PLASTICITY VERSUS ENVIRONMENTAL CANALIZATION: POPULATION DIFFERENCES IN THERMAL RESPONSES ALONG A LATITUDINAL GRADIENT IN DROSOPHILA SERRATA

O R I G I NA L A RT I C L E doi:10.1111/j.1558-5646.2009.00683.x

PLASTICITY VERSUS ENVIRONMENTAL CANALIZATION: POPULATION DIFFERENCES IN THERMAL RESPONSES ALONG A LATITUDINAL GRADIENT IN DROSOPHILA SERRATA Maartje Liefting,1,2 Ary A. Hoffmann,3,4 and Jacintha Ellers1,5 1

Department of Ecological Sciences, Section of Animal Ecology, VU University Amsterdam, De Boelelaan 1085, 1081 HV,

Amsterdam, The Netherlands 2 3

E-mail: [email protected]

Centre for Environmental Stress and Adaptation Research, Department of Genetics, Bio21 Institute, The University of

Melbourne, 30 Flemington Road, Parkville VIC 3010 Australia 4

E-mail: [email protected]

5

E-mail: [email protected]

Received August 31, 2008 Accepted February 23, 2009 The phenotypic plasticity of traits, defined as the ability of a genotype to express different phenotypic values of the trait across a range of environments, can vary between habitats depending on levels of temporal and spatial heterogeneity. Other traits can be insensitive to environmental perturbations and show environmental canalization. We tested levels of phenotypic plasticity in diverse Drosophila serrata populations along a latitudinal cline ranging from a temperate, variable climate to a tropical, stable climate by measuring developmental rate and size-related traits at three temperatures (16◦ C, 22◦ C, and 28◦ C). We then compared the slopes of the thermal reaction norms among populations. The 16–22◦ C part of the reaction norms for developmental rate was flatter (more canalized) for the temperate populations than for the tropical populations. However, slopes for the reaction norms of the two morphological traits (wing size, wing:thorax ratio), were steeper (more plastic) in the temperate versus the tropical populations over the entire thermal range. The different latitudinal patterns in plasticity for developmental rate and the morphological traits may reflect contrasting selection pressures along the tropical–temperate thermal gradient.

KEY WORDS:

Countergradient variation, developmental rate, phenotypic plasticity, reaction norm, temperature, wing centroid

size, wing:thorax ratio.

Organisms have evolved various strategies to cope with environmental variation, including genetic adaptation and phenotypic plasticity (Lynch and Gabriel 1987; Stearns 1989; Meyers and Bull 2002). Phenotypic plasticity, the ability of one genotype to express different phenotypes in different environments (Bradshaw 1965), can be measured as the slope of a reaction norm (de Jong 1990). Plasticity can be adaptive if it increases an organism’s fitness under a particular circumstance compared to organisms that  C

1954

are not plastic. Plasticity can also be nonadaptive if it results from a passive response to environmental conditions. A large number of studies have demonstrated that there is genetic variation in the slope of reaction norms (Gutteling et al. 2007; van Asch et al. 2007; Winterhalter and Mousseau 2007). Therefore, the level of phenotypic plasticity in a trait is expected to be under selection and dependent on the environmental regime encountered (Scheiner 1993; Price et al. 2003). Indeed, empirical studies

C 2009 The Society for the Study of Evolution. 2009 The Author(s). Journal compilation  Evolution 63-8: 1954–1963

O P P O S I T E L AT I T U D I NA L C L I N E S I N P L A S T I C I T Y

have shown variation in the slope of reaction norm between habitats with different levels of temporal and spatial heterogeneity (Ellers and van Alphen 1997; Pfennig and Murphy 2002; Hassall et al. 2005; Winterhalter and Mousseau 2007; Liefting and Ellers 2008). Understanding how environmental variation may shape plasticity of a trait is not straightforward because the adaptive value of plasticity is different for fitness traits and nonfitness traits (sensu Richards et al. 2006). For many traits that are not directly related to fitness, such as morphological or physiological traits, predictable environmental variation is known to select for increased phenotypic plasticity (Bradshaw 1965; de Jong 1995; Roff 2002). On the other hand, fitness traits such as reproduction and survival should be stable despite environmental change because this will enable organisms to maintain high fitness levels across environments. Fitness traits are therefore expected to show environmental canalization in variable environments, that is to show insensitivity of traits to environmental perturbations (Stearns et al. 1995; Wagner et al. 1997; Debat and David 2001), possibly through compensatory evolution (countergradient variation) (Conover and Schultz 1995). In more stable environments, however, the absence of compensatory evolution in fitness traits may result in higher plasticity levels (Conover and Schultz 1995), although canalization and plasticity are not mutually exclusive (Stearns and Kawecki 1994). Making the distinction between fitness traits and nonfitness traits may not always be straightforward (Nylin 1992), but this is important when we try to understand differences in levels of plasticity across environments. Unfortunately, an explicit test of these predictions has been lacking so far, because very few studies compare the plasticity levels of fitness and nonfitness traits under different regimes of environmental variation. Temperature variation provides a good framework for testing predictions on the evolution of phenotypic plasticity because many traits are highly sensitive to changes in thermal conditions. Particularly in ectotherms, temperature is a key environmental factor that affects a large number of morphological traits as well as fitness traits (Roff 2002). The typical thermal reaction norm shows an increase in trait value with temperature until an optimum is reached, after which the performance of the animal quickly declines (Huey and Kingsolver 1989; Angilletta et al. 2002). The slope of thermal reaction norms shows sufficient genetic variation between individuals and between species to evolve in response to variation in thermal regime (Huey and Kingsolver 1989; Gilchrist 1995; Driessen et al. 2007). Evidence for habitat-specific thermal reaction norms comes mostly from studies comparing populations from contrasting thermal regimes (Cohet et al. 1980; Noach et al. 1996; Morin et al. 1999; Trotta et al. 2006; Liefting and Ellers 2008). Only a few studies compare the strength of phenotypic plasticity along a latitudinal cline (Robinson and Partridge 2001; Gilchrist and Huey 2004) even though clinal data would provide

