phenotypes of hatchling lizards, regardless of overall mean incubation temperature

Functional Ecology 2004 Seasonal shifts in nest temperature can modify the Blackwell Publishing, Ltd. phenotypes of hatchling lizards, regardless of overall mean incubation temperature R. SHINE* Biological Sciences A08, University of Sydney, New South Wales 2006, Australia Summary 1. The thermal regimes experienced by a reptilian egg can influence phenotypic traits (size, sex, shape, locomotor performance, etc.) of the hatchling that emerges from that egg. Natural nests of the oviparous scincid lizard Bassiana duperreyi in the Brindabella Range of south-eastern Australia show strong seasonal shifts in temperature: first and last weeks of incubation often differ by >5 C. 2. Eggs of B. duperreyi were incubated under thermal regimes with identical overall mean values for average temperature and diel range, but differing in the sequence of temperatures. Some eggs were kept at 18 ± 5 C throughout incubation; others went gradually from cool (16 ± 5 C) to warm (20 ± 5 C); and others from warm (20 ± 5 C) to cool (16 ± 5 C). 3. These treatments significantly modified not only incubation periods (stable mean temperatures delayed hatching), but also hatchling traits: progressively decreasing temperatures yielded hatchlings with a higher incidence of deformities, smaller body size, relatively longer tails, and reduced locomotor performance than siblings from increasing temperatures. 4. Seasonal shifts in incubation temperatures are widespread, and may generate important variation in hatchling phenotypes. Sensitivity to such shifts may influence phenomena such as nest-site selection, the seasonal timing of nesting, and the evolution of viviparity. Key-words: Nest temperatures, oviparity, phenotypic plasticity, reptile, seasonality Functional Ecology (2004) Ecological Society Introduction The environment can exert a powerful direct effect on animal phenotypes at a variety of stages within the life history: for example, changes in the type or abundance of food, or the presence of predators, can induce major morphological variations (Wilbur, Tinkle & Collins 1974). Environmental influences that act early in ontogeny are likely to be the most significant in this respect. Because development typically proceeds by way of stepwise differentiation, phenotypic changes induced early in development can have later substantial effects (Taning 1952; Rhen & Lang 1995). For example, medical research provides abundant examples of the long-term and sometimes devastating consequences of externally mediated disruptions to embryogenesis (Henry & Ulijaszek 1996). These effects occur despite the considerable protection afforded to embryos of viviparous species, which are buffered from environmental *Author to whom correspondence should be addressed. E-mail: rics@bio.usyd.edu.au disturbance by physical and physiological mechanisms within the mother s body (Albon, Guinness & Clutton- Brock 1983). As in most groups of animals, the majority of reptile species are oviparous (Tinkle & Gibbons 1977; Blackburn 1995) and, thus, most embryonic development occurs external to the mother s body. In consequence, the physical conditions experienced during embryogenesis depend upon the nest environment, with no opportunity for maternal buffering apart from judicious selection of an appropriate nesting site. Thermal and hydric conditions during incubation exert powerful influences on the phenotypic traits of hatchling reptiles. For example, higher mean incubation temperatures can accelerate development (and thus hasten the date of hatching), and modify the size, sex, body shape, locomotor performance and behaviour of hatchlings (Deeming & Ferguson 1991). Although the general proposition that incubation temperatures influence the phenotypes of hatchling reptiles is now well established, the details of that influence remain obscure. Initial experimental work was 43

