either fish species was notas devastating as other fish studies have shown (Rieger et al. 2004;Teplitsky et al. 2003, 2005) and originally hypothesized. The onlysignificant result, time to metamorphosis, indicates that the larvalperiod of Wood Frogs is shorter when fish are absent (Fig. 1C). Thesedifferences are biologically relevant because in dynamic environ-ments (e.g., ephemeral wetlands) even minor changes in time tometamorphosis can translate into consequences for individual sur-vival and overall fitness (Semlitsch et al. 1988; Wilbur 1997). Thereduction in time to metamorphosis in fish-free treatments couldindicate two different mechanisms. First, Wood Frog tadpoles oftenreduce activity in the face of predation and can perceive fish as apredation risk (Relyea 2001). Wood Frogs may have reduced activityin the enclosures in response to fish predators. All enclosures expe-rienced chemical cues from fish, because brook and ninespine stick-lebacks are present in the pond. Therefore, this would indicate thatthe physical presence of fish may be important to perception ofpredation risk. The reduction in activity may have led to atrade-off in foraging activity and time to metamorphosis thathas been documented in other amphibians (Relyea and Werner1999). Second, the reduction could indicate that individualswere able to quickly get enough energy to metamorphose (Fig. 1C).For this mechanism, the nonsignificant lower survival in fish-freecontrols may have been similar to a predator thinning effect (Van

868 Can. J. Zool. Vol. 91, 2013

Published by NRC Research Press

Can

. J. Z

ool.

Dow

nloa

ded

from

ww

w.n

rcre

sear

chpr

ess.

com

by

SUN

Y A

T S

TO

NY

BR

OO

K o

n 11

/11/

14Fo

r pe

rson

al u

se o

nly.

Buskirk and Yurewicz 1998). Predators often can lead to higher mor-tality of prey; however, the few individuals that do survive can expe-rience higher growth towing to additional access to availableresources (Davenport and Chalcraft 2012).

These results do not support our original hypothesis of strongpredatory fish effects and is not congruent with the Eaton et al.(2005) study. Our study did differ from the Eaton et al. (2005) studybecause we explicitly tested the effects of sticklebacks on WoodFrog survival and growth in one subarctic wetland. One potentialreason for these differences may be that lakes with small-bodiedfish assemblages in the boreal region of Alberta were composed ofmore species than sticklebacks. Specifically, small-bodied fishlakes were composed of minnows (family Cyprinidae), daces (fam-ily Cyprinidae), and sticklebacks (Eaton et al. 2005). It is entirelypossible that the population effects on Wood Frogs and BorealToads (Bufo boreas boreas Baird and Girard, 1852) observed in thesmall-bodied fish lakes were driven by the other small-bodied fish(dace and minnows). Sticklebacks are the only fish present inwetlands of the Churchill region.

It is surprising that Wood Frog survival was not lower with fish.Although there was a trend for differences in survival data, thistrend was the opposite of expected. Instead, survival tended to belower in fish-free treatments, but this trend was not significant(Fig. 1A). Although we cannot fully explain the reduction in sur-vival of fish-free treatments, we do have several hypotheses thatmay explain this variability in survival. First, Wood Frog tadpoleshave been known to alter behavior in the presence of fish (Chiversand Mirza 2001). Wood Frog tadpoles with fish may have reducedforaging for resources to reduce nonlethal encounters with fish.Although Wood Frog tadpoles can reach a size refuge to preventconsumption by sticklebacks, tadpoles can still be vulnerable tononlethal attacks (Figiel and Semlitsch 1991). Specifically, one ofus (P.A.S.) has witnessed tail and limb damage to tadpoles fromstickleback strikes. These nonlethal strikes can induce stress intadpoles and lead to changes in development (Maher et al. 2013).Thus, a potential nonconsumptive effect (via visual cues becauseall individuals in our study were subject to fish chemical cues) ofboth fish species could have led to a slower depletion of resourcesand higher survival of tadpoles with fish (Fig. 1A). In contrast,Wood Frog tadpoles in the fish-free treatment may not have re-duced foraging and could have led to a faster depletion of re-sources. The strength of intraspecific competition could havebeen just as important in the fish-free treatment as has beenreported with hylid frogs and fish (Resetarits et al. 2004). Second,Wood Frogs in the fish-free treatment could have become canni-balistic. Wood Frogs could become cannibalistic under certainconditions and those conditions may have manifested in only thefish-free treatment. For example, fish may have prevented canni-balism via nonconsumptive effects and actually facilitated WoodFrog survival in fish treatments (Werner and McPeek 1994). This isan unlikely scenario, however, for the reason that cannibalism isusually found in Wood Frogs when breeding is asynchronized(Petranka and Thomas 1995). All tadpoles in our study hatched atthe same time and were of the same cohort. Nonetheless, with ourdeployed experimental design, we do not have the ability to dis-cern between the intraspecific competition and cannibalism.Lastly, the impacts of stickleback predation may be regulatedpredominantly to the earlier life stages (eggs, embryos, or hatch-lings). Previous work in temperate regions has shown that am-phibians can reach a size refuge from gape-limited predators(Urban 2007). While the consumptive effect of sticklebacks may beregulated to earlier life stages, nonlethal injuries can still be det-rimental to tadpole survival and growth (Figiel and Semlitsch1991; Nunes et al. 2010).

