SEARCH

SEARCH BY CITATION

Keywords:

  • amphibians;
  • countergradient variation;
  • developmental rate;
  • egg size;
  • latitudinal gradient;
  • local adaptation

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Study populations
  6. Laboratory experiment
  7. Experimental procedures
  8. Field data
  9. Statistical analyses
  10. Results
  11. Laboratory experiment
  12. Egg size
  13. Field data
  14. Discussion
  15. Conclusions
  16. Acknowledgments
  17. References

Adaptive genetic differentiation along a climatic gradient as a response to natural selection is not necessarily expressed at phenotypic level if environmental effects on population mean phenotypes oppose the genotypic effects. This form of cryptic evolution – called countergradient variation – has seldom been explicitly demonstrated for terrestrial vertebrates. We investigated the patterns of phenotypic and genotypic differentiation in developmental rates of common frogs (Rana temporaria) along a ca. 1600 km latitudinal gradient across Scandinavia. Developmental rates in the field were not latitudinally ordered, but displayed large variation even among different ponds within a given latitudinal area. In contrast, development rates assessed in the laboratory increased strongly and linearly with increasing latitude, suggesting a genetic capacity for faster development in the northern than the southern larvae. Experiments further revealed that environmental effects (temperature and food) could easily override the genetic effects on developmental rates, providing a possible mechanistic explanation as to why the genetic differentiation was not seen in the samples collected from the wild. Our results suggest that the higher developmental rates of the northern larvae are likely to be related to selection stemming from seasonal time constrains, rather than from selection dictated by low ambient temperatures per se. All in all, the results provide a demonstration of environmental effects concealing substantial latitudinally ordered genetic differentiation understandable in terms of adaptation to clinal variation in time constrains.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Study populations
  6. Laboratory experiment
  7. Experimental procedures
  8. Field data
  9. Statistical analyses
  10. Results
  11. Laboratory experiment
  12. Egg size
  13. Field data
  14. Discussion
  15. Conclusions
  16. Acknowledgments
  17. References

Environmental heterogeneity in abiotic and biotic factors influencing fitness is expected to result in selection promoting local adaptation and genetic divergence among different populations of the same species (e.g. Mayr, 1963; Endler, 1977). Climatic selection stemming from temperature differences is an important factor promoting genetic differentiation in growth and development of ectotherms both in natural (Berven & Gill, 1983; Lonsdale & Levinton, 1985; Blanckenhorn & Fairbairn, 1995; Parsons, 1997) and laboratory populations (James & Partridge, 1995; Partridge et al., 1995). Likewise, time constraints in terms of growth season length may also contribute to the shaping of development and growth patterns (Roff, 1980; Rowe & Ludwig, 1991; Abrams et al., 1996; Arendt, 1997). However, while both low temperature and strong time constraints may select for genetically fast growth and development rates with increasing latitude and altitude, their relative importance in explaining clinal differentiation remains elusive (Arendt, 1997; Yamahira & Conover, 2002).

Countergradient variation refers to a geographic pattern of variation in which genetic influences on a trait oppose environmental influences across an environmental gradient, thereby reducing phenotypic change across the gradient (Levins, 1969; Conover & Schultz, 1995). Such a negative covariance between environmental and genetic influences may give a superficial impression of lack of local adaptation, and conceal even substantial genetic differentiation along the gradient. Countergradient variation in response to climatic variation has been demonstrated in a number of ectotherms (e.g. Berven et al., 1979; Berven, 1982a,b; Conover & Present, 1990; Blanckenhorn & Fairbairn, 1995; Parsons, 1997; Jonassen et al., 2000), but, because of difficulties in detecting such a pattern in the field, it has often been overlooked in studies of local adaptation (Conover & Schultz, 1995).

Maternal effects are important determinants of offspring performance in a wide range of taxa (Mousseau & Fox, 1998). One path of maternal effects influencing offspring phenotype is via egg size: larger eggs produce larger and faster developing offspring with potentially far-reaching consequences for fitness (see Mousseau & Fox, 1998 for reviews). In most ectotherms, egg size tends to increase with decreasing temperature or with increasing altitude and latitude (e.g. Berven, 1982; Azevedo et al., 1996; Yampolsky & Scheiner, 1996). Consequently, unless egg size effects are accounted for, any genetic differentiation along a climatic gradient may become confounded with maternal effects acting through egg size. Nevertheless, the role of egg size in creating geographic gradients in offspring phenotypes has remained largely unexplored (Bernardo, 1996; but see: Laugen et al., 2002).

The common frog (Rana temporaria) is the most widespread anuran in Europe (Fog et al., 1997), and hence subject to widely differing environmental conditions in different parts of its distribution range (Miaud et al., 1999; Miaud & Merilä, 2000). Consequently, the opportunity for local differentiation among populations should be large. In accordance with this expectation, a number of common garden studies (Meriläet al., 2000; Laurila et al., 2001, 2002; Laugen et al., 2002) have demonstrated that individuals from Scandinavian high latitude populations are faster developing than those from low latitude populations. However, the evidence for a possible clinal and countergradient nature of this differentiation remains elusive since all of these studies are based on comparisons of single low and high latitude populations, and because no field data have been available.