44 R. Shine based on incubation of eggs at constant temperature in the laboratory, but more recent studies have shown that (a) temperatures within natural nests are far from constant, especially in shallow-nesting temperate-zone squamates, and (b) thermal variance, as well as mean temperatures, can significantly modify phenotypic traits of hatchling reptiles. For example, sex determination in some turtle species is influenced by diel thermal variation within the nest as well as by mean incubation temperature (Bull 1980; Georges 1989). In montane lizards, an increase in the diel range of temperature variation, without changing the mean, can strongly modify developmental rates and hatchling phenotypes (Shine & Harlow 1996; Shine, Elphick & Harlow 1997). Although the strong diel thermal variation in shallow nests clearly plays a role in shaping hatchling phenotypes, variation in temperature at other time-scales also may be important. For example, brief periods of cooler-than-usual or hotter-than-usual weather conditions can modify temperatures inside lizard nests, and influence the phenotypes of hatchling lizards that emerge from those nests (Shine & Elphick 2001). The magnitude of this phenotypic effect depends on the stage of embryogenesis at which the thermal fluctuation occurs (Shine & Elphick 2001). Similarly, rapid cooling even close to the end of incubation can modify phenotypic traits of hatchling lizards (Shine 2002). Studies on this topic increasingly recognize the importance of measuring actual nest temperatures and then replicating these patterns in laboratory experiments, rather than relying on simplified overall average temperature. Nonetheless, one of the most obvious forms of thermal variation within natural nests has so far attracted little attention. In many parts of the world, soil temperatures show strong seasonal variation and, thus, the thermal conditions in a nest at the time the eggs are laid may differ considerably from those present when the eggs hatch 2 or 3 months later. For example, eggs laid in early spring when the soil is cool will experience a progressive increase in mean temperature during incubation, whereas eggs laid later in the year in midsummer (perhaps by the same female) will experience declining temperatures during development (e.g. Webb 1986; Webb & Cooper-Preston 1989). Because the embryo s phenotypic response to a specific set of incubation conditions will differ during ontogeny (e.g. it is temperatures during mid-incubation that determine sex in many taxa: Webb et al. 1987; Georges 1989), such a seasonal shift in incubation regimes might directly modify hatchling phenotypes. Importantly, this might be true even if two nests (or the same nest-site at different times of the year) have identical overall mean temperatures and diel thermal ranges, calculated over the entire duration of incubation. For example, suppose that early stage embryos are most sensitive to temperature, as may often be the case ( Taning 1952). A regime of rising temperatures will expose the most sensitive period in embryogenesis to relatively cool conditions, whereas the less-sensitive later stages are kept warmer. In contrast, a regime of falling temperatures will expose the sensitive early embryos to warmer conditions, and the less sensitive older embryos to lower temperatures. Thus, seasonal cycles in nest temperatures might significantly influence reptilian embryogenesis, even if these cycles do not modify the thermal variables (the overall mean incubation temperature and the diel range of temperatures, calculated over the entire incubation period) that previous studies have identified as significant influences on hatchling phenotypes. The present paper describes my attempt to test this hypothesis. Materials and methods SPECIES AND STUDY AREA The Three-lined Skink, Bassiana duperreyi, is a mediumsized (to 80 mm snout vent length, SVL) insectivorous scincid lizard widely distributed through cool-climate (especially, montane) habitats in south-eastern Australia (Cogger 2000; Hutchinson, Swain & Driessen 2001). Females lay a single clutch of three to nine eggs early in summer, under rocks or logs in open areas exposed to high levels of solar radiation (Shine & Harlow 1996). Communal oviposition is frequent (Pengilley 1972). My colleagues and I have conducted extensive research on the reproductive biology of B. duperreyi in the Brindabella Range 40 km west of Canberra, Australian Capital Territory (Shine & Harlow 1996). Hatchling phenotypes are strongly influenced by thermal regimes during incubation, both in the field and in the laboratory, but not by hydric variation over the range of soil waterpotentials recorded in natural nests (Flatt et al. 2001). FIELD STUDY I used miniature thermal data-loggers (thermochron ibuttons, Dallas Semiconductor, Dallas, TX; diameter 15 mm, height 6 mm, mass 3 3 g) to monitor the seasonal shift in temperatures within ten natural nests of B. duperreyi at two sites (six at 1240 and four at 1615 m asl: see below). The sites