Australian Journal of Zoology

Transcription

1 CSIRO PUBLISHING Australian Journal of Zoology Volume 46, 1998 CSIRO Australia 1998 A journal for the publication of the results of original scientific research in all branches of zoology, except the taxonomy of invertebrates All enquiries and manuscripts should be directed to Australian Journal of Zoology CSIRO PUBLISHING PO Box 1139 (150 Oxford St) Collingwood Telephone: Vic Facsimile: Australia david.morton@publish.csiro.au Published by CSIRO PUBLISHING for CSIRO Australia and the Australian Academy of Science

2 CSIRO 1998 Australian Journal of Zoology, 1998, 46, Oviposition behaviour of post-diapause Hypolimnas bolina (L.) (Lepidoptera : Nymphalidae) in tropical Australia Darrell J. Kemp Department of Zoology, James Cook University, PO Box 6811, Cairns, Qld 4870, Australia. Abstract Observations were made on the oviposition behaviour of post-diapause adults of Hypolimnas bolina (L.) at a site in Townsville, Australia (19 15 S, E). Females most often laid one or two eggs on freshly emerged seedlings (<10 mm height) of Synedrella nodiflora (L.) (Asteraceae). Mature foodplants of S. nodiflora, Sida rhombifolia (L.), Sida acuta (Burm. f.) and Malvastrum coromandelianum (L.) (all Malvaceae) were present, and occasionally sampled by females, but were ignored as oviposition substrates. Females were present at the site from 0900 to 1500 hours, and were more persistent than their male counterparts under adverse environmental conditions (i.e. reduced temperature and solar radiation). The relatively high number and apparent selectiveness of ovipositing females in this situation indicates the importance placed on utilising fresh growth of the larval foodplant. This behaviour may serve to maximise the rate of return per unit reproductive effort of post-diapause females. Introduction Butterflies, like many insect species, provide little in the way of parental care for their offspring (although see Rothschild 1979; Nafus and Schreiner 1988). They generally produce large numbers of eggs so that some will survive to reproductive age. Although an individual female has little control over the fate of her offspring once eggs are laid (or abandoned: Chew and Robbins 1984), she can make a significant contribution to their potential survival by choosing appropriate oviposition substrates. In this sense, female butterflies attempt to place their offspring in the most hospitable environment possible, to promote offspring survival and maximise their own reproductive success (Thompson and Pellmyr 1991). This requires the ability to identify and predict when conditions will be most suitable for developing juveniles, and a degree of selectivity with regard to where and when eggs are laid. Because the growth and survival of early instar larvae is often critically dependent on foodplant quality (Scriber 1977; Scriber and Slansky 1981; Taylor 1984), it is important that eggs be deposited at the time when, upon hatching, the foodplant is of suitable condition. Suitability of hostplant to a first-instar larva generally equates to the availability of juvenile leaves, with many species favouring the growing tips or fresh regenerating shoots of their respective foodplants (e.g. Forsberg 1987; Damman and Feeny 1988; Floater 1997), which are often softer and higher in both water and nitrogen content than mature leaves (Scriber and Slansky 1981). In butterfly species that utilise annual host-plants, the optimum time for egglay may be close to the germination time of their host, especially if this foodplant germinates en masse in response to favourable environmental conditions and grows rapidly. Females of these species may need to be the most opportunistic with regard to their choice of when to begin oviposition, as the window of opportunity may be particularly small. In this study, the behaviour of the nymphalid butterfly Hypolimnas bolina (L.) was studied at an oviposition site in tropical Australia. This species, like many Australian tropical pierids and nymphalids, undergoes a reproductive dormancy in the adult stage during the winter dry season (Jones 1987). This species is known to utilise a range of herbaceous plants (Common and Waterhouse 1981) that grow typically in disturbed sites, with at least one of its tropical /ZO X/98/010000

