SEARCH

SEARCH BY CITATION

Keywords:

  • body size;
  • Drosophila;
  • genetic;
  • latitudinal variation;
  • life history

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

Large amounts of genetic variation for wing length and wing area were demonstrated both within and between Drosophila melanogaster populations along a latitudinal gradient in South America. Wing length and wing area showed a strong positive correlation with latitude in both wild flies and laboratory-raised descendants. Large population differences were observed for heritability and coefficient of variation of these two traits, whereas relatively small population differences were found for development time, viability, pupal mortality, sex ratio and their norms of reaction to four developmental temperatures. No clear-cut latitudinal clines were established for these life-history characters. These results are discussed in the light of Bergmann's Rule and the relation between larval development and adult body size.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

The genetic analysis of quantitative variation for a character may reveal information about the functional significance of the genetic variation, its maintenance over time, and it may further show the occurrence of natural selection on the trait (e.g. Coyne & Beecham, 1987; Barton & Turelli, 1989). Moreover, heritabilities and genetic correlations might be tools to reconstruct patterns of natural selection in the past (Charlesworth, 1987; Barton & Turelli, 1989). An intriguing example of genetic variation is geographical, and especially clinal variation in traits, of which body size is arguably the best studied. In Drosophila, this research can be classified as studies of geographical variation of natural populations (Sokoloff, 1965; Louis et al., 1982; Pfriem, 1983; Robertson, 1987; Imasheva et al., 1994; James et al., 1995; Long & Singh, 1995; Pegueroles et al., 1995), of phenotypic plasticity (Capy et al., 1993; Thomas & Barker, 1993; David et al., 1994; Via et al., 1995; Noach et al., 1996), and of heritability estimates (Reeve & Robertson, 1953; Tantawy et al., 1964; Coyne & Beecham, 1987; Robertson, 1987; Noach et al., 1997; see Roff & Mousseau, 1987, for a review). All the field studies on D. melanogaster mentioned above report the existence of latitudinal clines for wing length (or body size in general), with wing length increasing with latitude. Body size is genetically and phenotypically correlated with life-history variables such as development time (+), viability (−), fecundity (+), mating success (+) and longevity (+/0) (symbols indicate sign of correlation; see Roff & Mousseau, 1987, for a review on correlated traits, while others include Partridge et al., 1987; Hillesheim & Stearns, 1992; Zwaan et al., 1992, 1995; Partridge & Fowler, 1993; James et al., 1995). Despite the importance of body size in the life-history of an organism, the factors that influence it in the evolutionary long- and short-term are still poorly understood. In the case of clinal variation, the identity of the evolutionary agents moulding the cline remains open.

The repeatability of clines has been taken as evidence that natural selection is at play (Partridge et al., 1994a; James et al., 1995). Cavicchi et al. (1989) and Partridge et al. (1994a) have shown that laboratory populations kept at various temperatures for long periods of time (> 4 years) evolved towards body sizes as observed in the clines. This gave support to a role for temperature in clinal body size evolution. Moreover, if natural selection is acting on body size through temperature, it could be expected that other life-history traits would be affected too, because of their genetic correlation with body size. In addition, other life-history traits (e.g. developmental time and longevity) can be influenced directly by temperature (Zwaan et al., 1992; Partridge et al., 1995).

Variation in developmental time in Drosophila has been found among natural populations (Gupta & Lewontin, 1982; Cavener, 1983; Marinkovic & Ayala, 1986). James & Partridge (1995) reported the existence of a latitudinal cline, with more slowly developing flies closer to the equator, while the eclosed adults were smaller. This is in contrast to what has been found in selection experiments for developmental time in D. melanogaster: a strong positive genetic correlation between developmental time and adult size was reported there (e.g. Zwaan et al., 1995; Nunney, 1996). Therefore, the explanation for the developmental time cline is not straightforward. Van Delden & Kamping (1991) have suggested that slower development and smaller size of individuals in populations of D. melanogaster near the equator are correlated with the occurrence of the cosmopolitan inversion In(2L)t (but see Knibb et al., 1987). Interestingly, developmental time was shown to be reduced in the cold temperature laboratory lines (Partridge et al., 1994b; James & Partridge, 1995).

An additional or alternative way for organisms to adapt to variable or changing environments is phenotypic plasticity. It can either pre-empt the genetic response, or genetic variation for plasticity itself could contribute to latitudinal adaptation (de Jong, 1990). In the latter case, the (adaptive) phenotypic response to the environment will be able to evolve in populations that encounter environmental change (Via et al., 1995). Given the likely important role of temperature in establishing the size cline, measuring life-history characters of different populations over a range of developmental temperatures could reveal (co-) adaptation of traits to the prevailing temperature regime.

