Edinburgh Research Explorer

Transcription

1 Edinburgh Research Expler Plasticity of the duration of metamphosis in the African clawed toad Citation f published version: Walsh, PT, Downie, JR & Monaghan, P 0, 'Plasticity of the duration of metamphosis in the African clawed toad' Journal of Zoology, vol, no., pp. -. DOI:./j x Digital Object Identifier (DOI):./j x Link: Link to publication recd in Edinburgh Research Expler Document Version: Early version, also known as pre-print Published In: Journal of Zoology General rights Copyright f the publications made accessible via the Edinburgh Research Expler is retained by the auth(s) and / other copyright owners and it is a condition of accessing these publications that users recognise and abide by the legal requirements associated with these rights. Take down policy The University of Edinburgh has made every reasonable efft to ensure that Edinburgh Research Expler content complies with UK legislation. If you believe that the public display of this file breaches copyright please contact openaccess@ed.ac.uk providing details, and we will remove access to the wk immediately and investigate your claim. Download date:. Nov.

2 0 0 (BWUK JZO.PDF -Jul-0 : Bytes PAGES n operat=ravishankar) JZO Plasticity of the duration of metamphosis in the African clawed toad P. T. Walsh, J. R. Downie & P. Monaghan Division of Environmental and Evolutionary Biology, Institute of Biomedical and Life Sciences, University of Glasgow, Glasgow, UK Keywds metamphic climax; metamphic duration; Xenopus laevis; locomot perfmance; lifehisty variation. Crespondence Patrick T Walsh, Division of Environmental and Evolutionary Biology, Institute of Biomedical and Life Sciences, Graham Kerr Building, University of Glasgow, Glasgow, G QQ, UK. P.Walsh.@research.gla.ac.uk Received May 0; accepted June 0 doi:./j x Introduction Abstract Organisms with complex life cycles often have a number of distinct life-histy stages, each with substantially different requirements and risks (Wilbur, 0). The durations of different stages have been shown to vary with environmental circumstances. Variation in the duration of the larval period, f example, has been demonstrated in several studies covering a wide range of taxa (e.g. marine invertebrates: Twombly, ; insects: Blakley, ; fish: Policansky, ; amphibians: Werner, ; Harris, ). The duration of the transition between stages, often involving a metamphosis, has received considerable attention in terms of developmental mechanisms, but much less in terms of the effects of environmental circumstances. The duration of metamphosis is presumed to be minimized by selection because of the high vulnerability to predats during this period (Williams, ), due to reduced locomot ability and hence to be insensitive to environmental circumstances (Rose, 0). Contrary to Rose s (0) statement that metamphosis is a developmental phase of fixed duration, with no intraspecific variation, a recent analysis of metamphic duration has shown that it varies considerably among and within anuran species. This analysis also suggested that minimizing predation risk is not the sole fact in determining the duration of metamphosis, but that it is related to local growth conditions, as indicated by body Journal of Zoology. Print ISSN 0- In ganisms with complex life cycles, such as amphibians, selection is thought to have minimized the duration of metamphosis, because this is the stage at which predation risk is presumed to be highest. Consequently, metamphic duration is often assumed to show little if any environmentally induced plasticity, because the elevation in the extrinsic mtality risk associated with prolonging metamphosis is presumed to have selected f a duration as sht as is compatible with nmal development. We examined the extent to which metamphic duration in the anuran amphibian Xenopus laevis was sensitive to environmental temperature. Metamphic duration was influenced by body size, but independent of this effect, it was strongly influenced by environmental temperature: the duration at C was me than double that at and C. We also compared the vulnerability of larval, metamphosing and post metamphic Xenopus with predats by measuring their burst swimming speeds. Burst swim speed increased through development and while we found no evidence that it was reduced during metamphosis, it did increase sharply on completion of metamphosis. We therefe found no evidence of a substantial increase in vulnerability to predats during metamphosis compared with larval stages, and hence the slowing of metamphosis in response to temperature may not be as costly as has been assumed. condition and size, and to environmental