3 452 D. J. Kemp foodplants, Synedrella nodiflora (L.) (Asteraceae), representing a fast-growing annual species with mass germination of seedlings. Here I describe the oviposition behaviour of female H. bolina immediately following the cessation of reproductive dormancy, along with an inventory of both male and female occurrence and the influence of weather on butterfly activity regimes. Methods Study area Observations were made within an area of park at Townsville (19 15 S, E) in Queensland, Australia. The study site consisted of a stretch of partially landscaped streambank, approximately 300 m in length, where several larval foodplants, S. nodiflora, Sida rhombifolia (L.) and Sida acuta (Burm. f.) (Malvaceae), grew throughout. Mature plants of Malvastrum coromandelianum (L.) (Malvaceae) were present at the site, and last-instar immatures of H. bolina have developed successfully on this species under laboratory conditions (Kemp, unpublished data). The relative biomass (visually estimated) of each foodplant species at the study site was S. nodiflora seedling (<10 mm height; 35%), S. nodiflora mature (30%), S. acuta (20%), M. coromandelianum (10%) and S. rhombifolia (5%). A relatively open tree canopy (50 70% closure) at a height of approximately 30 m provided mostly shaded or dappled light conditions at ground level. Daily activity regimes Sampling of daily butterfly activity was based on a transect count method modified from Pollard (1977). A single transect of approximately 180 m was established and walked hourly (where possible) on eight consecutive days (10 17 September 1997). The presence of individual butterflies 10 m either side of the transect was recorded, along with notes regarding their sex, immediate behaviour, colour form (using the classifications of Clarke and Sheppard 1975), general condition, and the distance along the transect at which they were encountered (to the nearest metre). The occurrence and outcome of any interactions between individuals (with both conspecifics and heterospecifics) was noted. Environmental variables including temperature, wind speed, cloud cover, and light conditions were recorded at the time of sampling. Temperature was measured with a standard mercury thermometer, suspended 0.5 m above the ground in a shaded position, to the nearest degree (C); wind speed was classified as none, slight, moderate, steady, strong or extreme; cloud cover was visually estimated to the nearest 20%; and ambient light conditions were classified as low, intermediate or high. Wind speed and ambient light data were treated as numerical ordinal variables in subsequent analyses. Care was taken to avoid disturbing individuals along the transect, as preliminary observations revealed that startled butterflies often ceased or at least modified their immediate activity. Speed of travel of the observer along the transect was therefore slow at all times, and the clothing worn on sampling occasions was consistent and drab. Individual butterflies were identified where possible on the basis of distinctive wing wear (Rutowski 1992) and colour patterning (females only) to reduce the likelihood of recounting the same butterfly on each transect occasion. Preliminary mark and recapture was attempted; however, subjects appeared to quickly exit the area (a result also observed by Rutowski 1992). Although wing wear and colour pattern allowed effective identification of most individuals, not all butterflies could be consistently identified (especially fast-moving males). Therefore, transect counts represent a correlate of abundance rather than actual counts of discrete individuals (refer to Pollard 1977). Focal sampling Individual females (n = 16) within the sampling area were randomly selected and their behaviour observed for timed intervals of up to 20 min each in order to ascertain general oviposition behaviour, rates, and favoured substrates. This focal sampling was conducted during the periods between transect counts, in the middle of the day when butterflies were most active (i.e hours). Focal individuals were selected by drawing a random number corresponding to a length in metres along the established transect, and selecting the female closest to that point. This butterfly was then watched from a distance of approximately 1.5 m to determine the timing and the precise position of oviposition. All places where butterflies alighted and arched their abdomens were inspected for the presence of eggs. Further untimed observations of 14 females were included in the analysis of oviposition substrates and the number of eggs laid at each ovipositional event.