Here we report on a study of 10 populations of D. melanogaster collected along a range of 4416 km and 39 latitudinal degrees in South America. We investigated differences in wing length between wild populations. One would expect possible latitudinal variation for body size to correlate with a latitudinal cline for development time. We also tested latitudinal variation in phenotypic plasticity. For each population, heritability (based on father–son regression) and additive genetic coefficient of variation for wing length and wing area were estimated on laboratory-reared descendants. Developmental time, viability, pupal mortality and sex ratio were measured for the populations at four experimental temperatures.

Materials and methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

Populations

During February and March 1995, wild flies were collected at one location in Ecuador (Guayaquil, 2°13’S), and nine locations in Chile (Arica, 18°28′S; Iquique, 20°13′S; Antofagasta, 23°38′S; Copiapó, 27°20′S; Coquimbo, 29°56′S; Valparaíso, 33°05′S; Linares, 35°48′S; Valdivia, 39°48′S; and Puerto Montt, 41°30′S). An overview of the positions of the sampling sites is presented in Fig. 1. The number of flies collected at each location is presented in Table 2.

image

Figure 1 Map of South America. The names and positions of the 1. 0 sampling sites are indicated.

Download figure to PowerPoint

Table 2.    Number of collected adult flies at the respective locations. The second and third columns present the estimated heritabilities (h2) based on parent–offspring regression for (transformed) wing lengths and wing areas, respectively. The total number (n) of father–son pairs per population is given in parentheses. The additive genetic coefficient of variation, CVa, ±SE is presented in the last two columns. Thumbnail image of

In total, almost 40 latitudinal degrees were covered, while the longitudinal range of these locations was less than 10°. All sites, except Linares (140 m) and Copiapó (350 m) were situated at sea level. The average distance between two adjacent Chilean locations was 325 km. All flies were caught by sweeping a net over boxes containing rotten fruit in fruit stores or at market places. Number of flies collected at each site ranged from 22 to 334 (Table 2). Twenty offspring (equally divided over the sexes) of each gravid wild female were used to found the F1. Numbers in subsequent generations were between 2000 and 3000 (see below). The F3 generation was taken to the Netherlands for further research.

Rearing conditions

After arrival in the Netherlands, each population was divided over, and kept in, 10 half-pint bottles containing 30 mL standard medium. Standard medium consisted, per litre of water, of 26 g dead yeast, 54 g sugar, 17 g agar and 13 mL ethanol in which 1.3 mg nipagine was dissolved. Flies were kept at 25 °C, and at 24-h light conditions. The density per bottle was controlled and varied between 200 and 300 larvae. After three generations in the laboratory, the experiments were started.

Reaction norms

Eggs were obtained by allowing 100 randomly chosen females per population (from the third laboratory-reared generation) to lay eggs for 4 h on dishes containing standard medium. Per population and temperature, five replicates of 75 eggs were transferred to plastic vials containing 8 mL standard medium. Each of the five replicates/vials was then put at the respective temperatures.

Four characters were measured: developmental time, (egg-to-adult) viability, sex ratio and pupal mortality, at four experimental temperatures: 15, 20, 25 and 29 °C. When the first flies were about to eclose, the vials were checked for adults every 4 h (for the vials at 29 °C), every 6 h (for those at 25 °C), 8 h (20 °C) and 12 h (15 °C). After eclosion, the dead pupae were counted and pupal mortality was calculated. The viability was calculated as the number of adult flies divided by the initial number of eggs (i.e. 75 per vial). Developmental time was computed for each cohort as the time between average laying time and the midpoint of the interval (4, 6, 8 or 12 h) in which the flies emerged. Per vial, developmental time was computed as the weighted mean eclosion time.

Wing measurement

Left wings were taken from males only (although the total phenotypic variance is equal in the two sexes, the environmental variance is about 20% less in males than in females [Reeve & Robertson, 1953]), and were embedded in a drop of Euparal on a microscope slide. Wing lengths were determined using a light microscope (10 × 8) with an ocular micrometer. The distance from anterior cross vein to the wing tip was taken as a measure of wing length (Prout, 1958).

The wing areas were measured at 50× magnification using a microscope with a camera lucida attachment and graphics tablet. The tablet was connected to a computer to store the data for later analysis.

To obtain a measure for allometrical relationships within the wing, in other words a character that provides information about wing shape or roundness, we calculated the aspect ratio: wing length2/wing area.

Heritability and coefficient of variation

The flies that had been used for the reaction norm experiment at 25 °C were kept separately (after eclosion and measurement of the development time) in plastic vials and allowed to age for 1 week at 25 °C.

For each population and from each replicate (i.e. vial), 14 single-pair crosses were performed. After 3 days, the male was removed from the vial and wing length and wing area were measured. The mother was left in the vial in order to produce a sufficient number of eggs. When the progeny of these crosses eclosed, five sons per cross were randomly sampled and their left wings were measured.

