Reproductive expenditure affects utilization of thoracic and abdominal resources in male Pieris napi butterflies


†Author to whom correspondence should be addressed. E-mail:


  • 1One route to a deeper understanding of the life history of organisms is to identify how resources are acquired and used for reproduction. This may be particularly relevant for insects such as nectar-feeding butterflies, which change diets during the life cycle.
  • 2Nitrogenous resources used for reproduction in nectar-feeding butterflies come principally from the juvenile diet and are stored in abdominal reserves. Juvenile resources are also used to build the soma of the adult. Consequently somatic and reproductive investment will trade off and constrain the amount of resources available for egg and spermatophore production. Recent findings show a pronounced decrease in thorax resource content and suggest that nitrogen from somatic tissue can be reallocated to reproduction and thus alleviate the resource limitation upon reproduction.
  • 3In this paper we test the prediction that the observed decrease in thorax nitrogen content is related to the expenditure of resources in spermatophore production in Pieris napi males. By comparing thorax nitrogen content over the life span between males mated 0–3 times, we show that mating history is an important factor in explaining the observed decrease in nitrogen content.
  • 4These results support the hypothesis that thoracic resources are used for reproduction in male nectar-feeding butterflies.


One of the fundamental factors in determining the life history of organisms is the pattern of acquisition and expenditure of resources (e.g. Boggs 1981a; Roff 1992; Stearns 1992). Resources are required both for reproduction and for somatic functions such as growth, metabolism and maintenance. In insects that change diets between the different stages in the life cycle, specific resources may be available only to the juvenile and thus have to be stored for future needs (Boggs 1981a, 1997). Consequently, resources that are scarce during the reproductive stage may limit the reproductive output.

This situation seems to exist in nectar-feeding butterflies, which have a nutritionally balanced juvenile diet, but a carbohydrate-biased adult diet, deficient in amino acids and proteins (Baker & Baker 1973; Mattson 1980; Boggs 1981a, 1987; O’Brien, Fogel & Boggs 2002). Since females of many species emerge as adults with few mature eggs, i.e. low ovigeny index (Jervis, Boggs & Ferns 2005), the availability of resources during the reproductive stage has a large impact on the reproductive output (egg and spermatophore production). During metamorphosis the juvenile resources are partitioned between somatic tissue and reproductive reserves (Boggs 1981a; Wheeler et al. 2000; Pan & Telfer 2001). Since amino acids are scarce or unavailable in the adult diet, but important for reproduction, the allocation of resources between these pools strongly influences the reproductive potential (Boggs 1981a). With this lifestyle there should be strong selection on factors influencing the availability of the limiting resources. One such mechanism, of potential importance in butterflies, and the focus of this paper, is reallocation of nitrogenous resources from somatic tissue to production of eggs and spermatophores.

Studies on nectar-feeding butterflies have shown that the resource content of the thorax decreases during the reproductive period, when measured as a change both in mass (Karlsson 1994; Norberg & Leimar 2002; Stjernholm, Karlsson & Boggs 2005) and total nitrogen content (Karlsson 1998; Stjernholm & Karlsson 2000). Furthermore, Karlsson (1998) showed that in order to balance the output of nitrogen in eggs in Pieris napi (L.) females, resources from the thorax need to be included in the resource budget. These studies indicate that nitrogenous resources from the thorax are reallocated to egg and spermatophore production, thereby supplementing the reproductive reserves.

The function of the reduction in thorax resource can be examined by focusing on the relationship between reproductive expenditure, throughout this paper defined as incorporation of nitrogenous resources in eggs or spermatophores, and the decrease in thorax resources. A strong relationship between incorporation of nitrogenous resources in reproduction and the decrease in thorax nitrogen content would support the hypothesis that resources from the thorax are used to improve the reproductive potential. Karlsson (1998) investigated the relationship between egg production and the decrease in thorax nitrogen content in female P. napi butterflies. In his study thorax nitrogen content decreased with increasing reproductive expenditure, but since egg production and life span covaried, presumably because of the influence of nuptial gifts on both reproductive investment and longevity, the effect of the reproductive investment on the ‘use’ thorax resource was not clear. Two different scenarios can account for this result. In the first, the thorax nitrogen content decreases at a constant or age-related rate, irrespective of reproductive expenditure, thereby spuriously creating a correlation between reproductive expenditure and egg production. In the second scenario the decrease in thorax resource levels is functionally related to the investment of resources in reproduction. Under this scenario, the decrease in thorax mass would be large in individuals with high reproductive expenditure and small in individuals with low reproductive expenditure. Both scenarios are compatible with the view that resources from the thorax are reallocated to egg or spermatophore production, but only the second scenario establishes a functional relationship between resource reallocation and reproduction.

