• geographical races;
  • phenotypic plasticity;
  • thorax length;
  • wing length;
  • wing/thorax ratio


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

Reaction norms to growth temperature of two size-related traits, wing and thorax length, were compared in tropical (West Indies) and temperate (France) populations of the two sibling species, Drosophila melanogaster and D. simulans. A major body size difference was found in D. melanogaster, with much smaller Caribbean flies, while D. simulans exhibited little size variation between geographical populations. The concave norms of reaction were adjusted to second- or third-degree polynomials, and characteristic points calculated i.e. maximum value (MV) and temperature of maximum value (TMV). TMVs were confirmed to be higher for thorax than for wing length, higher in D. melanogaster than in D. simulans, and higher in females than in males. For both traits Caribbean populations exhibited higher TMVs in the two species, strongly suggesting an adaptive shift of the reaction norms toward higher temperature in warm-adapted populations. The wing/thorax ratio was also analysed, and found to be significantly lower in tropical populations of both species. This ratio, which is related to wing loading and flight capacity, might evolve independently of body weight itself.


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

In several Drosophila species, body size varies regularly according to geographical gradients, and especially according to latitude. Latitudinal clines have been described in the American D. robusta (Stalker & Carson, 1947, 1948), in the European D. subobscura (Prevosti, 1955; Misra & Reeve, 1964), and in cosmopolitan species such as D. melanogaster and D. simulans (Tantawy & Mallah, 1961; David & Bocquet, 1975a,b; Capyet al., 1993) and D. virilis (David & Kitagawa, 1982). In all these cases a larger body size has been observed with increasing latitude, and average temperature is considered a key selective factor (David et al., 1983). We do not know, however, why it is better for an ectotherm such as Drosophila to be bigger in a colder climate (Powell, 1974; David et al., 1994; Partridge et al., 1994).

The interpretation of such variations is still complicated by the occurrence of phenotypic plasticity according to developmental temperature. Size exhibits an overall decrease when temperature increases, and the reaction norm (i.e. response curve) is somewhat parallel to the genetical variations observed between geographical populations (David et al., 1983, 1994; Noach et al., 1996). When size-related traits are investigated over a broad thermal range, the complete reaction norms are, however, nonlinear but exhibit a maximum at low or medium temperature (David et al., 1983, 1990, 1994; Morin et al., 1996, 1997). This temperature of maximum value (TMV) is different according to the trait investigated. Such complex shapes raise a major, unsolved question: is the reaction norm adaptive, i.e. shaped by natural selection, or is it only some by-product of internal constraints related to development? Numerous papers have considered the evolution of reaction norm and discussed this problem (Via & Lande, 1985, 1987; Scheiner & Lyman, 1989, 1991; De Jong, 1990, 1995; Scheiner et al., 1991; Gomulkiewicz & Kirkpatrick, 1992; Weber & Scheiner, 1992; Gavrilets & Scheiner, 1993; Scheiner, 1993a,b; Via, 1993, 1994; Van Tienderen & Koelewijn, 1994; Via et al., 1995), but experimental data remain fairly scarce.

Since size differences at the extremes of a latitudinal cline are assumed to be adaptive, we could also expect a modification of the reaction norms if the norm itself responds to selection. This problem has already been addressed for ovariole number in D. melanogaster (Delpuech et al., 1995). A major average difference was found in mean trait value between European (48 ovarioles) and Equatorial African (37 ovarioles) populations. The shapes of the reaction norms, however, were found to be quite similar, with a TMV of 22.2 °C in Europe and 22.7 °C in Africa.

In the present paper, we investigated two body-size-related traits, wing and thorax length. As an example of a tropical population of D. melanogaster we chose a Caribbean one, since these flies are among the smallest found in the world (Capy et al., 1993). Data were compared to those of a French D. melanogaster population, which was previously studied (David et al., 1994). The analysis was extended in a parallel way to the sibling D. simulans: Caribbean flies were compared to French ones (Morin et al., 1996). We confirm a major size divergence between temperate and tropical populations of D. melanogaster, and also a large difference in the wing/thorax ratio. Variations in D. simulans were much less. In the two species, significant differences were observed in the shape of the reaction norms and especially in TMVs, suggesting an adaptive response to ambient temperature.