where I placed thermochrons had been used as lizard nests in previous years, and eight of the ten were also used as nests during the period when the thermochrons were in place. My data show not only nest temperatures during the incubation period, but also thermal regimes in the same sites both before and after the eggs were present. Monitoring of thermal regimes in 95 natural nests over previous years provides additional data on temperatures actually experienced by eggs during the incubation period. LABORATORY STUDY I collected recently oviposited (<1 week incubation, based on extensive previous records of oviposition timing) eggs from Picadilly Circus (1240 m elevation,

45 Thermal effects on lizard phenotypes 148 50 E, 35 21 S, n = 324 eggs from 20 nests), Mount Ginini (1615 m, 148 46 E, 35 32 S; n = 86 eggs from 8 nests) and Mount Gingera (1670 m, 148 47 E, 35 34 S; n = 176 eggs from 5 nests) on 17 and 18 December 2001. The eggs were transported to the University of Sydney, where they were placed in individual glass vials containing moist vermiculite (water potential 200 kpa), sealed with plastic foodwrap to prevent evaporation (Shine & Harlow 1996). Using cycling-temperature incubators, the eggs from each site were pooled and then randomly allocated among three treatments (thus, total n = 151 153 eggs per treatment) with identical mean temperatures (18 C) and daily thermal ranges (10 C), but differing in the temporal sequence of thermal regimes, as follows: 1. Stable (18 ± 5 C throughout development). 2. Rising treatment. These eggs were incubated for 28 days at 16 ± 5 C, then at 18 ± 5 C for the next 28 days, then at 20 ± 5 C for the next 28 days. The eggs (which were more than 85% of the way through development by this stage) were then transferred to the intermediate temperature (18 ± 5 C), and kept there until hatching. 3. Falling (reverse of the rising treatment, i.e. 28 days at 20 ± 5 C, then 28 days at 18 ± 5 C, then 28 days at 16 ± 5 C, then the remainder at 18 ± 5 C). The final brief retention at the intermediate temperature regime (mean = 14 days of the average 97-day total incubation period) was used to ensure that all eggs experienced identical mean temperatures during incubation. If I had extended the period for each of the first three components of the sequence to take up the full incubation period, some eggs hatching slightly earlier than expected would then have spent less time at the final phase (e.g. cooler regime) than earlier phases, and hence would have experienced slightly different mean temperatures. I did not have enough incubators to replicate units at each temperature regime within the same year, but did so at the stable regime; no significant (P < 0 05) differences were apparent for mean values of hatchlings from these two stable -regime incubators. Previous work has shown that phenotypic variation induced by among-incubator differences is very small relative to thermal effects (Flatt et al. 2001). Eggs were checked daily for hatching, and any hatchlings were measured (SVL, tail length, mass) and their sex determined by manual eversion of hemipenes. The young lizards were housed in individual cages (22 13 7 cm 3 ) in a constant-temperature room at 20 C. Under-cage heating allowed the hatchlings to select body temperatures between 20 and 38 C during daylight hours, with the heating switching off at dusk. The lizards were fed mealworms and crickets dusted with reptile vitamins; water was available ad libitum. At 1 week of age, hatchlings were re-measured and had their locomotor capacity tested in a 1-m raceway maintained at 25 C (mean of three runs per lizard; see Shine et al. 1997 for details of methods). For analysis, I examined mean speeds (over 1 m) and sprint speeds (over the fastest 25 cm). All hatchlings were then returned to the Brindabella Range and released at the site of egg collection. DATA ANALYSIS I used the software program Statview 5 (SAS Institute 1998) on an Apple Macintosh G4 computer to analyse these data. Data sets were checked to ensure that they conformed to the requirements of statistical tests. For all analyses of hatchling phenotypic traits, I included the sex of the animal as a factor because of strong sex differences in responses to incubation temperature in B. duperreyi. Low-temperature incubation overrides heteromorphic sex chromosomes to produce sexreversed males (Shine, Elphick & Donnellan 2002), and the phenotypic traits of genetically male and female offspring respond