Our study highlights the importance of experimental fieldworkbecause our large-scale surveys would lead to the conclusion thatfish are likely a major factor affecting amphibian subarctic popu-lations. This was not the case in the subarctic wetland of our studyand may not be the case in the subarctic region of the Hudson BayLowlands with larval Wood Frogs. With over one-third of theworld’s amphibians in decline, it is important to know what fac-tors are influencing amphibian populations. In our study, fish

Fig. 1. (A) Proportion of surviving Wood Frog (Lithobates sylvaticus)tadpoles among three treatments. (B) Mass at metamorphosis ofsurviving Wood Frog tadpoles among three treatments. (C) Time tometamorphosis for surviving Wood Frog tadpoles among threetreatments. Treatments are control (fish-free, tadpole only), ninespine(presence of one ninespine stickleback (Pungitius pungitius)), and brook(presence of one brook stickleback (Culaea inconstans)). Means are pooledfrom replicates. Bars are means ± 1 SE.

0

0.2

0.4

0.6

0.8

1

Surv

ival

(pro

porti

on)

A

0

0.1

0.2

0.3

0.4

Mas

s (g

)

B

54

56

58

60

62

Larv

al p

erio

d (d

ays)

control ninespine brook

A

B BC

Davenport et al. 869

Published by NRC Research Press

Can

. J. Z

ool.

Dow

nloa

ded

from

ww

w.n

rcre

sear

chpr

ess.

com

by

SUN

Y A

T S

TO

NY

BR

OO

K o

n 11

/11/

14Fo

r pe

rson

al u

se o

nly.

predation does not appear to be a significant impact on larvalWood Frog survival and growth in a subarctic wetland. Althoughtadpoles were small enough to be consumed by either fish speciesat the beginning of the experiment, we acknowledge that theeffect of fish predators may exist but be limited to early periods ofdevelopment (i.e., egg or embryo stages). Additional work isneeded to confirm this. An interesting result from our experimentsuggests that intraspecific competition or cannibalism may be asimportant in explaining Wood Frog subarctic population fluctua-tions. It is also possible that other predators may be responsiblefor changes in Wood Frog population dynamics. Invertebratepredators (e.g., dragonfly and beetle larvae) are also present inwetland communities, but are not as numerically common assticklebacks (W.B. Cash, L.A. Fishback, and J.M. Davenport, unpub-lished data). Additional research is needed to recognize what fac-tors are regulating larval amphibian survival and growth athigher latitudes.

AcknowledgementsWe thank D. Gibson for assistance in the field and the labora-

tory at the Churchill Northern Studies Centre (CNSC). We alsothank the late Clifford Paddock and K. Cash for their assistancewith construction and transportation of enclosures. Fundingwas provided by CNSC and the University of Central Arkansas.Two anonymous reviewers also greatly improved the quality ofthe manuscript. Animals were collected and handled in accor-dance with Manitoba Conservation and Water Stewardship per-mit 34-12 to W.B.C.

ReferencesAngelon, K.A., and Petranka, J.W. 2002. Chemicals of predatory mosquitofish

(Gambusia affinis) influence selection of oviposition site by Culex mosquitoes.J. Chem. Ecol. 28(4): 797–806. doi:10.1023/A:1015292827514. PMID:12035927.

Berven, K.A. 1990. Factors affecting population fluctuations in larval and adultstages of the wood frog (Rana sylvatica). Ecology, 71(4): 1599–1608. doi:10.2307/1938295.

Binckley, C.A., and Resetarits, W.J., Jr. 2005. Habitat selection determines abun-dance, richness and species composition of beetles in aquatic communities.Biol. Lett. 1: 370–374. doi:10.1098/rsbl.2005.0310. PMID:17148209.

Bleakney, S.C. 1958. Cannibalism in Rana sylvatica tadpoles. Herpetologica,14: 34.