temperature (Downie, Bryce & Smith, 0). Anuran amphibians are widely used in studies of the effect of environmental facts on life cycles because they go through clear and distinct stages. Following hatching from the egg, there is a larval period characterized by rapid growth. Metamphosis then occurs, during which time individuals do not feed and change from a tail-driven tadpole to a four-legged froglet; this is then nmally followed by further growth and then by sexual maturation. It has been suggested that metamphosis is a particularly vulnerable period in this group. F example, Arnold & Wassersug () found that chus frogs Pseudacris triseriata were me frequently captured by garter snakes during metamphosis than as pre-metamphic larvae post-metamphic juveniles. Additionally, it was established that the impaired locomot perfmance (measured as swimming endurance) of metamphosing individuals compared with tadpoles observed in this species was responsible f the increased vulnerability (Wassersug & Sperry, ). The locomot impairment thought to be typical of metamphosis is of critical imptance to the hypothesis that the duration of anuran metamphosis has been minimized by selection. While there have been several studies evaluating the effects of conditions during the larval period on the speed of larval and juvenile movement (Van Buskirk & Saxer, 0; Alvarez & Nicieza, 0; Altwegg & Reyer, Journal of Zoology ]] (0) c 0 The Auths. Journal compilation c 0 The Zoological Society of London J Z O B Dispatch:..0 Journal: JZO CE: Chandirka Journal Name Manuscript No. Auth Received: No. of pages: TE: Rathna Op: gsravi

3 Plasticity of metamphic duration P. T. Walsh, J. R. Downie and P. Monaghan (BWUK JZO.PDF -Jul-0 : Bytes PAGES n operat=ravishankar) 0 0 JZO 0), and differences in locomotion between pre- and postmetamphic urodeles (Shaffer, Austin & Huey, ; Azizi & Landberg, 0; Wilson, 0), the locomot perfmance of metamphosing individuals has only been examined in two other anuran species since Wassersug & Sperry s () iginal study. Watkins () did not find a significant difference in maximum burst swim speed between premetamphic individuals (Gosner stage ; Gosner, 0) and individuals at the start of metamphic climax (Gosner stage ) in Hyla regilla. However, Huey (0) demonstrated that burst speed in Bufo beas decreased during metamphosis, from a peak just befe the onset, in tandem with a decrease in tail length. In this study, we examined the plasticity of metamphic duration in the fully aquatic Xenopus laevis in relation to conditions during metamphosis. Our aims were twofold. Firstly, we examined the effect of experimentally imposed temperatures during metamphosis, taking into account variation in body size due to differential growth during the larval period. Secondly, we examined the predat avoidance capability of individuals at different stages in their development to determine whether this is impaired during metamphosis compared with the pre-metamphic larval and post-metamphic juvenile stages. There are several methods f assessing the predat avoidance capability in anurans, such as turning speed, endurance swimming, maximum attainable swim speed burst swim speed (Wassersug & Sperry, ; Huey, 0; Wassersug, ). Burst swim speed, the starting velocity from a stationary position, has been shown to be imptant f avoiding predation across all developmental stages (Miller, ; Azizi & Landberg, 0; Wilson, Kraft & Van Damme, 0). Therefe, we measured predat avoidance capability using burst swim speed. Methods and materials Animals and rearing conditions Approximately 0 wild-type eggs of X. laevis Daudin were obtained from St Andrews University (St Andrews, Fife, Scotland, UK) in 0. Tadpoles were kept in a single 0 L holding tank f days befe being transferred to smaller holding tanks, by which time they had reached approximately Nieuwkoop and Faber (NF) stage (Nieuwkoop & Faber, ). Each of the smaller holding tanks ( cm) contained tadpoles and was filled with c. L of aerated, de-chlinated, copper-free water. The temperature in the tanks was controlled by heaters, which were switched off during the night, and maintained the temperature at C during daytime and C at night. This range is representative of the temperatures that X. laevis would experience under natural conditions (Tinsley & Kobel, ). Thus, all tadpoles were exposed to the same range of experimental temperatures befe metamphosis; that this range encompassed the full range of temperatures to be used in the experiments during metamphosis avoided any acclimatization problems. Perspex sheets were placed over the tanks to reduce water loss from evapation, and a constant water level