Testing the beneficial acclimation hypothesis: temperature ... - CiteSeerX

31 downloads 0 Views 140KB Size Report
Apr 24, 2007 - We here set out to test acclimation effects on reproductive performance using the tropical butterfly Bicyclus anynana. (Butler 1879) as model ...
Behavioral Ecology doi:10.1093/beheco/arm024 Advance Access publication 24 April 2007

Testing the beneficial acclimation hypothesis: temperature effects on mating success in a butterfly Thorin L. Geister and Klaus Fischer Department of Animal Ecology I, University of Bayreuth, PO Box 101 251, D-95440 Bayreuth, Germany Traditionally, it has been assumed that all acclimation changes to the phenotype enhance the performance of an individual organism in the environment in which those changes were induced (beneficial acclimation hypothesis [BAH]), a theory that has been repeatedly challenged in recent years. We here use a full-factorial design with 2 developmental and 2 acclimation temperatures to test their effects on reproductive performance in the tropical butterfly, Bicyclus anynana. Competition experiments among virgin males from different thermal groups revealed that, at 20 C, both groups acclimated to 20 C achieved more than twice as many matings as those acclimated to 27 C, whereas at 27 C, only one group (acclimated to 27 C) outperformed all others. Chill-coma recovery times were also longer for butterflies that developed at higher temperatures, indicating that butterflies responded physiologically to the temperatures at which they were reared. Our results support the BAH at least in part, and do not support any alternative hypotheses. Key words: Bicyclus anynana, chill-coma recovery time, female mate choice, phenotypic plasticity, temperature adaptation. [Behav Ecol 18:658–664 (2007)]

ecent developments in evolutionary physiology, in particular the use of more hypothesis-driven approaches in physiological research, have stimulated a renewed interest in the magnitude and nature of nongenetic effects on an organism’s phenotype (Nylin and Gotthard 1998; Feder et al. 2000; Wilson and Franklin 2002; Woods and Harrison 2002). Only recently, we have begun to understand the mechanistic basis of phenotypic plasticity and its potential role in organismal ecology and evolution (Stearns 1989; Scheiner 1993; Via et al. 1995). A fundamental question now is whether plasticity is adaptive or merely a biochemical or physiological interaction of the organism with its environment (Gotthard and Nylin 1995). As a result, the significance of acclimatory changes has received substantial attention, primarily through the examination of the acclimatory responses of ectotherms to temperature (Leroi et al. 1994; Zamudio et al. 1995; Huey and Berrigan 1996; Bennett and Lenski 1997; Huey et al. 1999; Gibert, Huey, Gilchrist 2001; Wilson and Franklin 2002; Woods and Harrison 2002). Traditionally, it has been assumed that all acclimation changes to the phenotype act to enhance the performance of an individual organism in the environment where those changes were induced (beneficial acclimation hypothesis [BAH]; Leroi et al. 1994). Such responses can be expected in all cases where the costs of acclimation do not outweigh their benefits (e.g., Hoffmann 1995; Scott et al. 1997; Scheiner and Berrigan 1998). Although intuitively appealing, only a few studies support the BAH (e.g., Nunney and Cheung 1997; Fischer, Brakefield, Zwaan 2003; Fischer, Eenhoorn, et al. 2003), whereas the majority of experimental analyses to date have rejected its generality (reviewed in Wilson and Franklin 2002; Woods and Harrison 2002). The negative results are surprising because many studies showed a survival advan-

R

Address correspondence to T.L. Geister. E-mail: thorin.geister@ uni-bayreuth.de. Received 17 May 2006; revised 23 November 2006; accepted 5 March 2007.  The Author 2007. Published by Oxford University Press on behalf of the International Society for Behavioral Ecology. All rights reserved. For permissions, please e-mail: [email protected]

tage at extreme temperatures for animals that had been previously exposed to less extreme temperatures (e.g., Sejerkilde et al. 2003; Sørensen et al. 2003). This raises an important but currently unresolved question: why have so many studies rejected such an intuitive hypothesis? Three potential problems with studies testing the BAH have been recognized: 1) most studies thus far should be considered elegant tests of developmental plasticity (a usually nonreversible cascade of phenotypic changes due to differences in the developmental environment) rather than of acclimation (a reversible, facultative response to changes in a single environmental variable in the adult stage; Leroi et al. 1994; Benett and Lenski 1997; Willmer et al. 2000; Wilson and Franklin 2002); 2) tests frequently included stressful environments, and long-term exposure to these will depress performance in all subsequently experienced environments (Leroi et al. 1994; Benett and Lenski 1997; Huey et al. 1999; Wilson and Franklin 2002; Woods and Harrison 2002); and 3) some of the traits examined were only indirectly related to fitness (Gibert, Huey, Gilchrist 2001; Wilson and Franklin 2002). We here set out to test acclimation effects on reproductive performance using the tropical butterfly Bicyclus anynana (Butler 1879) as model organism. We used a full-factorial design with 2 developmental and 2 adult (i.e., acclimation) temperatures (20 and 27 C). In our experiments, males with different thermal histories competed for a single virgin female (cf., Wilson and Franklin 2002). The mating success of males may depend on locomotor or general activity, which has been shown to acclimate to temperature in a variety of taxa (Prosser 1991). Based on the BAH, it is predicted that, at the low temperature (reduced activity levels), cool-acclimated males should be more successful than warm-acclimated ones and vice versa. Note, however, that the latter might not necessarily be the case. The higher temperature chosen here should allow high levels of activity for all butterflies, not only for warmacclimated ones (one could even imagine overcompensation of cold-acclimated ones). A clear advantage of warm-acclimated individuals may not be visible unless test conditions include stressfully high temperatures.