Materials and methods

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

The Caribbean D. melanogaster population was collected in 1995 in Gosier, a locality on Guadeloupe island (latitude 16°N). D. simulans is completely absent from Guadeloupe (David & Capy, 1983). It was, however, discovered and abundantly collected in 1995 on Saint Martin island, about 200 km north of Guadeloupe (latitude 18°N). On Saint Martin, D. melanogaster was also present, but in less abundant numbers than D. simulans, possibly because of some ecological competition between the two species. We verified that the D. melanogaster from Saint Martin had a body size similar to that of the population from Guadeloupe, and we chose to study the latter since more flies were available. Wild collected adults were set in groups of 10 pairs and for each species we investigated five such groups, called isogroups (see Moreteau et al., 1995). The F1 laboratory-grown flies were used as parents in our experiments. Ten pairs of already mated flies were randomly chosen in each group, and allowed to oviposit at 25 °C, in one culture vial per group. Each vial had a volume of 28 mL, and contained around 10 mL of a high-nutrient medium based on corn and killed yeast (David & Clavel, 1965). This was largely enough to prevent future adult size from any crowding effect, since in each case we let flies oviposit for 1–3 h until egg density ranged between 100 and 200 per vial. For each species, one vial for each group was transferred to one of seven experimental temperatures, covering the thermal living range of the species, i.e. 12, 14, 17, 21, 25, 28 and 32 °C in D. melanogaster. D. simulans proved to be very difficult to rear at 32 °C so the maximum temperature was 31 °C only.

Ten adult males and 10 adult females were randomly taken for each group at each temperature, killed with ether, then preserved in conservative liquid (alcohol + glycerine + acetic acid). For measurement, flies were put in a small Petri dish, with a thin layer of liquid, so that wing and thorax length were easily observed. Left wing length was measured from the thoracic articulation to the distal tip. Thorax length was measured from the neck to the tip of the scutellum, on a left side view. Measurements were made under a binocular microscope equipped with an eyepiece micrometer. Results were expressed in mm × 100.

Data were analysed with the STATISTICA (STATISTICA, 1997) and SAS (SAS, 1985) software. The major aim of this investigation was to analyse the shape of the reaction norms. A convenient way to study the shape is to adjust the norms of each isogroup to a polynomial (Gavrilets & Scheiner, 1993; Via et al., 1995). The problem is to choose the appropriate degree of polynomial adjustment. For this purpose, slope variations, i.e. empirical derivatives of the norms, were considered (see David et al., 1990, 1994). More complex slope variation leads to higher polynomial degree. In addition, the adjusted coefficients of determination (R2) for successive polynomial degrees, as well as the plausibility of particular points calculated from the adjustments, were taken into account. Quadratic (for thorax length) and cubic (for wing length and wing/thorax ratio) polynomials were used. Calculations for a second-degree adjustment are explained below. If P(t) is the investigated phenotype at temperature t, and gi are the polynomial parameters, we have

inline image

From this we can calculate two characteristic values:

inline image

inline image

where TMV is the temperature of maximum value calculated for each concave norm by resolving P ′(t) = 0, and MV is the maximum value of the phenotype at TMV.

Data of the Caribbean populations of D. melanogaster and D. simulans were compared to those of French populations of the same species collected in 1993 in a vineyard near Bordeaux (David et al., 1994; Morin et al., 1996). The French populations were analysed with the isofemale line technique (10 lines for each species). The isogroup method was chosen in the present paper to investigate the tropical populations, since the main interest was to compare the average values and not the genetic variability within populations. The advantage of this method is that five groups provide about the same precision on mean trait values and on norm shape as 10 lines (see Moreteau et al., 1995).


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

Reaction norms of wing and thorax length

Average curves

Data of the Caribbean populations were submitted to ANOVA testing (not shown). In both species, temperature and sex explained 90% (thorax) or 95% (wing) of the total variation. A small (less than 1% of the total variation) but significant group effect demonstrated a slight heterogeneity, which might reflect genetic differences between groups, related to founder effects. The significant temperature–group interaction indicated that reaction norms of the groups were not parallel but exhibited different shapes. A slight but significant temperature–sex interaction was also observed. As usual in Drosophila, males are smaller than females and, for clarity, only curves of females are presented (Fig. 1). Comparing the reaction norms of males would lead to similar conclusions.


Figure 1. Average reaction norms of wing and thorax length of females according to growth temperature: comparison of Caribbean and French populations of D. melanogaster and D. simulans. Lengths are in mm × 100. Filled circles (Caribbean) and squares (France) indicate the position of the calculated maximum (TMV, MV). Vertical bars indicate the 95% confidence interval of the means.