was maintained in all tanks. Tadpoles were exposed to a h: h light/dark cycle throughout the larval period. They were fed daily on c. 0. g of dried algal pellets per tank, ground and hydrated befe being suspended in water. Experimental protocol Checks were made three times a day f individuals showing emergence of one both felimbs, indicative of the commencement of metamphosis (stage 0). In total, tadpoles were transferred to separate, individual experimental tanks ( cm) at the onset of metamphosis. The remainder were not included due to natural mtality befe the onset of metamphosis because they did not commence metamphosis within the time span of the study. Individuals were checked three times a day f completion of metamphosis (stage ), identified by complete tail absption (tail length o0. mm). Experimental tanks were maintained at constant water temperatures of, and C, and tadpoles were allocated sequentially to the three temperature treatments at the onset of metamphosis. The and C treatments each had individuals, while the C treatment had individuals. Wet mass, after removal of surface water ( 0.00 g), snout vent length (SVL), head width, tail length and maximum tail depth measurements ( 0. mm, using callipers) were taken at the commencement of metamphosis. There were no significant differences in mass, SVL, head width, tail length, tail depth (Table ) body condition (mean: SEM; F, =., P=0.) at the start of metamphosis between the three temperature treatments. Tail length and depth ( 0. mm) measurements were taken every days during metamphosis, to determine whether metamphosis progressed with the same pattern in all three temperature treatments. At the completion of metamphosis, mass, SVL and head width were measured. Body condition at the start and end of metamphosis was calculated from body measurements using the fmula provided in Veith (), defined as condition=(mass/svl ) 00. Newly metamphosed juvenile individuals to be tested f locomot perfmance were maintained at their metamphic temperature until testing; after testing, they were transferred to a stock tank maintained at C. Table Starting measurements of different temperature treatment groups (ANOVA results are displayed; NS indicates non-significant difference) C C C Mass NS (g) SVL NS (mm) Head width NS (mm) Tail length NS (mm) Tail depth NS (mm) SVL, snout vent length. Journal of Zoology ]] (0) c 0 The Auths. Journal compilation c 0 The Zoological Society of London

4 P. T. Walsh, J. R. Downie and P. Monaghan Plasticity of metamphic duration 0 0 (BWUK JZO.PDF -Jul-0 : Bytes PAGES n operat=ravishankar) Q JZO Swimming speed data collection Measurements of swimming speed were taken from NF stages, comprising the larval stages (stages ), the middle of metamphosis (stage, when 0% of the tail was reabsbed) and soon after metamphosis was complete (stage +). Pre-metamphic tadpoles were grouped as follows: stages (small hind-limb buds); stages (differentiation and flattening of hind foot); and stages (separation of toes and development of hind legs). Mid-metamphosis (stage ) was selected because it is likely to represent the most critical period during which metamphosis would be impaired, because the tail is being rapidly absbed and the hind limbs are still developing. Owing to the sensitivity of the equipment being used, filming of swimming had to be perfmed in one room (air temperature C) f all temperature treatments. Individuals to be filmed were brought into the room in a shallow dish h befe being filmed to acclimatize to the temperature (the water temperature ranged between and C, due to heat from the lighting equipment required f filming). The h time period was selected as an acceptable period f acclimatization, as a longer period may have affected the developmental processes being investigated. Water temperature at the time of trials was recded and included in the analysis as a covariate. This did not significantly affect locomot perfmance at any of the three developmental stages (tadpoles: F, =., P=0.; metamphs: F, =., P=0.; and juveniles: F, =., P=0.). A Photron FASTCAM-PCI high-speed camera was used to capture video footage. The camera was placed facing directly down into a tank cm long, with laminated gridpaper along the bottom; there was a cm wide crid through the middle of the grid, filled with c. cm depth of water. The camera was mounted sufficiently far above the water surface (0. m) relative to water depth ( cm) to minimize parallax errs in determining position. Each individual was placed at one end of the tank and allowed to settle, and then gently prodded at the base of the tail with a glass stirring rod wrapped in red electrical tape to elicit