Support for the second scenario comes from Stjernholm & Karlsson (2000). They showed that for field-caught P. napi females, the reduction in thorax nitrogen content increased with number of matings (which is correlated with reproductive expenditure; Karlsson 1998), irrespective of age. This suggests that the use of thorax resources is at least to some extent controlled by reproductive expenditure and not only dependent on life span.

In this paper we further examine the effect of age and reproduction on the decrease in thorax resource content in order to obtain a clearer picture of how resources are used for reproduction in butterflies. Specifically, we test the prediction that the investment of resources in spermatophores in male P. napi butterflies increases the use of thoracic as well as abdominal resources. P. napi males acquire little or no nitrogenous resources during the adult stage but have a large investment of nitrogen in spermatophores (Svärd & Wiklund 1989; Bissoondath & Wiklund 1995). This means that use of thorax resources could conceivably improve the reproductive output. Compared with females, the investment of resources in reproduction in males is easy to manipulate and quantify; this is done here by varying the number of spermatophores produced, i.e. the number of matings. By using males we can more accurately manipulate reproductive expenditure and also include a treatment with zero reproductive investment, which may be difficult in females that can lay unfertilized eggs. Furthermore, mating is not predicted to affect life span as it does in females, at least not under favourable conditions (e.g. Svärd 1985; Oberhauser 1989; Cordero 2000). Thus, we manipulate the expenditure of resources in spermatophore production and test how this affects the absolute as well as relative decrease in thorax and abdomen nitrogen content. This will show whether the ‘utilization’ of thoracic resources is regulated by processes related to the reproductive expenditure, or is simply a function of age, independent of reproduction. With this approach we also estimate the efficiency of nitrogen utilization, which is largely unknown among nectar-feeding butterflies.

Materials and methods

experimental design

The butterflies in the experiment were the offspring of five females collected in early June 2003, in the vicinity of Stockholm, Sweden. The larvae were reared in pairs on garlic mustard (Alliaria petiolata (Bieb.)) in 0·5 l plastic cups. During the larval stage the photoperiod was 22:2 h light : dark to induce direct development and the temperature was 20 °C. On emergence the butterflies were sexed and weighed once they had released the meconial wastes and become active. The males were individually numbered and divided among five treatments. The males in the first treatment, henceforth referred to as the newly emerged males, were frozen at the start of the experiment and used as a reference of the resource levels in recently emerged butterflies (n = 25). The males in the four remaining treatments were mated zero, one, two and three times, respectively. The numbers of males in each treatment at the start of the experiment were 25 (newly emerged), 11 (0 matings), 11 (1 mating), 15 (2 matings) and 19 (3 matings).

The experiment was started on 12 July and the last male died on 8 August. During the experiment the males were kept in five 0·55 × 0·75 × 0·80 m3 cages until the time of death. Three of the cages were used for matings, each with the same proportion of males from the three mating treatments, and included a surplus of virgin females. In the remaining two cages the males not allowed to mate were kept without access to females. Since the presence of females per se could potentially affect resource reallocation, a small mesh walled cage with a live female to provide visual and olfactory stimulation was kept in each larger cage. Throughout the experiment, cages were monitored for mating events every 20 min. After mating, males were allowed to recuperate for 2 days in the two cages without females and were then returned to the mating cages. The large size of the spermatophores produced after this period of recuperation ensures that there is a difference in reproductive expenditure between the treatments. When the males had mated the assigned number of times, they were transferred permanently to the cages without females for the remainder of their lives. Throughout the experiment the photoperiod was 9:15 h L:D and the butterflies were fed on a combination of sugar solution, ad libitum, and Chrysanthemum sp. and Cirsium arvense (L.) nectar.

After each mating the female was frozen and the spermatophore dissected out. Spermatophores and the dead males were stored frozen (−25 °C). After termination of the experiment, males and spermatophores were dried to constant weight at 60 °C and the total nitrogen content (mg) of spermatophores, and abdomen and thorax (excluding wings, legs and the head) of each male were measured by flash combustion on a LECO CHNS-932 elemental analyser (Leco Corp, 3000 Lakeview Av., St. Joseph, USA). Entire body parts were analysed.