4 Oviposition of post-diapause Hypolimnas bolina 453 Statistical procedures Firstly, to determine the spatial distribution of butterfly counts, the overall transect length was divided into nine discrete segments of 20 m length. Counts of males and females in each segment were tallied across all sampling occasions and compared using a 2 9 contingency table. Owing to there being no significant difference between the spatial distribution of male and female counts 2 (c 2 = 11.33, d.f. = 8, P > 0.05), data were pooled and compared to uniformly expected frequencies for each segment (total butterfly counts divided by nine). This was done to determine whether the distribution of butterflies was uniform along the transect. Data were compared using Chi-square analysis. To examine whether butterfly counts were related to environmental conditions, census data from 1100, 1200 and 1300 hours on each day were averaged to provide a mean measure for midday butterfly activity and weather. Variables describing mean butterfly activity and environmental conditions for each day were then compared using linear correlation analysis (Pearson product-moment) to determine relationships within and between these sets. Due to significant (P < 0.05) covariance between the variables describing temperature, cloud cover, and ambient light (Table 1), these were combined using principle components analysis into a single variable (PC1). The extracted component (PC1) had an eigenvalue of 2.6, and hence explained approximately 87% of the variance of the original three (refer to Table 1 for component loadings of individual variables). PC1 was thereafter used as an integrated measure of ambient light and temperature conditions. Statistical analyses were conducted using STATISTICA, and all sample means throughout this paper are quoted ± 1 standard error. Table 1. Pearson product correlations between environmental variables, and their correlation (component loading) with the extracted temperature-light principal component PC1 *, marked correlations are significant (P < 0.05); n = 8 for all comparisons Ambient light Cloud cover Ambient temp. PC1 Ambient temp. r = 0.71* r = 0.88* r = 0.93* Cloud cover r = 0.88* r = 0.97* Ambient light r = 0.90* Wind speed r = 0.10 r = 0.20 r = 0.53 r = 0.30 Results Butterfly activity A total of 201 females and 88 males were counted during sampling (Table 2). Observed females were of the colour forms nerina, naresi, euploeoides, and various intermediates between these forms classified as euploeoides nerina and nerina pallascens (as per Clarke and Sheppard 1975). The largest single transect count of H. bolina females was 13, representing an unusually high number of individuals (compared to the few recorded along a larger transect in a nearby area by Rutowski 1992). Analysis of the spatial distribution of counts (Table 3) indicated that they were not uniformly distributed across all transect segments (c 2 = 43.89, d.f. = 8, P < 0.05). Certain segments of the transect were disproportionately favoured by both sexes. Female activity was less related to weather conditions than male activity. Whilst male activity throughout the census area was significantly (at a = 0.10) negatively correlated with PC1 (r = 0.67, n = 8, P = 0.07), female counts were independent of this variable (r = 0.19, n = 8, P = 0.67). Higher values of PC1 were correlated with low temperatures, low light levels and high degrees of cloud cover (Table 1), hence males preferred to fly in periods with relatively high temperatures and solar radiation. This corresponds with general observation males were rarely seen under adverse conditions, and never seen under the coolest conditions when only few females were present. Neither male counts (r = 0.00, n = 8, P = 1.0) nor female counts (r = 0.09, n = 8, P = 0.83) were significantly related to wind speed. Temperature at the time of sampling

