Dr. Klaus Fischer: Life-history plasticity in the life-cycles of the copper butterflies Lycaena hippothoe and L. tityrus (Lepidoptera: Lycaenidae) (2000)

Abstract

The present study investigates plasticity in the life-cycles of the two declining copper butterflies Lycaena hippothoe and L. tityrus. Variability in life-history traits (and thus the ability to respond plastically) comprises a major prerequisite for the survival in fluctuating or changing environments. To analyse life-history plasticity the concept of "reaction norms" is used. Hypotheses were predicted from life-history theory, which aims at explaining differences in traits of reproduction, growth and development as adaptive in the sense that they have been shaped by natural selection. This approach is founded on optimality models. As a powerful means in this context I used a combined approach, including experimental and comparative aspects.

Rearing experiments over a range of constant temperatures considered four populations of copper butterflies: L. hippothoe hippothoe from western Germany (Westerwald), L. hippothoe eurydame from the central Alps (Senales valley), L. hippothoe sumadiensis from western Hungary (Örség), and L. tityrus subalpinus from the Senales valley as well. Because the use of optimality approaches relies on a profound knowledge of the organism under study, field work was carried out in addition to the laboratory experiments. The following results were obtained:

A. Population biology and behavioural ecology of L. hippothoe

A mark-recapture analysis showed a strong decline in butterfly numbers during the study period (1995-1999). Neither land-use practices nor larval or adult mortality could be accounted for this development. The hypothesis is advanced that the observed decline basically reflects a series of flight periods with adverse weather conditions. These should constrain the time available for feeding and oviposition. Based on the assumption of a high importance of adult-derived resources for a successful reproduction, a lowered realised fecundity is regarded as the proximate reason for the population decline. Evidence supporting this hypothesis comes from experiments on the reproductive biology (see below), from the mate-location behaviour (see below) and a strong correlation between L. hippothoe numbers and the mean cloud cover per day during the previous flight period. These findings may be important for the conservation of this and other species, as a reduced availability of nectar sources caused by modern agricultural techniques can play an important role for regional declines.

Effects of adult feeding on fecundity. When fed highly concentrated sucrose solution, females laid significantly more eggs (mean 464) than those individuals given water only (mean 65). Longevity was also three to five times greater. Hence, fecundity of L.hippothoe butterflies depends far more on adult-derived resources than in other nectar-feeding butterflies for which comparable data exist.

Mate-location behaviour.L. hippothoe males exhibit aggressive territorial behaviour. For mate-location they use a combined strategy of perching and patrolling, with the respective portion of time allocated to each behaviour depending on the weather conditions. This strategy seems to be connected with enhanced mating opportunities and relatively large territories. Territories are located within rich aspects of flowering nectar plants (resource-based territoriality). Other parameters (such as vegetation height, larval host-plant density, and distance to the next bushes) were insignificant in this context. In this study, defending nectar sources in butterflies could be confirmed for the second time only. The evolution of such a rare territorial system is attributed to the widely dispersed occurrence of receptive females, which only concentrate at rich nectar sources. Given the strong dependence of the reproductive output on adult resources it is a straightforward strategy for males to monopolise nectar sources in spite of their high investments.

B. Life-history plasticity in L. tityrus subalpinus

Sex-related reaction norms. To analyse sexual differences in reaction norms trade-offs were deliberately enforced in directly developing individuals of L. tityrus. Therefore, I tried to drive growth rates, using a range of temperatures, close to their physiological upper limit and thus disclosing otherwise potentially concealed responses. At all temperatures larval development time of males was significantly shorter compared to females, as was predicted by protandry theory. Development time strongly depended on temperature, leading in accordance with a central assumption in life-history theory to a larger size at low temperatures, and vice versa.

However, sexes responded differently to time constraints imposed by a short development time. Despite the shorter larval development of males, both sexes achieved similar body sizes at lower temperatures, because males avoided a reduction in weight due to plastic growth. At high temperatures, in contrast, males were forced to make a trade-off in which they favoured early emergence over large size, leading to a dramatic weight loss (about 38.4 %). Weight of females, however, remained similar throughout showing no trade-off. These different reaction norms (shown here for the first time) reflect divergent selective pressures acting on males and females, which can be explained in relation to the reproductive system. The strong selection for early emergence in males is likely to be due to monandry, discrete non-overlapping generations, and territoriality, because prior ownership of a territory seems to be a major advantage for successful reproduction. On the other hand, the preference of females for large body size was expected due to the generally close relationship between this trait and fecundity.

A comparison between direct and indirect development (at 25 °C) revealed a broad conformity in sex-related reaction norms between both developmental pathways. The only major difference concerned higher pupal (25 %) and adult (28 %) weights (owing to a longer larval development) in connection with diapause development. This does not at all change the extent of the sex-related dimorphism. Hence, males were not able to avoid a weight loss relative to the females, in spite of the longer development (as compared to direct development). This is explained as a consequence of the persistent selection for protandry. In the direct pathway, males are able to accumulate a developmental advantage during the whole development time, whereas they rely in this respect on the post-hibernation part only in the diapause cycle.