Brooks, J.L., and Dodson, S.I. 1965. Predation, body size, and composition of plank-ton. Science, 150: 28–35. doi:10.1126/science.150.3692.28. PMID:17829740.

Chivers, D.P., and Mirza, R.S. 2001. Importance of predator diet cues in responsesof larval wood frogs to fish and invertebrate predators. J. Chem. Ecol. 27(1):45–51. doi:10.1023/A:1005663815856. PMID:11382066.

Davenport, J.M., and Chalcraft, D.R. 2012. Evaluating the effects of trophic com-plexity on a keystone predator by disassembling a partial intraguild preda-tion food web. J. Anim. Ecol. 81: 242–250. doi:10.1111/j.1365-2656.2011.01906.x.PMID:21950407.

Duguay, C.R., and Lafleur, P.M. 2003. Determining depth and ice thickness ofshallow sub-Arctic lakes using space-borne optical and SAR data. Int. J.Remote Sens. 24: 475–489. doi:10.1080/01431160304992.

Eaton, B.R., Tonn, W.M., Paszkowski, C.A., Danylchuk, A.J., and Boss, S.M. 2005.Indirect effects of fish winterkills on amphibian populations in boreal lakes.Can. J. Zool. 83(12): 1532–1539. doi:10.1139/z05-151.

Figiel, C.R., Jr., and Semlitsch, R.D. 1991. Effects of nonlethal injury and habitatcomplexity on predation in tadpole populations. Can. J. Zool. 69(4): 830–834.doi:10.1139/z91-125.

Gaston, K.J. 2003. The structure and dynamics of geographic ranges. OxfordUniversity Press, Oxford, UK.

Gotelli, N.J. 2008. A primer of ecology. 4th ed. Sinauer Associates Inc., Sunder-land, Mass.

Gray, S.M., and Robinson, B.W. 2002. Experimental evidence that competitionbetween stickleback species favours adaptive character divergence. Ecol.Lett. 5: 264–272. doi:10.1046/j.1461-0248.2002.00313.x.

Gray, S.M., Robinson, B.W., and Parsons, K.J. 2005. Testing alternative explana-tions of character shifts against ecological character displacement in brooksticklebacks (Culaea inconstans) that coexist with ninespine sticklebacks(Pungitius pungitius). Oecologia, 146(1): 25–35. doi:10.1007/s00442-005-0184-3.PMID:16151862.

Hardie, D.C., and Hutchings, J.A. 2010. Evolutionary ecology at the extremes ofspecies’ ranges. Environ. Rev. 18: 1–20. doi:10.1139/A09-014.

Hecnar, S.J., and M’Closkey, R.T. 1997. The effects of predatory fish on amphibianspecies richness and distribution. Biol. Conserv. 79(2–3): 123–131. doi:10.1016/S0006-3207(96)00113-9.

Hjernquist, M.B., Söderman, F., Jönsson, K.I., Herczeg, G., Laurila, A., andMerilä, J. 2012. Seasonality determines patterns of growth and age structureover a geographic gradient in an ectothermic vertebrate. Oecologia, 170:641–649. doi:10.1007/s00442-012-2338-4. PMID:22565493.

Hopey, M.E., and Petranka, J.W. 1994. Restriction of wood frogs to fish-freehabitats: how important is adult choice? Copeia, 1994: 1023–1025. doi:10.2307/1446726.

Kats, L.B., Petranka, J.W., and Sih, A. 1988. Antipredator defenses and the per-sistence of amphibian larvae with fishes. Ecology, 69(6): 1865–1870. doi:10.2307/1941163.

Kawahara, R., Miya, M., Mabuchi, K., Near, T.J., and Nishida, M. 2009. Sticklebackphylogenies resolved: evidence from mitochondrial genomes and 11 nucleargenes. Mol. Phylogenet. Evol. 50: 401–404. doi:10.1016/j.ympev.2008.10.014.PMID:19027080.

Laurila, A., Lindgren, B., and Laugen, A.T. 2008. Antipredator defenses along alatitudinal gradient in Rana temporaria. Ecology, 89(5): 1399–1413. doi:10.1890/07-1521.1. PMID:18543632.

Lawler, S.P., Dritz, D., Strange, T., and Holyoak, M. 1999. Effects of introducedmosquitofish and bullfrogs on the threatened California red-legged frog.Conserv. Biol. 13(3): 613–622. doi:10.1046/j.1523-1739.1999.98075.x.