Geister and Fischer • Testing the beneficial acclimation hypothesis

However, our design allows us to explore the following alternative, though not mutually exclusive hypotheses (Leroi et al. 1994; Zamudio et al. 1995; Huey and Berrigan 1996; Huey et al. 1999): 1) random mating (null hypothesis); 2) BAH: best performance when test temperature matches acclimation temperature; 3) developmental plasticity: best performance when test temperature matches developmental temperature; 4) best performance at warmer developmental temperature; and 5) best performance at colder developmental temperature. In addition, we investigate whether mating success is affected by relative eyespot size or male body size. To investigate whether the selected temperatures were sufficiently divergent to induce clear physiological changes, we also tested for differences in chill-coma recovery times (defined as the time elapsed until a butterfly was able to stand up after cold exposure) between temperature groups. This recently developed method has proved to be a rapid and sensitive index of climatic adaptation for Drosophila and at least one butterfly species (David et al. 1998; Macdonald et al. 2004; Zeilstra and Fischer 2005). Our study system is appropriate for testing the BAH for several reasons. First, B. anynana lives in a seasonal environment with a colder dry season and a warmer wet season. The temperatures chosen for our experiments are similar to the ones experienced by the butterflies in the field during the dry and wet season, respectively (Brakefield and Reitsma 1991; Brakefield and Mazzotta 1995). Thus, we did not include marginal temperatures, but ones the butterflies should be well adapted to. Second, the trait examined (male mating success) is undoubtedly closely related to fitness (Wilson and Franklin 2002). Third, male competition experiments have been successfully used in this species in other contexts (Beldade 2002; Joron and Brakefield 2003; Robertson and Monteiro 2005). METHODS Study organism and experimental population B. anynana is a tropical, fruit-feeding butterfly distributed from southern Africa to Ethiopia (Larsen 1991). This species exhibits striking phenotypic plasticity with 2 seasonal morphs, which is thought to function as an adaptation to alternate wet–dry seasonal environments and the associated changes in resting background and predation (Brakefield 1997; Lyytinen et al. 2004). Reproduction is essentially confined to the warmer wet season when host plants are abundantly available and where 2–3 generations occur. Toward the end of the wet season, there is a marked decrease in temperature, starting several weeks before larval food plants dry out. During the colder dry season, reproduction ceases and butterflies do not mate before the first rains at the beginning of the next wet season (Windig 1994; Brakefield 1997). Morphs are gradually replaced during seasonal transitions. Thus, both phenotypes may occur simultaneously (Brakefield and Reitsma 1991). For mate location, male B. anynana butterflies employ a perch-and-chase strategy. Close-range courtship signals involve the flickering of the male’s wings and the release of pheromones, followed by attempts by the male to latch onto the female’s abdomen. As perching males can be found in high densities, they are frequently involved in circuit and chasing flights (Brakefield and Reitsma 1991; Joron and Brakefield 2003). For this study, butterflies from the Bayreuth stock population were used. A laboratory stock population of B. anynana was established in 2003 from several hundred eggs derived from a well-established stock population at the University of Leiden, the Netherlands. The Leiden population was founded in 1988 from more than 80 gravid females collected at Nkhata Bay, Malawi. In each generation, several hundred individuals