5 454 D. J. Kemp Table 2. Frequencies and relative proportions of H. bolina counts along the established assessment transect Colour form Count frequency Proportion of total counts (%) Male Female nerina euploeoides nerina nerina pallascens naresi euploeoides Total Table 3. Results of transect survey and analysis of the spatial distribution of male and female H. bolina Observed frequencies Male Female Total Segment 1 20 m Segment m Segment m Segment m Segment m Segment m Segment m Segment m Segment m Total Expected frequencies (i.e. uniform distribution) c 2 value P value ranged from 20 to 29 C (mean = ± 0.36 C), light levels ranged from low to high, and cloud cover ranged from 0 to 100% (mean = 57 ± 7%). The relative homogeneity of female activity between sample days (due to the minimal effect of weather) enabled these to be pooled to illustrate daily activity regimes (Fig. 1). This was not attempted for male activity due to the confounding effect of weather, and using clear days only (as per Rutowski 1992) resulted in too few data points. Female behaviour All females sighted in the study area were engaged in oviposition behaviour. This behaviour appeared very consistent across females and consisted of periodic bursts of egg laying interspersed with periods of basking (with wings open) or resting (with wings closed). Most activity was centered on or near the ground, with individuals rarely rising above several metres. Observed females all began by basking in sunlit patches for a period of time before taking flight and assuming a slow gliding posture low to the ground ( mm) around their immediate area. Butterflies would then alight on the ground or directly onto the S. nodiflora carpet and

6 Oviposition of post-diapause Hypolimnas bolina 455 Mean female count Fig. 1. The activity of H. bolina females as indicated by transect counts during each hourly transect. The bars represent ± 1 standard error of the sample mean, and the number of sampling days is given above each point. begin to search for a suitable oviposition substrate. Freshly emerged S. nodiflora seedlings (i.e. <10 mm height) were most preferred, with this substrate utilised by 97% of all females observed to oviposit (n = 30 females), and receiving 85% of all eggs laid (n = 205 eggs). Additional substrates included grass blades (4% of total eggs laid), leaf litter (4% of eggs), dry twigs (3% of eggs) and semi-mature S. nodiflora plants (3% of eggs). Where apparent mistakes were made, and eggs deposited on non-foodplants, the substrate was always near ground level and in close proximity (i.e. <50 mm) to S. nodiflora seedlings. Mature S. nodiflora plants and both S. rhombifolia and S. acuta appeared unsuitable for oviposition, as focal females readily contacted these but did not oviposit. No female contact was witnessed with M. coromandelianum. Once grounded, ovipositing butterflies would orient themselves sequentially in several different directions and walk small distances (~ mm), all the while tapping with their antennae and likewise drumming with their forelegs. Once a suitable substrate was located, individuals positioned themselves so that their arched abdomen contacted the underside of the chosen substrate, then laid their eggs. Females most often deposited eggs singly (69% of all ovipositional events) or in twos (25% of all events). The mean number of eggs laid per ovipositional event by each focal individual (n = 30 females) was 1.56 ± 0.13 eggs, and no individual was observed to lay more than four eggs at once during focal sampling. On a separate occasion a female (form nerina) flew into the area and deposited 13 eggs on a mature S. nodiflora plant (six of which later proved to be infertile). This individual had been caught and tagged (during a previous study) at an overwintering site in Townsville on 8 August 1997, and did not appear significantly older than other females on site (judging from general wing condition). Whilst searching on the ground for oviposition sites, females continuously displayed their dorsal coloration by slowly alternating their wings. This display continued up to the precise moment of egglay, when the wings were closed until completion. If individuals with alternating wings were approached by another butterfly, they would always respond by immediately closing

7 456 D. J. Kemp their wings. Grounded females responded with closed wings most often to the investigations of conspecific males, although other conspecific females and heterospecifics (Euploea core, Tirumala hamatus; both Nymphalidae, and Eurema spp.; Pieridae) also elicited this response. On numerous occasions the alternating wing display of a grounded female attracted a conspecific male, although no subsequent interaction ever led to copulation. Rates of egglay Females (n = 16) were watched continuously for a total of 174 min (mean = ± 1.60 min), during which time 138 eggs were laid (mean = 8.62 ± 1.62 eggs). Rates of egglay for focal individuals ranged from 0.2 to 3.0 eggs per min, providing a mean rate of 1.07 ± 0.21 eggs min 1. Only three females were observed for the total pre-designated time of 20 min, with the remainder either departing the area of their own accord prior to completion of sampling, or being interrupted by