Heritabilities (h2) were estimated using offspring-one parent regression analysis (Falconer & Mackay, 1996). Per population, a general factorial ANOVA was carried out, with the wing measurements of the son as dependent variable, the wing measurements of the father as covariable, and with the replicates as random factor. Heritabilities were calculated by taking twice the regression coefficient of the mean wing length and mean wing area of five sons on the wing length and wing area of the father.

As an additional measure for genetic variability, coefficients of variation (CVa) were calculated (Houle, 1992): CVa=100√Va/X where Va is the additive genetic variance (=total variance [Vp] × h2) and X represents the population mean. The standard error of CVa is calculated as CVa/√2n * (√(1 + 2(CVa/100)2)).

Statistical analysis

Data distributions were tested for normality and variances were tested for heterogeneity using the Bartlett's test of equal variances. As a result, wing lengths and wing areas were ln-transformed, and viabilities and pupal mortalities were angularly transformed.

To test the effects of population and rearing temperature, and interactions between these two factors, on measured variables, model I ANOVAs were performed with temperature and population as fixed treatment effects. A Student–Newman–Keuls test was used to test for significant differences among means of populations and temperatures. Product-moment correlation coefficients (r) were tested for their difference from zero by using a t-test.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

Reaction norms

Figure 2 presents the measured development time, egg-to-adult viability, sex ratio and pupal mortality for each of the four temperatures used. Two-way ANOVAs were performed to determine the effects of population and temperature on the traits measured (Table 1).

image

Figure 2.   Summary for all 10 populations of the reaction norms for five parameters at four different temperatures. Depicted are a. female development time, b. male development time, c. sex ratio of the adult flies, d. viability and e. pupal mortality. Each line represents one population and connects the average values obtained at the four experimental temperatures.

Download figure to PowerPoint

Table 1.    ANOVA for the effects of population and temperature on the traits measured. Thumbnail image of

For developmental time, a three-way ANOVA, with population, temperature and sex as main effects, indicated significant sex effects, but no sex × population effects: males developed more slowly irrespective of population. The differences between sexes became more pronounced at lower temperatures (temperature–sex interaction). Differences were also observed between populations. The Antofagasta and Arica populations clearly showed a longer developmental time than all other populations, while these two populations also differed from each other (SNK test, < 0.05). The difference between these two populations and the rest was most obvious at 15 °C.

Sex ratio did not yield any significant population, temperature and interaction effects.

The ANOVA for viability showed temperature and population effects. A significant difference over all temperatures (SNK test, < 0.05) was observed only between the highest (Valparaíso) and lowest population (Iquique). When tested within each temperature, the lower viability of the Iquique population compared to the other populations was significant at 15 °C and 29 °C only. For all populations, a higher viability was found at 20 °C and 29 °C (when compared to 15 °C and 25 °C).

Pupal mortality was significantly different for different populations and temperatures (Table 1). Mortality was highest at 29 °C, lowest at 25 °C (SNK test, < 0.05). Iquique had over all temperatures the highest pupal mortality, differing significantly from Valparaíso and Puerto Montt (i.e. two populations with the lowest mortality). The significant interaction term indicates that the difference between populations depended on the temperature. At 15 °C, only a difference between the populations with the highest (Iquique) and the lowest mortality (Puerto Montt) was observed, whereas at 20 °C, the pupae of Puerto Montt, Iquique and Antofagasta all differed from the pupae of the population with the lowest mortality (Copiapó). At 25 °C, no significant difference between populations was found, but at 29 °C all populations could be divided into three groups which differed significantly from each other.

Latitudinal clines for life-history traits

No significant correlations with latitude were detected for sex ratio, pupal mortality and viability at any one temperature (single regression) or when combined over all temperatures (ANCOVA, see below). Furthermore, no significant correlation between developmental time and latitude was observed for both sexes at a single temperature (Fig. 3). Even when the Antofagasta population is considered as an outlier (Dixon's test for outliers: at 15 °C and 25 °C significant at α=0.05), omission of this population for developmental time analysis did not yield a significant correlation with latitude. However, an analysis of covariance for developmental time, with temperature and sex as independent variables and latitude as a covariate, showed a significant negative regression between development time and latitude (all populations: t=−2.445, < 0.015). However, the effect of latitude in this analysis explains only 0.1% of the total observed variation.

image

Figure 3 Relationship between development time and latitude. The mean values (± standard deviation), and linear regression lines for each temperature are presented. For each temperature, data from females (closed circles; hatched line) and males (open circles; con-tinuous line) are given. The data points for the Antofagasta/15° combination lie outside the graph boundaries (i.e. 945.14 ± 27.42 [♀♀] and 972.30 ± 42.83 [♂♂]), and are excluded for the regression lines of the 15 °C and 25 °C data (see text). The slopes of the regression lines are: 15 °C, −1.163(♀♀) and −1.121(♂♂); 20 °C, −0.414(♀♀) and −0.413(♂♂); 25 °C, −0.103. (♀♀) and −0.207(♂♂); 29 °C, −0.126(♀♀) and −0.116(♂♂).