659

are reared, maintaining high levels of heterozygosity at neutral loci (Van’t Hof et al. 2005). Butterfly rearing B. anynana eggs collected from several hundred adults were randomly divided among a high (27 C) and a low rearing temperature (20 C; relative humidity 70% and 12:12 h light: dark throughout). The temperatures chosen are similar to the daily highs during the wet and dry season (Brakefield and Reitsma 1991). To synchronize adult eclosion, 2 generations were reared at 27 C prior to experiments but only one at 20 C (development at 20 C takes roughly twice as long as it does at 27 C; Fischer, Brakefield, Zwaan 2003). Prior to this experiment, the stock population was reared for several generations at 20 C. Larvae were kept in population cages and were fed on young maize plants. The cages were checked daily for pupae, which were weighed to the nearest 0.01 mg on the day after pupation. Afterward, pupae were kept individually until adult eclosion. After eclosion, all butterflies were individually marked and once again randomly divided among 20 and 27 C, resulting in 4 treatment groups with different thermal histories (developmental temperature/adult temperature): 27/27 C, 27/20 C, 20/20 C, and 20/27 C. Throughout, male and female butterflies were kept apart until the experiments started. Butterflies were fed daily with moist banana throughout all experiments. Cold tolerance To test for physiological responses to temperatures, butterflies from all thermal groups were exposed (after varying acclimation periods of 1, 2, 3, or 4 days at 20 or 27 C) for 19 h to 1 C. For testing, butterflies were placed individually in small translucent plastic cups (125 ml), which were arranged on a small tray in a randomized block design. Each block consisted of 18 butterflies (4 thermal groups 3 4 acclimation periods, plus 2 freshly eclosed butterflies being reared at 20 and 27 C, respectively, to test for effects of developmental temperature). After cold exposure, butterflies were transferred to an environmental cabinet with a constant temperature of 20 C and chill-coma recovery times (defined as the time between taking the tray out of the refrigerator until a butterfly was able to stand on its legs) were measured. Butterflies were monitored for a maximum of 90 min. A total of 31 blocks were analyzed, with those missing more than 2 values (due to recovery times longer than 90 min or death) being discarded. For cold tolerance assays, all butterflies were only used once. An excess of male butterflies was needed for the mating experiments described below, and thus cold tolerance experiments tested mostly female butterflies. However, pilot experiments demonstrated that there is no sex difference in recovery times (Table 1B; see also Zeilstra and Fischer 2005). As in the above experiment, the 27/20 C group responded remarkably fast to the temperature change (Figure 2A), the first 24 h of the acclimation response was investigated in more detail. Therefore, chill-coma recovery time was analyzed after acclimation periods of 0, 6, 12, 18, and 24 h at 20 C. Mating experiments Each mating trial involved competition among 4 randomly chosen 3- to 5-day-old virgin males (one from each thermal group) for a single female in a cylindrical net cage (diameter 38 cm, height 10 cm) at either 20 or 27 C. Such assay conditions have been successfully used before in B. anynana (Beldade 2002; Robertson and Monteiro 2005). Direct competition among males frequently occurs in the field in this

Behavioral Ecology

660

species (see above). As females show low levels of polyandry only and do reject courting males (Brakefield and Reitsma 1991; Brakefield et al. 2001), female choice should prevail in this system. Nevertheless, male mating success seems to depend strongly on persistence and aggressiveness in latching onto the female’s abdomen. To ensure equal opportunities for all males, the ones not being kept at the respective test temperature were introduced into the cages first in order to allow for adjustment of body temperature. We tested for the time period needed by measuring thoracic temperature of male butterflies in 15 s intervals after a temperature transfer. This was done by inserting a hypodermic needle probe with a copper-constantan thermocouple (Omega Hyp0-33-1-T-G-60SMP-M), connected to a digital thermometer (Omega HH21), into the males’ thorax. According to these results (Figure 1), the other 2 males were introduced 5 min later. Finally, a single, 2-day-old virgin female (from the 27/27 C group at 27 C and from the 27/20 C at 20 C) was released into each cage to start the trial. To minimize variation in female phenotype, only females that had developed at 27 C were used in the experiment. Cages were checked for matings every 15 min for the following 8 h (mating in B. anynana lasts about 30 min; Joron and Brakefield 2003). The first male to mate was scored as ‘‘winner,’’ and the trial was terminated thereafter. Butterflies were only used once for mating trials. At the 2 test temperatures, 60 and 61 matings were recorded in 85 (27 C) and 132 (20 C) trials. In addition to pupal mass (see above), forewing length (from basal margin to apex) and the size of the posterior eyespot (horizontal diameter relative to forewing length) were measured for the males tested at 20 C. To do this, the right, ventral forewings were photographed using a digital camera (Leica DC300) mounted on a binocular microscope (Leica MZ 7.5). The resulting images were analyzed using the freeware Scion Image (release Beta 4.0.2, www.scioncorp.com). Data analysis The data from the cold shock experiment were analyzed using a 3-way analysis of covariance (ANCOVA), with developmental temperature, acclimation temperature, and acclimation time as fixed factors and pupal mass as covariate. All data were transformed into percentage deviations from block means to even out slight temperature differences across experimental blocks caused by position within the refrigerator. Such varia-

tion can largely affect recovery time (data not shown). To meet ANCOVA requirements, data were log-transformed prior to analysis. The statistical analysis of the mating experiments is complicated by the nonindependence of data (if one male wins, the other 3 lose and thus their scores are not independent). Therefore, we used 2 approaches. First, distributions of successful matings across thermal groups were tested against even distributions using v2 tests. To adjust for multiple pairwise comparisons, a sequential Bonferroni correction was applied. Second, to simultaneously test for effects of developmental temperature, acclimation temperature, and different covariates (pupal mass, forewing length, and/or relative eyespot size) on the mating success of B. anynana males (binary data: successful or unsuccessful), we calculated generalized nonlinear models (GNLM) with a binomial error distribution and a logit-link function. This was problematic due to the nonindependence of data. However, all group differences are supported by the above v2 tests. We decided to present the GNLM results here because they give a much more complete picture due to the inclusion of covariates and the possibility to compute interaction terms (see also Zamudio et al. 1995). All statistical tests were performed using Statistica 6.1. Unless otherwise stated, all means are given 61 standard error.