a conspecific male (as observed by Odendaal et al in Euphydryas anica; Nymphalidae). Focal sampling was biased to when females were actively ovipositing and may overestimate the functional rate of egglay by females of this species over longer periods. Male female interactions A total of 30 male female interactions were observed in their entirety during the sampling period, with none leading to copulation. Females consistently and effectively rejected males by one of three methods. The most common method (n = 20 observations) was to simply perch on the ground with wings closed. This posture was adopted most often by females that were detected either on the ground or flying near the ground. If detected on the wing, females instigated ascending flights (n = 8), whereby they flew directly to the top of the trees and beyond at full speed with the male in hot pursuit. Sometimes females would perch in the tree canopy; at other times they would continue flight until out of sight. The third method was observed only twice, and consisted of females suddenly dropping out of flight into the foliage of a shrub or groundcover. Discussion There are few quantitative data available on the oviposition patterns of butterflies of the genus Hypolimnas in Australia, and none concerning H. bolina immediately following reproductive dormancy. Rutowski (1992) observed female H. bolina during the wet season laying single eggs on the undersides of S. rhombifolia leaves in Townsville. He did not record the age of selected oviposition substrates or the occurrence of other foodplant species. In Guam, Nafus (1993) reported that this species lays eggs singly or in small batches of up to 20 eggs, and Common and Waterhouse (1981) suggest that eggs are laid in irregular batches beneath the leaves of the foodplant. Whilst females observed in this study almost exclusively laid eggs singly or in twos on the underside of S. nodiflora seedlings, one isolated instance of egg clustering was observed. These results, coupled with (the few) published accounts, suggest that clutch size in this species has the potential to vary significantly. It is interesting to note that two closely related tropical congeners, H. anomala and H. antelope, always lay their eggs in large batches, which are guarded thereafter by the female butterfly (Rothschild 1979; Nafus and Schreiner 1988; Schreiner and Nafus 1991). The fact that most eggs were laid on a single foodplant species (S. nodiflora) suggests that, in the studied habitat at least, H. bolina represents a polyphagous species with a monophagous strategy (Wiklund 1977). This apparent monophagy may merely reflect the relative availability of each foodplant species at the study site, and not active discrimination by the females involved. Although not specifically quantified, S. nodiflora was clearly most abundant of all species growing in the study area, and would have been most accessible to ovipositing females in search of suitable substrate. Irrespective of any hostplant specificity, females did appear to discriminate clearly in favour of newly germinated seedlings amongst the available specimens of

8 Oviposition of post-diapause Hypolimnas bolina 457 S. nodiflora. Stereotyped searching behaviours, such as flying low, landing directly onto the ground, and walking along the ground, served to consistently place females in direct contact with recently emerged seedlings growing at ground level, rather than the foliage of more mature plants growing mm above ground. Further evidence for this is provided by the fact that almost all of the eggs not placed on seedlings were deposited onto nearby ground-layer substrates (see Chew and Robbins 1984 for discussion of oviposition mistakes in butterflies). Where focal butterflies did contact mature plants they rarely oviposited, with only a small number (3%) of eggs being deposited on semi-mature plants. Similar behaviour has been observed in Pieris rapae (Pieridae; Jones 1977), where butterflies were more likely to alight on larger, older plants than young plants, but once alighted, the probability of oviposition was much higher on younger plants. The advantages of utilising freshly germinated seedlings for earlyinstar feeding may relate to nutritional benefits (Scriber and Slansky 1981; Taylor 1984; Damman 1987; Ravenscroft 1994), higher leaf water content (Scriber 1977; Scriber and Feeny 1979; Floater 1997) or availability of softer leaves (Scriber and Feeny 1979; Coley 1983). Additionally, selective oviposition on seedlings may be a way of maximising the potential availability