The goal of this study was to explore whether there is evidence for adaptive genetic differentiation in developmental rates among R. temporaria populations along a 1600 km latitudinal gradient across Scandinavia. We were especially interested in testing whether the divergence in developmental rate conforms to the pattern of countergradient variation, namely that it would increase with increasing latitude under laboratory conditions, but this would not be seen in the field data because of overriding environmental effects (Conover & Schultz, 1995). To this end, we also assessed, using a factorial experiment, how different environmental factors (temperature, food availability) and maternal investment (egg size) influenced developmental patterns of tadpoles from different latitudes.

Study populations

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Study populations
  6. Laboratory experiment
  7. Experimental procedures
  8. Field data
  9. Statistical analyses
  10. Results
  11. Laboratory experiment
  12. Egg size
  13. Field data
  14. Discussion
  15. Conclusions
  16. Acknowledgments
  17. References

The five populations included in the laboratory study – situated along a latitudinal gradient from southern Sweden to northern Finland – are shown in Fig. 1. All of these populations bred in medium-sized ponds (max depth <1.6 m) with populations of 50–150 breeding females. The onset of breeding season between the southern and northernmost localities differs by ca. 60 days, and there is a twofold difference in the length of growth season between these localities (Table 1). Mean daily ambient temperatures over the past 26 years (1976–2002; data from the Swedish Meteorological and Hydrological Institute and the Finnish Meteorological Institute) during 30, 60 and 90 day periods following onset of egg laying in a given locality are shown in Fig. 2. Mean ambient temperatures during the first 30 days following the onset of breeding increase as a function of latitude, but by the time the first metamorphs appeared in the majority of the ponds (mean time to emergence of first metamorphs in 15 ponds = 60.2 days; see results), mean ambient temperatures peak at mid-latitudes and are about equal at both ends of the latitudinal gradient (Fig. 2). When metamorphosis in nearly all ponds has passed its culmination (90 days after breeding onset), mean ambient temperatures over the developmental period are already much lower in the north than in the south (Fig. 2). Consequently, when viewed over the entire period of development, the northernmost frogs experience similar or lower mean ambient temperatures than the more southern frogs.

image

Figure 1. Map showing the geographic origin of the animals used in the laboratory experiments in 1998. Populations for the 1999 field studies are underlined. See Table 1 for detailed coordinates and additional information about the localities.

Download figure to PowerPoint

Table 1.  Descriptive information about the five different Rana temporaria populations used in this study. GS = growth season length (number of days where mean air temperature exceeds +5 °C; Odin et al., 1983), T = mean (±SE) average ambient temperature during a 90 day period (1976–2002) following the onset of spawning, Min = minimum average temperature over the same period, Max = maximum average temperature over the same period, CV = coefficient of variation yearly mean temperatures over the same period, NF = the number of females used per population, NH = the number of tadpoles at the start of the experiment (stage 25; Gosner, 1960) and NM = the number of tadpoles that survived through metamorphosis. Field = the number of ponds where development was followed in the field.
PopulationLatitude (°N)Longitude (°E)Altitude (m)GS (d)T (°C)MinMaxCVNFNHNMField
Lund55°40′13°27′2021711.17 ± 0.179.4912.697.687195456
Uppsala59°40′17°15′4518411.59 ± 0.179.7812.727.282672133
Umeå63°50′20°25′515812.20 ± 0.1710.6113.727.016706560 
Kiruna67°52′20°29′42511310.30 ± 0.188.6912.388.68500388 
Kilpisjärvi69°04′20°50′485989.32 ± 0.197.4410.6910.185294226
Total         6069472815
image

Figure 2. Mean (±SE) average (1976–2001) daily temperatures over 30, 60 and 90 day periods following onset of spawning in given population as a function of latitude. Each latitudinal point corresponds to data from localities listed in Table 1. The periods considered for each of the populations are Lund: starting 1 April; Uppsala: 15 April; Umeå: 5 May; Kiruna: 25 May and Kilpisjärvi: 1 June.

Download figure to PowerPoint

Owing to logistic constrains, field data was obtained for only three of the latitudinal areas included in the laboratory study. These were the most southern and northern localities, as well as one (Uppsala) of the ‘mid-latitude’ localities (Fig. 1). Within each of these three areas, larval development from hatching to metamorphosis was monitored in 3–6 different ponds (Table 1). These were all small to medium-sized ponds that lacked shading by canopy cover. The maximum distance between the ponds within each latitudinal area was 25 km.

Experimental procedures

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Study populations
  6. Laboratory experiment
  7. Experimental procedures
  8. Field data
  9. Statistical analyses
  10. Results
  11. Laboratory experiment
  12. Egg size
  13. Field data
  14. Discussion
  15. Conclusions
  16. Acknowledgments
  17. References

Tadpoles used in the experiment were obtained using laboratory crossings of adults collected from spawning sites at the onset of breeding season. This procedure was adopted in the context of a larger quantitative genetic study (A. T. Laugen et al. in prep.), but it also ensured that all tadpoles at the start of the experiment were of equal age and subject to the same early environmental effects. In all but one (Umeå) population we created 16 maternal half-sib families where eggs from each of eight females were fertilized by sperm from two of 16 males. The Umeå tadpoles originate from 32 maternal half-sib families (16 females and 32 males used). Owing to the large difference in the onset of spawning among the populations (Table 1), the starting dates for the experiment also differed. In case of the southernmost population (Lund), the fertilizations were performed on the 9th of April, whereas in the case of the northernmost population (Kilpisjärvi) the corresponding date was the 4th of June. However, the rearing conditions were identical for all populations (see below).