Download figure to PowerPoint

For wing length, the curves are always concave with a maximum at low temperature. In D. melanogaster, a major size difference is observed between French and Caribbean populations. The curves are not parallel and the divergence seems to be more pronounced at low temperature. In D. simulans, Caribbean flies have also shorter wings, but the difference is small and quite stable with temperature.

Reaction norms of thorax length are also concave with a maximum value at a higher temperature than does wing length. In D. melanogaster, the major size difference between French and Caribbean populations is confirmed. In D. simulans, on the other hand, the response curves are almost identical. Significant differences are observed only at extreme, low or high temperatures, and in these cases the French population is bigger.

D. simulans is usually considered a smaller species than D. melanogaster (Capy et al., 1993), but Fig. 1 shows this is not as clear when geographical populations are compared. In France, D. melanogaster is much bigger than D. simulans at comparable temperatures, especially for wing length. In the West Indies, on the other hand, D. melanogaster appears to be smaller than D. simulans except at high temperatures (see Fig. 1). More precise comparisons are given in the next section.

The shape of reaction norms and characteristic values

Thorax length in Caribbean populations was adjusted to second-degree polynomials in both sexes for the two species. Adjusted R2 values were high, with a mean over the four cases of 0.92 ± 0.01, and the calculated TMVs were included within the thermal living range of the two species. Wing length variations, on the other hand, are less symmetrical than thorax length variations (see Fig. 1). An analysis of slope variation (not shown) demonstrated that wing length varied less regularly than thorax length. A quadratic adjustment was not convenient, especially for male wing. For reasons of homogeneity, all wing length norms were adjusted to a third-degree polynomial, providing a very high mean adjusted R2 (0.98 ± 0.00).

The values of the gi polynomial parameters for wing length (not shown) were found to be highly variable between groups, with average CVs of 31.7% in D. melanogaster and 36.3% in D. simulans. For thorax length variability was less, on average 8.2% in D. melanogaster and 7.6% in D. simulans, presumably due to the lower level of the polynomial. The biological significance of these parameters is, however, not obvious, and as in previous studies (Delpuech et al., 1995; Morin et al., 1996, 1997) we used these values to calculate the coordinates of the maximum size, i.e. MVs (maximum values) and TMVs.

Average MVs are given in Table 1, and illustrated in Fig. 1 for females. As expected, MVs are always larger in females than in males. In D. melanogaster, the Caribbean population exhibits much smaller MVs than the French one, both for wing and for thorax length. In D. simulans, on the other hand, the differences are less pronounced, especially for thorax length, which is similar in females. Finally, a comparison of sympatric populations shows that in the West Indies, D. simulans has generally greater MVs than its sibling species, especially for thorax length, and thus appears to be larger. The reverse is true for temperate French populations.

Table 1.  Maximum values (MVs) in mm × 100 of wing and thorax length in males and females of the geographical populations in the two species. Thumbnail image of

Average TMVs (Table 2 and Fig. 1) characterize the position of the reaction norm along the thermal axis. TMVs are higher in females than in males, higher for thorax than for wing length and higher in D. melanogaster than in D. simulans. More interestingly, in all cases higher values were observed in Caribbean flies, i.e. all differences were positive (Table 2), and three out of four were significantly higher than zero in D. melanogaster. In D. simulans, differences between French and Caribbean populations were similar to those found in D. melanogaster, but none was significantly superior to zero. The problem was further analysed with ANOVA (Table 3) for each species. In D. melanogaster all three major effects (geographical origin, sex and trait) were highly significant, as was the origin–trait interaction. In D. simulans, the major effects were also significant but to a lesser degree, owing to the greater error variance (4.08 versus 0.59 in D. melanogaster).

Table 2.  Temperatures of maximum value (TMVs) in °C of wing and thorax length in males and females of the geographical populations in the two species. Thumbnail image of
Table 3.  Results of ANOVA on TMVs in D. melanogaster and D. simulans. Thumbnail image of

It seemed important to quantify the magnitude of the differences observed between geographical populations, sexes, traits and also between the two species. Mean differences were calculated from Table 2 data and are given in Table 4. Each difference is the average of four values (see legend to Table 4) and is compared to zero. In all cases, mean differences are significant. With regard to the geographical origin, the TMVs are higher in Caribbean flies and the difference is more pronounced for the thorax (2.2 °C) than for the wing (1.0 °C). Females exhibit higher TMVs than males, and the sex difference is similar in both species, on average 1.5 °C. The TMVs for thorax length are above those for wing length, and again the two species exhibit a similar difference (on average 3.6 °C). Finally the TMVs are highly different between species, but similar for the two traits: on average, TMVs in D. simulans are 3.1 °C below those in D. melanogaster.