an escape response. These procedures limited the tadpoles to swimming fward in response to the stimulus. Filming was carried out at 0 frames per second (fps) and the camera recded up to s of swimming; the swim speed of each individual was measured five times with c. min break between each. There was no evidence of habituation through the series of trials in any of the three developmental groupings (pre-metamphic: r =0.0, P=0.0; metamphic: r =0.00, P=0.; juvenile: r =0.0, P=0.). As in other studies (e.g. Wilson et al., 0), the first 0 ms ( frames) following initial movement were used representing burst speed, a critical variable in fleeing from a predat (Watkins, ; Dayton et al., 0). Data analysis Data are presented as mean SEM. F-tests were used to determine whether variation in metamphic duration differed between temperature treatments. Principal components analysis was carried out on the four body measurements (SVL, head width, tail length and depth) f use in analysis on metamphic duration. A single fact was extracted (PC), which explained.% of the variation, representing predominately body rather than tail measurements. w was used to determine differences in metamphic mtality between temperature treatments. Tail loss during metamphosis was assessed using -parameter, sigmoid regression (SigmaPlot, Systat Software Inc.). The remaining analyses were perfmed using analysis of variance with post hoc Tukey tests (SPSS, SPSS Inc.). F film analysis, Photron Motion Tools software was used, which allowed digital analysis, giving the distance per frame (velocity) and the distance travelled. The maximum measurement achieved from the five trials was used f data analysis. Befe stage 0, individual tadpoles were taken at random from the small holding tanks f swim testing; they were then returned to the tank, and so it was not possible to track them as individuals. Therefe, the data on locomot perfmance between larval and metamphic stages are treated as independent samples. Comparisons were made between four developmental groups: three pre-metamphic stages (all experiencing the fluctuating temperature environment) and one at mid-metamphosis (stage ). Three such analyses were perfmed to allow comparisons between larval and metamph swim speeds at the three different temperature treatments experienced only after metamphosis commenced. Comparative analysis of locomot perfmance between metamphic and juvenile stages (stages and ) was perfmed on known individuals using a linear mixed model with repeated measures. Results Metamphic duration Metamphic duration differed significantly between the three experimental treatments. It was the slowest at the lowest temperature, taking twice as long at C than at C (Fig. ). The difference between the two higher temperatures, while significant, was relatively small (c. day). The variability in metamphic duration differed significantly among the temperature treatments, with the C treatment being the least and the C treatment the most variable (F=., P=0.000; C: range.0.0 days, CV:.%; C:..00 days, CV:.%; C:.00. days, CV:.%). As the body size (PC) increased, the duration of metamphosis increased (F, =., P=0.00). Metamphic duration was not influenced by starting body condition (F, =.0, P=0.0; Fig. ). The effects of temperature were significant (F, =., P=0.0) and there was no interaction between temperature and body size (F, =0., P=0.) temperature and condition (F, =0.0, P=0.). Journal of Zoology ]] (0) c 0 The Auths. Journal compilation c 0 The Zoological Society of London

5 Plasticity of metamphic duration P. T. Walsh, J. R. Downie and P. Monaghan (BWUK JZO.PDF -Jul-0 : Bytes PAGES n operat=ravishankar) 0 0 JZO Duration of metamphosis (days) Temperature treatment ( C) Figure Duration of metamphosis in days ( SEM) f the three temperature treatments. Mtality levels were significantly different among treatments. No mtality occurred in the C treatment. The C treatment had the highest mtality rate ( mtalities,.%), while the C treatment had less (six mtalities, %) (w =.0, d.f.