a strong test of the predicted changes in the slope of reaction norms. Here we examine patterns in plasticity and environmental canalization of Drosophila serrata populations along a natural gradient in response to their local thermal regime. The gradient ranges from a temperate climate with cold and warm periods to a tropical climate with a stable warm environment. Drosophila serrata is an endemic Australian species distributed along the eastern coast, and is found in native vegetation rather than anthropogenic habitats (Jenkins and Hoffmann 2000). This species shows clinal variation in stress resistance (Hallas et al. 2002; Magiafoglou et al. 2002), development time (Sgr`o and Blows 2003), and morphological traits (Hallas et al. 2002) as well as in traits associated with sexual selection (Blows and Higgie 2002). Development time decreases toward tropical areas (Sgr`o and Blows 2003), similarly body size and wing length decrease in a nonlinear manner toward the tropics (Hallas et al. 2002). Differences in mean development time and size among D. serrata populations have an additive genetic basis (Schiffer et al. 2006). We tested the level of phenotypic plasticity in a fitness trait and two morphological traits. Traits were measured over three temperatures in the laboratory and we compared the slopes of the thermal reaction norms among populations. The fitness trait was developmental rate, whereas morphological traits involved wing size and thorax length and the ratio between these two parameters (referred to as wing:thorax ratio). Wing size is an indicator for body size, whereas the ratio between wing and thorax is related to wing loading (body weight per wing area) and flight capacity (David et al. 1994; Morin et al. 1999; Gilchrist and Huey 2004). We expect developmental rate to show the lowest levels of plasticity in the temperate populations, due to environmental canalization of this fitness trait in such a variable thermal regime. In the tropical populations, we expect higher levels of plasticity in developmental rate because of lack of compensatory evolution in a thermally more stable regime. The levels of plasticity for the morphological traits are expected to show the opposite pattern, with steeper slopes in temperate populations and more flat reaction norms in tropical populations.

Materials and Methods STOCKS AND EXPERIMENTAL SET UP

The distribution of D. serrata in eastern Australia ranges from just below Sydney up to Cairns. Populations were sampled in 2006 between Wollongong (34◦ 20 S) in New South Wales and Townsville (19◦ 22 S) in Queensland (Fig. 1). Flies were collected via sweep netting over buckets with fermenting banana at three collection sites per location with a total of eight locations along the transect. Field inseminated females were set up individually in vials in the laboratory. Up to 10 isofemale lines were

EVOLUTION AUGUST 2009

1955

M A A RT J E L I E F T I N G E T A L .

Figure 1.

Map of Australia showing the collection sites for the D.

serrata populations.

maintained for four to five weeks at 25◦ C under continuous light on a dead yeast–sucrose–agar medium at a population size of 50–100 individuals per line. To establish mass-bred populations for each location along the gradient, equal numbers of secondgeneration progeny from 10 isofemale lines were combined and maintained at a population size of 500–1000 flies. REACTION NORM EXPERIMENT

From each mass-bred population representing one of the eight locations along the cline, 10 groups of 10 nonvirgin females were collected and maintained at 25◦ C under continuous light. From each group of females (subsequently referred to as a subpopulation), eggs were collected by placing a plastic spoon containing a treacle–yeast–agar medium into the vial and allowing the females to lay eggs for 12 h. The exact number of males and females contributing to the eggs of each subpopulation was not determined, but because females of D. serrata are known to have high frequency of multiple mating, it is likely that their offspring are sired by different fathers (Frentiu and Chenoweth 2008). Eggs from each subpopulation were distributed over nine vials with 10 eggs each and the vials were placed randomly at 16◦ C, 22◦ C, and 28◦ C to develop (three replicate vials per temperature).Vials were randomized daily within each incubator. Vials were checked for emerging flies every 6 h at 28◦ C, 8 h at 22◦ C, and 12 h at 16◦ C to measure development time. Developmental rate was calculated as the inverse of the number of days it took for a fly to eclose (development time−1 ) and was averaged per vial. Viability was also scored as the percentage of hatched eggs per vial. Analyses were done on these vial means. The relationship between developmental rate and temperature was close to linear, yet the slopes of the lower part of the reaction norm were clearly steeper than the slopes from the upper part (paired t-test; t = 9.815, df = 7, P < 0.001). The magnitude of directional selection can vary for 1956