The crossings were carried out following the principles outlined in Laugen et al. (2002). One hour after fertilization, eggs were detached from the dish and a sample of 15–30 eggs from each cross was stored in 4% formaldehyde for later determination of egg size. The remaining eggs were divided into three different temperature treatments [14, 18 and 22 °C (±1 °C), two bowls per cross in each temperature], where they were kept until hatching. Water was changed every third day during embryonic development. When the majority of the embryos in a given temperature treatment had reached Gosner stage 25, eight randomly chosen tadpoles from each cross were placed individually in 0.9 L opaque plastic containers in each of two food levels (restricted and ad libitum). This procedure was repeated for each population in the three temperature treatments resulting in 48 experimental tadpoles per cross. However, as a result of mortality during the experiment, the number of tadpoles for which developmental rate could be determined was typically less than this (Table 1). The tadpoles were fed a finely ground 1 : 3 mixture of fish flakes (TetraMin, Ulrich Baensch GmbH, Germany) and rodent pellets (AB Joh. Hansson, Uppsala, Sweden) every seventh day. The amount of food given to each tadpole was 15 (restricted) and 45 (ad libitum) mg for the first week, 30 and 90 mg for the second week, and 60 and 180 mg per week thereafter until metamorphosis. The ad libitum level was selected to be such that the individuals did not consume all the food before the next feeding event in any of the temperature treatments. In the restricted food treatment, the tadpoles in the two highest temperature treatments consumed all of their food resources before the next feeding, indicating food limitation, but in the low temperature treatment, tadpoles frequently had food left even after 7 days following feeding. Tadpoles were raised in dechlorinated tap water, which was aerated and aged for at least 24 h before use. The water was changed every seventh day in conjunction with feeding. The light rhythm was 16L : 8D. Close to metamorphosis, the tadpoles were checked every day, and individuals that had reached Gosner stage 42 were noted. Developmental rate, or age at metamorphosis, was defined as the number of days elapsed between reaching Gosner stages 25 and 42.

To control for the effect of initial egg size on development, the diameter of 30–60 newly fertilized eggs from all females was measured (to the nearest 0.01 mm) using a stereo microscope fitted with an ocular micrometer. Two measurements from each egg were taken, and the mean value of these measurements was used to calculate the female-specific egg size. The storage in formaldehyde may affect the size of eggs, but as eggs from all populations were subject to the same treatment, the values should be comparable among the populations.

Field data

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Study populations
  6. Laboratory experiment
  7. Experimental procedures
  8. Field data
  9. Statistical analyses
  10. Results
  11. Laboratory experiment
  12. Egg size
  13. Field data
  14. Discussion
  15. Conclusions
  16. Acknowledgments
  17. References

The fieldwork was carried out 3rd April–9th July, 25th April–15th July and 1st June–30th September in Lund, Uppsala and Kilpisjärvi, respectively. In each of the localities, the development of eggs and hatchlings was monitored to record when the first larvae reached Gosner stage 25. Thereafter, the ponds were sampled every 7 days. During every visit, we dip-netted the ponds until 20–30 tadpoles were sampled. The tadpoles were killed using MS-222 and stored in 70% ethanol. From each preserved larva, we determined the developmental stage according to Gosner (1960). The developmental rate for each pond was defined as the number of days elapsed between the day when the first hatchlings reached Gosner stage 25 (complete gill absorption) and the day when the first larvae reached Gosner stage 42 (emergence of at least one foreleg). Since this method relies heavily on observations of single individuals, we also used the mean developmental stage ca. 40–50 days after the larvae had reached Gosner stage 25 as an additional measure of developmental rate in the wild. Although this measure is not directly comparable to our laboratory measure of developmental rate, it allows a more powerful analysis of developmental rates as it utilises data on 547 individuals as compared with 15 in the approach described above.

Statistical analyses

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Study populations
  6. Laboratory experiment
  7. Experimental procedures
  8. Field data
  9. Statistical analyses
  10. Results
  11. Laboratory experiment
  12. Egg size
  13. Field data
  14. Discussion
  15. Conclusions
  16. Acknowledgments
  17. References

Our laboratory experiment was a 5 × 3 × 2 factorial design with population (five levels), temperature (three levels) and food level (two levels) as factors. To control for possible egg size-mediated maternal effects in larval traits, we introduced the female-specific mean egg size as a covariate in a separate set of analyses. The data were analysed with fixed effects anova and ancova using type III mean squares as implemented in PROC GLM of SAS 6.12 statistical package (SAS Institute, 1996). The mean value of the two maternal half-sib families in each treatment combination was used as response variable in the analyses. Hence, the per population sample size for each treatment combination was eight, except for the Umeå population for which it was 16. Developmental rate (= age at metamorphosis) was log-transformed before the analyses to meet the assumptions of normality and homogeneity of variances. However, for ease of interpretation, graphed means and SE are given in untransformed scale. Linear regressions were used for illustrative purposes displaying the latitudinal trends, and hence, the details of these analyses are not presented. The field data on emergence of the first metamorphs was analysed with a one-way anova treating the latitudinal area (three levels) as a factor. The data on developmental stages in different ponds was analysed with a linear mixed model using PROC MIXED of SAS. The latitudinal area (three levels) was fitted as fixed effect, population (nested within area) as a random effect. Age (time elapsed between Gosner stage 25 and sampling) was fitted as covariate to control for variation consequent to differences in timing of sampling in different populations. The residuals from the models were normally distributed.