Table 4.  Comparisons of mean TMVs according to geographical origin, species, sex and trait. Values are given in °C. For each comparison, four differences were calculated from data in Table 2, and averaged. For instance, to compare thorax length between geographical races, differences were calculated between Caribbean and French populations, separately for males and females of the two species. Thumbnail image of

Wing/thorax correlation and ratio

Within each group of Caribbean flies, wing and thorax length were found to be highly correlated, as previously observed in isofemale lines of French populations. Phenotypic correlations, averaged over groups and temperatures, were 0.74 ± 0.02 for males and 0.77 ± 0.03 for females in D. melanogaster, and 0.83 ± 0.02 and 0.71 ± 0.05 in D. simulans. Within-line correlations were similar in French populations, with average values ranging between 0.71 and 0.77.

Previous investigations on French populations (David et al., 1994; Morin et al., 1996) showed that the wing/thorax ratio could be considered as an original trait with a specific reaction norm. The ratio was investigated in Caribbean flies and ANOVA (not shown) demonstrated that all direct effects and most double interactions were significant. While sex was one of the two main sources of variation for wing and thorax length, it had only a slight effect for the ratio (about 0.5% of the total variation). Data of males and females were thus pooled to visualize the reaction norms (Fig. 2).


Figure 2. Variations of the wing/thorax ratio (both sexes averaged) according to growth temperature in geographical populations of D. melanogaster and D. simulans. Vertical bars show the 95% confidence interval of the means.

Download figure to PowerPoint

A major difference between geographical populations of the two species is the lower ratio in the Caribbean flies. The reaction norms also seem to exhibit a quite different shape in the West Indies (decreasing parabolic curves) and in France (sigmoid norms). Finally the ratio is clearly less in D. simulans than in D. melanogaster, and the variation amplitude is higher in D. melanogaster.

For a more precise shape analysis, the reaction norms of Caribbean populations were adjusted to second- and third-degree polynomials. As with French populations, the latter degree was found to be more convenient, with an adjusted R2 value of 0.98 in both species. The norms were characterized by the value of the ratio at the inflexion point and the temperature of inflexion point (TIP), i.e. the temperature at which the second derivative of the third-degree polynomial function becomes null, and also by the coordinates of minimum value (calculated in the same way as TMV and MV for wing length). Owing to a particular curvature, two tropical D. melanogaster groups exhibited an aberrant TIP for males. We thus decided to analyse only the average curve of the wing/thorax ratio and, for reasons of homogeneity, average curves were considered in all cases (ANOVA, Table 5).

Table 5.  Results of ANOVA on the coordinates of minimum value (TmV: temperature of minimum value; mV: minimum value) and of inflexion point for the wing/thorax ratio. Average reaction norms were used for calculations. Thumbnail image of

For the four analysed characters, significant differences were observed between species and between populations. Sex had no effect, except for TIP. These differences are illustrated in Fig. 3. The minimum value of the ratio and the ratio at the inflexion point are highly positively correlated (= 0.99). Values in D. melanogaster are superior to those in D. simulans. In each species, values from Caribbean populations are less than those from French populations. In each population, male and female values are very close. Characteristic temperatures (Fig. 3) clearly differentiate the norms of the two species: D. melanogaster exhibits higher temperatures of minimum values and lower temperatures of inflexion point. Caribbean flies in both species exhibit lower characteristic temperatures: in the West Indies the reaction norms are shifted toward the left (see also Fig. 2).


Figure 3. Relationships between characteristic ratio values (left) and characteristic temperature values (right) of reaction norms of wing/thorax ratio in geographical populations of D. melanogaster and D. simulans. For each population the two points correspond to male and female data.

Download figure to PowerPoint

Sexual dimorphism

Sexual dimorphism for body size was estimated as the female/male ratio, calculated in each group at each temperature. Reaction norms of this ratio are illustrated in Fig. 4 (values for wing and thorax were similar and averaged). In all cases, the female/male ratio exhibits a significant plasticity and increases according to temperature, with a maximum at about 28 °C. A significant decrease above 28 °C is observed in D. simulans and in the Caribbean population of D. melanogaster. Males and females are more similar when reared at low temperature. Sexual dimorphism is consistently lower in D. simulans than in D. melanogaster. The two populations of D. simulans are remarkably similar. In D. melanogaster, on the other hand, a significant difference exists between Caribbean and French populations: the sex dimorphism is more pronounced in the smaller flies from Guadeloupe.