EVOLUTION AUGUST 2009

different parts of the reaction norm (Kingsolver et al. 2007) and small deviations from linearity in reaction norms can have profound effects on overall performance (Kingsolver et al. 2004). We therefore chose to calculate the slope of the reaction norm over the whole temperature range, that is 16–28◦ C, and separately over the low (16–22◦ C) and high (22–28◦ C) part of the thermal range. The slope of the reaction norm for developmental rate was calculated at the subpopulation level without separating the sexes. In addition to the slope, a reaction norm can also differ in its mean response, which is described by elevation. These aspects of the reaction norm were analyzed separately. Two morphological traits were measured, thorax length and wing centroid size. Thorax length was not included as a separate size factor in the analyses because it produced similar patterns to wing size (results not shown). The wing:thorax ratio of flies was instead computed as it is thought to be related to flight ability and to be positively correlated to the distance flies move in release experiments (Hoffmann et al. 2007). From the flies that eclosed and that were used to determine developmental rate, 10 female and 10 male flies were randomly selected per location (evenly drawn from the several subpopulations per location). This was done at each of the three temperatures, resulting in a total of 480 measured flies. The response of size to the different temperatures also gives a reaction norm that could then be calculated per location. In this case, the lower and upper parts of the reaction norm did not differ in slope, and we calculated the slope of the reaction norm over the entire thermal range.

WING ANALYSES

Thorax length was measured as the distance from the anterior point of the thorax to the posterior tip of the scutellum (James et al. 1997; Hoffmann et al. 2007). Measurements were taken to the nearest 0.02 mm with an eyepiece graticule under a dissection microscope (Leica MZ 12.5, Leica, Wetzler, Germany) at 40× magnification. We also examined centroid size (the square root of the sum of the squared interlandmark distances), which provides an overall measure of wing size (Hoffmann and Shirriffs 2002). To determine wing centroid size, the left wing was removed and mounted on a microscope slide and images were taken with a digital camera (Nikon DS-M5-L1, Nikon, Natori, Japan) mounted on a dissection microscope (Leica MZ FLIII, Leica). A centroid value was then obtained after landmarks were placed on the images with tpsDig version 1.11 written by F. James Rohlf. Eight landmarks were used (Fig. 2). The wing:thorax size ratio was computed for each fly. To check for measurement error, repeatability for the measurements was estimated by taking multiple measures for both thorax and wing centroid size. Repeat measures were found to be highly correlated (r > 0.96, N = 50 for thorax and r > 0.99, N = 50 for centroid size) for both sexes in all cases.

O P P O S I T E L AT I T U D I NA L C L I N E S I N P L A S T I C I T Y

Results of the ANCOVA with effects of location (along the gradient) and rearing temperature on developmental rate (de-

Table 1.

velopment time−1 ), with viability of the eggs as a covariate. + Error df = 2.91 for location, 588 for other terms. ∗∗∗ P < 0.001 ∗∗ P < 0.01 ∗P

< 0.05.

df

Figure 2.

Wing of D. serrata with the positions of the eight land-

marks.

MS

F+

Location 7 0.028 6.01∗∗∗ Subpopulation (within location) 72 0.005 2.29∗∗∗ Temperature 2 59.483 29417.37∗∗∗ Temperature × location 14 0.004 1.84∗ Viability 1 0.018 8.75∗∗ Error 588 0.002

STATISTICAL ANALYSES

To test for differences in developmental rate among locations, a nested analysis of covariance (ANCOVA) was performed in STATISTICA 7.0 (StatSoft, Tulsa, OK). The model included location (the eight locations along the gradient) and temperature (at which eggs developed) as fixed factors, subpopulation nested within location as random factor, and viability (percentage of hatched eggs, calculated per vial) as a covariate. Subpopulation refers to the set of males and females that produced the offspring on one spoon. In addition, correlations were performed between developmental rate and viability per temperature. Developmental rate was ln transformed to meet normality assumptions of ANCOVA; no other variables were needed to be transformed. The morphological measurements on the wing and thorax were also analyzed using ANCOVA in SPSS 15.0 (SPSS, Chicago, IL). The model included location, temperature, and sex as fixed factors. Again, viability was included as a covariate. Latitudinal patterns in slope and elevation of the reaction norms for each trait were tested with regression analysis in SPSS 15.0 with latitude as a continuous factor. Latitude (in degrees South) had a tight relation with average annual temperature (b = −2.70 ± 0.66, P < 0.001), annual average high temperature (b = −0.66 ± 0.25, P = 0.011), annual average low temperature (b = 1.42 ± 0.49, P = 0.005), and annual average precipitation (b = −0.07 ± 0.01, P < 0.001). Climatic data for these regressions were obtained from Climatic Averages Australia (www.bom.gov.au). Recognizing that latitude reflects different climatic variables, we used latitude in further regression analyses.