Download figure to PowerPoint

Genetic parameters

When computing the heritability (h2) based on parent–offspring regression of wing length, we detected significant variation among the 10 populations, with h2 values ranging from 0.099 to 0.863 (Table 2). Over all populations, similar heritabilities were found when using wing area instead of wing length (correlation coefficient r=0.976, < 0.0001). The exceptions were the Iquique and the Valdivia populations for which significantly different heritabilities were detected when taking wing lengths or wing areas as the point of departure.

No significant correlation between the genetic parameters and the number of flies collected at each location was detected. This is potentially an important relationship, because the number of founders directly influences the amount of genetic variation trapped in the established laboratory populations. Inbreeding and drift in these populations can be considered minimal because of equal founder representation and the extensive breeding procedures used (see Materials and Methods). Moreover, the effective population size is likely to be higher than the number of collected females, because multiple mating occurs frequently in nature (e.g. Partridge et al., 1987). Therefore, it can be expected that the laboratory populations closely reflect the natural source population with respect to genetic variation for the traits under study.

For wing length and wing area, the additive genetic coefficients of variation (CVa) were calculated in each population (Table 2). The CVa values strongly correlated with heritabilities (wing length: r=0.88; P < 0.001; wing area: r=0.90; P < 0.0001) and, like the h2 values, significant differences between populations were found. For almost all populations (Guayaquil and Iquique being the exceptions with similar CVas for wing lengths and wing areas), the CVa for wing area was 2–3 times higher than the CVa for wing length.

No significant linear correlation between heritabilities and coefficient of variation, and latitude was found. But when applying an ad hoc nonlinear curve-fitting, we observed a significant second-order polynomial relation (i.e. nonlinear; > 0.95, < 0.0001 for all four combinations), suggesting a more complex relationship between heritability and CVa on the one hand, and latitude on the other.

Wing morphology

When raised under standard laboratory conditions, both the wing lengths and the wing areas of the males were significantly and positively correlated to latitude, with correlation coefficients (r) around 0.93 (Fig. 4).

image

Figure 4.   Relationship between a. wing length, b. wing area and c. aspect ratio of males, and the latitude of the original population. Data from three generations are given: wild males (open triangles; continuous line), laboratory-raised males, i.e. ‘lab fathers’ (open circles; hatched line) and their laboratory-raised sons, i.e. ‘lab sons’ (closed circles; continuous line). For wing area and aspect ratio, no data from wild males were available. The lines represent the regression for each category.

Download figure to PowerPoint

Wing area and wing length were also correlated with each other (= 0.96 [fathers], respectively, 0.95 [sons]). However, the slopes of the three wing length regression lines (Fig. 4a) differed significantly from each other (< 0.05 [father vs. son], < 0.0001 [wild vs. lab]), indicating a strong environmental component in the wild. The significantly higher values for wing area of the sons compared to the fathers for each population (Fig. 4b) might be attributed to the fact that, although the actual number of eggs per vial was similar, the laying-interval for the sons (i.e. several days due to single-pair crosses) was much longer than for the fathers (i.e. several hours). This could have resulted in a lower level of larval competition due to a more spread development, and consequently in a higher body weight and larger wings. In fruit flies, fathers and sons will always be raised at potentially slightly different environments because they develop at consecutive times. Therefore, minor variations in food quality, temperature and humidity can occur even under highly controlled conditions. However, this environmental variation applies to all populations, and thus these effects do not invalidate our comparison of h2 over latitude (see above). Moreover, heritability estimates should be considered robust since it has been shown that estimations of h2 in the field and in the laboratory are highly correlated, despite vastly different rearing environments (Coyne & Beecham, 1987; for review, Weigensberg & Roff, 1996).

Figure 4(c) shows the relationship between aspect ratios and latitude. A second-order polynomial regression between the two variables was significant for both fathers (r=0.926, < 0.001) and sons (r=0.778, < 0.05). The only tropical population (Guayaquil) showed a substantially higher aspect ratio for fathers than all other populations (ANOVA, F=35.86, < 0.0001; SNK test, < 0.05), indicating more elongated wings than the other (Chilean) populations. The fathers from the most northern Chilean population (Arica) differed to a lesser degree from the other populations and also from Guayaquil (SNK test, < 0.05).

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

In this paper, we used the descendants of 10 Latin American populations of D. melanogaster taken from the wild to demonstrate the large amount of genetic variation within and between populations along a geographical scale. Wing length and wing area, being measures for body size, showed a strong positive correlation with latitude for both wild and laboratory-raised flies. Large population differences were observed for h2 and CVa for these traits, which were not correlated to the number of flies collected at each site. Compared with wing length and area, relatively small population differences were found for developmental time, viability, pupal mortality and sex ratio and between their norms of reaction to four experimental temperatures. No clear-cut latitudinal cline was established for these life-history characters, except for a weak negative correlation between developmental time and latitude when temperatures and sexes were combined.