RESULTS Cold tolerance Recovery after cold exposure was significantly affected by developmental temperature, acclimation temperature, and acclimation period (Table 1A), with animals being reared or maintained at the higher temperature needing longer time to recover from chill coma than those at the lower temperature (Figure 2A). Effects of the developmental temperature were also visible on the day of eclosion: animals reared at 27 C (25:40 6 0:53 min) showed significantly longer recovery times than those reared at 20 C (13:44 6 0:37 min, Mann–Whitney U test: U ¼ 234.5, n ¼ 58, P ¼ 0.004). On subsequent days, however, acclimation temperature was the most important factor affecting cold tolerance. Animals kept at the cooler temperature recovered on average 23 min earlier than their counterparts at the higher temperature. Although recovery times increased with the length of the acclimation period when adults were kept at 27 C, a comparable pattern did

20°C to 27°C

31

27°C to 20°C

Control 27°C

Control 20°C

30

Figure 1 Thorax temperature over time (means 6 1 standard deviation) for Bicyclus anynana butterflies after transfer from 27 to 20 C (filled symbols) and vice versa (open symbols; n ¼ 10 each). Values for control individuals are 29.45 6 0.39 C at 27 C and 21.75 6 0.64 C at 20 C (indicated by horizontal lines). From 135 s onward, the transfer groups did not differ significantly from their respective controls (Tukey HSD after analysis of variance).

Temperature [°C]

29 28 27 26 25 24 23 22 21

0

15

30

45

60

75

90 105 120 135 150 165 180 195 210 225 240 255 270

Time [s]

Geister and Fischer • Testing the beneficial acclimation hypothesis

661

Table 1 ANCOVAs for the effects of various factors on recovery times of Bicyclus anynana after 19 h exposure to 1 C: (A), effects of pupal mass (covariate), developmental temperature (20, 27 C), acclimation temperature (20, 27 C), and acclimation time (1–4 days) on female recovery times; (B), effects of acclimation temperature (20, 27 C) and sex. Butterflies had developed at 27 C and were tested after acclimation periods of 2 days Effect

df

MS

F

P

(A) Pupal mass Developmental temperature Acclimation temperature Acclimation time Developmental temperature 3 acclimation temperature Developmental temperature 3 acclimation time Acclimation temperature 3 acclimation time Developmental temperature 3 acclimation temperature 3 acclimation time Error

1 1 1 3 1 3 3

152.4 1119.4 18358.2 810.3 110.8 8.4 872.0

2.02 14.82 242.98 10.73 1.47 0.11 11.54

0.156 ,0.001 ,0.001 ,0.001 0.226 0.954 ,0.001

3 452

99.3 75.5

1.31

0.269

(B) Acclimation temperature Sex Acclimation temperature 3 sex Error

1 1 1 128

5020.4 12.6 37.4 64.4

78.00 0.12 0.58

,0.001 0.659 0.448

Significant P values are given in bold. df ¼ degrees of freedom.

not occur at 20 C (causing a significant interaction between acclimation temperature and period; Table 1A). Note that an acclimation response to a temperature change occurs within 1 day when individuals are transferred from 27 to 20 C (Figure 2B) or 2 days when individuals are transferred from 20 to 27 C (Figure 2A). Males and females had very similar recovery times (P ¼ 0.66) when tested after acclimation periods of 2 days (Table 1B; females 27 C: 47:07 6 3:30 min, males 27 C: 45:48 6 4:46 min, females 20 C: 20:22 6 2:53 min, males 20 C: 20:37 6 2:14 min). In summary, these results demonstrate that B. anynana does show physiological responses to developmental and acclimation temperature, with the latter being induced within 48 h at the latest. Consequently, the acclimation periods between

At the higher test temperature, the group reared at 20 C but acclimated to 27 C (20/27 C) was the most successful group that gained 45% of all matings. The other groups succeeded in only 17–20% of the successful mating trials, resulting in a significant deviation from an even distribution across temperature groups (v23 ¼ 12:93, n ¼ 60, P ¼ 0.005; 20/27 C . 27/27 C ¼ 27/20 C ¼ 20/20 C, pairwise post hoc comparisons of mating frequencies with sequential Bonferroni correction; Figure 3). The fact that one group outperformed all

B

45 40

Recovery time [min]

Mating experiments

35

50

20 °C/ 27 °C

45

30 25 20 27 °C/ 20 °C

15

55

27 °C/ 27 °C

Recovery time [min]

A

3 and 5 days for males used in the mating experiments below were sufficient to allow for physiological responses.

40 35 30 25

20 °C/ 20 °C

10 5

20

0

1

2

3

Acclimation time [d]

4

15

0

6

12

18

24

Acclimation time [h]

Figure 2 Chill-coma recovery times (means 6 1 standard error) of Bicyclus anynana females after 19 h exposure to 1 C in relation to developmental temperature, acclimation temperature, and acclimation time. Squares and circles refer to a developmental temperature of 27 and 20 C and filled and open symbols to an acclimation temperature of 27 and 20 C, respectively. (A) Change in recovery times after 0 (controls), 1, 2, 3, or 4 days of acclimation. (B) Change in recovery times within the first 24 h after transfer from 27 to 20 C. After only 6 h at 20 C, butterflies recover significantly quicker than controls (acclimation time 0 h; Tukey HSD after analysis of variance; sample sizes range between 14 and 18).