Results DEVELOPMENTAL RATE

All parameters in the ANCOVA had a significant effect on developmental rate (Table 1). The covariate viability (defined as percentage of eggs hatched) was included in the analysis of de-

velopmental rate as a covariate to correct for possible confounding effects. Viability and developmental rate were correlated across temperature, but there was no correlation between developmental rate and viability within temperature (r = 0.238, N = 8, P = 0.570 for 16◦ C, r = 0.762, N = 8, P = 0.280 for 22◦ C, and r = 0.515, N = 8, P = 0.192 for 28◦ C). As expected, developmental rate was faster at higher temperatures (Figs. 3 and 4). Location also had a significant effect on developmental rate, as well as on subpopulation, which refers to the set of eggs picked from one spoon (from different parents) that were divided over the temperatures. This indicates that there is variation in developmental rate between locations and subpopulations irrespective of rearing temperature (Fig. 3). For example, flies from the location Tewantin at 26◦ 23 S develop relatively slow at all temperatures compared to flies from the other locations. The interaction between location and temperature also had a significant effect on developmental rate, so flies from the locations respond differently to changes in temperature. No correlation was found between the slope of the thermal reaction norm for developmental rate and latitude when the entire temperature range (16–28◦ C) was considered (r = −0.619, N = 8, P = 0.102). Between 16◦ C and 22◦ C, the mean slope was negatively correlated with latitude (r = −0.738, N = 8, P = 0.037, Fig. 5A), so that the southern temperate populations had flatter norms of reaction than the northern tropical populations. There was no such relation between slope and latitude over the higher (22–28◦ C) part of the reaction norm (r = 0.238, N = 8, P = 0.570, Fig. 5B). Flies from the Port Macquarie (31◦ 27 S) location had relatively steep slopes over both the 16–22◦ C and 22–28◦ C parts of the thermal range and deviated from the observed pattern. To test for differences in elevation between the reaction norms, mean values of the reaction norm were correlated with latitude. No relation between mean reaction norm and latitude was found (16–28◦ C range: r = −0.095, N = 8, P = 0.824, 16– 22◦ C range: r = −0.376, N = 8, P = 0.359, and 22–28◦ C range: r = −0.275, N = 8, P = 0.511). EVOLUTION AUGUST 2009

1957

M A A RT J E L I E F T I N G E T A L .

16 °C developmental rate (day-1)

0.037

0.108

0.074 0.037

0.105

0.072 0.036

0.102

0.070

0.036

0.099 18

22

26

30

34

1.04

centroid size (mm)

0.111

0.076

0.035

1.00

0.068 18

22

26

30

18

34

0.96

0.84

0.92

0.80

0.88

0.76

0.84

0.72

0.80

0.68

22

26

30

34

18

22

26

30

34

18

22

26

30

34

0.96

0.92

0.88

0.76 18

22

26

30

34

1.00

wing : thorax ratio

28 °C

22 °C

0.038

0.64 18

22

26

30

34

0.86

0.90

0.84

0.98

0.88 0.82

0.96

0.86 0.80

0.94

0.84

0.92

0.90

0.78

0.76

0.82 18

22

26

30

34

18

22

latitude °S

26

latitude °S

30

34

latitude °S

Developmental rate, wing centroid size, and wing:thorax ratio plotted against latitude. For developmental rate, each point represents the mean of 10 families per location (±SE). For wing centroid size and wing:thorax ratio, each point represents the mean of 8–10 males (◦) and females (•) per location (±SE).

Figure 3.

SIZE TRAITS

Wing centroid size and the wing:thorax ratio showed significant effects of temperature with larger flies at lower temperatures, and a significant effect of sex (Table 2). Location had a significant effect on both wing centroid size and wing:thorax ratio, with smaller flies and smaller ratios toward the tropics (Fig. 3). There was a significant interaction between location and temperature, hence populations from the geographic locations respond differentially to changes in temperature. The relation between wing centroid size and temperature is best described by a linear model (b = −17.25 ± 0.77, P < 0.001 and r2 = 0.958 for females with N = 24, b = −17.54 ± 0.71, P < 0.001 and r2 = 0.965 for males with N = 24, Fig. 6A) with smaller wings at higher temperatures. Also the relation of wing:thorax ratio with temperature was linear for both sexes (b = −0.27 ± 0.02, P < 0.001 and r2 = 0.918 for females with 1958

EVOLUTION AUGUST 2009

N = 24, b = −0.27 ± 0.01, P < 0.001 and r2 = 0.970 for males with N = 24, Fig. 6B) with lower ratios at higher temperatures. The regression analysis showed a strong negative relationship between the mean negative slope of the response of wing centroid size to temperature with latitude (b = −0.13 ± 0.06, P = 0.034 and r2 = 0.312 with N = 16) independent of sex (b = −0.29 ± 0.58, P = 0.626, Fig. 7A). Flies from Tannum Sands (23◦ 57 S) do not fit the linear pattern, as this population demonstrates a very steep mean slope when compared to the typical flat slopes of the other tropical populations. The regression analysis also showed a strong negative relationship for the association between mean wing:thorax ratio slope with latitude (b = −0.006 ± 0.002, P = 0.008 and r2 = 0.428 with N = 16, Fig. 7B) and again no differences between the sexes (b = 0.004 ± 0.020, P = 0.857). This means that slopes for both the reaction norms are steeper (sensitivity is higher) in the temperate and more variable regions.