differently to thermal regimes over a broad range of nest temperatures (Elphick & Shine 1999). The temperatures used in the current study were high enough that we would not expect to see sex reversal, but differential sex-specific effects might confound interpretation if sex is not included as a factor in analyses. Results THERMAL REGIMES IN NATURAL NESTS Temperatures inside natural nests in the Brindabella Range show a strong seasonal cycle, attaining their highest levels during midsummer (December January: Fig. 1). Nests at a relatively low elevation (1240 m) were warmer than those at higher elevation (1615 m), especially late in the incubation period (Fig. 1). At the low-elevation site (1240 m), eggs are laid in early December (week 0) and hatch about 14 weeks later (mean temperature over this period = 19 3 C, and incubation period is similar to that recorded for eggs kept at the stable 18 ± 5 C regime in the experimental study, about 100 days: Fig. 1). Nest temperatures at week 14 are similar to those at the beginning of incubation (Fig. 1). In contrast, oviposition is delayed 2 weeks at the higher elevation site and incubation temperatures are lower (Fig. 1). Eggs laid at this elevation typically hatch after about 18 weeks incubation (R. Shine, unpublished data). The combination of later oviposition and a more rapid seasonal decline in nest temperatures at the higher elevation means that eggs experience conditions about 5 C warmer at the time of laying than at the time of hatching (Fig. 1). Figure 2 shows data from selected individual nests during the incubation period, with examples of nests that exhibited each of the three thermal regimes simulated in the experiment (falling, stable and rising). Thus, some natural nests exhibited thermal shifts similar to those simulated in the laboratory experiment. This

46 R. Shine Fig. 1. Thermal regimes in nest-sites used by the scincid lizard Bassiana duperreyi at two elevations in the Brindabella Range, showing the seasonal cycle in nest temperatures of natural nests, including time periods before laying and after hatching. Data are shown for nests at 1240 m asl (Picadilly Circus) and 1615 m asl (Ginini Flats). The bars indicate the time during which eggs were present in the nests; data on nest temperatures missing for weeks 13 and 14 postlaying. variation occurred among years at the same location (note that two of the graphs in Fig. 2 are from the same site) as well as with elevation (Fig. 1). The change in mean incubation temperature from week 1 to week 9 (the time-scale of Fig. 2) offers a simple index of thermal patterns. This value remained stable (<1 C change) in 34 of 95 natural nests (34%), fell in 37 (39%) and rose in 24 (25%). EFFECTS OF INCUBATION REGIMES Responses of eggs from the three sites were similar in all major respects (i.e. no significant interaction terms between site and either incubation treatment or sex), and hence data were pooled for analysis. All three sites were equally represented in all three incubation treatments. Because of communal oviposition, I could not determine the clutch of origin and thus did not include it as a factor. This leaves two main putative determinants of hatchling phenotype that can be included in the analysis: sex and experimental treatment. Initial egg mass did not differ significantly among treatments or between sexes (two-factor ANOVA with sex and incubation treatment as factors, egg mass as the dependent variable: effect of sex, F 1,368 = 2 21, P = 0 13; treatment, F 2,368 = 0 23, P = 0 80; sex treatment, F 2,368 = 0 55, P = 0 58). Likewise, incubation treatments did not affect offspring sex ratios over the range of thermal regimes used in this experiment (all treatments yielded 50 53% male offspring; χ 2 = 0 32, 2 df, P = 0 85). Hatching success overall was 78% (358 of 458 eggs) and did not differ significantly among treatments (range 20 3 24% mortality: χ 2 = 0 69, 2 df, P = 0 71) or between sexes (χ 2 = 0 02, 1 df, P = 0 89). Deformities (generally malformed heads and kinked tails) were seen in 8 8% of males and 6 1% of female hatchlings (χ 2 = 0 65, 1 df, P = 0 42). These deformities were recorded Fig. 2. Examples of thermal regimes measured within individual natural nests of the scincid lizard Bassiana duperreyi in the Brindabella Range, beginning from the first week after eggs were laid in the nest. The graphs were selected to show cases where mean nest temperatures (a) increased during incubation (nest 88, Picadilly