Laboratory experiment

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Study populations
  6. Laboratory experiment
  7. Experimental procedures
  8. Field data
  9. Statistical analyses
  10. Results
  11. Laboratory experiment
  12. Egg size
  13. Field data
  14. Discussion
  15. Conclusions
  16. Acknowledgments
  17. References

There was a strong negative relationship between latitude and age at metamorphosis in all temperatures and in both food treatments (Fig. 3a, b; Table 2). In general, larval period was roughly 10 days shorter in the northernmost population as compared with the southernmost population (Fig. 3a, b). However, the exact magnitude of this difference depended slightly but significantly on the temperature as indicated by the significant population × temperature interaction (Table 2). Acceleration of development from 14 to 18 °C was larger in the two southernmost populations (Lund, Uppsala, Sweden) than in the more northern populations (Fig. 3a, b). In contrast, acceleration of development from 18 to 22 °C was more pronounced in the two northernmost populations as compared to the three southernmost populations (Fig. 3a, b). In general, increasing temperature had a large accelerating effect on developmental rate in both food levels and in all populations (Table 2; Fig. 3a, b): larval period was 30–40 days longer at 14 °C than at 22 °C.

image

Figure 3. Effects of population of origin, temperature and food treatments on age at metamorphosis. Mean age (±SE) as a function of latitude of origin under (a) ad libitum and (b) restricted food levels. Filled dots represent means at 14 °C, open dots means at 18 °C and grey dots means at 22 °C.

Download figure to PowerPoint

Table 2.  Results of anovas examining the effects of population origin, food (ad libitum, restricted) and temperature (14, 18, and 22 °C) treatments on developmental rate (age at metamorphosis) of Rana temporaria tadpoles. MS = mean square, r2 = coefficient of determination.
Sourced.f.MSF
  1. ***P < 0.001.

Population40.12208.72***
Temperature22.474309.26***
Food10.13234.76***
Population × Temperature80.0813.60***
Population × Food40.000.80
Temperature × Food20.0354.46***
Population × Temperature × Food80.001.66
Model290.21367.97***
Error2190.00
r20.98

Ad libitum food level accelerated development in all populations (Table 2, Fig. 3). As indicated by the significant food × temperature interaction (Table 2), this effect was mostly confined to the two highest temperatures, whereas at 14 °C, ad libitum food level did not lead to faster development (Fig. 3a, b). The populations did not differ in their responses to food treatment as revealed by the nonsignificant population × food and population × temperature × food interactions (Table 2; Fig. 3a, b).

Egg size

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Study populations
  6. Laboratory experiment
  7. Experimental procedures
  8. Field data
  9. Statistical analyses
  10. Results
  11. Laboratory experiment
  12. Egg size
  13. Field data
  14. Discussion
  15. Conclusions
  16. Acknowledgments
  17. References

Mean egg size differed significantly between populations (F4,43 = 85.82, P < 0.0001; Fig. 4a). However, the regression of mean egg size on latitude was not significant (b = 0.032 ± 0.17; F1,4 = 3.39, n.s.), mostly because the eggs in Uppsala (59° N) were as large as those in Kilpisjärvi (69° N; Fig. 4a). Although egg size contributed significantly to the variation in developmental rate (F1,218 = 16.57, P < 0.0001), all effects in Table 2 also remained highly significant when egg size was included as a covariate in the models. This is illustrated in Fig. 4b with the regressions of age at metamorphosis on latitude in different treatments, both with and without accounting for the influence of egg size. Accounting for egg size variation thus tends to reduce the population differences, but not to the point of nonsignificance (Fig. 4b; Table 3).

image

Figure 4. Mean egg size and impact of egg size on age at metamorphosis in different populations as a function of latitude. (a) Mean egg size (±SE) as a function of latitude, and (b) comparison of regression slopes of age at metamorphosis on latitude accounting for egg size variation (solid lines) and not accounting for egg size variation (dotted lines). In (a) each mean is based on mean egg size of mothers of experimental tadpoles. In (b) the pairs of lines from top to bottom are: restricted food – 14 °C, ad libitum food – 14 °C, restricted food – 18 °C, restricted food – 22 °C, ad libitum food – 18 °C, ad libitum food 22 °C.