O P P O S I T E L AT I T U D I NA L C L I N E S I N P L A S T I C I T Y

developmental rate (days-1)

0,12 KA KC KF KH KK KM KP SA

0,10

0,08

0,06

0,04

0,02 16

22

28

temperature (°C)

Reaction norms of developmental rate across three

Figure 4.

temperatures. Each point represents the mean value per location (± SE, falls within marker). The symbols refer to the sampled locations, from South to North; KA, Wollongong; KC, Terrigal; KF, Port Macquarie; KH, Byron Bay; KK, Tewantin; KM, Tannum Sands; KP, Finch Hatton; SA, Townsville.

Discussion We studied the relationship between phenotypic plasticity and environmental variation in a fitness trait and traits indirectly related to fitness. Theory predicts that increasing environmental variation can select for a higher level of phenotypic plasticity in morphological traits (Bradshaw 1965; de Jong 1995), but for environmental canalization in fitness traits (Stearns and Kawecki 1994; Wagner et al. 1997). These predictions are supported by our two main findings. First, populations indeed exhibited different levels of

A

thermal sensitivity of traits depending on the thermal regime of origin. We found clinal variation in the level of phenotypic plasticity for developmental rate and size traits, suggesting that the populations had evolved in response to the more stable thermal environment in the tropics and the thermally variable environment in the temperate regions. Second, opposing clines in the degree of phenotypic plasticity for morphological and fitness traits were observed (Figs. 5 and 7). Therefore, evolution of phenotypic plasticity is not only dependent on environmental variation but also on the trait under consideration. Developmental rate of D. serrata at high temperatures was higher in tropical populations than in temperate populations (Fig. 3), consistent with the results of a previous study on this species along a similar latitudinal cline (Sgr`o and Blows 2003). However, at low temperatures, flies from tropical populations showed a much larger reduction in developmental rate than flies from temperate populations. In the temperate populations, compensatory evolution of developmental rate most likely underlies the observed environmental canalization (countergradient variation). Tropical flies lack compensatory evolution because they hardly experience colder temperatures; hence the reaction norm for developmental rate of these flies is much steeper. Compensatory evolution in variable environments is important because it allows ectotherms to carry out their life cycle under highly variable conditions and short growing seasons (Conover and Schultz 1995; Conover et al. 1997). This pattern is consistent with other studies. In a comparative study on the plasticity of life-history traits of tropical and temperate D. melanogaster populations, productivity of tropical flies was higher in a warm environment but lower in a cold environment when compared to temperate flies (Trotta et al. 2006). Also, in the widely distributed springtail species

B 0.0058

mean slope dev. rate reaction norm 22-28°C

mean slope dev. rate reaction norm 16-22°C

0.0066

0.0064

0.0062

0.0060

0.0058

0.0056

0.0056

0.0054

0.0052

0.0050

0.0048

0.0046

0.0044 18

20

22

24

26

28

latitude °S Figure 5.

30

32

34

36

18

20

22

24

26

28

30

32

34

36

latitude °S

The mean slopes of the developmental rate reaction norm for (A) 16–22◦ C and (B) 22–28◦ C across the latitudinal range. Each

point represents the mean of the 10 families per location (±SE). The linear regression of mean slope on latitude was significant only for 16–22◦ C.

EVOLUTION AUGUST 2009

1959

M A A RT J E L I E F T I N G E T A L .

Results of the ANCOVA with the effects of location (along the gradient) and rearing temperature on wing centroid size and wing:thorax ratio, with viability of the eggs as a covariate. ∗∗∗ P < 0.001 ∗∗ P < 0.01.

Table 2.

Wing centroid size df Sex Temperature Location Viability Temperature × location Temperature × sex

error df

1 2 7 1 14 2

wing:thorax ratio MS

20 18.63 14.46 20 20 20

F

0.41 0.061 0.002 1.09 · 10−4 1.86 · 10−4 6.03 · 10−5

∗∗∗

1142.96 523.85∗∗∗ 9.97∗∗∗ 3.04 5.20∗∗∗ 1.68

A

df

error df

MS

F

1 2 7 1 14 2

20 18.74 14.46 20 20 20

0.001 0.04 2.25 · 10−4 9.41 · 10−5 4.16 · 10−4 2.52 · 10−4

10.63∗∗ 138.78∗∗∗ 0.58 1.15 5.07∗∗ 3.08

B

1.05

1.00

0.95

0.95

wing:thorax ratio

wing centroid size (mm)

1.00

0.90 0.85 0.80

0.90

0.85

0.75 female male

0.70

0.80

0.65

female male

0.75

16 16

22 22

28 28

16 16

22 22

temperature (°C) Figure 6.

28 28

temperature (°C)

Reaction norms of (A) wing centroid size and (B) wing:thorax ratio across temperature. Each point represents the mean value

per location (separate for males [◦] and females [•]), based on circa 10 flies. For reasons of clarity, locations are not indicated with separate symbols.

A

B -0.40

mean slope wing:thorax ratio reaction norm

mean slope centroid size reaction norm

-20 female male

-18

-16

-14

18

20

22

24

26

28

30

latitude °S

32

34

36

female male

-0.35

-0.30

-0.25

-0.20

-0.15

-0.10 18

20

22

24

26

28

30

32

34

36

latitude °S

Figure 7. The mean slope of the wing centroid size reaction norm for temperature (A) and the wing:thorax ratio reaction norm for temperature (B) plotted against latitude. Each point represents the mean of 8–10 males (◦) and females (•) per location (±SE). The linear regression of mean slope on latitude is significant for both sexes in both traits. Note that the y-axis is reversed for visual interpretation,

a higher negative value for mean slope translates in a stronger response of the original negative reaction norm.