Figure 4. Variations of sexual dimorphism for body size (female/male ratio) according to growth temperature in the two geographical populations of D. melanogaster and D. simulans. Values for wing and thorax were very similar and averaged. For clarity, confidence intervals are not shown.

Download figure to PowerPoint

Discussion and conclusions

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

This paper compared the reactivity of size traits to developmental temperatures in tropical and temperate populations of two sibling species. In the West Indies, average year temperature is high, above 25 °C, and seasonal variations are low. In France (Bordeaux), the average annual temperature is much less: 13.7 °C, and 15.6 °C if we exclude the winter season. We thus expected that morphometrical differences should reflect some temperature adaptation.

Before discussing the significance of changes in reaction norms, we would like to address some possible methodological problems. Firstly, French populations were investigated with the technique of isofemale lines while Caribbean populations were analysed with isogroups, as was also the case for recent studies in D. ananassae (Morin et al., 1997) and D. subobscura (Moreteau et al., 1997). Isogroups are independent samples from a natural population. Since only slight, if any, genetic differences among them are expected, they will provide a better estimate of the mean value of the population. Also, the initial genetic sampling is bigger with isogroups: only 10 females with 10 lines, but 50 females with five groups. A detailed comparison of the two techniques was presented by Moreteau et al. (1995): significant differences in variances among groups or lines were evidenced, as expected, but no difference in mean values, which only are compared in the present paper, were noted.

A second problem is that measurements were not made simultaneously, and might be affected either by uncontrolled changes in the experimental laboratory conditions or by genetic modifications of the natural populations over time. We are confident that these two possible sources of bias are not important. We have measured body size of laboratory strains over different years and generally have found an excellent stability, meaning that neither drift nor experimental fluctuations occurred. More importantly, in D. melanogaster we have collected populations in different French localities or over successive years, and they have proved to be very stable (Capy et al., 1993; many unpublished results). In some cases, slight differences were observed, which might reflect either local adaptations or founder effects. However, the amplitude of variations remained small: in females, for example, wing length at 25 °C ranged between 2.72 and 2.80 mm. We also have several samples from Martinique, Guadeloupe and Saint Martin, collected in different years, and the range of mean wing length variation was 2.40–2.48 mm.

The third possible objection is that we sampled only two populations of each species. Of course these populations were chosen because we knew, owing to previous studies (Capy et al., 1993, 1994), that, at least for D. melanogaster, they were close to the ends of latitudinal clines and were thus expected to exhibit a major adaptive divergence. In other words, the probability of finding differences in the shape of reaction norms, if any, was maximized by choosing such populations. Nevertheless, we recognize that it should be interesting to investigate more numerous populations, along a latitudinal transect, looking for a progressive variation of reaction norm parameters. We failed, however, to find clear size differences between French and Caribbean D. simulans, and thorax dimensions, in particular, were similar. We already knew (Capy et al., 1993, 1994) that, with respect to morphometrical traits, D. simulans is far less differentiated into geographical races than its sibling. Thus, while D. simulans is usually regarded as a smaller species than D. melanogaster, in the Caribbean the reverse seems to be true.

When reaction norms were adjusted to polynomials, the gi parameters were found to be highly variable. Polynomial coefficients were thus used to calculate peculiar values (e.g. the coordinates of a maximum), for characterizing the norm shape and comparing geographical populations. The hypothesis was that, if reaction norms were able to react to natural selection in an adaptive way, different thermal regimes should modify the norms. Our data confirmed that expectation. Indeed, TMVs were clearly higher in tropical than in temperate flies, by 1.0 °C for wing length and by 2.2 °C for thorax length. In a previous study (Delpuech et al., 1995) the TMV of ovariole number in Afrotropical flies was found to be only 0.5 °C above that of French flies. Thus, at least in D. melanogaster, the latitudinal cline seems to result mainly in a difference of mean trait value for ovariole number, whereas it results in a modification of both norm shape and average value in size-related traits. The adaptive significance of such lateral shifts in the reaction norms is strongly enforced if we consider data on more distantly related species. In the tropical, cold-sensitive D. ananassae, TMVs of wing length, thorax length and ovariole number were significantly higher than in D. melanogaster (Morin et al., 1997). Reciprocally, the European, cold-adapted D. subobscura exhibited much lower TMVs than D. melanogaster for the same traits (Moreteau et al., 1997).