Download figure to PowerPoint

Table 3.  Regressions (b ± SE) of mean age at metamorphosis on latitude of origin in different temperature and food level treatments. (a) without accounting for egg size differences among localities, (b) accounting for egg differences among localities. n = number of populations. r2 = coefficient of determination.
Temperaturen(a) Without correction for egg size(b) With correction for egg size
RestrictedR2Ad libitumr2Restrictedr2Ad libitumr2
  1. *P < 0.05, **P < 0.01, ***P < 0.001.

14 °C5−0.94 ± 0.12**0.95−0.99 ± 0.26*0.83−1.10 ± 0.15*0.97−1.41 ± 0.07**0.99
18 °C5−0.47 ± 0.07**0.94−0.64 ± 0.08**0.96−0.39 ± 0.09*0.96−0.52 ± 0.01***1.00
22 °C5−0.94 ± 0.08**0.98−0.65 ± 0.06**0.98−1.02 ± 0.11*0.99−0.74 ± 0.05**0.99

Field data

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Study populations
  6. Laboratory experiment
  7. Experimental procedures
  8. Field data
  9. Statistical analyses
  10. Results
  11. Laboratory experiment
  12. Egg size
  13. Field data
  14. Discussion
  15. Conclusions
  16. Acknowledgments
  17. References

Mean estimated age at metamorphosis did not differ significantly among the three latitudinal areas (F2,12 = 1.52, n.s.; Fig. 5). In fact, the among-pond variation in estimated age at metamorphosis was huge, and the range of values recorded for ponds from the southernmost latitudinal area all fall into the range of values recorded in the northernmost area (Fig. 5). The results were qualitatively similar if the developmental stage roughly 40–50 days after reaching Gosner stage 25 was analysed: in spite of significant area (F2,13 = 11.19, P < 0.01), age (F1,13 = 88.38, P < 0.001) and pond (z = 2.06, P < 0.05) effects on mean developmental stage, post-hoc tests revealed that only mid-latitude [x = 42.91 ± 0.98 (SE)] larvae differed significantly from southern (x = 37.39 ± 0.64; t13 = −4.71, P < 0.01) and northern (x = 39.39 ± 0.70; t13 = −2.93, P < 0.05) larvae, but that there was no significant difference in developmental rate of northern and southern larvae (t13 = 2.11, n.s.).

image

Figure 5. Age at metamorphosis determined from field caught Rana temporaria larvae during the 1999 growth season. Each dot represents one pond.

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Study populations
  6. Laboratory experiment
  7. Experimental procedures
  8. Field data
  9. Statistical analyses
  10. Results
  11. Laboratory experiment
  12. Egg size
  13. Field data
  14. Discussion
  15. Conclusions
  16. Acknowledgments
  17. References

The results demonstrate a clear latitudinal cline in developmental rates of common frogs grown in a common garden situation, but show no such relationship in the field data. Ignoring potential maternal effects and historical factors for the time being (see below), the results of the common garden experiment have an unambiguous interpretation: genetic capacity for fast development is a positive function of latitude, and this type of clinal differentiation is difficult to interpret otherwise than in the light of local adaptation to some selective factor that is closely associated with latitude (Endler, 1977). What could this selective factor be?

Both low temperature and short growth period may impose time constraints on ectotherms and select for faster growth and development rates (Roff, 1980; Rowe & Ludwig, 1991; Abrams et al., 1996; Arendt, 1997). At higher latitudes selection to complete development before the onset of winter should be strong, and increased uptake rate and digestion efficiency has been demonstrated in northern fish populations (Present & Conover, 1992; Jonassen et al., 2000). Accordingly, when raised in a common environment, northern populations are expected to develop faster (Conover & Schultz, 1995). Our results are consistent with this expectation, and suggest that time constraints, rather than temperature per se, is the causal agent that has selected for faster development in the north as compared to the south. The mean ambient temperature over most of the developmental period is higher or similar in the northernmost as compared with the southernmost populations – and always highest at mid-latitudes – meaning that the variation in thermal environment is not a simple linear function of latitude. Consequently, mean ambient temperatures experienced by developing tadpoles are not correlated with variation in developmental rates across the latitudinal gradient as determined in our common garden experiments (r = 0.72, n = 5, P = 0.16; calculated using mean residual development rate from the linear model of developmental rate on temperature and food treatments using data from Fig. 3a, b). However, in contrast, they are strongly correlated with the length of the thermal growth season (r = 0.98, n = 5, P < 0.002), suggesting that time constraint – posed by declining ambient temperatures in north towards the end of the larval period (cf. Fig. 2) – is the main selective factor behind the faster developmental rates in north. However, a genetically based latitudinal cline could also evolve as a consequence of restricted gene flow among previously isolated populations (Endler, 1977), and many Northern Scandinavian species have colonized Sweden from two directions (i.e. south and north) after the last glaciations (Hewitt, 2000). However, our preliminary analyses of variation in eight microsatellite loci across these populations do not suggest that common frogs would have colonized Sweden from multiple directions (J. Palo; D. Schmeller; C. R. Primmer & J. Merilä, unpublished data).