1960

EVOLUTION AUGUST 2009

O P P O S I T E L AT I T U D I NA L C L I N E S I N P L A S T I C I T Y

Orchesella cincta, thermal reaction norms for growth rate showed a greater degree of environmental canalization in habitats with high thermal variation (Liefting and Ellers 2008). We only found this effect for the 16–22◦ C part of the reaction norm. Most Drosophila species perform best at 25–27◦ C (Delpuech et al. 1995), and this range of temperatures was experienced by all of our D. serrata populations. In contrast, the lower temperatures (and selection for performance under these temperatures) are only experienced by the temperate populations. Geographical patterns for cold shock resistance in D. serrata are consistent with the notion that there is only selection for responses to the lower thermal range at high latitudes (Jenkins and Hoffmann 2000; Hallas et al. 2002). Assuming that there is a cost involved in keeping traits plastic (DeWitt et al. 1998; Meril¨a et al. 2004), selection will also quickly remove plasticity in traits when there is no benefit of high levels of plasticity. Wing size of flies decreased toward the tropics (Fig. 3), a trend frequently observed in previous work on other Drosophila species (James and Partridge 1995; Azevedo et al. 1998) and D. serrata (Hallas et al. 2002). We found that D. serrata decreased in size in a nonlinear pattern toward the tropics, with a sharp decrease at latitudes of 20◦ S, which is consistent with patterns found earlier in D. serrata (Hallas et al. 2002). Size also showed a phenotypic response to temperature, with different levels of plasticity over the latitudinal gradient. The slopes of the reaction norms for wing size and wing:thorax ratio increased toward the temperate regions (Fig. 7), which is opposite to the trend observed in developmental rate slopes (Fig. 5). In other words, flies from temperate populations are more flexible in adjusting their size traits to environmental temperature than flies from the tropical populations for which size is more canalized. Do these patterns of morphological reaction norms make sense in terms of changing thermal conditions along a gradient? The relative size of wings and body have functional significance: increased wing area relative to body mass (i.e., decreased wing loading) will facilitate flight performance (Gilchrist and Huey 2004). The ratio between wing size and body weight is therefore expected to be the direct target of natural selection (David et al. 1994). Low wing loading is thought to be particularly important under cool conditions, when wing beat frequency decreases (Unwin and Corbet 1984) and perhaps when resources are more scattered (Hoffmann et al. 2007). Flies from temperate populations demonstrated a strong plastic response in wing:thorax ratio. They emerged with relatively larger wings (high ratio) at low temperatures versus relatively smaller wings (smaller ratio) at high temperatures. When considering foraging and dispersal behavior under cool conditions, a larger wing:thorax ratio is thought to be adaptive because the longer wings favored at low temperatures will assist in dispersal (Gilchrist and Huey 2004; Hoffmann et al. 2007; Frazier et al. 2008). In fact, the thermal response of size in

temperate populations paralleled the genetic cline in size along the latitudinal gradient (Fig. 1). Most likely this is due to similar selection pressures on wing loading acting locally as well as over a larger scale (Gilchrist and Huey 2004). These latitudinal patterns in the degree of plasticity for the morphological traits therefore corresponded to the gradient in environmental variation (co-gradient variation). There are a number of outliers that do not fit the general latitudinal patterns. For example, the Tannum Sands population stands out in Figure 7B representing plastic responses in the wing:thorax ratio, whereas the Port Macquarie population stands out in Figure 5A, B representing the developmental rate reaction norm. Although much care was taken in sampling at the same altitudes and in comparable habitats, additional factors are likely to play a role over such a large gradient. Annual differences in precipitation might be important, as dry periods could select for compensatory growth, creating similar selection pressures to those found in cool environments. These outliers emphasize the importance of considering multiple populations when establishing trends rather than just two end points along a gradient. In conclusion, contrasting latitudinal patterns in plasticity and environmental canalization of life history and morphological traits were found. Contrasting patterns may be common because of the different types of selection pressures acting on traits along gradients. This suggests that statements about overall levels of plasticity of a population are probably meaningless. Instead, testing specific hypotheses about the role of phenotypic plasticity in adaptation along gradients is likely to be a more productive approach. ACKNOWLEDGMENTS We thank V. M. Kellermann and B. Van Heerwaarden for help during field collections, J. Shirriffs and other staff at CESAR (The University of Melbourne) for their assistance during laboratory work and experiments, and A. Rebocho at the Department of Genetics (VU University) for access to the digital imaging equipment and technical assistance. We are also grateful to V. Debat and one anonymous reviewer for their constructive comments on earlier drafts of the manuscript. JE and ML were supported by the Netherlands Organisation for Scientific Research, VIDI grant no. 864.03.003, and AAH was supported by the Australian Research Council.