=, P=0.00). Tail loss during metamphosis Tail re-absption followed a sigmoid pattern, with a rapid decrease in tail length occurring between days and in the two higher temperatures ( and C) and between day and in the C treatment (Fig. ). Inflection points (x 0 ) and steepness of curves (b) were significantly different between the three treatments (x 0 : F, =., P=0.000; b: F, =0., P=0.000); post hoc analysis showed that this difference was between C and the two higher temperature treatments. Duration of metamphosis (days)... Ln (body condition)..0 Figure Duration of metamphosis in days in relation to body condition at the onset of metamphosis ( C:, solid line; C:, dashed line, C:., dotted line). Tail length (mm) 0 0 Days since commencement of metamphosis Locomot perfmance C C C Figure Decrease in mean tail length ( SEM) through metamphosis ( C: y ¼ :=½ þ e ððx :Þ=0:Þ Š r =0., P=0.000; C: y ¼ :=½ þ e ððx :Þ=0:Þ Š r =0., P=0.000; C: y ¼ :=½ þ e ððx :Þ=0:Þ Š r =0.0, P=0.000). SVL had a significant positive effect on burst swim speed during the larval (F, =., P=0.00) and juvenile stages (F, =., P=0.000), but only marginally during metamphosis (F, =.000, P=0.0). The temperature treatment experienced during metamphosis had a significant effect on metamph burst swim speed (F, =., P=0.00), but this was not found in juveniles (F, =0., P=0.). During metamphosis, individuals from the C treatment had a significantly faster burst swim speed than the higher and lower temperature treatment groups, which were not significantly different from one another (Fig. ). To assess whether locomotion is impaired during metamphosis, comparisons were made between larvae and metamphs and between metamphs and juveniles. The comparisons between the three larval stage groups and metamphic individuals at the three temperature treatments all showed a significant relationship between maximum swimming speed and stage ( C: F, =.0, P=0.000; C: F, =., P=0.000; C: F, =., P=0.00; Fig. ). We found no evidence of locomot impairment during metamphosis: at and C, metamphic burst swim speeds did not differ from burst speeds attained by pre-metamphic larvae. In fact, metamphs at C actually swam significantly faster than the late-stage larvae. The maximum swimming speed increased as individuals progressed through larval development and continued to increase up to mid-metamphosis (NF stage ). Therefe, swimming speed at metamphosis was not reduced compared with late-stage, premetamphic larvae of similar sizes and was significantly faster than smaller, early-stage larvae. Crecting f metamphic temperature treatment (F,0.0 =0., P=0.) and SVL (F,0. =.0, P=0.0), at the end of metamphosis individuals swam Journal of Zoology ]] (0) c 0 The Auths. Journal compilation c 0 The Zoological Society of London

6 P. T. Walsh, J. R. Downie and P. Monaghan Plasticity of metamphic duration 0 0 (BWUK JZO.PDF -Jul-0 : Bytes PAGES n operat=ravishankar) Burst swimming speed (cm s ) JZO N F stage significantly faster than they did during metamphosis (F,0. =., P=0.000). Discussion N F stage Metamphic duration N F stage Developmental stages N F stage N F stage Figure Burst swimming speed ( SEM) at different developmental stages. At mid-metamphosis, the three temperature treatments are represented individually (: C, m: C, : C). Our results show that metamphic duration is sensitive to environmental temperature. At temperatures commonly experienced by X. laevis (c. C), metamphosis takes c. days (Huang et al., 0). Metamphosing in cold temperatures resulted in me than double the standard length of time to complete metamphosis. Metamphosis at the two higher temperature treatments ( and C) was much closer in duration, with the duration at C taking about a day less than at C. This result indicates that above a certain temperature threshold, the speed of metamphosis is optimized, as is suggested by Downie et al. (0) f other species. There was also substantial variability in metamphic duration within the temperature treatment groups. Body size was a main component explaining this variation. Contrary to the prediction of Downie et al. (0), metamphic duration was not related to the index of body condition. However, Downie et al. (0) did not use the term body condition in the precise manner defined by Veith (). The lack of relation with body condition might be a result of the rearing conditions. All individuals were reared at relatively low densities, with an abundance of food on a daily basis. Therefe, the condition at the start of metamphosis was relatively high and less variable than would be expected in the wild. Rearing individuals at different densities lower food availabilities may allow the effects of body condition on metamphic duration to be studied in me detail. Within-group variability in metamphic duration differed between the three temperatures. At C, metamphosis was completed in the fastest time, but the degree of variation was the highest. Additionally, the mtality rate was the highest in this group. The higher mtality indicates that there may be costs associated with such a rapid rate of metamphosis, as has been found in rapid growth rates (Arendt, ). At C, the temperature treatment closest to what would be experienced by X. laevis in the field (Tinsley & Kobel, ), the variation was the smallest. This indicates