Although developing faster in the laboratory, the northern frogs did not realize their genetic capacity for fast development in the field. This is to be expected if the local environmental effects, such as variations in temperature and food availability, override the genotypic effects. In fact, results on growth and development rates from field data are notoriously difficult to interpret, because they can be sensitive to several other abiotic (e.g. oxygen content, pH, desiccation) and biotic (e.g. competition, predation) factors, as well as to their interactions (reviewed in Newman, 1992; Alford, 1999). This is also clearly illustrated by the results of our laboratory experiment: even small (1–2 °C) differences in temperature – if random in respect to latitude, can effectively conceal the clinal nature of variation observed in laboratory (cf. Fig. 3a, b). Hence, whatever the proximate cause for the large variation in developmental rates in the field, it effectively concealed the genotypic differentiation, resulting in a pattern of variation better known as ‘countergradient variation’ (Conover & Schultz, 1995). However, further studies utilizing a larger number of populations are required to understand causes of variation in developmental rates in the field.

The fact that the developmental rates within the latitudinal areas in the field were so variable raises an important question though: how could selection have created the genetic cline in developmental rates observed in the laboratory experiments in the first place? In other words, is there really selection for faster development in the north if the environment provides (e.g. thermally) favourable refugia allowing fast development even in many of the northern ponds? We think that the answer to this question is affirmative on the following grounds. First, the field-situation that we have described is just a snapshot both in time and space and does not capture all the variation relevant for evolutionary adaptation. For instance, the frequency of summers with suboptimal temperatures for development is higher in the north than in the south, as reflected in the fact that the minimum average temperatures over the season are clearly lowest in the northernmost localities (Table 1). Also, coefficients of variation for mean yearly average temperatures are highest for the northernmost frogs, suggesting that they are facing more variable temperature regimes over different years than the more southern frogs. Hence, such cold summers (absence of thermal refugia) could be chiefly responsible for the evolution of the observed cline in common frog developmental times. Indeed, catastrophic mortality of common frog larvae occurs frequently in northern Scandinavia (Gislen & Kauri, 1959; J. Merilä, personal observations) where the majority of the tadpoles in certain ponds can fail to metamorphose before winter.

Levinton (1983) proposed that genetic divergence in growth rate with respect to latitude could represent adaptation to temperatures most frequently encountered in nature. According to this view, our result that northern larvae developed faster than southern larvae in all temperatures (and food levels) may be surprising, as one would have expected southern populations to perform better in higher temperatures and northern populations in lower temperatures. However, larvae in northern populations frequently encounter temperatures as high as in the present experiment (J. Merilä, unpublished data; see below), and their fast development in high temperature could represent an adaptation to exploit temperature peaks and, hence, a wider range of temperatures (see also Ståhlberg et al., 2001). Superior developmental and growth performance of northern forms over southern conspecifics across a range of temperatures has also been observed in some other studies (e.g. Conover & Present, 1990; Blanckenhorn, 1991; Jonassen et al., 2000; Yamahira & Conover, 2002, but see Ståhlberg et al., 2001). The general conclusion emerging from these studies is that the time constraint imposed by the short growth season in the north may have favoured the capability of northern populations to exploit higher temperatures when they become available. Our results seem to conform this general picture.

Finally, three potentially confounding factors in the interpretation of the laboratory data should be mentioned. First, the interpretation that the clinal variation in developmental rates in the laboratory is due to genetic effects is made under the assumption that maternal effects were negligible. Although we cannot entirely rule out a role for maternal effects, two lines of evidence suggest that they are not overly important. First, subjecting two (Lund & Umeå) of the populations included in the present study to hybridization experiments (Laugen et al., 2002), we have shown that although maternal effects on developmental rate are present, they do not seem to override the genetic effects. Second, as egg size is considered to be the major pathway for maternal effects in amphibians (Kaplan, 1998), controlling for egg size effects should remove most of the influences of the maternal effects. Egg size explained a significant amount of variation in developmental rate, but even after these effects were statistically partitioned out, the observed cline in developmental rates was obvious and statistically significant. Hence, we conclude that the differentiation in developmental rates among these populations is mostly on account of genetic differentiation, and not because of egg size related maternal effects.

Second, since the experiments were conducted in a photoperiod typical for the mid-latitude populations, local adaptation in respect to photoperiod could, at least in theory, explain the observed cline in developmental rate. However, northern and southern larvae do not seem to respond differently to a wide range of photoperiods, including the one used in this study (Laurila et al. 2001). In fact, northern R. temporaria larvae do show very weak responses to any photoperiodism, and southern larvae respond differently only to extremely long day lengths. Hence, our results cannot be explained in terms of photoperiod responses.

Third, our experiments could be criticised as not being relevant for field situations because the range of temperatures utilized (14–22 °C) exceeded the average ambient temperatures (ca. 9–12 °C; Fig. 2) in the wild. However, actual temperatures encountered by the larvae in the field are extremely hard to estimate, as they can thermoregulate by moving to shallow water and experience temperatures higher than 20–24 °C even when the ambient temperatures are much lower. Indeed, the fact that developmental rates measured in the field fell within the range (40–60 days) which in the laboratory were attained in between 14 and 18 °C, suggests that the temperature range used in our experiments was relevant for the field situations, at least in terms of speed in which the temperature sum required for completing development is acquired: further studies examining effects of diel patterns of temperature fluctuations on developmental rates in different populations are an obvious area where further studies are needed.