Only animals with a life span of 6 or more days, which is the shortest life span in which a male could have mated three times, were included in the study. For this reason, one male each was omitted from the treatments with one and two matings and two males were omitted from the treatment with three matings. Furthermore, in mating treatments 2 and 3, four and nine males, respectively, did not mate the assigned number of times and were excluded from the analysis.

Differences between treatments in life span, eclosion mass, body and spermatophore nitrogen and C/N ratio were tested with anova or ancova and Tukey HSD post hoc tests. Since variances were not equal for total abdomen nitrogen content, data were ln-transformed. Eclosion mass was included in the model for body and spermatophore nitrogen as a covariate (ancova) in order to remove the influence of size differences between individuals. All tests were performed using Statistica 5·5 (StatSoft 1999).


life span

Life span was not significantly different between any of the four treatments with live butterflies (anova F3·35 = 0·32, P = 0·8, Table 1). This is in agreement with previous studies recording life span between males with different reproductive efforts (e.g. Svärd 1985; Oberhauser 1989; Karlsson & Wickman 1990; Cordero 2000), and means that the influence of reproductive expenditure on change in body nitrogen content will not be confounded by differences in life span between treatments. It also indicates that there are no or negligible costs of reproductive expenditure in terms of reduced life span in P. napi males under favourable conditions.

Table 1.  Number of males (n), eclosion mass, life span, reproductive expenditure and loss of nitrogen over the life span in the five treatments
Treatment (no. matings)nEclosion mass (mg) Mean ± SELife span (days) Mean ± SECumulative spermatophore nitrogen (mg) Mean ± SEExcretory loss of nitrogen
Newly emerged2564·1 ± 1·6   
01167·2 ± 1·5 9·5 ± 1·317%
11063·7 ± 2·710·0 ± 1·40·20 ± 0·0221·5%
21064·9 ± 2·8 8·4 ± 1·40·36 ± 0·0219%
3 863·9 ± 2·110·1 ± 1·50·48 ± 0·0217·5%

effectiveness of treatments

In order to test whether the treatments were successful we compared the amounts of nitrogen invested in spermatophores between the three mating treatments. The average cumulative amounts of nitrogen invested in spermatophores in the groups mating one, two and three times were 0·20, 0·36 and 0·48 mg, respectively (Table 1).

There was a significant effect of treatment (ancova: F2,24 = 69·0, P < 0·001), and the amount of nitrogen invested in spermatophores differed significantly between all groups (Tukey HSD test, α = 0·05). Thus the treatments were successful insofar that the cumulative investment of nitrogen in spermatophores increased with the number of matings.

comparison of remaining nitrogen reserves between groups

There was no significant difference in size (eclosion mass) between males used in the five treatments (anova F4·59 = 0·40, P = 0·8, Table 1). Thus, assuming that the resource levels at the start of the experiment in the four live treatments was accurately predicted by the levels in newly emerged males, this allowed us to determine how reproductive expenditure influence the decrease in thorax and abdomen resource levels.


For the thorax there was a significant effect of treatment on the amount of nitrogen remaining at death (ancova: F4,58 = 35·9, P < 0·001). Post-hoc comparisons (Tukey HSD test, α = 0·05) revealed that the thorax nitrogen content was greater in the butterflies killed at eclosion than in any of the other treatments. Among the latter, the males allowed to mate had significantly less nitrogen remaining in the thorax than those prohibited from mating, but there were no difference between the three mating treatments (Fig. 1).

Figure 1.

Abdomen nitrogen content in the five treatments (mean ± SE). Values are ln transformed and adjusted for eclosion mass.


For abdomen nitrogen content, there was a significant effect of treatment on the amount of nitrogen remaining in the abdomen at death (ancova: F4,58 = 239, P < 0·001), and the level of nitrogen was significantly different between all groups (Tukey HSD test, α = 0·05, Fig. 2).

Figure 2.

Thorax nitrogen content in the five treatments (mean ± SE). Values are ln transformed and adjusted for eclosion mass.

c/n ratio

For the thorax, the C/N ratio did not change between treatments (anova: F4,59 = 0·6, P = 0·7, Fig. 3). Relative abdomen nitrogen content, on the other hand, changed with increasing reproductive expenditure (anova: F4,59 = 4·1, P = 0·005). The proportion of nitrogen remaining at death was significantly lower in the treatments with two and three matings compared with the males that were not allowed to mate (Tukey HSD test, α = 0·05, Fig. 4).