Latitudinal clines

In flies collected at a comparable latitudinal range in Australia (16°S to 43°S), James et al. (1995) reported a similar cline in body size. Differences for developmental time between these populations were also reported: larvae from higher latitudes developed faster at intermediate experimental temperatures (James & Partridge, 1995). One of the tentative suggestions put forward by these authors is the involvement in tropical populations of the cosmopolitan inversion In(2L)t which shows slower development and smaller body sizes (Van Delden & Kamping, 1989, 1991). We were not able to confirm this hypothesis as no evidence for a latitudinal cline for the frequency of In(2L)t was found in a broader set of Latin American populations (van’t Land, 1997). Interestingly, when relating developmental times to inversion frequencies of the same populations (data on In(2L)t frequency from van’t Land, 1997), a significant positive correlation (most prominent at 25 °C) between development time and frequency of the cosmopolitan inversion In(2L)t was detected (r = 0.878 [males] and 0.773 [females], respectively; Fig. 5). No correlation was found between In(2 L)t frequency and wing length.

image

Figure 5 Relationship between development time and frequency of the cosmopolitan inversion In(2L)t at 25 °C. Female (closed circles; hatched line) and male (open circles; continuous line) data (mean ± SD) are presented separately. The two lines represent a linear regression. The Antofasta data (In[2L]t freq.=0.43; male devt.time =284.29 h.; female devt.time=272.54 h) are omitted in this graph and for the regression analysis (Dixon's test for outliers at α=0.05. ). The frequency of the inversion was determined using males in a backcrossing design described in van’ Delden & Kamping (1989) and van ’t Land (1997).

Download figure to PowerPoint

The correlation described by James & Partridge (1995) depends heavily on one population measured at low latitude (16°53′). When this population is left out, the picture becomes less clear, with the correlation between time to pupation and latitude only being significant in one of the two experiments (A. C. James, personal communication). Obviously, it is important to measure developmental time at more than one temperature. Our results for single temperatures showed no significant correlation between developmental time and latitude for the eight tests we performed (Fig. 3). Only an ANCOVA was able to detect a small effect of latitude on developmental time.

The persistence of a latitudinal cline for wing size, even under standardized laboratory conditions, illustrates the differences in genetic constitution of the populations. This is especially true when bearing in mind the absence of a significant correlation between developmental time and wing length in the present study (Fig. 6).

image

Figure 6.   Relationship between developmental time and wing length at 25 °C. The mean values (± standard deviation) for wing length and developmental time are from the laboratory fathers.

Download figure to PowerPoint

Generally, within populations, body size and developmental rate are genetically negatively correlated (e.g. Nunney, 1996; for review, Zwaan et al., 1995). One could argue that, in less predictable environments (temperate climate zones), strong selection for rapid development is operating on the one hand, and there is independent selection for increased body size on the other. An alternative and possibly additional explanation could be that variation in acquisition loci between the populations, caused by the environments in which they live, is responsible for obscuring the fundamental trade-off between developmental rate and body size (Houle, 1991; Chippindale et al., 1996). Artificial selection experiments could be designed to test this hypothesis.

Genetic parameters

The heritabilities reported in this study showed a large range of values across the populations. The averages of these values (0.48 [WL] and 0.42 [WA]) are well within the range of h2 values for wing length reported in other studies (0.30–0.76; Reeve & Robertson, 1953; Tantawy et al., 1964; Coyne & Beecham, 1987; Robertson, 1987; Thomas & Barker, 1993; David et al., 1994; Noach et al., 1996, 1997). Note that nothing is known about the genetic history of the populations in terms of selection and drift, irrespective of possible natural selection due to geography. Therefore, it is important to investigate whether different alleles and/or different genes are contributing to Va in the different populations. Estimating h2 at various temperatures could be a first step in this and in unravelling the genetics of latitudinal variation. Selection experiments for body size and developmental time in the separate populations may reveal underlying patterns of variation. This could also indicate whether evolution for phenotypic plasticity as a trait in itself has occurred in the populations. Our data show that at least two populations (Antofagasta and Arica) have reaction norms for developmental time at various temperatures, which differ from the other populations (Fig. 2a,b). At the moment, no definitive answer can be given about these matters and the particular coadaptation of traits in these populations, which painfully points to our lack of knowledge of which life-history set-up is important at different latitudes.

Wing morphology

The repeatability of the latitudinal cline for body size suggests that it is caused by natural selection (Partridge et al., 1994a; James et al., 1995). However, it is not exactly clear what the target of selection is (Coyne & Beecham, 1987; Capy et al., 1993; Imasheva et al., 1994; James et al., 1995). Body size itself could be this target (e.g. the ratio between surface area and body volume), or this character has been altered as a correlated response to selection on another trait. That trait can certainly not be developmental time and no other candidates have appeared in the literature so far.