Conclusions

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Study populations
  6. Laboratory experiment
  7. Experimental procedures
  8. Field data
  9. Statistical analyses
  10. Results
  11. Laboratory experiment
  12. Egg size
  13. Field data
  14. Discussion
  15. Conclusions
  16. Acknowledgments
  17. References

In conclusion, our analyses demonstrate that northern frogs may have a genetic capacity for faster development than southern frogs, and this differentiation can be understood in terms of adaptation to seasonal time constraints that increase in their severity with increasing latitude. The data on developmental rates from the wild show that the genetically faster development of the northern tadpoles is not realized in field conditions where environmental effects seem to override the genetic effects, rendering the observed divergence cryptic as predicted by the countergradient variation hypothesis. Hence, our results reinforce the conclusion that intraspecific genetic heterogeneity in the young northern European ecosystems may be more widespread than previously anticipated (Pamilo & Savolainen, 1999).

Acknowledgments

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Study populations
  6. Laboratory experiment
  7. Experimental procedures
  8. Field data
  9. Statistical analyses
  10. Results
  11. Laboratory experiment
  12. Egg size
  13. Field data
  14. Discussion
  15. Conclusions
  16. Acknowledgments
  17. References

We thank J. Ågren, S. Pakkasmaa, J. Höglund, J. Stone and two anonymous referees for comments on the manuscript. S. Andersson, P.-A. Crochet, J. Elmberg, A. Järvinen, I. Jönsson, O. Kalttopää, B. Lardner, J. Loman, L. Muhonen, Karoliina Räsänen, G. Sahlén, F. Söderman, R. Tramontano and H. Virtanen provided help in locating populations or/and with field work. E. Karvonen, S. Karttunen, N. Kolm, J.K. Larsson, M. Pahkala, F. Söderman and G. Løe assisted in the lab. The experiments were performed with the permission (C21/98) from the Ethical Committee of Uppsala University. Our research was supported by grants from the Swedish Natural Science Research Council (JM), the European Union (AL), Academy of Finland (AL, JM), NorFA (AL, JM, KR) and Uppsala University (Zoological Foundation & Sederholms Nordiska Resestipendium; both to ATL).