Figure 3.

Thorax C/N ratio (mg/mg) as a function of treatment (mean ± SE). No differences in C/N between treatments (Tukey HSD test).

Figure 4.

Abdomen C/N ratio (mg/mg) as a function of treatment (mean ± SE). C/N of treatment 2 and 3 are significantly different from treatment 0 (Tukey HSD test).

nitrogen utilization efficiency

In addition to our study on the use of thorax and abdomen resources, we also examined the efficiency of nitrogen utilization. When comparing the accountable amounts of nitrogen, i.e. nitrogen remaining in the thorax and the abdomen at death, together with the nitrogen expended in spermatophores (when relevant), there was a significant effect of treatment on accountable nitrogen (ancova: F4,58 = 33·4, P < 0·001). Males killed at eclosion had higher accountable amounts of nitrogen than the males in the live treatments, which between them were not significantly different (Tukey HSD test, α = 0·05), Fig. 5, Table 1. This means that a similar amount of nitrogen (approximately 20%) was lost, presumably through excretion, in the four treatments with live butterflies.

Figure 5.

Total amount of nitrogen in abdomen, thorax and spermatophore(s) (where relevant) in the five treatments (mean ± SE). Values are adjusted for eclosion mass.


results and patterns

Our results show that the nitrogen content of both thorax and abdomen decrease over the adult life span of male P. napi butterflies. For the thorax the decrease in nitrogen content can be divided into two components. Some of the decrease is coupled to reproductive activity, but our data suggest it is decoupled from the actual reproductive expenditure. Additionally some of the decrease in nitrogen content is not directly related to reproductive activity, but can better be explained by age-dependent processes. These findings confirm the existence of a connection between the investment of resources in reproduction and the decrease in thorax nitrogen content, and support the proposition that nitrogenous resources are reallocated from thorax tissue to reproduction (Karlsson 1998).

Much of our understanding of how resources from the adult diet and nuptial gifts interact to shape patterns of reproduction have been acquired through radiolabel and stable isotope experiments (e.g. Boggs 1981b, 1997; Boggs & Gilbert 1979; Wiklund et al. 1993; O’Brien, Schrag & del Rio 2000; O’Brien et al. 2002; O’Brien, Boggs & Fogel 2003, 2004; Wedell & Karlsson 2003; Fischer, O’Brien & Boggs 2004). Unfortunately, these methods are not generally applicable to studies on the use of endogenous resources from different pools in the adult body. Thus, direct evidence that thorax resources are used for reproduction in butterflies is scant. Nonetheless, recent studies examining the use of thorax resources have revealed circumstantial evidence for reallocation of resources from thorax to reproduction. In P. napi females, Karlsson (1998) showed that abdominal and spermatophore nitrogen did not suffice to explain the incorporation of nitrogen in eggs. The production of eggs in long-lived females could only be accounted for on the assumption that nitrogen from the thorax was used. In males, there are indications that the decrease in thorax mass is positively correlated with the degree of polyandry, on the interspecific level (Stjernholm et al. 2005). This supports the hypothesis that, given an equal intake of resources as adults, use of thorax resources should increase with increasing reproductive expenditure (degree of polyandry). Moreover, thorax mass and nitrogen content decrease more in female than male butterflies (Karlsson 1994; Norberg & Leimar 2002; Stjernholm et al. 2005), where females generally have a greater expenditure of resources in eggs than males have in spermatophores.

In the majority of butterflies, the adult diet provides little nitrogenous resources for egg and spermatophore production (Boggs 1981a), even though it has recently been shown that amino acids in the adult diet may have a significant effect on fecundity in some conditions (Mevi-Schütz & Erhardt 2005). As a rare exception Heliconius butterflies use amino acids that they extract from pollen for reproduction (Gilbert 1972). This devalues the importance of juvenile resources for reproduction (Boggs 1981a), particularly the use of resources not primarily allocated to reproductive reserves. In male and female H. hecale (Hewitson) provided with pollen, thorax mass increased substantially over the first 6–8 weeks of adult life (Karlsson 1994). This is in contrast to the decline or (primarily for males) constancy of thorax mass over the life span in nectar-feeding butterflies (e.g. Stjernholm et al. 2005), and implies that limited availability of nitrogen for nectar-feeding butterflies has led to utilization of thorax resources for reproduction.