Lindgren, B., and Laurila, A. 2010. Are high-latitude individuals superior com-petitors? A test with Rana temporaria tadpoles. Evol. Ecol. 24: 115–131. doi:10.1007/s10682-009-9294-4.

Ludwig, D., and Rowe, L. 1990. Life-history strategies for energy gain and preda-tor avoidance under time constraints. Am. Nat. 135: 686–707. doi:10.1086/285069.

Macrae, M.L., Bello, R.L., and Molot, L.A. 2004. Long-term carbon storage andhydrological control of CO2 exchange in tundra ponds in the Hudson BayLowland. Hydrol. Process. 18: 2051–2069. doi:10.1002/hyp.1461.

Maher, J.M., Werner, E.E., and Denver, R.J. 2013. Stress hormones mediatepredator-induced phenotypic plasticity in amphibian tadpoles. Proc. R. Soc. BBiol. Sci. 280(1758). doi:10.1098/rspb.2012.3075. PMID:23466985.

Morin, P.J. 1984. The impact of fish exclusion on the abundance and speciescomposition of larval odonates: results of short-term experiments in a NorthCarolina farm pond. Ecology, 65(1): 53–60. doi:10.2307/1939457.

Nunes, A.L., Cruz, M.J., Tejedo, M., Laurila, A., and Rebelo, R. 2010. Nonlethalinjury caused by an invasive alien predator and its consequences for ananuran tadpole. Basic Appl. Ecol. 11(7): 645–654. doi:10.1016/j.baae.2010.09.003.

Petranka, J.W., and Thomas, D.A.G. 1995. Explosive breeding reduces egg andtadpole cannibalism in the wood frog, Rana sylvatica. Anim. Behav. 50(3):731–739. doi:10.1016/0003-3472(95)80133-2.

Petranka, J.W., Kats, L.B., and Sih, A. 1987. Predator–prey interactions amongfish and larval amphibians: use of chemical cues to detect predatory fish.Anim. Behav. 35(2): 420–425. doi:10.1016/S0003-3472(87)80266-X.

Petranka, J.W., Hopey, M.E., Jennings, B.T., Baird, S.D., and Boone, S.J. 1994.Breeding habitat segregation of wood frogs and American toads: the role ofinterspecific tadpole predation and adult choice. Copeia, 1994: 691–697. doi:10.2307/1447185.

Petranka, J.W., Rushlow, A.W., and Hopey, M.E. 1998. Predation by tadpoles ofRana sylvatica on embryos of Ambystoma maculatum: implications of ecologicalrole reversals by Rana (predator) and Ambystoma (prey). Herpetologica, 54(1):108–115.

Redmer, M., and Trauth, S.E. 2005. Rana sylvatica LeConte, 1825. Wood frog. InAmphibian declines: the conservation status of United States species. Editedby M. Lannoo. University of California Press, Berkeley and Los Angeles.pp. 590–593.

Relyea, R.A. 2001. Morphological and behavioral plasticity of larval anurans inresponse to different predators. Ecology, 82(2): 523–540. doi:10.1890/0012-9658(2001)082[0523:MABPOL]2.0.CO;2.

Relyea, R.A. 2007. Getting out alive: how predators affect the decision to meta-morphose. Oecologia, 152: 389–400. doi:10.1007/s00442-007-0675-5. PMID:17356812.

Relyea, R.A., and Werner, E.E. 1999. Quantifying the relation between predator-induced behavior and growth performance in larval anurans. Ecology, 80:2117–2124. doi:10.1890/0012-9658(1999)080[2117:QTRBPI]2.0.CO;2.

Resetarits, W.J., Jr., and Wilbur, H.M. 1989. Choice of oviposition site by Hylachrysoscelis: role of predators and competitors. Ecology, 70(1): 220–228. doi:10.2307/1938428.

Resetarits, W.J., Jr., Rieger, J.F., and Binckley, C.A. 2004. Threat of predationnegates density effects in larval gray treefrogs. Oecologia, 138: 532–538. doi:10.1007/s00442-003-1466-2. PMID:14722747.

Reznick, D.A., Bryga, H., and Endler, J.A. 1990. Experimentally induced life-history evolution in a natural population. Nature, 346(6282): 357–359. doi:10.1038/346357a0.

Rieger, J.F., Binckley, C.A., and Resetarits, W.J., Jr. 2004. Larval performance andoviposition site preference along a predation gradient. Ecology, 85(8): 2094–2099. doi:10.1890/04-0156.

Rowe, L., and Ludwig, D. 1991. Size and timing of metamorphosis in complex lifecycles: time constraints and variation. Ecology, 72: 413–427. doi:10.2307/2937184.