Effects of starvation. A starvation period of two days during the last larval instar led to longer development times. Males appeared to suffer more from food shortage, because they invested more time in compensatory feeding as compared to females. However, they compensated the longer larval time through a faster development in the non-growing pupal phase. Consequently, total (larval and pupal) development time was exactly two days longer in both sexes, although this was due to different mechanisms.

Effects of increased leaf nitrogen in natural food-plants. In accordance with previous studies, high nitrogen levels in the natural food-plant increased larval growth rates and concomitantly decreased development times. However, because of high pupal (and larval) mortality (overall 73.0 %) as well as a reduction in adult size (by ca. 8 %) this was, overall, not beneficial to the butterflies. Thus, the results were not consistent with the broad interspecific trend that insect herbivore performance is positively correlated with leaf nitrogen. These findings undermine the general applicability of the nitrogen limitation hypothesis. As the detrimental effects were largely confined to the pupal and adult stages, results obtained from the larval phase only may not yield reliable results and must therefore be interpreted with caution. If negative effects of nitrogen enrichment will be found more frequently in declining species inhabiting nutrient poor grassland, this would have major implications for the conservation of those species.

C. Life-history plasticity in L. hippothoe

Egg size and fecundity. A comparison between the three populations of L. hippothoe revealed strong evidence for trade-offs between egg size and egg number as well as between body size and fecundity. L. h. eurydame, which is of the same size as L. h. hippothoe, laid larger but significantly fewer eggs compared to the population from western Germany (mean 245 vs. 464). In L. h. sumadiensis, however, the small egg number (mean 244) in spite of a small egg size seemed to be caused by the small body size of these animals. Furthermore, a selective advantage of large eggs could be confirmed. The significantly larger eggs of L. h. eurydame (as compared to the other populations) showed the highest hatching rates at all temperatures from 15 to 30 °C. The selective advantage of large eggs in Lepidoptera remained unclear for a long time, because most studies on butterflies to date failed to indicate positive correlations between egg size and quite a number of life-history traits. Based on the present results it is now possible to explain the enhanced allocation of reproductive reserves to early eggs (egg size and oviposition rate were frequently found to decrease over oviposition period). As a straightforward explanation it is suggested that late eggs contribute less to offspring production than earlier ones owing to mortality of reproducing females.

Life-history plasticity on a population level. Pronounced differences in reaction norms and life-cycles were found between the three populations of L. hippothoe, which (based on optimality models) can be interpreted as being local adaptations. The life-cycles of the two univoltine populations L. h. hippothoe and eurydame appeared to be very similar. (Partial) direct development could be induced at high temperatures and long photoperiod only. As was predicted in connection with the short vegetation period in the higher altitudes of the Alps, there was a much reduced tendency to develop directly in L. h. eurydame compared to L. h. hippothoe. Furthermore, evidence for a facultatively biennial life-cycle and a shorter preimaginal development time at ecologically relevant temperatures (up to 15 days at 15 °C) are attributed as being adaptations to the adverse climatic conditions in the alpine environment. In spite of the shorter development, no differences in body size occurred due to a plastic adjustment of growth rates in L. h. eurydame. Thus, the central assumption of a trade-off between development time and size in life-history theory was not supported. This result highlights the triangular nature of the relationship between size, development time and growth rate. The apparent importance of a large size in the alpine environment is explained by thermal constraints.

Compared to the above mentioned populations, L. h. sumadiensis showed completely different reaction norms. Under laboratory conditions, almost all individuals developed directly into adults at all temperatures. This is the only bivoltine population of L. hippothoe, obviously as a consequence of the favourable climate in lowland western Hungary. Time constraints imposed by the realisation of a second generation a year demand specific adaptations, which could be found as a strongly reduced development time compared to the other two populations (up to 10 days compared to L. h. eurydame). This was partly achieved by the facultative omission of one (out of five) larval instar, partly by a reduction (15-20 %) of the adult size.

Sex-related reaction norms. The similar life-cycles of L. h. sumadiensis and L. tityrus led to strikingly convergent reaction norms including an increasing sex-dimorphism with increasing temperature (see above). In contrast, no changes in the relationship between male and female weight could be detected in either of the univoltine populations over the range of temperatures. Body size was found to be a fairly canalised trait in both sexes of L. h. hippothoe and eurydame. These results confirm that a trade-off between body size and development time can only be expected in connection with severe time constraints. The convergent results between different species (L. h. sumadiensis and L. tityrus) as well as the geographic diversity within one species (L. hippothoe) clearly favour an adaptive (ecological) over a constraint (phylogenetic) explanation.

D. On the costs of a fast development

An increasingly large body of evidence indicates that growth rates can vary adaptively. Frequently, they were not found to be maximised but rather optimised by natural selection. This suggests the existence of costs attached to a faster development. As in many insects males typically exhibit higher growth rates than females (due to selection for protandry), strictly protandrous species might provide particularly useful models to investigate possible costs related to high developmental rates. Higher proportional weight losses at metamorphosis (ca. 62 compared to 58 %; for which passive effects, based on differences in pupal size, could be ruled out) as well as a reduced adult longevity (10 vs. 14 days) of males compared to females provide evidence for such costs. However, the latter can be explained as being an adaptive consequence of the mating opportunities as well. A direct correlation between weight loss at metamorphosis and larval growth rates could be established for the first time, further supporting the notion that rapid growth carries some physiological costs.