A more functional approach is to consider wing length not as a derivative of body size or as a consequence of, for example, developmental time, but as a more or less independent trait connected to flight capacity, wing load and thermoregulation. As long ago as 1942, it was demonstrated that wing-beat frequency is positively correlated to air temperature as well as to geographical origin of the strains tested. Flies adapted to high temperature therefore have short wings relative to thorax volume (Reed et al., 1942; Cavicchi et al., 1991; David et al., 1994; Van ‘t Land et al., 1994). Flies reared at high temperatures are expected to have a relatively high wing-load index (i.e. unit of body volume per unit wing area) and a more rapid wing beat (Stalker, 1980). These findings could explain the results of the present study that tropical flies have smaller wings and a higher aspect ratio (>1.3 for ‘lab fathers’ from Guayaquil and Arica, Fig. 4c; higher aspect ratios are predicted for fast-frequency hovering animals [Norberg, 1995]). It could be worthwhile to concentrate more on wing-load and flight-capacity when investigating correlations of wing morphology with latitude.

On the other hand, given the developmental plasticity of body size for temperature (e.g. Zwaan et al., 1992) it is surprising that any evolutionary response to temperature occurred at all (see also Partridge et al., 1994a). Indeed, the wing length data of wild as compared to lab-reared males shows that the regression line on latitude is much steeper for the former than for the latter group (Fig. 4a; see also James et al., 1997). Therefore, it appears that environmental temperature variation may conceal part of the genetic contribution to variation in body size in these populations in nature: the absolute difference in wing length between wild males and lab fathers is 0.124 mm and 0.129 mm for Guayaquil and Puerto Montt, respectively, while the difference between Guayaquil and Puerto Montt lab fathers is 0.123 mm. Recently, Van Voorhies (1996) proposed a general explanation for the occurrence of size clines in ectotherms independent of species-specific ecology. He argued that at lower temperatures cells grow bigger and this in turn is responsible for increased size, which he illustrated on Caenorhabditis elegans data. However, he failed to address the issue as to how this effect is translated into genetic differences. It could be that, under different temperature regimes, enzyme activity and cellular processes are tuned to different optima which, as a side-effect, would cause either the production of more or less cells, respectively, and/or the cells to grow bigger or smaller, respectively (e.g. Van der Have & de Jong, 1996).

Although this may be a universal effect for all ectotherms, it could be counteracted or reinforced by other factors like natural selection acting to coadapt life-history traits. The fact that the cellular basis of the body size variation in natural (James et al., 1995) and laboratory populations (Partridge et al., 1994a) is so different could be an indication of this. Moreover, it has been shown that direct selection on wing area or thorax length caused an asymmetrical response in the cellular basis: big flies are bigger because of more cells, small flies are smaller because of smaller cells with some insignificant variation in cell number (L. Partridge, K. Fowler, V. French, R. Langelan and B. Zwaan, in preparation). If both adaptive and nonadaptive (i.e. side-effects of physiological adaptation) factors play a role in thermal evolution, then finding a general pattern could be a very strenuous job. A start is made by unravelling thermal evolution in D. melanogaster. Identifying the genes underlying variation in body size seems a necessary step to achieve this.

Acknowledgments

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

This research was funded by the Netherlands Foundation for the Advancement of Tropical Research (WOTRO) by grant W82-182. B.J.Z. acknowledges the BBSRC for financial support. R. Bijlsma and L. Partridge gave valuable comments on an earlier draft of this paper. Dr H. Villarroel from the University of Playa Ancha, Valparaíso, Chile, was very helpful with collecting the flies and providing laboratory facilities. We would like to thank L. Hoeksema-du Pui for preparing the Drosophila medium.

Footnotes
  1. Present address: Medical Council/Organization for The Advancement of Tropical Research, Netherlands Organization for Scientific Research (NWO), PO Box 93138, 2509 AC Den Haag, The Netherlands