that at the temperature a species is adapted f, a thermal developmental optimum is established (Stahlberg, Olsson & Uller, 0). Locomot perfmance The temperature experienced during metamphosis had different effects on swimming speed during and after metamphosis. At stage, the C temperature treatment resulted in individuals having a faster maximum swim speed, with either extreme having slower speeds. Alvarez & Nicieza (0) demonstrated a similar result in juvenile jumping perfmance in the Iberian painted frog Discoglossus galganoi, with perfmance peaking at an optimal temperature and decreasing as the temperature increased decreased. Miller () found a similar trend in adult Xenopus, with the highest locomot perfmance occurring at C and perfmance decreasing with elevated lowered temperatures. Additionally, in our study, size was found to have only a marginally significant effect on swimming speed during metamphosis, possibly due to other facts, such as the progression of hind limb development, being me critical. This result suggests that the temperature experienced during metamphosis may have some effect on musculoskeletal development, such as the type capacity of hind limb musculature being developed during metamphosis, which influences swimming perfmance. On completion of metamphosis, temperature no longer had an effect on swim speed, but size, specifically SVL, did have an effect. This result contrasts with the rept of a temperature effect on juvenile locomotion repted by Alvarez & Nicieza (0). However, in their study, the temperature treatments were experienced throughout the larval period, while in this study individuals were only subjected to different temperatures from the start of metamphosis. Thus, the temperature effects found by Alvarez & Nicieza (0) may be a consequence of effects operating during larval development. Comparisons between larval and juvenile locomot perfmance may be complicated by differences in thermal capacity between the musculature that the two stages use f locomotion. Sherman (0) showed that, in Bufo woodhousii fowleri, the developing hind limbs were less able to cope with thermal stress than the tail, suggesting that extremely high temperatures would favour tadpole speed over juvenile speed. However, this is not likely to be the case in this study, because the thermal range used was well within the species tolerance and well below the thermal maxima of Xenopus (Sherman & Levitis, 0). Additionally, differences in thermal preference between larval and adult amphibians have not been shown in all species (e.g. Triturus cristatus: Journal of Zoology ]] (0) c 0 The Auths. Journal compilation c 0 The Zoological Society of London

7 Plasticity of metamphic duration P. T. Walsh, J. R. Downie and P. Monaghan (BWUK JZO.PDF -Jul-0 : Bytes PAGES n operat=ravishankar) 0 0 JZO Wilson, 0), and it has even been suggested that the thermal environment of adult and larval Xenopus is likely to be the same, as they occur in the same ponds (Sherman & Levitis, 0), but this will be dependent on pond depth. Comparison of metamph predat avoidance In this study, even when crected f size, individuals did not experience the decrease in locomot perfmance predicted and observed by Wassersug & Sperry (), using chus frogs. In X. laevis, metamphic swimming speed was slightly, but not significantly, faster than speeds displayed by larval individuals of a similar size. There was a dramatic increase in swimming perfmance on completion of metamphosis. There could be several explanations f this: () this species is fully aquatic, and so the transition costs between terrestrial and aquatic locomotion are not present. () Research on Xenopus development has shown that during metamphic climax, individuals are able to use both tail and hind limb-based locomotion (Combes et al., 0), which may be different from other species (e.g. species with rapid tail loss; see Downie et al., 0), and could improve metamphs locomotion perfmance. () Hind limb development, which has been shown during the larval stage to aid swimming (Park et al., 0), may confer a greater advantage to metamphosing Xenopus because of the large, well-developed webbed feet found in this species. It has been shown that drag fces are considerably increased during metamphic climax due to felimb emergence (Dudley, King & Wassersug, ), but the advantage of the thrust generated by the developing hind limbs could offset this cost. However, Dudley et al. () used Rana catesbeiana tadpoles, which have a narrower, me streamlined anteri end compared with X. laevis tadpoles. () Finally, this study examined locomotion perfmance as burst speed, whereas Wassersug & Sperry () examined