of foliage for the life of the larva. Because S. nodiflora germinates en masse and is fast growing, potentially large amounts of foliage will be locally available to the growing larva at all life stages, especially for eggs laid at the beginning of the tropical wet season. Regardless of which of these proximate factor(s) is most important, it seems that female H. bolina adopt this specific oviposition strategy in order to maximise their reproductive success. The hypothesis that ovipositing H. bolina target freshly emerged seedlings of their foodplant generates the key prediction that females must be opportunistic with regard to when eggs are laid; that is, they must posses the ability to delay oviposition until conditions are favourable for germination. Female H. bolina have this ability since they undertake reproductive dormancy in the adult stage and apparently respond quickly to environmental cues (see below). Up to ten days prior to this study all females were behaviourally dormant and positioned at local overwintering sites. Not a single female was present in the transect area, and dissection of several females revealed that their ovaries were regressed (Kemp, unpublished data). Following the mass germination of S. nodiflora (an event triggered by rainfall), not a single female was recorded from any of the recognised overwintering sites either in Townsville or Cairns (Kemp, unpublished data). Butterflies had resumed their reproductive activity at this time. An interesting finding of this study was that females appeared to be less sensitive to adverse weather conditions than males. Generally in butterflies, intrasexual male competition is expected to result in greater tolerance for adverse conditions in this sex, as demonstrated by Thymelicus lineola (Hesperiidae; Pivnick and McNiel 1986, 1987). That female activity exceeded male activity suggests either the presence of different thermal requirements between the sexes, or greater selective pressure on female butterflies to remain active during suboptimal conditions. Nonetheless, climatic conditions were important to ovipositing females, as indicated by their diurnal activity regime and the incidence of basking during oviposition. It is also likely that microclimatic conditions (e.g. available sunlight) accounted for the similar non-uniform distribution of both male and female activity along the transect, although this point requires further investigation. Episodes of sexual harassment were common along the transect, and observed methods of refusal were similar to those outlined for H. missipus by Stride (1956, 1958) and Edmunds (1969). The fact that no observed male female interaction concluded with copulation suggests that most ovipositing females present along the transect were unreceptive to matings. This point is reinforced by the suggestion of Rutowski (1992) that females of this species mate only once in their lifetime. If this is true, ovipositing females should attempt to minimise their contact with males, as the process of rejection consumes time that could otherwise be spent laying eggs. On this basis, it is not known why females searching on the ground did not behave in the most cryptic manner possible, but chose to display their relatively conspicuous dorsal colouration. This behaviour presented a cost to females in terms of detection by males and associated loss of functional oviposition time.

9 458 D. J. Kemp In summary, this study has reported field observations of the natural behaviour of H. bolina individuals immediately following reproductive dormancy in tropical Australia. Whilst limited, these observations provide valuable information on the natural behaviour of this species and key directions for useful further study. Since during this study S. nodiflora seedlings were used almost exclusively by females as oviposition substrates, experimentation is now underway to investigate the effect of foodplant age and species upon aspects of offspring fitness. This type of research may prove highly relevant to the causes and timing of reproductive seasonality in tropical species. This study has also raised interesting questions about clutch size variation in this species, and investigation into the ecological correlates of such variation may constitute a fruitful line of further inquiry. Acknowledgments Thanks are due to Professor R. E. Jones, Dr J. E. Seymour, and an anonymous reviewer for critically reviewing the manuscript and providing helpful comments. I would also like to acknowledge the assistance of Associate Professor B. R. Jackes