All these differences, similarly observed between and within species, suggest that the MVs of the investigated traits are more or less related to some functional optimum (Stearns, 1992). The position of the optimum is shifted towards higher temperature in warm-adapted flies, and towards lower temperature in cold-adapted ones. Two major problems, however, remain within this general adaptive interpretation of norm evolution. Firstly, wing and thorax length exhibit different TMVs, and MVs are observed at different temperatures for the different investigated traits in all species so far analysed, i.e. D. melanogaster (David et al., 1994; Delpuech et al., 1995), D. ananassae (Morin et al., 1997) and D. subobscura (Moreteau et al., 1997). For the three species, the same order was observed among TMVs: ovariole number > thorax length > wing length. For the moment, in the absence of any other interpretation, we may consider that such differences reflect unknown developmental constraints. The difference between TMVs of thorax and wing length is quite surprising if we consider that these two traits are positively correlated (Falconer, 1989; David et al., 1994; Barker & Krebs, 1995; Morin et al., 1996, 1997) and arise from the same imaginal disk. It remains possible, however, that natural selection acts more directly on the wing/thorax ratio than either on wing, thorax or body mass (Pétavy et al., 1997a,b).

A second interpretation difficulty concerns sexual dimorphism. In all populations and species investigated so far, males exhibit lower TMVs than females and there is no reason to believe that males are better adapted to cold than females. In Drosophila, a strong size dimorphism exists, and the present study confirmed that dimorphism increased with temperature, thus being also a plastic trait. In that case thorax and wing length provide almost identical results. We do not have any evolutionary interpretation for such observations (Charnov, 1993) and more descriptive data are needed, while this topic has already been dealt with in other phyla such as Spiders (Vollrath & Parker, 1992; Coddington et al., 1997).

The comparison of the two sibling species consistently showed that TMVs in D. melanogaster were higher than those of D. simulans. If we follow the adaptive hypothesis developed above, we should conclude that D. simulans is better adapted to cold than D. melanogaster. In favour of this hypothesis, we may recall that Caribbean D. melanogaster could be easily grown at 32 °C, which was not the case for D. simulans. Other data, however, and especially those concerning the lower wing/thorax ratio found in D. simulans, point to an opposite interpretation.

As already discussed by Morin et al. (1996), the wing/thorax ratio appears as a fitness-related trait inversely proportional to wing loading, and thus related to flight and possibly to dispersal capacity (Thomas, 1993; David et al., 1994; Barker & Krebs, 1995; Pétavy et al., 1997a,b). A higher wing loading implies a higher wing beat frequency which could be considered as an adaptation to a warmer environment. In the two species, we have found that Caribbean populations, compared to temperate ones, were characterized by a lower wing/thorax ratio and thus a higher wing loading. It might be argued that wing loading is somehow constrained by overall size. Indeed, in D. melanogaster, the wing/thorax ratio increases with body weight and latitude (see also Capy et al., 1993). Data on D. simulans are especially interesting in showing that the ratio can evolve independently of size and body weight, and thus may be a direct target of natural selection.

We have also demonstrated that the reaction norms of the ratio were modified, and the difference can be interpreted as a lateral shift to the left, i.e. toward colder temperatures in the tropical flies. A similar trend was observed in D. ananassae (Morin et al., 1997) but not in D. subobscura (Moreteau et al., 1997). For the moment it is difficult to propose an adaptive interpretation to the changes in the shape of the ratio reaction norm, and to the shift toward lower temperatures observed in tropical flies. Characteristic values of the wing/thorax ratio were highly correlated, whereas such was not the case for the characteristic temperatures (Fig. 3). These observations suggest that stronger constraints could act on the values of the wing/thorax ratio than on the shape of its reaction norm, but more extensive comparative investigations are needed.


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

This work benefited from grants from the Ministère de l’Environnement (EGPN Committee) and from the Direction Générale de la Recherche et de la Technologie (Project Environmental stress).