Behavioral Ecology

662

Effect

df

Wald statistics

P

(A) 27 C (n ¼ 237) Pupal mass Developmental temperature Acclimation temperature Developmental temperature 3 acclimation temperature

1 1 1

3.038 1.973 2.967

0.081 0.160 0.085

1

6.450

0.011

(B) 20 C (n ¼ 242) Pupal mass Relative eyespot size Developmental temperature Acclimation temperature Developmental temperature 3 acclimation temperature

1 1 1 1

3.070 0.512 0.013 16.696

0.080 0.474 0.908 ,0.001

1

0.759

0.384

(C) 20 C (n ¼ 244) Forewing length Relative eyespot size Developmental temperature Acclimation temperature Developmental temperature 3 acclimation temperature

1 1 1 1

11.523 2.039 0.004 18.052

,0.001 0.153 0.950 ,0.001

1

0.572

0.450

50

40

Mating success [%]

Table 2 Effects of different covariates (pupal mass, relative eyespot size, and/or forewing length), developmental temperature (20, 27 C), and acclimation temperature (20, 27 C) on the mating success of Bicyclus anynana males at test temperatures of 27 C (A) and 20 C (B/C); tested by GNLM (binomial error distribution, logit-link function)

27°C / 27°C 27°C / 20°C 20°C / 20°C 20°C / 27°C

others (suggesting that neither developmental nor acclimation temperature is most important) is also indicated by a significant interaction between developmental and acclimation temperature in the GNLM analysis (Table 2A). Statistical trends for acclimation temperature and pupal mass indicate tendencies toward a higher mating success of warm-acclimated and larger males. The largest out of the 4 competing males within a trial won in 20 of 60 matings, 12 of which belonged to the 20/27 C group (Figure 3, Table 3A). Thus, there was no general advantage of being particularly large. Only the smallest male within in a given trial had a significant disadvantage, succeeding in only 5 trials (v23 ¼ 8:94, n ¼ 57, P ¼ 0.03). Mating success in the remaining 3 body rank groups was closely similar (v23 ¼ 0:50, n ¼ 52, P ¼ 0.78; Table 3A). At the lower test temperature, both groups acclimated to 20 C achieved more than twice as many matings (27/20 C: 33%; 20/20 C: 39%) as their counterparts acclimated to 27 C (27/27 C: 15%; 20/27 C: 13%). This again results in a significant deviation from an even distribution (v23 ¼ 12:51, n ¼ 61, P ¼ 0.006; 20/20 C ¼ 27/20 C . 20/27 C ¼ 27/27 C, pairwise post hoc comparisons; note that groups 27/20 C and 27/27 C do not differ significantly: P ¼ 0.019, adjusted significance threshold P ¼ 0.017; Figure 3). Thus, when tested at 20 C, the single most important factor affecting mating success is a priori acclimation to the test temperature (see also Table 2B). Similar to the results at 27 C, 16 out of 61 matings were won by the largest male within a trial. Three of these originated from the 27/20 C group and 10 from the 20/20 C group (Table 3). The distribution of winners across body ranks did not differ significantly from an even distribution (v23 ¼ 2:47; n ¼ 59, P ¼ 0.49). A comparable pattern was found when using wing size as a proxy for body size, though

24 20

30

20

12 10

11 9

8

10

0

Note that forewing length and eyespot size were only measured in animals tested at 20 C. Significant P values are given in bold. df ¼ degrees of freedom.

27

27 °C

20°C

Test temperature

Figure 3 Mating success of 4 groups of Bicyclus anynana males with different thermal histories (developmental temperature/acclimation temperature) at 2 test temperatures. Bars add up to 100% for the matings at 20 and 27 C, respectively. Figures atop bars represent the number of matings won by members of the respective group.

the smallest males succeeded in 8 trials only (v23 ¼ 5:00; n ¼ 61, P ¼ 0.17; Table 3C). Also, relative eyespot size did not affect mating success (v23 ¼ 0:18; n ¼ 61, P ¼ 0.98; Table 3). Table 3 Body size (A–C) and eyespot size (D) ranks of Bicyclus anynana males with different thermal histories (developmental temperature/ acclimation temperature) at 2 test temperatures 27/27 C

27/20 C

20/20 C

20/27 C

Sum

(A) At 27 C (pupal mass) Largest Second largest Second smallest Smallest

2 2 3 2

1 3 7 1

4 4 2 0

12 6 6 2

19 15 18 5

(B) At 20 C (pupal mass) Largest Second largest Second smallest Smallest

0 3 2 3

3 6 4 7

10 8 4 1

3 2 1 2

16 19 11 13

(C) At 20 C (wing size) Largest Second largest Second smallest Smallest

2 0 5 2

3 5 6 6

10 10 4 0

3 4 1 0

19 18 16 8

(D) At 20 C (eyespot size) Largest Second largest Second smallest Smallest

2 4 3 0

10 8 2 0

2 4 8 10

0 0 3 5

14 16 16 15

Figures refer to the number of males per thermal group that had won a mating trial involving competition of 4 virgin males for a single virgin female (see text). At 27 C, only pupal mass was measured as proxy for body size (A), whereas at 20 C, forewing length and eyespot size were measured in addition (C, D).