Circus, 1999 2000 season); (b) remained relatively stable throughout incubation (nest 5, Ginini Flats, 2001 02 season); or (c) declined over this period (nest 3, Ginini Flats, 1997 98 season). Week 1 represents the first week after eggs were laid in the nests (early December), and week 9 corresponds to the earliest recorded time of hatching. Dotted lines show linear regressions fitted to the weekly mean values. Error bars show standard errors. more frequently in hatchlings from eggs exposed to falling temperatures (13 0%) than rising temperatures (3 3%; χ 2 = 5 36, 1 df, P < 0 025). Stable temperatures generated an intermediate incidence of deformities (8 2%; comparing all three treatments, χ 2 = 6 42, 2 df, P < 0 035). Deformed hatchlings ran more slowly than their non-deformed siblings (over 1 m, F 1,356 = 8 06,

47 Table 1. Effects of sex and thermal regimes during incubation on phenotypic traits of Thermal hatchling effects lizards, on Bassiana duperreyi showing output from two-factor ANOVAs or lizard ANCOVAs phenotypes with sex and incubation regime as the factors. Fig. 3 shows mean values for each sex and treatment combination. For traits correlated with hatchling snout vent length, the latter variable was incorporated as a covariate in two-factor ANCOVAs. These are shown as relative measures; interaction terms involving the covariate were all nonsignificant and thus were deleted. Bold face shows significant results (P < 0 05) Effect of incubation treatment Effect of sex Interaction sex treatment Variable F 2,363 P F 1,363 P F 2,363 P Incubation period (days) 19 81 0 0001 2 40 0 12 0 99 0 37 Snout vent length (mm) 4 58 0 01 85 49 0 0001 0 36 0 70 Relative tail length 2 81 0 06 36 45 0 0001 1 71 0 18 Relative body mass 0 53 0 58 59 09 0 0001 1 38 0 25 Relative speed (m s 1 ) over 1 m 4 52 0 02 35 01 0 0001 0 68 0 51 over 25 cm 1 94 0 15 43 01 0 0001 0 66 0 52 need to be taken into account before analysing effects on other traits. Regression analyses showed that a hatchling s SVL was correlated with its mass (n = 358, r = 0 69, P < 0 001), tail length (r = 0 48, P < 0 001) and speed (over 1 m, r = 0 12, P < 0 035; over 25 cm, r = 0 17, P < 0 002). Thus, I incorporated hatchling SVL as a covariate in analyses of these traits to remove the confounding influence of absolute body size. I used two-factor ANOVA and ANCOVA, with incubation treatment and sex as factors, to analyse continuous variables such as the duration of incubation and the phenotypic traits of hatchling lizards. Both sex and incubation treatment generated significant phenotypic variation in all of the traits that I examined (Table 1). Sex effects were clear-cut, with male hatchlings having shorter SVLs than females, but relatively longer tails. Males were also more heavy-bodied (mass relative to SVL) and ran significantly faster over distances of both 1 m and 25 cm. No significant interactions were observed between sex and incubation treatment or with the covariate (Table 1), greatly simplifying interpretation. Hatchlings from the three experimental treatments differed significantly in incubation periods, body lengths, relative tail lengths and running speeds (Table 1). Eggs maintained under the stable regime incubated an average of about 7 more days before hatching than did eggs from either the rising or falling treatments (Fig. 3; Fishers PLSD post hoc comparisons show that stable differs from both the other treatments at P < 0 05). Hatchlings from eggs exposed to decreasing temperatures were smaller than those exposed to rising temperatures ( post hoc, P < 0 05) with hatchlings from the stable treatment intermediate in this respect (Fig. 3). Tails were shorter, relative to SVL, in stable -temperature hatchlings than in falling -temperature hatchlings ( post hoc, P < 0 05; Fig. 3). Body mass relative to SVL was not significantly affected by the incubation treatments (Table 1, Fig. 3). Hatchlings from the falling treatment were slower over 1 m than were either of the other groups ( post hoc, P < 0 05); the same pattern was evident but not statistically significant for speeds over a shorter distance (Fig. 3). Discussion Fig. 3. Effects of hatchling sex and incubation treatment (stable, rising or falling temperatures) on incubation periods and phenotypic traits of hatchling scincid lizards Bassiana duperreyi. The graphs show mean values and associated standard errors. For traits correlated with SVL, the graphs show values for residual scores from the general linear