  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion and conclusions
  7. Acknowledgments
  8. References
  • 1
    Barker, J.S.F. & Krebs, R.A. 1995. Genetic variation and plasticity of thorax length and wing length in Drosophila aldrichi and D. buzzatii. J. Evol. Biol. 8: 689 709.
  • 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
    Capy, P., Pla, E., David, J.R. 1994. Phenotypic and genetic variability of morphometrical traits in natural populations of Drosophila melanogaster and D. simulans. II. Within-population variability. Genet. Sel. Evol. 26: 15 28.
  • 4
    Charnov, E.L. 1993. Life History Invariants. Oxford University Press, Oxford, UK.
  • 5
    Coddington, J.A., Hormiga, G., Scharff, N. 1997. Giant female or dwarf male spiders? Nature 385: 687 688.
  • 6
    David, J.R., Allemand, R., Van Herrewege, J., Cohet, Y. 1983. Ecophysiology: abiotic factors. In: The Genetics and Biology of Drosophila, Vol. 3D (M. Ashburner, H.L. Carson, & J.N. Thompson, eds), pp. 105–170. Academic Press, London, UK.
  • 7
    David, J.R. & Bocquet, C. 1975a. Evolution in a cosmopolitan species: genetic latitudinal clines in Drosophila melanogaster wild populations. Experientia 31: 164 166.
  • 8
    David, J.R. & Bocquet, C. 1975b. Similarities and differences in latitudinal adaptation of two Drosophila sibling species. Nature 257: 588 590.
  • 9
    David, J.R. & Capy, P. 1983. Drosophila community in domestic habitats of Martinique island, and some specialized breeding sites of native species. Acta Oecol., Oecol. Gener. 4: 265 270.
  • 10
    David, J.R., Capy, P., Gauthier, J.P. 1990. Abdominal pigmentation and growth temperature in Drosophila melanogaster: similarities and differences in the norms of reaction of successive segments. J. Evol. Biol. 3: 429 445.
  • 11
    David, J.R. & Clavel, M.F. 1965. Interaction entre le génotype et le milieu d’élevage. Conséquences sur les caractéristiques du développement de la Drosophile. Bull. Biol. Fr. Belg. 99: 369 378.
  • 12
    David, J.R. & Kitagawa, O. 1982. Possible similarities in ethanol tolerance and latitudinal variations between Drosophila virilis and D. melanogaster. Jap. J. Genet. 57: 89 95.
  • 13
    David, J.R., Moreteau, B., Gauthier, J.P., Pétavy, G., Stockel, J., 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.
  • 14
    De Jong, G. 1990. Genotype-by–environment interaction and the genetic covariance between environments: multilocus genetics. Genetica 81: 171 177.
  • 15
    De Jong, G. 1995. Phenotypic plasticity as a product of selection in a variable environment. Am. Nat. 145: 493 512.
  • 16
    Delpuech, J.M., Moreteau, B., Chiche, J., Pla, E., Vouidibio, J., David, J.R. 1995. Phenotypic plasticity and reaction norms in temperate and tropical populations of Drosophila melanogaster: ovarian size and developmental temperature. Evolution 49: 670 675.
  • 17
    Falconer, D.S. 1989. Introduction to Quantitative Genetics, 3rd edn. Longman, New York, USA.
  • 18
    Gavrilets, S. & Scheiner, S.M. 1993. The genetics of phenotypic plasticity. V. Evolution of reaction norm shape. J. Evol. Biol. 6: 31 48.
  • 19
    Gomulkiewicz, R. & Kirkpatrick, M. 1992. Quantitative genetics and the evolution of reaction norms. Evolution 46: 390 411.
  • 20
    Misra, R.K. & Reeve, E.C.R. 1964. Clines in body dimensions in populations of Drosophila subobscura. Genet. Res. 5: 240 256.
  • 21
    Moreteau, B., Capy, P., Alonso-Moraga, A., Munoz-Serrano, A., Stockel, J., David, J.R. 1995. Genetic characterization of geographic populations using morphometrical traits in Drosophila melanogaster: isogroups versus isofemale lines. Genetica 96: 207 215.
  • 22
    Moreteau, B., Morin, J.P., Gibert, P., Pétavy, G., Pla, E., David, J.R. 1997. Evolutionary changes of nonlinear reaction norms according to thermal adaptation: a comparison of two Drosophila species. C.R. Acad. Sci. Paris 320: 833 841.
  • 23