Geister and Fischer • Testing the beneficial acclimation hypothesis

In accordance with the above results, a GNLM showed a statistical trend toward an effect of pupal mass, while neither relative eyespot size nor developmental temperature were significant predictors for mating success (Table 2). Replacing pupal mass by forewing length shows a significant effect of that trait on mating success, with successful males having, on average, 2.6% longer forewings than unsuccessful ones (Table 2). Nevertheless, acclimation temperature remains a significant factor. Note that, if body size were of prime importance, the groups reared at 20 C would be expected to be most successful, because these males attained higher pupal masses than the groups reared at 27 C (156.85 6 1.23 mg, n ¼ 238 vs. 141.09 6 1.39 mg, n ¼ 241; t-test: t477¼ 8.50, P , 0.001; in accordance with the temperature-size rule; Atkinson 1994; Chown and Gaston 1999). Likewise, the eyespots of B. anynana responded plastically to rearing temperature (27 . 20 C, Mann–Whitney U test: U ¼ 1238.0, n ¼ 244, P , 0.001; Brakefield and Reitsma 1991), but their size had no detectable effect on mating success. In conclusion, mating success was not evenly distributed across thermal groups at either test temperature, with the respective patterns differing strikingly between 20 and 27 C (v23 ¼ 57:88; n ¼ 121, P , 0.001; Figure 3).

DISCUSSION Cold tolerance Chill-coma recovery times as a nonlethal index of cold tolerance have been shown to be a sensitive and reliable tool for demonstrating climatic adaptation of different Drosophila species and populations (David et al. 1998; Gibert, Moreteau, et al. 2001; David et al. 2003; Macdonald et al. 2004). Most studies using this technique have concentrated on geographic variation in cold tolerance (e.g., David et al. 2003) or on the underlying mechanisms of cold resistance (e.g., Macdonald et al. 2004). In this study, we used this technique to demonstrate plastic responses after a temperature change. Both developmental and acclimation temperature influenced recovery times of B. anynana in a similar way (Ayrinhac et al. 2004; Rako and Hoffmann 2006). Colder temperatures generally led to shorter recovery times than warmer ones. Recovery times changed quickly after a transfer to a different temperature: adjustments occurred within 24 and 48 h in the groups transferred from 27 to 20 C and 20 to 27 C, respectively. The increase in recovery time with the length of the acclimation period at 27 C but not at 20 C may indicate accelerated rates of aging at the higher temperature, although longevity has not been observed to differ between individuals at both temperatures (Fischer, Brakefield, Zwaan 2003). In any case, the above results indicate that the acclimation temperatures and periods chosen for the mating experiments were sufficient to induce pronounced physiological responses. These results were fully supported measuring heat knockdown times in a comparable set up: butterflies acclimated to 27 C had a significantly enhanced heat stress resistance than the ones acclimated to 20 C (Fischer K, Pflicke C, Geister TL, in preparation). Mating experiments Overall mating success was much higher at 27 C as compared with 20 C, as was expected based on the much higher activity levels of the butterflies at the higher temperature. Our experiments suggest that temperature acclimation may greatly affect mating success in B. anynana. Either one (at 27 C) or both (at 20 C) groups acclimated to the respective test temperature

663

had a clearly increased mating success. In contrast to some other studies (e.g., Leroi et al. 1994; Zamudio et al. 1995; Huey and Berrigan 1996; Huey et al. 1999), there was no general advantage from developing at any specific temperature here, but only 2 temperatures were tested (see Huey and Berrigan 1996). This difference may at least in part be due to the inclusion of stressful temperatures in earlier studies (Woods and Harrison 2002). Despite differences in mating patterns between the 2 test temperatures, the BAH is fully supported at 20 C only. The outcome at 27 C, in which the 20/27 C group outperformed all others, was not predicted by any hypothesis. We do not really have an explanation for this pattern, though differences in body size are likely to be involved. Particularly small males had a disadvantage in the competition experiments, but the opposite tendency (advantage of particularly large ones) was absent. Considering the 2 groups acclimated to 27 C, the group of larger size was on average most successful (see also below). The outcome at 27 C was more difficult to predict because butterflies acclimated to 20 C should also be able to maintain high levels of activity at a warmer temperature. In contrast to temperature, the relative size of the posterior eyespot had no detectable effect on mating success in our study (at 20 C). Using an artificially increased spectrum of dorsal eyespots, Breuker and Brakefield (2002) demonstrated that female B. anynana refused to mate with males having extremely small dorsal eyespots, whereas Beldade (2002) found no effect of male dorsal and ventral eyespot size on female mate choice. A recent study suggests that the only feature under selection by females might be the males’ central white pupil (Robertson and Monteiro 2005). Clearly, more data are needed to fully elucidate the function and importance of wing patterns on mate choice in B. anynana. Body size, however, did affect mating success to some extent (see also above): also at 20 C, if nonsignificantly, the smallest males per trial (measured as forewing length) seemed to have a disadvantage (also note the trends toward a higher mating success of males with a higher pupal mass at 20 and 27 C in the GNLM). This was expected because male size is known to be important in territorial contests of butterflies and other insects (Wickmann 1985; Archer 1988; Rosenberg and Enquist 1991; Kemp and Wiklund 2001). However, temperature remains a significant and the most important predictor regardless of variation in body size. Based on the fact that body size ranks of winning males differed only at 27 C from an even distribution (and on the effect sizes in the GNLM, Table 2), body size is of subordinate importance compared with acclimation temperature. Further, the patterns found suggest a disadvantage for the smallest individual within a contest, rather than a general advantage of being large. If size had been of prime importance, the males from the low developmental temperature would be expected to be the most successful ones throughout due to their larger size. This, however, is not the case. To summarize, analyses of chill-coma recovery times showed that acclimation responses to temperature occur quickly in B. anynana. Further, mating success of B. anynana males was affected in part by acclimation to test conditions. With regard to the hypotheses outlined above, we found no support for hypotheses: 1) random matings, 3) developmental plasticity, 4) better performance at warmer developmental temperatures, 5) better performance at colder developmental temperature, and 6) mating success is determined by relative eyespot size. There was some support for the bigger is better hypothesis. Our results partly favor the BAH, but more studies that address developmental and acclimation effects are needed before any general conclusions will be possible.