in plant identification, and the Townsville Council Parks division for providing permission to conduct research and collect specimens. This research was supported by an Australian Postgraduate Research Award. References Chew, F. S., and Robbins, R. K. (1984). Egg laying in butterflies. In The Biology of Butterflies. (Eds R. I. Vane-Wright and P. R. Ackery.) pp (Academic Press: London.) Clarke, C., and Sheppard, P. M. (1975). The genetics of the mimetic butterfly Hypolimnas bolina (L.). Philosophical Transactions of the Royal Society of London (B) 272, Coley, P. D. (1983). Herbivory and defensive characteristics of tree species in a lowland tropical rainforest. Ecological Monographs 53, Common, I. F. B., and Waterhouse, D. F. (1981). 'Butterflies of Australia.' (Angus and Robertson: Sydney.) Damman, H. (1987). Leaf quality and enemy avoidance by the larvae of a pyralid moth. Ecology 68, Damman, H., and Feeny, P. (1988). Mechanisms and consequences of selective oviposition by the zebra swallowtail butterfly. Animal Behaviour 36, Edmunds, M. (1969). Evidence for sexual selection in the mimetic butterfly Hypolimnas missipus L. Nature 221, 488. Floater, G. J. (1997). Rainfall, nitrogen and host plant condition: consequences for the processionary caterpillar, Ochrogaster linifer. Ecological Entomology 22, Forsberg, J. (1987). Size discrimination among conspecific hostplants in two pierid butterflies; Pieris napi L. and Pontia daplidice L. Oecologia 72, Jones, R. E. (1977). Movement patterns and egg distribution in cabbage butterflies. Journal of Animal Ecology 46, Jones, R. E. (1987). Reproductive strategies for the seasonal tropics. Insect Science and its Application 8, Nafus, D. M. (1993). Movement of introduced biological control agents onto nontarget butterflies, Hypolimnas spp. (Lepidoptera : Nymphalidae). Environmental Entomology 22, Nafus, D. M., and Schreiner, I. H. (1988). Parental care in the tropical nymphalid butterfly, Hypolimnas anomala. Animal Behaviour 36, Odendaal, F. J., Turchin, P., and Stermitz, F. R. (1989). Influence of host-plant density and male harassment on the distribution of female Euphydryas anica (Nymphalidae). Oecologia 78, Pivnick, K. A., and McNeil, J. N. (1986). Sexual differences in the thermoregulation of Thymelicus lineola adults (Lepidopter a : Hesperiidae). Ecology 67, Pivnick, K. A., and McNeil, J. N. (1987). Diel patterns of activity of Thymelicus lineola adults (Lepidoptera : Hesperiidae) in relation to weather. Ecological Entomology 12, Pollard, E. (1977). A method for assessing changes in the abundance of butterflies. Biological Conservation 12, Ravenscroft, N. O. M. (1994). The ecology of the chequered skipper butterfly Carterocephalus palaemon in Scotland. II. Foodplant quality and population range. Journal of Applied Ecology 31, Rothschild, M. (1979). Female butterfly guarding eggs. Antenna 3, 94.

10 Oviposition of post-diapause Hypolimnas bolina 459 Rutowski, R. L. (1992). Male mate-locating behaviour in the common eggfly, Hypolimnas bolina (Nymphalidae). Journal of the Lepidopterists Society 46, Schreiner, I. H., and Nafus, D. M. (1991). Evolution of sub-social behaviour in the nymphalid butterfly Hypolimnas anomala. Ecological Entomology 16, Scriber, J. M. (1977). Limiting effects of low leaf-water content on the nitrogen utilisation, energy budget, and larval growth of Hyalophora cecropia (Lepidoptera : Saturiidae). Oecologia 28, Scriber, J. M., and Feeny, P. (1979). Growth of herbivorous caterpillars in relation to feeding specialisation and to the growth form of their foodplants. Ecology 60, Scriber, J. M., and Slansky, F. (1981). The nutritional ecology of immature insects. Annual Review of Ecology and Systematics 26, Stride, G. O. (1956). On the courtship behaviour of Hypolimnas missipus L., (Lepidoptera, Nymphalidae), with notes on the mimetic association with Danaus chrysippus L. (Lepidoptera, Danaidae). British Journal of Animal Behaviour 4, Stride, G. O. (1958). Further studies on the courtship of African mimetic butterflies. Animal Behaviour 6, Taylor, M. F. J. (1984). The dependence of development and fecundity of Samea multiplicalis on early larval nitrogen intake. Journal of Insect Physiology 30, Thompson, J. N., and Pellmyr, O. (1991). Evolution of oviposition behaviour and host preference in Lepidoptera. Annual Review of Entomology 36, Wiklund, C. (1977). Oviposition, feeding and spatial separation of breeding and foraging habitats in a population of Lepidea sinapsis (Lepidoptera). Oikos 28, Manuscript received 24 March 1998; accepted 20 November