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Study populations
  6. Laboratory experiment
  7. Experimental procedures
  8. Field data
  9. Statistical analyses
  10. Results
  11. Laboratory experiment
  12. Egg size
  13. Field data
  14. Discussion
  15. Conclusions
  16. Acknowledgments
  17. References
  • Abrams, P.A., Leimar, O., Nylin, S. & Wiklund, C. 1996. The effect of flexible growth rates on optimal sizes and development times in a seasonal environment. Am. Nat. 147: 381395.
  • Alford, R.A. 1999. Ecology: resource use, competition, and predation. In: Tadpoles (R. W.McDiarmid & R.Altig eds), pp. 240278. University of Chicago Press, Chicago.
  • Arendt, J.D. 1997. Adaptive intrinsic growth rates: an integration across taxa. Q. Rev. Biol. 72: 149177.
  • Azevedo, R.B.R., French, V. & Partridge, L. 1996. Thermal evolution of egg size in Drosophila melanogaster. Evolution 50: 23382345.
  • Bernardo, J. 1996. Maternal effects in animal ecology. Am. Zool. 36: 83105.
  • Berven, K.A. 1982a. The genetic basis of altitudinal variation in the wood frog Rana sylvatica. I. An experimental analysis of life history traits. Evolution 36: 962983.
  • Berven, K.A. 1982b. The genetic basis of altitudinal variation in the wood frog Rana sylvatica. II. An experimental analysis of larval development. Oecologia 52: 360369.
  • Berven, K.A. & Gill D.E. 1983. Interpreting geographical variation in life-history traits. Am. Zool. 23: 8597.
  • Berven, K.A., Gill, D.E. & Smith-Gill, S.J. 1979. Countergradient selection in the green frog, Rana clamitans. Evolution 33: 609623.
  • Blanckenhorn, W.U. 1991. Life-history differences in adjacent water strider populations: phenotypic plasticity or heritable responses to stream temperature? Evolution 45: 15201525.
  • Blanckenhorn, W.U. & Fairbairn D.J. 1995. Life history adaptation along a latitudinal cline in the water strider Aquarius remigis (Heteroptera: Gerridae). J. Evol. Biol. 8: 2141.
  • Conover, D.O. & Present, T.M.C. 1990. Countergradient variation in growth rate: Compensation for length of the growing season among Atlantic silversides from different latitudes. Oecologia 83: 316324.
  • Conover, D.O. & Schultz, E.T. 1995. Phenotypic similarity and the evolutionary significance of countergradient variation. Trends Ecol. Evol. 10: 248252.
  • Endler, J.A. 1977. Geographic Variation, Speciation, and Clines, 1st edn. Princeton University Press, Princeton.
  • Fog, K., Schmedes, A. & de Lasson, D.R. 1997. Nordens padder og krybdyr. G.E.C. Gad, Copenhagen.
  • Gislén, T. & Kauri, H. 1959. Zoogeography of the Swedish amphibians and reptiles. Acta Vertebratica 1: 270395.
  • Gosner, K.L. 1960. A simplified table for staging Anuran embryos and larvae, with notes on identification. Herpetologica 16: 183190.
  • Hewitt, G. 2000. The genetic legacy of quaternary ice ages. Nature 405: 907913.
  • James, A.C. & Partridge, L. 1995. Thermal evolution of rate of larval development in Drosophila melanogaster in laboratory and field populations. J. Evol. Biol. 8: 315330.
  • Jonassen, T.M., Imsland, A.K., Fitzgerald, R., Bonga, S.W., Ham, E.V., Nævdal, G., Stefánsson, M.O. & Stefansson, S.O. 2000. Geographic variation in growth and food conversion efficiency of juvenile Atlantic halibut related to latitude. J. Fish Biol. 56: 279294.
  • Kaplan, R.H. 1998. Maternal effects, developmental plasticity, and life history evolution. In: Maternal Effects as Adaptations (T. E.Mousseau & C. W.Fox eds), pp. 244260. Oxford University Press, New York.
  • Laugen, A.T., Laurila, A. & Merilä, J. 2002. Maternal and genetic contributions to geographic variation in larval life-history traits in common frogs (Rana temporaria). Biol. J. Linn. Soc. 76: 6170.
  • Laurila, A., Pakkasmaa, S. & Merilä, J. 2001. Influence of seasonal time constraints on growth and development of common frog tadpoles: a photoperiod experiment. Oikos 95: 451460.
  • Laurila, A., Karttunen, S. & Merilä, J. 2002. Adaptive phenotypic plasticity and genetics of larval life histories in two Rana temporaria populations. Evolution 56: 617627.
  • Levins, R. 1969. Thermal acclimation and heat resistance in Drosophila species. Am. Nat. 103: 483499.
  • Levinton, J.S. 1983. The latitudinal compensation hypothesis: growth data and a model of latitudinal growth differentiation based upon energy budgets I: interspecific comparison based on Ophrytrocha puerlis (Polychaeta: Dorvilleidae). Biol. Bull. 165: 699707.
  • Lonsdale, D.J. & Levinton, J.S. 1985. Latitudinal differentiation in copepod growth: an adaptation to temperature. Ecology 66: 13971407.
  • Mayr, E. 1963. Animal Species and Evolution, 1st edn. Belknab Press, Cambridge.
  • Merilä, J., Laurila, A., Laugen, A.T., Räsänen, K. & Pahkala, M. 2000. Plasticity in age and size at metamorphosis in Rana temporaria– comparison of high and low latitude populations. Ecography 23: 457465.
  • Miaud, C., Guyetant, R. & Elmberg, J. 1999. Variations in life-history traits in the common frog Rana temporaria (Amphibia: Anura): a literature review and new data from the French Alps. J. Zool. 249: 6173.
  • Miaud, C. & Merilä, J. 2000. Local adaptation or environmental induction? Causes of population differentiation in alpine amphibians. Biota 2: 3150.
  • Mousseau, T.A. & Fox, C.W. 1998. The adaptive significance of maternal effects. Trends Ecol. Evol. 13: 403407.
  • Newman, R.A. 1992. Adaptive plasticity in amphibian metamorphosis. BioScience 42: 671678.
  • Odin, H., Eriksson, B. & Perttu, K. 1983. Temperature Climate Maps for Swedish Forestry. Department of Forest Soils, Swedish University of Agriculture Science, Uppsala.
  • Pamilo, P. & Savolainen, O. 1999. Post-glacial colonization, drift, local selection and conservation value of populations: a northern perspective. Hereditas 130: 229238.
  • Parsons, K.E. 1997. Contrasting patterns of heritable geographic variation in shell morphology and growth potential in the marine gastropod Bembicium vittatum: evidence from field experiments. Evolution 51: 784796.
  • Partridge, L., Barrie, B., Barton, N.H., Fowler, K. & French, V. 1995. Rapid laboratory evolution of adult life-history traits in Drosophila melanogaster in response to temperature. Evolution 49: 538544.
  • Present, T.M.C. & Conover, D.O. 1992. Physiological basis of latitudinal growth differences in Menidia menidia: variation in consumption or efficiency? Funct. Ecol. 6: 2331.
  • Roff, D.A. 1980. Optimizing development time in a seasonal environment: the ‘‘ups and downs’’ of clinal variation. Oecologia 45: 202208.
  • Rowe, L. & Ludwig, D. 1991. Size and timing of metamorphosis in complex life cycles: time constraints and variation. Ecology 72: 413427.
  • SAS Institute, I. 1996. SAS Proprietary Software Release 6.12, 6.12. SAS Inst. , Inc., Cary, NC.
  • Ståhlberg, F., Olsson, M. & Uller, T. 2001. Population divergence of developmental thermal optima in Swedish common frogs, Rana temporaria. J. Evol. Biol. 14: 755762.
  • Yamahira, K. & Conover, D.O. 2002. Intra- vs. interspecific latitudinal variation in growth: adaptation to temperature or seasonality? Ecology 83: 12521262.
  • Yampolsky, L.Y. & Scheiner, S.M. 1996. Why larger offspring at lower temperatures? A demographic approach. Am. Nat. 147: 86100.