664

We thank Sven Arnold, Claudia Pflicke, Sebastian Schindler and Anett Starkloff for help with the practical work, Konrad Fiedler for statistical advice, and Joseph Woodring for improving the manuscript linguistically. Comments by Naomi Pierce and 2 unknown reviewers considerably improved the quality of this paper. Financial support was provided by the German Research Council (DFG grant no. Fi 846/ 1-2 to KF).

REFERENCES Archer J. 1988. The behavioural biology of agression. Cambridge (UK): Cambridge University Press. Atkinson D. 1994. Temperature and organism size—A biological law for ectotherms? Adv Ecol Res. 25:1–58. Ayrinhac A, Debat V, Gibert P, Kister AG, Legout H, Moreteau B, Vergilino R, David JR. 2004. Cold adaptation in geographical populations of Drosophila melanogaster: phenotypic plasticity is more important than genetic variability. Funct Ecol. 18:700–706. Beldade P. 2002. Evolution and development of butterfly wing patterns—the eyespots of Bicyclus anynana [dissertation]. Leiden (the Netherlands): University of Leiden. 125 p. Bennett AF, Lenski RE. 1997. Evolutionary adaptation to temperature. Phenotypic acclimation and its evolution in Escherichia coli. Evolution. 51:36–44. Brakefield PM. 1997. Phenotypic plasticity and fluctuating asymmetry as responses to environmental stress in the butterfly Bicyclus anynana. In: Bijlsma RR, Loeschcke V, editors. Environmental stress: adaptation and evolution. Basel (Switzerland): Birkha¨user. p. 65–78. Brakefield PM, El Filali E, Van der Laan R, Breuker CJ, Saccheri IJ, Zwaan B. 2001. Effective population size, reproductive success and sperm precedence in the butterfly, Bicyclus anynana, in captivity. J Evol Biol. 14:148–156. Brakefield PM, Mazzotta V. 1995. Matching field and laboratory environments—effects of neglecting daily temperature-variation on insect reaction norms. J Evol Biol. 8:559–573. Brakefield PM, Reitsma N. 1991. Phenotypic plasticity, seasonal climate and the population biology of Bicyclus butterflies (Satyridae) in Malawi. Ecol Entomol. 16:291–303. Chown SL, Gaston KJ. 1999. Exploring links between physiology and ecology at macro-scales: the role of respiratory metabolism in insects. Biol Rev. 74:87–120. David JR, Gibert P, Moreteau B, Gilchrist GW, Huey RB. 2003. The fly that came in from the cold: geographic variation of recovery time from low-temperature exposure in Drosophila subobscura. Funct Ecol. 17:425–430. David RJ, Gibert P, Pla E, Petavy G, Karan D, Moreteau B. 1998. Cold stress tolerance in Drosophila: analysis of chill coma recovery in D. melanogaster. J Therm Biol. 23:291–299. Feder ME, Bennett AF, Huey RB. 2000. Evolutionary physiology. Annu Rev Ecol Syst. 31:315–341. Fischer K, Brakefield PM, Zwaan BJ. 2003. Plasticity in butterfly egg size: why larger offspring at lower temperatures? Ecology. 84:3138– 3147. Fischer K, Eenhoorn E, Bot ANM, Brakefield PM, Zwaan BJ. 2003. Cooler butterflies lay larger eggs: developmental plasticity versus acclimation. Proc R Soc Lond B Biol Sci. 270:2051–2056. Gibert P, Huey RB, Gilchrist GW. 2001. Locomotor performance of Drosophila melanogaster : interactions among developmental and adult temperatures, age, and geography. Evolution. 55:205–209. Gibert P, Moreteau B, Petavy G, Karan D, David JR. 2001. Chill-coma tolerance, a major climatic adaptation among Drosophila species. Evolution. 55:1063–1068. Gotthard K, Nylin S. 1995. Adaptive plasticity and plasticity as an adaptation: a selective review of plasticity in animal morphology and life history. Oikos. 74:3–17. Hoffmann AA. 1995. Acclimation—increasing survival at a cost. Trends Ecol Evol. 10:1–2. Huey RB, Berrigan D. 1996. Testing evolutionary hypotheses of acclimation. In: Bennett AF, editor. Animals and temperature: phenotypic and evolutionary adaptation to temperature. Cambridge (UK): Cambridge University Press. p. 205–237. Huey RB, Berrigan D, Gilchrist GW, Herron JC. 1999. Testing the adaptive significance of acclimation: a strong inference approach. Am Zool. 39:323–336.