swimming endurance. However, burst speed is me accurate than endurance as an indicat of predat evasion (see Dayton et al., 0). In two studies on locomot perfmance, where burst speed was examined, conflicting results were observed. Huey (0) did show a decrease in burst speed in B. beas during metamphosis, with individuals with longer tails at stage (c. NF stage ) having faster swim speeds. Watkins (), wking on the Pacific tree frog, found that the maximum burst speed was not significantly different between premetamphic and stage individuals, but argued that using the mean of some of the speed trials, there was a significant decrease in locomot perfmance in metamphic individuals. Watkins () did not investigate post-metamphic locomot perfmance in the Pacific tree frog. In the present study, using either mean maximum, burst speed did not demonstrate a decrease in perfmance during metamphosis in any of the three temperature treatments. In conclusion, our data show that metamphic duration varies with both environmental temperature and body size, and that locomotion is not impaired during metamphosis compared with the pre-metamphic stage. These findings present some fundamental complications f the idea that selection fuelled by high predation risk has minimized the duration of metamphosis. It is possible that variation in natural predation risk results in different strengths of selective pressure f minimizing metamphic duration, and it would be interesting to compare populations from high and low predation risk habitats. It would also be interesting to know me about the costs associated with rapid metamphosis. Acknowledgements We would like to thank Isobel Maynard f providing the eggs used in this experiment and two anonymous referees f helpful comments. P.T.W. would like to thank the Carnegie Trust f the Universities of Scotland f providing the PhD studentship, during which this research was conducted; the Louise Hiom Award provided funds f purchasing equipment. References Altwegg, R. & Reyer, H.U. (0). Patterns of natural selection on size at metamphosis in water frogs. Evolution,. Alvarez, D. & Nicieza, A. (0). Effects of induced variation in anuran larval development on postmetamphic energy reserves and locomotion. Oecologia,. Arendt, J.D. (). Adaptive intrinsic growth rates: an integration across taxa. Q. Rev. Biol.,. Arnold, S.J. & Wassersug, R.J. (). Differential predation on metamphic anurans by garter snakes (Thamnophis): social behaviour as a possible defence. Ecology,. Azizi, E. & Landberg, T. (0). Effects of metamphosis on the aquatic escape response of the two-lined salamander (Eurycea bislineata). J. Exp. Biol.,. Blakley, N. (). Life-histy significance of size-triggered metamphosis in milkweed bugs (Oncopeltus). Ecology,. Combes, D., Merrywest, S.D., Simmers, J. & Sillar, K.T. (0). Developmental segregation of spinal netwks driving axial- and hindlimb-based locomotion in metamphosing Xenopus laevis. J. Physiol.-(Lond.),. Dayton, G.H., Saenz, D., Baum, K.A., Langerhans, R.B. & DeWitt, T.J. (0). Body shape, burst speed and escape behaviour of larval anurans. Oikos,. Downie, J.R., Bryce, R. & Smith, J. (0). Metamphic duration: an under-studied variable in frog life histies. Biol. J. Linn. Soc.,. Dudley, R., King, V.A. & Wassersug, R.J. (). The implications of shape and metamphosis f drag fces on a generalized pond tadpole (Rana catesbeiana). Copeia,. Journal of Zoology ]] (0) c 0 The Auths. Journal compilation c 0 The Zoological Society of London

8 P. T. Walsh, J. R. Downie and P. Monaghan Plasticity of metamphic duration 0 0 (BWUK JZO.PDF -Jul-0 : Bytes PAGES n operat=ravishankar) JZO Gosner, K.L. (0). A simplified table f staging anuran embryos and larvae with notes on identification. Herpetologica,. Harris, R.N. (). The anuran tadpole: evolution and maintenance. In Tadpoles: the biology of anuran larvae:. McDiarmid, R.W. & Altig, R. (Eds). Chicago: University of Chicago Press. Huang, H., Cai, L., Remo, B.F. & Brown, D.D. (0). Timing of metamphosis and the onset of the negative feedback loop between the thyroid gland and the pituitary is controlled by type II iodothyronine deiodinase in Xenopus laevis. Proc. Natl. Acad. Sci. USA,. Huey, R.B. (0). Sprint velocity of tadpoles (Bufo beas) through metamphosis. Copeia 0, 0. Miller, K. (). Effect of temperature on sprint perfmance in the frog Xenopus laevis and the salamander Necturus maculosus. Copeia,. Nieuwkoop, P.D. & Faber, J. (). Nmal table of Xenopus laevis (Daudin). New Yk: Garland Publishing Inc.. Park, J.C., Kim, H.S., Yamashita, M. & Choi, I. (0). Development of contractile and energetic capacity in anuran hindlimb muscle during