LITERATURE CITED Angilletta, M. J., P. H. Niewiarowski, and C. A. Navas. 2002. The evolution of thermal physiology in ectotherms. J. Therm. Biol. 27:249–268. Azevedo, R. B. R., A. C. James, J. McCabe, and L. Partridge. 1998. Latitudinal variation of wing: thorax size ratio and wing-aspect ratio in Drosophila melanogaster. Evolution 52:1353–1362. Blows, M. W., and M. Higgie. 2002. Evolutionary experiments on mate recognition in the Drosophila serrata species complex. Genetica 116:239– 250. Bradshaw, A. D. 1965. Evolutionary significance of phenotypic plasticity in plants. Adv. Genet. 13:115–155.

EVOLUTION AUGUST 2009

1961

M A A RT J E L I E F T I N G E T A L .

Cohet, Y., J. Vouidibio, and J. R. David. 1980. Thermal tolerance and geographic distribution—a comparison of cosmopolitan and tropical endemic Drosophila species. J. Therm. Biol. 5:69–74. Conover, D. O., and E. T. Schultz. 1995. Phenotypic similarity and the evolutionary significance of countergradient variation. Trends Ecol. Evol. 10:248–252. Conover, D. O., J. J. Brown, and A. Ehtisham. 1997. Countergradient variation in growth of young striped bass (Morone saxatilis) from different latitudes. Can. J. Fish. Aquat. Sci. 54:2401–2409. David, J. R., B. Moreteau, J. P. Gauthier, G. Petavy, A. Stockel, and A. G. Imasheva. 1994. Reaction norms of size characters in relation to growth temperature in Drosophila melanogaster—an isofemale lines analysis. Genet. Sel. Evol. 26:229–251. Debat, V., and P. David. 2001. Mapping phenotypes: canalization, plasticity and developmental stability. Trends Ecol. Evol. 16:555–561. de Jong, G. 1990. Quantitative genetics of reaction norms. J. Evol. Biol. 3:447–468. ———. 1995. Phenotypic plasticity as a product of selection in a variable environment. Am. Nat. 145:493–512. Delpuech, J. M., B. Moreteau, J. Chiche, E. Pla, J. Vouidibio, and J. R. David. 1995. Phenotypic plasticity and reaction norms in temperate and tropical populations of Drosophila melanogaster—ovarian size and developmental temperature. Evolution 49:670–675. DeWitt, T. J., A. Sih, and D. S. Wilson. 1998. Costs and limits of phenotypic plasticity. Trends Ecol. Evol. 13:77–81. Driessen, G., J. Ellers, and N. M. van Straalen. 2007. Variation, selection and heritability of thermal reaction norms for juvenile growth in Orchesella cincta (Collembola: Entomobryidae). Eur. J. Entomol. 104:39–46. Ellers, J., and J. J. M. van Alphen. 1997. Life history evolution in Asobara tabida: plasticity in allocation of fat reserves to survival and reproduction. J. Evol. Biol. 10:771–785. Frazier, M. R., J. F. Harrison, S. D. Kirkton, and S. P. Roberts. 2008. Cold rearing improves cold-flight performance in Drosophila via changes in wing morphology. J. Exp. Biol. 211:2116–2122. Frentiu, F. D., and S. F. Chenoweth. 2008. Polyandry and paternity skew in natural and experimental populations of Drosophila serrata. Mol. Ecol. 17:1589–1596. Gilchrist, G. W. 1995. Specialists and generalists in changing environments. I. Fitness landscapes of thermal sensitivity. Am. Nat. 146:252–270. Gilchrist, G. W., and R. B. Huey. 2004. Plastic and genetic variation in wing loading as a function of temperature within and among parallel clines in Drosophila subobscura. Integr. Comp. Biol. 44:461–470. Gutteling, E. W., J. A. G. Riksen, J. Bakker, and J. E. Kammenga. 2007. Mapping phenotypic plasticity and genotype-environment interactions affecting life-history traits in Caenorhabditis elegans. Heredity 98:28– 37. Hallas, R., M. Schiffer, and A. A. Hoffmann. 2002. Clinal variation in Drosophila serrata for stress resistance and body size. Genet. Res. 79:141–148. Hassall, M., A. Helden, A. Goldson, and A. Grant. 2005. Ecotypic differentiation and phenotypic plasticity in reproductive traits of Armadillidium vulgare (Isopoda: Oniscidea). Oecologia 143:51–60. Hoffmann, A. A., and J. Shirriffs. 2002. Geographic variation for wing shape in Drosophila serrata. Evolution 56:1068–1073. Hoffmann, A. A., E. Ratna, C. M. Sgr`o, M. Barton, M. Blacket, R. Hallas, S. De Garis, and A. R. Weeks. 2007. Antagonistic selection between adult thorax and wing size in field released Drosophila melanogaster independent of thermal conditions. J. Evol. Biol. 20:2219–2227. Huey, R. B., and J. G. Kingsolver. 1989. Evolution of thermal sensitivity of ectotherm performance. Trends Ecol. Evol. 4:131–135. James, A. C., and L. Partridge. 1995. Thermal evolution of rate of larval devel-