regression of that trait against SVL. See Table 1 for results from two-factor ANOVA and ANCOVA on these variables, and the text for results of post hoc tests. P < 0 005; over 25 cm, F 1,356 = 10 26, P < 0 002) and died sooner after hatching (age at death, F 1,96 = 34 50, P < 0 0001). Many phenotypic traits of hatchlings were correlated with the animal s body size, so body-size differences The clear result from this study is that seasonal shifts in incubation temperatures, as commonly occur in the nests of many reptile species (Packard & Packard 1988; see Figs 1 and 2), can significantly modify both the time of hatching and the developmental trajectories of developing embryos. This sensitivity is reflected in the incidence of deformities as well as the morphology and locomotor performance of hatchlings. This conclusion means that in order to understand the relationship between a reptilian embryo and its nest environment, we need to consider not only mean temperatures and associated diel variances, but also the ways in which these attributes shift through time within the incubation period. Even if the overall means

48 R. Shine and variances of thermal regimes are held constant, subtle temporal shifts around those values can have substantial influences on hatchlings. In the present study, the sex of embryos influenced hatchling morphology and locomotor performance, but did not interact with incubation treatments in this respect ( Table 1). This reflects the thermal range used in the current study, in which mean temperatures were above those that induce strong effects on offspring phenotypes and sex determination (Shine et al. 2002). However, sex differences were generally stronger than incubation-induced effects (compare F-values in Table 1), so that including sex as a factor in these analyses substantially enhanced the power to detect treatment effects. Data from natural nests confirm that the thermal regimes used in the laboratory study are realistic simulations of conditions that occur in nature. For example, the three cases in Fig. 2 are similar to (but more extreme than) the three experimental treatments. More generally, the seasonal cycle in incubation temperatures at high-elevation sites is so pronounced ( Fig. 1) that eggs may experience a massive shift in thermal conditions over the course of incubation. At 1615 m (Ginini Flats), many eggs commence incubation with a mean temperature of around 18 C and finish it more than 5 C cooler (Fig. 1). Eggs may fail to hatch at all from some of these nests, especially in cool summers. Nonetheless, it is clear that my experimental treatments fall well within the range of thermal conditions experienced by eggs in successful natural nests. In keeping with this result, the phenotypic traits of hatchlings in the current study fell within the range observed from hatchlings emerging from natural nests (Shine et al. 1997). To compare the magnitude of effects induced by seasonal shifts in incubation temperature with that induced by other factors, I calculated mean values for six traits (incubation period, hatchling SVL, mass, tail length and speeds over 1 m and 25 cm) for each treatment group of hatchlings. I then calculated effect sizes by dividing the largest value by the smallest value for that factor (e.g. male vs female). The phenotypic effect of a seasonal shift in nest temperatures was similar in magnitude to the difference between sexes, and smaller than the difference in mean values between localities (4 4, 4 5, 7 4%, respectively). Pooling my previous experimental studies on this system (n = 3372 eggs), effects of sex and location are similar to those in the current work (4 4, 11 2%) but the much greater range of thermal conditions tested means that the incubation effect is vastly greater (57 3%). Thus, a relatively small seasonal shift in nest temperatures generates about 8% as much of a change in hatchling phenotypes as does the entire range of mean incubation temperatures over which Bassiana hatch successfully. Whether or not such effects are biologically significant (i.e. affect organismal fitness) is an important challenge for future work. Temporal shifts in nest temperatures within the incubation period will be particularly evident in regions with a cold winter and a hot summer (as occurs in the Brindabella Range: see Fig. 1), but the same phenomenon is widespread in other areas also. Thus, embryonic sensitivity to this time-scale of thermal variation could engender variation