metamphosis. Physiol. Biochem. Zool.,. Policansky, D. (). Size, age and demography of metamphosis and sexual-maturation in fishes. Am. Zool.,. Rose, C.S. (0). Integrating ecology and developmental biology to explain the timing of frog metamphosis. Trends Ecol. Evol.,. Shaffer, H.B., Austin, C.C. & Huey, R.B. (). The consequences of metamphosis on salamander (Ambystoma) locomot perfmance. Physiol. Zool.,. Sherman, E. (0). Ontogenetic change in thermal tolerance of the toad Bufo woodhousii fowleri. Comp. Biochem. Phys. A,. Sherman, E. & Levitis, D. (0). Heat hardening as a function of developmental stage in larval and juvenile Bufo americanus and Xenopus laevis. J. Therm. Biol.,. Stahlberg, F., Olsson, M. & Uller, T. (0). Population divergence of developmental thermal optima in Swedish common frogs, Rana temparia. J. Evol. Biol.,. Tinsley, R.C. & Kobel, H.R. (). The biology of Xenopus. Oxfd: Clarendon Press. Twombly, S. (). Timing of metamphosis in a freshwater crustacean: comparison with anuran models. Ecology,. Van Buskirk, J. & Saxer, G. (0). Delayed costs of an induced defense in tadpoles? Mphology, hopping, and development rate at metamphosis. Evolution,. Veith, M. (). Weight condition in Triturus alpestris methodological and ecological aspects. In. Van Gelder, J., Strijbosch, H. & Berger, P. (Eds). Hijmegen: Societas Herpetologica Europaea. Wassersug, R.J. (). Locomotion in amphibian larvae ( Why aren t tadpoles built like fishes? ). Am. Zool.,. Wassersug, R.J. & Sperry, D.G. (). The relationship of locomotion to deifferential predation on Pseudacris triseriata (Anura: Hylidae). Ecology,. Watkins, T.B. (). Predat-mediated selection on burst swimming perfmance in tadpoles of the Pacific tree frog, Pseudacris regilla. Physiol. Zool.,. Watkins, T.B. (). The effect of metamphosis on the repeatability of maximal locomot perfmance in the Pacific tree frog Hyla regilla. J. Exp. Biol. 0,. Werner, E.E. (). Amphibian metamphosis: growth rate, predation risk, and the optimal size at transfmation. Am. Nat.,. Wilbur, H.M. (0). Complex life-cycles. Annu. Rev. Ecol. Syst.,. Williams, G. (). Adaption and natural selection. Princeton, NJ: Princeton University Press. Wilson, R.S. (0). Consequences of metamphosis f the locomot perfmance and thermal physiology of the newt Triturus critatus. Physiol. Biochem. Zool.,. Wilson, R.S., Kraft, P.G. & Van Damme, R. (0). Predat-specific changes in the mphology and swimming perfmance of larval Rana lessonae. Funct. Ecol.,. Journal of Zoology ]] (0) c 0 The Auths. Journal compilation c 0 The Zoological Society of London Q

9 Auth Query Fm Journal JZO Article Dear Auth, During the copy-editing of your paper, the following queries arose. Please respond to these by marking up your proofs with the necessary changes/additions. Please write your answers clearly on the query sheet if there is insufficient space on the page proofs. If returning the proof by fax do not write too close to the paper's edge. Please remember that illegible mark-ups may delay publication. Query No. Description Auth Response AQ: Please provide manufacturer infmation f Photron FASTCAM-PCI high-speed camera: Q company name, town, state (if USA) and country. Q AQ: Please provide book title f the reference Veith ().

10 MARKED PROOF Please crect and return this set Please use the proof crection marks shown below f all alterations and crections. If you wish to return your proof by fax you should ensure that all amendments are written clearly in dark ink and are made well within the page margins. Instruction to printer Leave unchanged Insert in text the matter indicated in the margin Delete Substitute character substitute part of one me wd(s) Change to italics Change to capitals Change to small capitals Change to bold type Change to bold italic Change to lower case Change italic to upright type Change bold to non-bold type Insert superi character Insert inferi character Insert full stop Insert comma Insert single quotation marks Textual mark under matter to remain through single character, rule underline through all characters to be deleted through letter through characters under matter to be changed under matter to be changed under matter to be changed under matter to be changed under matter to be changed Encircle matter to be changed through character where required Marginal mark New matter followed by new character new characters under character e.g. over character e.g. and/ Insert double quotation marks and/ Insert hyphen Start new paragraph No new paragraph Transpose Close up linking characters Insert substitute space between characters wds through character where required Reduce space between characters wds between characters wds affected