  2. Present address: NIOO-CTO, PO Box 40, 6666 ZG Heteren, The Netherlands.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References
  • 1
    Barton, N.H. & Turelli, M. 1989. Evolutionary quantitative genetics: how little do we know? Annu. Rev. Genet. 23: 337 370.
  • 2
    Capy, P., Pla, E., David, J.R. 1993. Phenotypic and genetic variability of morphometrical traits in natural populations of Drosophila melanogaster and D. simulans. I. Geographic variations. Genet. Sel. Evol. 25: 517 536.
  • 3
    Cavener, D.R. 1983. The response of enzyme polymorphisms to developmental rate selection in Drosophila melanogaster . Genetics 105: 105 113.
  • 4
    Cavicchi, S., Giorgi, G., Natali, V., Guerra, D. 1991. Temperature-related divergence in experimental populations of Drosophila melanogaster. III. Fourier and centroid analysis of wing shape and relationship between shape variation and fitness. J. Evol. Biol. 4: 141 159.
  • 5
    Cavicchi, S., Guerra, D., Natali, V., Pezzoli, C., Giorgi, G. 1989. Temperature-related divergence in experimental populations of Drosophila melanogaster. II. Correlation between fitness and body dimensions. J. Evol. Biol. 2: 235 251.
  • 6
    Charlesworth, B. 1987. The heritability of fitness. In: Sexual Selection: Testing the Alternatives (J. W. Bradbury and M. B. Andersson, eds), pp. 21–40. John Wiley & Sons, New York.
  • 7
    Chippindale, A.K., Chu, T.J.F., Rose, M.R. 1996. Complex trade-offs and the evolution of starvation resistance in Drosophila melanogaster . Evolution 50: 753 766.
  • 8
    Coyne, J.A. & Beecham, E. 1987. Heritability of two morphological characters within and among natural populations of Drosophila melanogaster . Genetics 117: 727 737.
  • 9
    David, J.R., Moreteau, B., Gauthier, J.P., Pétavy, G., Stockel, A., Imasheva, A.G. 1994. Reaction norms of size characters in relation to growth temperature in Drosophila melanogaster an isofemale lines analysis. Genet. Sel. Evol. 26: 229 251.
  • 10
    Falconer, D.S. & Mackay, T.F.C. 1996. Introduction to Quantitative Genetics, 4th edn. Longman Group Ltd, London.
  • 11
    Gupta, A.P. & Lewontin, R.C. 1982. A study of reaction norms in natural populations of Drosophila pseudoobscura . Evolution 36: 934 948.
  • 12
    Hillesheim, E. & Stearns, S.C. 1992. Correlated responses in life-history traits to artificial selection for body weight in Drosophila melanogaster . Evolution 46: 745 752.
  • 13
    Houle, D. 1991. Genetic covariance of fitness correlates: what genetic correlations are made of and why it matters. Evolution 45: 630 648.
  • 14
    Houle, D. 1992. Comparing evolvability and variability of quantitative traits. Genetics 130: 195 204.
  • 15
    Imasheva, A.G., Bubli, O.A., Lazebny, O.E. 1994. Variation in wing length in Eurasian natural populations of Drosophila melanogaster . Heredity 72: 508 514.
  • 16
    James, A.C., Azevedo, R.B.R., Partridge, L. 1995. Cellular basis and developmental timing in a size cline of Drosophila melanogaster . Genetics 140: 659 666.
  • 17
    James, A.C., Azevedo, R.B.R., Partridge, L. 1997. Genetic and environmental responses to temperature of Drosophila melanogaster from a latitudinal cline. Genetics 146: 881 890.
  • 18
    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: 315 330.
  • 19
    De Jong, G. 1990. Quantitative genetics of reaction norms. J. Evol. Biol. 3: 447 468.
  • 20
    Knibb, W.R., Oakeshott, J.G., Wilson, S.R. 1987. Chromosome inversion polymorphisms in Drosophila melanogaster. IV. Inversion and Adh allele frequency changes under selection for different development times. Heredity 59: 95 104.
  • 21
    Van’t Land, J. 1997. Latitudinal variation in Drosophila melanogaster. On the Maintenance of the World-Wide Polymorphisms for Adh, aGpdh and In (2L)T. PhD Thesis, University of Groningen.
  • 22
    Van‘t Land, J., Kamping, A., Van Delden, W. 1994. Differences in some parameters among two geographically distinct populations of Drosophila melanogaster . Dros. Inf. Serv. 75: 83 84.
  • 23
    Long, A.D. & Singh, R.S. 1995. Molecules versus morphology: the detection of selection acting on morphological characters along a cline in Drosophila melanogaster . Heredity 74: 569 581.
  • 24
    Louis, J., David, J., Rouault, J., Capy, P. 1982. Altitudinal variations of Afro-tropical D. melanogaster populations . Dros. Inf. Serv. 58: 100 101.
  • 25
    Marinkovic, D. & Ayala, F.J. 1986. Genetic variation for rate of development in natural populations of Drosophila melanogaster . Genetica 71: 123 132.
  • 26
    Noach, E.J.K., De Jong, G., Scharloo, W. 1996. Phenotypic plasticity in morphological traits in two populations of Drosophila melanogaster . J. Evol. Biol. 9: 831 844.
  • 27