in hatchling phenotypes across a series of comparisons, as follows: 1. Among years within a single area. Minor year-toyear differences in the time of oviposition (as occur in the Brindabella lizards: Pengilley 1972) could alter the rate and timing of thermal shifts during incubation, generating phenotypic differences between cohorts of hatchlings from successive years. This could occur even in many tropical areas, especially if soil temperatures fall considerably with the arrival of the monsoon season (e.g. Webb 1992). Frequently, there is substantial year-to-year variation in the exact timing of such events (Ridpath 1985). Thus, annual variation in phenotypic traits of hatchlings may partly result from climatic variation of this kind. 2. Among clutches within a single cohort. Even within a single population, females often lay their eggs over a relatively wide time-span (e.g. Olsson & Shine 1997). Time of oviposition may thus determine exact thermal regimes experienced in the nest, generating variation among clutches within a single population in a single season. 3. Between first and second clutches within a year. Many female reptiles lay more than one clutch per season, and the eggs from successive clutches may encounter very different thermal regimes. For example, eggs laid early in spring may experience rising temperatures during incubation, whereas those laid late in summer may experience falling temperatures. This difference may generate phenotypic variation within hatchlings from successive clutches of the same female. 4. With elevation within a montane area. Eggs are typically laid later in the season at higher elevations (when soil temperatures are close to their maximum values) but earlier at lower elevations (while temperatures are still increasing). Thus, in a site such as the Brindabella Range, low-elevation eggs may experience stable or rising temperatures during embryogenesis, whereas eggs laid at higher elevations experience falling temperatures (Fig. 1). 5. Between sympatric oviparous and viviparous species. Even if mean incubation temperatures are similar for oviparous and viviparous squamates living in the same area, embryos inside viviparous females will be buffered from weather conditions by their mother s behavioural thermoregulation (Tinkle & Gibbons 1977; Beuchat 1986, 1988). Thus, seasonal shifts in incubation regimes will be less marked than for eggs in a nest (i.e. viviparous embryos could experience the stable regime used in the present study, even when oviparous animals had rising or falling scenarios). This process could engender phenotypic differences between oviparous hatchlings and viviparous neonates.

49 Thermal effects on lizard phenotypes In conclusion, the sensitivity of reptilian embryogenesis to rates of change in incubation temperature, as well as to overall means and variances of thermal regimes, may generate significant phenotypic variation among hatchlings from different areas, or among clutches produced at different times of year or in different years, or between sympatric oviparous and viviparous taxa. This result emphasizes the need to look carefully, and in considerable detail, at the incubation conditions that actually apply in natural nests. The use of simplifying assumptions about incubation conditions, although logistically convenient, may exclude biologically significant influences. Details of changes during the incubation period, as well as overall mean values, need to be considered in this respect. The current study emphasizes the subtlety and complexity of the developmental pathways that link incubation environments to offspring phenotypes. Acknowledgements I thank Melanie Elphick and Elizabeth Barrott-Brown for their dedication to the field and laboratory work involved in this study. The work was conducted under approval #L04/3-2-001/3/3348 issued by the University of Sydney Animal Care and Ethics Committee. The study was funded by the Australian Research Council. References Albon, S.D., Guinness, F.E. & Clutton-Brock, T.H. (1983) The influence of climatic variation on the birth weights of red deer (Cervus elephas). Journal of Zoology (London) 200, 295 297. Beuchat, C.A. (1986) Reproductive influences on the thermoregulatory behaviour of a live-bearing lizard. Copeia 1986, 971 979. Beuchat, C.A. (1988) Temperature effects during gestation in a viviparous lizard. Journal of Thermal Biology 13, 135 142. Blackburn, D.G. (1995) Saltationist and punctuated equilibrium models