pattern of nitrogen mobilization

For thorax, C/N ratio remains constant among treatments, suggesting that the decline in nitrogen content is explained by the decrease in mass and that nitrogen is not preferentially reallocated from thoracic tissues. This is compatible with the finding that the decrease in thorax nitrogen content is to a large extent a result of breakdown of the flight muscles (F. Stjernholm, unpublished results). For the abdomen on the other hand relative nitrogen content decreases as the number of matings performed increases. This may either be a consequence of specific mobilization of nitrogen from abdominal tissues or could result from storage of carbohydrates from the adult diet, both leading to reduced relative nitrogen content.

regulation of thorax nitrogen decrease

As outlined in the introduction, use of thorax resources for reproduction could be regulated in two ways, i.e. dependent on age per se or dependent on reproductive expenditure. Our study suggests that both mechanisms interact to explain the pattern of decrease in thorax nitrogen content in P. napi males. The nitrogenous resources in the thorax decrease in the treatment with virgin males, indicating that there is a basal rate of decrease in thorax nitrogen content irrespective of reproductive investment. Whether this decrease is related to reproduction or not is unknown, but since nitrogenous resources bound for the accessory glands are transported in the haemolymph, it seems likely that amino acids or proteins from the thorax contribute to the haemolymph resource pool or are sequestered as storage proteins in the fat body of the adult. Thoracic amino acids may thus be incorporated into eggs and spermatophores in proportion to their abundance in the haemolymph or fat body, possibly with adjustment to the availability and need for different amino acids. Only if the nitrogen lost from the thorax is in a non-utilizable form should it be removed from the resource pool. In Danaus plexippus (Nymphalidae) males the size of the first spermatophore produced increases with time to mating (Oberhauser 1988). Thus, at least in this species, resources are gradually accumulated in the reproductive organs. Whether these resources originate solely from abdominal reserves or also from the thorax is unknown.

Compared to the treatment with virgin males there was a greater reduction in thorax nitrogen content over the life span of the males in the three mating treatments. This shows that reproduction affects the use of resources, most likely through an increase in the rate of thorax nitrogen depletion. In other insects, mating is known to increase the rate of protein synthesis in the accessory glands (Gillot 1995), and consequently enhance the depletion rate of stored resources (from whatever source). In Drosophila melanogaster the rate of protein synthesis in the accessory glands increases with the number of matings in a dose-dependent manner (Schmidt, Stumm-Zollinger & Chen 1985), suggesting that the rate of transfer of resources to the reproductive organs is also dependent on the number of matings. In our experiment there was no difference in the decrease in thorax nitrogen content between the three mating treatments, even though there was a difference in reproductive investment. Thus, it seems that reallocation of thorax resources is regulated in an all-or-none fashion in P. napi males, rather than being directly related to the reproductive expenditure in a dose-dependent manner. A maximum rate of resource accumulation in the reproductive organs, and consequently a fixed rate of depletion of resources from reproductive reserves and somatic (thoracic) tissues may optimize fitness after the first mating.


Nectarivorous butterflies, which have a finite supply of nitrogen available, should be efficient in their utilization of nitrogenous resources to minimize the loss. Nonetheless, some of the end products of protein turnover and metabolism may be inaccessible for further use and consequently eliminated through excretion. Adult butterflies are known to excrete nitrogen as uric acid, allantoic acid and to some extent as allantoin (Bursell 1967). In this experiment we show that approximately 20% of the nitrogen in each of the live treatments is unaccounted for (Table 1) and has presumably been lost through excretion. This substantial loss of nitrogenous resources from the butterflies may have consequences for how the results are interpreted. Since the decrease in thorax nitrogen content is less than the loss through excretion, it is possible that none of the thorax resources are used for reproduction. However, the finding that nitrogen loss was just as high for non-mating and mating males suggests that excretory nitrogen loss is related to activity and not mating per se, which is in agreement with the hypothesis that nitrogen from the thorax is used for reproduction. Thus, we conclude that not only resources from the reproductive reserves but also resources from the thorax are used for reproduction in male P. napi butterflies.


We thank Diane M. O’Brien, Mark A. Jervis, Christer Wiklund and Bertil Borg for providing valuable comments that significantly improved the quality of this paper. This work was financially supported by a grant from Alice and Lars Siléns foundation to F.S., and a grant from the Swedish research council to B.K.