    Noach, E.J.K., De Jong, G., Scharloo, W. 1997. Phenotypic plasticity of wings in selection lines of Drosophila melanogaster . Heredity 79: 1 9.
  • 28
    Norberg, U.M. 1995. How a long tail and changes in mass and wing shape affect the cost for flight in animals. Funct. Ecol. 9: 48 54.
  • 29
    Nunney, L. 1996. The response to selection for fast larval development in Drosophila melanogaster and its effect on adult weight: an example of a fitness trade-off. Evolution 50: 1193 1204.
  • 30
    Partridge, L., Barrie, B., Barton, N.H., Fowler, K., French, V. 1995. Rapid evolution of adult life-history traits in Drosophila melanogaster in response to temperature . Evolution 49: 538 544.
  • 31
    Partridge, L., Barrie, B., Fowler, K., French, V. 1994a. Evolution and development of body size and cell size in Drosophila melanogaster in response to temperature. Evolution 48: 1269 1276.
  • 32
    Partridge, L., Barrie, B., Fowler, K., French, V. 1994b. Thermal evolution of pre-adult life-history traits in Drosophila melanogaster . J. Evol. Biol. 7: 645 663.
  • 33
    Partridge, L. & Fowler, K. 1993. Responses and correlated responses to artificial selection on thorax length in Drosophila melanogaster . Evolution 47: 213 226.
  • 34
    Partridge, L., Hoffmann, A., Jones, J.S. 1987. Male size and mating success in Drosophila melanogaster and D. pseudoobscura under field conditions. Anim. Behav. 35: 468 476.
  • 35
    Pegueroles, G., Papaceit, M., Quintana, A., Guillén, A., Prevosti, A., Serra, L. 1995. An experimental study of evolution in progress: clines for quantitative traits in colonizing and Parearctic populations of Drosophila . Evol. Ecol.. 9: 453 465.
  • 36
    Pfriem, P. 1983. Latitudinal variation in wing size in Drosophila subobscura and its dependence on polygenes of chromosome O . Genetica 61: 221 232.
  • 37
    Prout, T. 1958. A rapid method for measuring wing length. Dros. Inf. Serv. 32: 170 171.
  • 38
    Reed, S.C., Williams, C.M., Chadwick, L.E. 1942. Frequency of wing-beat as a character for separating species, races and geographic varieties of Drosophila . Genetics 27: 349 361.
  • 39
    Reeve, E.C.R. & Robertson, F.W. 1953. Studies in quantitative inheritance. II. Analysis of a strain of Drosophila melanogaster selected for long wings . J. Genet. 51: 276 316.
  • 40
    Robertson, F.W. 1987. Variation of body size within and between wild populations of Drosophila buzzatii . Genetica 72: 111 125.
  • 41
    Roff, D.A. & Mousseau, T.A. 1987. Quantitative genetics and fitness: lessons from Drosophila . Heredity 58: 103 118.
  • 42
    Sokoloff, A. 1965. Geographic variation of quantitative characters in populations of Drosophila pseudoobscura . Evolution 19: 300 310.
  • 43
    Stalker, H.D. 1980. Chromosome studies in wild populations of Drosophila melanogaster. II. Relationship of inversion frequencies to latitude, season, wing-loading and flight activity. Genetics 95: 211 223.
  • 44
    Tantawy, A.O., Mallah, G.S., Tewfik, H.R. 1964. Studies on natural populations of Drosophila. II. Heritability and response to selection for wing length in Drosophila melanogaster and D. simulans at different temperatures. Genetics 49: 935 948.
  • 45
    Thomas, R.H. & Barker, J.S.F. 1993. Quantitative genetic analysis of the body size and shape of Drosophila buzzatii . Theor. Appl. Genet. 85: 598 608.
  • 46
    Van Delden, W. & Kamping, A. 1989. The association between the polymorphisms at the Adh and αGpdh loci and the In(2L)t inversion in Drosophila melanogaster in relation to temperature. Evolution 43: 775 793.
  • 47
    Van Delden, W. & Kamping, A. 1991.Changes in relative fitness with temperature among second chromosome arrangements in Drosophila melanogaster. Genetics 127: 507 514.
  • 48
    Van der Have, T.M. & De Jong, G. 1996. Adult size in ectotherms: temperature effects on growth and differentiation. J. Theor. Biol. 183: 329 340.
  • 49
    Van Voorhies, W.A. 1996. Bergmann size clines: a simple explanation for their occurrence in ectotherms. Evolution 50: 1259 1264.
  • 50
    Via, S., Gomulkiewicz, R., De Jong, G., Scheiner, S.M., Schlichting, C.D., Van Tienderen, P.H. 1995. Adaptive phenotypic plasticity: consensus and controversy. Trends Ecol. Evol. 10: 212 217.
  • 51
    Weigensberg, I. & Roff, D.A. 1996. Natural heritabilities: can they be reliably estimated in the laboratory? Evolution 50: 2149 2157.
  • 52
    Zwaan, B.J., Bijlsma, R., Hoekstra, R.F. 1992. On the developmental theory of ageing. II. The effect of developmental temperature on longevity in relation to adult body size in D. melanogaster . Heredity 68: 123 130.
  • 53
    Zwaan, B.J., Bijlsma, R., Hoekstra, R.F. 1995. Artificial selection for developmental time in Drosophila melanogaster in relation to the evolution of ageing: direct and correlated responses. Evolution 49: 635 648.