SAS Institute Inc. 2010. SAS. Version 9.0 [computer program]. SAS Institute Inc.,Cary, N.C.

870 Can. J. Zool. Vol. 91, 2013

Published by NRC Research Press

Can

. J. Z

ool.

Dow

nloa

ded

from

ww

w.n

rcre

sear

chpr

ess.

com

by

SUN

Y A

T S

TO

NY

BR

OO

K o

n 11

/11/

14Fo

r pe

rson

al u

se o

nly.

Scott, W.B., and Crossman, E.J. 1998. Freshwater fishes of Canada. Galt HousePublications, Oakville, Ont.

Semlitsch, R.D., Scott, D.E., and Pechmann, J.H.K. 1988. Time and size at meta-morphosis related to adult fitness in Ambystoma talpoideum. Ecology, 69: 184–192. doi:10.2307/1943173.

Skelly, D.K. 1995. Competition and the distribution of spring peeper larvae.Oecologia, 103(2): 203–207. doi:10.1007/BF00329081.

Smith, G.R., Rettig, J.E., Mittelbach, G.G., Valiulis, J.L., and Schaack, S.R. 1999.The effects of fish on assemblages of amphibians in ponds: a field experi-ment. Freshw. Biol. 41: 829–837. doi:10.1046/j.1365-2427.1999.00445.x.

Symons, C.C., Arnott, S.E., and Sweetman, J.N. 2012. Nutrient limitation ofphytoplankton communities in subarctic lakes and ponds in Wapusk Na-tional Park, Canada. Polar Biol. 35: 481–489. doi:10.1007/s00300-011-1092-0.

Teplitsky, C., Plénet, S., and Joly, P. 2003. Tadpoles’ responses to risk of fishintroduction. Oecologia, 134(2): 270–277. PMID:12647168.

Teplitsky, C., Plénet, S., Léna, J.-P., Mermet, N., Malet, E., and Joly, P. 2005. Escapebehaviour and ultimate causes of specific induced defences in an anurantadpole. J. Evol. Biol. 18: 180–190. doi:10.1111/j.1420-9101.2004.00790.x. PMID:15669975.

Tollrian, R. 1995. Predator-induced morphological defenses: costs, life historyshifts, and maternal effects in Daphnia pulex. Ecology, 76(6): 1691–1705. doi:10.2307/1940703.

Urban, M.C. 2007. Predator size and phenology shape prey survival in temporaryponds. Oecologia, 154: 571–580. doi:10.1007/s00442-007-0856-2. PMID:17891545.

Van Buskirk, J., and Yurewicz, K.L. 1998. Effects of predators on prey growth rate:relative contributions of thinning and reduced activity. Oikos, 82: 20–28.doi:10.2307/3546913.

Wassersug, R. 1971. On the comparative palatability of some dry-season tadpolesfrom Costa Rica. Am. Midl. Nat. 86(1): 101–109. doi:10.2307/2423690.

Wellborn, G.A., Skelly, D.K., and Werner, E.E. 1996. Mechanisms creating com-munity structure across a freshwater habitat gradient. Annu. Rev. Ecol. Syst.27: 337–363. doi:10.1146/annurev.ecolsys.27.1.337.

Wells, K.D. 2007. The ecology and behavior of amphibians. University of ChicagoPress, Chicago, Ill.

Werner, E.E. 1986. Amphibian metamorphosis: growth rate, predation risk, andthe optimal size at transformation. Am. Nat. 138: 319–341. doi:10.1086/284565.

Werner, E.E., and Anholt, B.R. 1993. Ecological consequences of the trade-offbetween growth and mortality rates mediated by foraging activity. Am. Nat.142: 242–272. doi:10.1086/285537. PMID:19425978.

Werner, E.E., and McPeek, M.A. 1994. Direct and indirect effects of predators ontwo anuran species along an environmental gradient. Ecology, 75(5): 1368–1382. doi:10.2307/1937461.

Wilbur, H.M. 1997. Experimental ecology of food webs: complex systems intemporary ponds. Ecology, 78(8): 2279–2302. doi:10.1890/0012-9658(1997)078[2279:EEOFWC]2.0.CO;2.

Wooton, R.J. 1976. The biology of the sticklebacks. Academic Press, London, UK.

Davenport et al. 871

Published by NRC Research Press

Can

. J. Z

ool.

Dow

nloa

ded

from

ww

w.n

rcre

sear

chpr

ess.

com

by

SUN

Y A

T S

TO

NY

BR

OO

K o

n 11

/11/

14Fo

r pe

rson

al u

se o

nly.