Behavioral Ecology

Joron M, Brakefield PM. 2003. Captivity masks inbreeding effects on male mating success in butterflies. Nature. 424:191–194. Kemp DJ, Wiklund C. 2001. Fighting without weaponry: a review of male-male contest competition in butterflies. Behav Ecol Sociobiol. 49:429–442. Larsen TB. 1991. The butterflies of Kenya and their natural history. Oxford (UK): Oxford University Press. Leroi AM, Bennett AF, Lenski RE. 1994. Temperature acclimation and competitive fitness: an experimental test of the beneficial acclimation assumption. Proc Natl Acad Sci USA. 91:1917–1921. Lyytinen A, Brakefield PM, Lindstrom L, Mappes J. 2004. Does predation maintain eyespot plasticity in Bicyclus anynana? Proc R Soc Lond B Biol Sci. 271:279–283. Macdonald SS, Rako L, Batterham P, Hoffmann AA. 2004. Dissecting chill coma recovery as a measure of cold resistance: evidence for a biphasic response in Drosophila melanogaster. J Insect Physiol. 50:695–700. Nunney L, Cheung W. 1997. The effect of temperature on body size and fecundity in female Drosophila melanogaster: evidence for adaptive plasticity. Evolution. 51:1529–1535. Nylin S, Gotthard K. 1998. Plasticity in life-history traits. Annu Rev Entomol. 43:63–83. Prosser CL. 1991. Comparative animal physiology. New York (NY): John Wiley & Sons Inc. Rako L, Hoffmann AA. 2006. Complexity of the cold acclimation response in Drosophila melanogaster. J Insect Physiol. 52:94–104. Robertson KA, Monteiro A. 2005. Female Bicyclus anynana butterflies choose males on the basis of their dorsal UV-reflective eyespot pupils. Proc R Soc Lond B Biol Sci. 272:1541–1546. Rosenberg RH, Enquist M. 1991. Contest behavior in Weidemeyer admiral butterfly Limenitis weidemeyerii (Nymphalidae)—the effect of size and residency. Anim Behav. 42:805–811. Scheiner SM. 1993. Genetics and evolution of phenotypic plasticity. Annu Rev Ecol Syst. 24:35–68. Scheiner SM, Berrigan D. 1998. The genetics of phenotypic plasticity. VIII. The cost of plasticity in Daphnia pulex. Evolution. 52: 368–378. Scott M, Berrigan D, Hoffmann AA. 1997. Costs and benefits of acclimation to elevated temperature in Trichogramma carverae. Entomol Exp Appl. 85:211–219. Sejerkilde M, Sørensen JG, Loeschcke V. 2003. Effects of cold- and heat hardening on thermal resistance and Hsp70 expression in Drosophila melanogaster. J Insect Physiol. 49:719–726. Sørensen JG, Kristensen TN, Loeschcke V. 2003. The evolutionary and ecological role of heat shock proteins. Ecol Lett. 6:1025–1037. Stearns SC. 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. Van’t Hof AE, Zwaan BJ, Saccheri IJ, Daly D, Bot ANM, Brakefield PM. 2005. Characterization of 28 microsatellite loci for the butterfly Bicyclus anynana. Mol Ecol Notes. 5:169–172. Via S, Gomulkiewicz R, Dejong G, Scheiner SM, Schlichting CD, Van Tienderen PH. 1995. Adaptive phenotypic plasticity—consensus and controversy. Trends Ecol Evol. 10:212–217. Wickmann P. 1985. Territorial defence and mating success in males of the small heath butterfly, Coenonympha pamphilus L. (Lepidoptera: Satyridae). Anim Behav. 33:1162–1168. Willmer P, Stone G, Johnston IA. 2000. Environmental physiology of animals. Oxford (UK): Blackwell. Wilson RS, Franklin CE. 2002. Testing the beneficial acclimation hypothesis. Trends Ecol Evol. 17:66–70. Windig JJ. 1994. Reaction norms and the genetic-basis of phenotypic plasticity in the wing pattern of the butterfly Bicyclus anynana. J Evol Biol. 7:665–695. Woods HA, Harrison JF. 2002. Interpreting rejections of the beneficial acclimation hypothesis: when is physiological plasticity adaptive? Evolution. 56:1863–1866. Zamudio KR, Huey RB, Crill WD. 1995. Bigger isn’t always better—body-size, developmental and parental temperature and male territorial success in Drosophila melanogaster. Anim Behav. 49: 671–677. Zeilstra I, Fischer K. 2005. Cold tolerance in relation to developmental and adult temperature in a butterfly. Physiol Entomol. 30: 92–95.