1962

EVOLUTION AUGUST 2009

opment in Drosophila melanogaster in laboratory and field populations. J. Evol. Biol. 8:315–330. James, A. C., R. B. R. Azevedo, and L. Partridge. 1997. Genetic and environmental responses to temperature of Drosophila melanogaster from a latitudinal cline. Genetics 146:881–890. Jenkins, N. L., and A. A. Hoffmann. 2000. Variation in morphological traits and trait asymmetry in field Drosophila serrata from marginal populations. J. Evol. Biol. 13:113–130. Kingsolver, J. G., G. J. Ragland, and J. G. Shlichta. 2004. Quantitative genetics of continuous reaction norms: thermal sensitivity of caterpillar growth rates. Evolution 58:1521–1529. Kingsolver, J. G., K. R. Massie, J. G. Shlichta, M. H. Smith, G. J. Ragland, and R. Gomulkiewicz. 2007. Relating environmental variation to selection on reaction norms: an experimental test. Am. Nat. 169:163–174. Liefting, M., and J. Ellers. 2008. Habitat-specific differences in thermal plasticity in natural populations of a soil arthropod. Biol. J. Linn. Soc. 94:265–271. Lynch, M., and W. Gabriel. 1987. Environmental tolerance. Am. Nat. 129:283– 303. Magiafoglou, A., M. E. Carew, and A. A. Hoffmann. 2002. Shifting clinal patterns and microsatellite variation in Drosophila serrata populations: a comparison of populations near the southern border of the species range. J. Evol. Biol. 15:763–774. Meril¨a, J., A. Laurila, and B. Lindgren. 2004. Variation in the degree and costs of adaptive phenotypic plasticity among Rana temporaria populations. J. Evol. Biol. 17:1132–1140. Meyers, L. A., and J. J. Bull. 2002. Fighting change with change: adaptive variation in an uncertain world. Trends Ecol. Evol. 17:551–557. Morin, J. P., B. Moreteau, G. P´etavy, and J. R. David. 1999. Divergence of reaction norms of size characters between tropical and temperate populations of Drosophila melanogaster and D. simulans. J. Evol. Biol. 12:329–339. Noach, E. J. K., G. deJong, and W. Scharloo. 1996. Phenotypic plasticity in morphological traits in two populations of Drosophila melanogaster. J. Evol. Biol. 9:831–844. Nylin, S. 1992. Seasonal plasticity in life history traits: growth and development in Polygonia calbum (Lepidoptera, Nymphalidae). Biol. J. Linn. Soc. 47:301–323. Pfennig, D. W., and P. J. Murphy. 2002. How fluctuating competition and phenotypic plasticity mediate species divergence. Evolution 56:1217– 1228. Price, T. D., A. Qvarnstr¨om, and D. E. Irwin. 2003. The role of phenotypic plasticity in driving genetic evolution. Proc. R. Soc. Lond. B 270:1433– 1440. Richards, C. L., O. Bossdorf, N. Z. Muth, J. Gurevitch, and M. Pigliucci. 2006. Jack of all trades, master of some? On the role of phenotypic plasticity in plant invasions. Ecol. Lett. 9:981–993. Robinson, S. J. W., and L. Partridge. 2001. Temperature and clinal variation in larval growth efficiency in Drosophila melanogaster. J. Evol. Biol. 14:14–21. Roff, D. 2002. Life history evolution. Sinauer Associates, Inc, Sunderland, MA. Scheiner, S. M. 1993. Genetics and evolution of phenotypic plasticity. Annu. Rev. Ecol. Syst. 24:35–68. Schiffer, M., A. S. Gilchrist, and A. A. Hoffmann. 2006. The contrasting genetic architecture of wing size, viability, and development time in a rainforest species and its more widely distributed relative. Evolution 60:106–114. Sgr`o, C. M., and M. W. Blows. 2003. Evolution of additive and nonadditive genetic variance in development time along a cline in Drosophila serrata. Evolution 57:1846–1851.

O P P O S I T E L AT I T U D I NA L C L I N E S I N P L A S T I C I T Y

Stearns, S. C. 1989. The evolutionary significance of phenotypic plasticity— phenotypic sources of variation among organisms can be described by developmental switches and reaction norms. Bioscience 39:436–445. Stearns, S. C., and T. J. Kawecki. 1994. Fitness sensitivity and the canalization of life-history traits. Evolution 48:1438–1450. Stearns, S. C., M. Kaiser, and T. J. Kawecki. 1995. The differential genetic and environmental canalization of fitness components in Drosophila melanogaster. J. Evol. Biol. 8:539–557. Trotta, V., F. C. F. Calboli, M. Ziosi, D. Guerra, M. C. Pezzoli, J. R. David, and S. Cavicchi. 2006. Thermal plasticity in Drosophila melanogaster: a comparison of geographic populations. BMC Evol. Biol. 6:1471– 2148.

Unwin, D. M., and S. A. Corbet. 1984. Wingbeat frequency, temperature and body size in bees and flies. Physiol. Entomol. 9:115–121. van Asch, M., P. H. Tienderen, L. J. M. Holleman, and M. E. Visser. 2007. Predicting adaptation of phenology in response to climate change, an insect herbivore example. Glob. Change Biol. 13:1596–1604. Wagner, G. P., G. Booth, and H. C. Bagheri. 1997. A population genetic theory of canalization. Evolution 51:329–347. Winterhalter, W. E., and T. A. Mousseau. 2007. Patterns of phenotypic and genetic variation for the plasticity of diapause incidence. Evolution 61:1520–1531.

Associate Editor: J. Feder

EVOLUTION AUGUST 2009

1963