    Morin, J.P., Moreteau, B., Pétavy, G., Imasheva, A.G., David, J.R. 1996. Body size and developmental temperature in Drosophila simulans: comparison of reaction norms with sympatric Drosophila melanogaster. Genet. Sel. Evol. 28: 415 436.
  • 24
    Morin, J.P., Moreteau, B., Pétavy, G., Parkash, R., David, J.R. 1997. Reaction norms of morphological traits in Drosophila: adaptive shape changes in a stenotherm circumtropical species? Evolution 51: 1140 1148.
  • 25
    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.
  • 26
    Partridge, L., Barrie, B., Fowler, K., French, V. 1994. Evolution and development of body size and cell size in Drosophila melanogaster in response to temperature. Evolution 48: 1269 1276.
  • 27
    Pétavy, G., David et, J.R., Moreteau, B. 1997a. Les variations induites par la température sur les paramètres de taille et du vol des Drosophiles sont-elles adaptatives? Bull. Soc. Zool. Fr. 122 : 13 20.
  • 28
    Pétavy, G., Morin, J.P., Moreteau, B., David, J.R. 1997b. Growth temperature and phenotypic plasticity in two Drosophila sibling species: probable adaptive changes in flight capacities. J. Evol. Biol. 10: 875 887.
  • 29
    Powell, J.R. 1974. Temperature related genetic divergence in Drosophila size. J. Hered. 65: 257 258.
  • 30
    Prevosti, A. 1955. Geographical variability in quantitative traits in populations of D. subobscura. Cold Spring Harb, Symp. Quant. Biol. 20: 294 299.
  • 31
    SAS INSTITUTE. 1985. SAS User's Guide: Statistics. Vers. 5. SAS Institute Inc., Cary, NC, USA.
  • 32
    Scheiner, S.M. 1993a. Genetics and evolution of phenotypic plasticity. Ann. Rev. Ecol. Syst. 24: 35 68.
  • 33
    Scheiner, S.M. 1993b. Plasticity as a selectable trait: reply to Via. Am. Nat. 142: 371 373.
  • 34
    Scheiner, S.M., Caplan, R.L., Lyman, R.F. 1991. The genetics of phenotypic plasticity. III. Genetic correlations and fluctuating asymmetries. J. Evol. Biol. 4: 51 68.
  • 35
    Scheiner, S.M. & Lyman, R.F. 1989. The genetics of phenotypic plasticity. I. Heritability. J. Evol. Biol. 2: 95 107.
  • 36
    Scheiner, S.M. & Lyman, R.F. 1991. The genetics of phenotypic plasticity. II. Response to selection. J. Evol. Biol. 4: 23 50.
  • 37
    Stalker, H.D. & Carson, H.L. 1947. Morphological variation in natural populations of Drosophila robusta Sturtevant. Evolution 1: 237 248.
  • 38
    Stalker, H.D. & Carson, H.L. 1948. An altitudinal transect of Drosophila robusta Sturtevant. Evolution 2: 295 305.
  • 39
    STATISTICA. 1997. Statistica. Rel. 5.1. Statistica Statsoft Inc., Tulsa, OK, USA.
  • 40
    Stearns, S.C. 1992. The Evolution of Life Histories. Oxford University Press, Oxford, UK.
  • 41
    Tantawy, A.O. & Mallah, G.S. 1961. Studies on natural populations of Drosophila. I. Heat resistance and geographical variation in D. melanogaster and D. simulans. Evolution 15: 1 14.
  • 42
    Thomas, R.H. 1993. Ecology of body size in Drosophila buzzatii. Untangling the effects of temperature and nutrition. Ecol. Entomol. 18: 84 90.
  • 43
    Van Tienderen, P.H. & Koelewijn, H.P. 1994. Selection on reaction norms, genetic correlations and constraints. Genet. Res. 64: 115 125.
  • 44
    Via, S. 1993. Adaptive phenotypic plasticity: target or by-product of selection in a variable environment? Am. Nat. 142: 352 365.
  • 45
    Via, S. 1994. The evolution of phenotypic plasticity: what do we really know? In: Ecological Genetics (L. A. Real, ed.), pp. 35–57. Princeton University Press, Princeton, NJ, USA.
  • 46
    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.
  • 47
    Via, S. & Lande, R. 1985. Genotype–environment interaction and the evolution of phenotypic plasticity. Evolution 39: 505 522.
  • 48
    Via, S. & Lande, R. 1987. Evolution of genetic variability in a spatially heterogeneous environment: effects of genotype–environment interaction. Genet. Res. 49: 147 156.
  • 49
    Vollrath, F. & Parker, G.A. 1992. Sexual dimorphism and distorted sex ratios in spiders. Nature 360: 156 159.
  • 50
    Weber, S.L. & Scheiner, S.M. 1992. The genetics of phenotypic plasticity. IV. Chromosomal localization. J. Evol. Biol. 5: 109 120.