for the evolution of viviparity and placentation. Journal of Theoretical Biology 174, 199 216. Bull, J.J. (1980) Sex determination in reptiles. Quarterly Review of Biology 55, 4 21. Cogger, H.G. (2000) Reptiles and Amphibians of Australia, 6th edn. Reed New Holland, Sydney. Deeming, D.C. & Ferguson, M.W.J. (1991) Physiological effects of incubation temperature on embryonic development in reptiles and birds. Egg Incubation: Its Effects on Embryonic Development in Birds and Reptiles (eds D.C. Deeming & M.W.J. Ferguson), pp. 147 171. Cambridge University Press, Cambridge. Elphick, M.J. & Shine, R. (1999) Sex differences in optimal incubation temperatures in a scincid lizard species. Oecologia 118, 431 437. Flatt, T., Shine, R., Borges-Landaez, P.A. & Downes, S.J. (2001) Phenotypic variation in an oviparous montane lizard (Bassiana duperreyi): the effects of thermal and hydric incubation environments. Biological Journal of the Linnean Society 74, 339 350. Georges, A. (1989) Female turtles from hot nests: is it duration of incubation or proportion of development at high temperatures that matters? Oecologia 81, 323 328. Henry, C.J.K. & Ulijaszek, S.J. (1996) Long-term Consequences of Early Environment. Cambridge University Press, Cambridge. Hutchinson, M., Swain, R. & Driessen, M. (2001) Snakes and Lizards of Tasmania. University of Tasmania, Hobart. Olsson, M. & Shine, R. (1997) The seasonal timing of oviposition in sand lizards (Lacerta agilis): why earlier clutches are better. Journal of Evolutionary Biology 10, 369 381. Packard, G.C. & Packard, M.J. (1988) The physiological ecology of reptilian eggs and embryos. Biology of the Reptilia, Vol. 16 (eds C. Gans & R.B. Huey), pp. 523 605. Alan R. Liss, New York. Pengilley, R. (1972) Systematic relationships and ecology of some lygosomine lizards from Southeastern Australia. PhD Thesis, Australian National University, Canberra. Rhen, T. & Lang, J.W. (1995) Phenotypic plasticity for growth in the common snapping turtle: effects of incubation temperature, clutch, and their interaction. American Naturalist 146, 726 747. Ridpath, M.G. (1985) Ecology in the wet-dry tropics: how different? Proceedings of the Ecological Society of Australia 13, 3 20. SAS Institute. (1998) Statview 5. SAS, Cary, IN. Shine, R. (2002) Eggs in autumn: responses to declining incubation temperatures by the eggs of montane lizards. Biological Journal of the Linnean Society 76, 71 77. Shine, R. & Elphick, M.J. (2001) The effect of short-term weather fluctuations on temperatures inside lizard nests, and on the phenotypic traits of hatchling lizards. Biological Journal of the Linnean Society 72, 555 565. Shine, R. & Harlow, P.S. (1996) Maternal manipulation of offspring phenotypes via nest-site selection in an oviparous reptile. Ecology 77, 1808 1817. Shine, R., Elphick, M.J. & Harlow, P.S. (1997) The influence of natural incubation environments on the phenotypic traits of hatchling lizards. Ecology 78, 2559 2568. Shine, R., Elphick, M.J. & Donnellan, S. (2002) Co-occurrence of multiple, supposedly incompatible modes of sex determination in a lizard population. Ecology Letters 5, 486 489. Taning, A.V. (1952) Experimental study of meristic characters in fishes. Biological Reviews 27, 169 193. Tinkle, D.W. & Gibbons, J.W. (1977) The distribution and evolution of viviparity in reptiles. Miscellaneous Publications of the Museum of Zoology, University of Michigan 154, 1 55. Webb, G.J.W. (1986) Nests, eggs and embryonic development of Carettochelys insculpta (Chelonia: Carettochelyidae) from northern Australia. Journal of Zoology (London) 208, 521 550. Webb, G.J.W. (1992) The influence of season on Australian crocodiles. Monsoonal Australia: Landscape, Ecology and Man in the Northern Lowlands (eds C.D. Haynes, M.G. Ridpath & M.A. J. Williams), pp. 125 131. A.A. Balkema, Rotterdam. Webb, G.J.W., Manolis, S.C., Dempsey, K.E. & Whitehead, P.J. (1987) Crocodilian eggs: a functional overview. Wildlife Management: Crocodiles and Alligators (eds G.J.W. Webb, S.C. Manolis & P.J. Whitehead), pp. 417 422. Surrey Beatty & Sons, Sydney. Webb, G.J.W. & Cooper-Preston, H. (1989) Effects of incubation temperature on crocodiles and the evolution of reptilian oviparity. American Zoologist 29, 953 971. Wilbur, H.M., Tinkle, D.W. & Collins, J.P. (1974) Environmental certainty, trophic level, and resource availability in life history evolution. American Naturalist 108, 805 816. Received 15 April 2003; revised 4 August 2003; accepted 19 August 2003