Territorial and Mating Success of Dragonflies That Vary in Muscle Power Output and Presence of Gregarine Gut Parasites

ARTICLE IN PRESS

ANIMAL BEHAVIOUR, 2004, 68, 657–665 doi:10.1016/j.anbehav.2003.09.019

Territorial and mating success of dragonﬂies that vary in muscle power output and presence of gregarine gut parasites

JAMES H. MARDEN & JASON R. COBB Department of Biology, Pennsylvania State University

(Received 9 May 2003; initial acceptance 7 August 2003; ﬁnal acceptance 26 September 2003; MS. number: A9609)

Competition for mating territories in libellulid dragonflies involves aerial contests that require high levels of exertion and locomotor performance. Here we test the hypothesis that success of male Libellula pulchella dragonflies in territoriality and mating is affected by muscle contractile performance, and we examine how gregarine gut parasites affect muscle performance, energy reserves and territorial behaviour of their hosts. At a pond where gregarine parasites are rare, long-term territorial and mating success of males showed a significant positive association with muscle power output. At a nearby pond that had a much higher incidence and intensity of gregarine parasitism, there was no relationship between muscle performance and short-term territorial success. Instead, males assorted themselves into aggressive territory holders and submissive satellites, with the large majority of territory holders having no parasites and nearly all of the satellites parasitized. Unparasitized males showed a tight positive relationship between muscle power and fat content, which suggests that they use a known phenotypic adjustment in muscle contractile performance to allow the energy consumption rate of the flight muscles to match the rate at which energy can be mobilized from storage pools. Parasitized dragonflies showed a small decrease in average fat content and a marked change in the relationship between fat content and muscle power output. The apparent loss of the ability to match muscle contractility to the size of the energy storage pool in parasitized dragonflies suggests that gregarines may have systemic effects on signalling pathways and energy homeostasis. By indirectly choosing males that had successfully competed for territories, females consistently mated with physiologically or immunologically superior males despite large between-pond differences in male behaviour and the incidence and intensity of parasitic infection. Ó 2004 The Association for the Study of Animal Behaviour. Published by Elsevier Ltd. All rights reserved.

Competition for mates and the resources necessary to contractile performance of muscles (Marden et al. 2001) obtain mates is a pervasive and well-studied feature of should enhance success in types of competitive interac- animal behaviour. Much is known regarding the role of tions that depend heavily on exertion and locomotion. externally apparent traits such as ornaments, vocaliza- Knowledge regarding the variability of such traits within tions and chemical signals, but sexual competition also populations, the source of the variation, and its conse- commonly involves contests or battles that require high quences for competitive success are important for un- levels of exertion and locomotor performance. In such derstanding how selection acts on physiological capacities cases, internal physiological traits are likely to play a strong and how they evolve. role in determining success or failure, yet for most types of A common source of performance variation in free- animals we know little about variability in the physiolog- living animals is parasitism. The behaviour and success of ical features that affect intraspecific competition. Attri- parasitized animals during competitive interactions with butes such as high levels of energy reserves (Marden & conspecifics often differ markedly from those of non- Waage 1990), ability to up-regulate metabolic flux and parasitized animals (Schall & Sarni 1987; Howard & produce ATP at a high rate (Harrison & Hall 1993), and Minchella 1990; Moore & Gotelli 1996), but the physiological mechanisms underlying these differences have seldom been elucidated. Parasites may cause specific Correspondence: J. H. Marden, Department of Biology, 208 Mueller lesions or deficiencies that affect performance-related Labs, Pennsylvania State University, University Park, PA 16802, U.S.A. traits (Harrison et al. 2001); they may deprive the host (email: [email protected]). of energy necessary for high levels of exertion (Munger & 657 0003–3472/03/$30.00/0 Ó 2004 The Association for the Study of Animal Behaviour. Published by Elsevier Ltd. All rights reserved. ARTICLE IN PRESS

658 ANIMAL BEHAVIOUR, 68, -

Karasov 1989; Siva-Jothy & Plaistow 1999), and they may concentration of their flights along boundaries between cause systemic physiological changes that affect many territories guarded by aggressive males, and their poor aspects of host metabolism and locomotion (Schall et al. ability to successfully copulate with females or guard the 1982). few females with which they copulated. In the particular case of flying animals it has been The focal observations and sampling described above shown that a high ratio of flight muscle mass to total body used males that had been marked individually with weight affects aerial performance and success in compet- coloured powder (Day-Glo Color Corp., Cleveland, Ohio, itive interactions (Marden 1989; Coelho & Holliday 2001; U.S.A.). The powder was applied using a blowgun tech- Berwaerts et al. 2002). The generality of this result nique (Marden 1989) that did not require capture or depends on there being fairly little variation in muscle handling of the dragonflies and had no discernable effect contractile performance (i.e. it requires that the ratio of on their behaviour. Marked males were observed for at muscle mass to body weight captures most of the be- least 1 h before determination of territorial or satellite tween-individual variation in force or power output status. Many of the nonterritorial males that we marked relative to total load). However, muscle contractile perfor- were not site specific; they were primarily feeding and mance can vary widely between individual conspecific their competitive status could not be determined. We animals (Marden et al. 2001), and contractility is a trait classified as satellites only those nonterritorial males that that is likely to be affected by parasites (i.e. perhaps by showed site fidelity, attempts to engage in opportunistic regulatory responses stimulated by metabolic changes or mating, and the submissive behaviours described above. decreased energy availability). Effects of intramuscular Darlington Pond measured approximately 300 m2 and parasites on contractility have been demonstrated (Har- all of the dragonflies present could be observed simulta- wood et al. 1996), but no studies have examined how neously and across many consecutive days, thereby muscle contractile performance is related to the presence allowing quantification of long-term mating success of of parasites in nonmuscle tissues. a large population of males. When each male arrived and Here we test the hypothesis that success of male Libel- began to engage in territorial interactions for the first time lula pulchella dragonflies in territoriality and mating is at Darlington Pond we captured it in an insect net. After affected by muscle contractile performance, and we netting, we used permanent ink to write a number on one examine how gregarine gut parasites affect muscle perfor- wing and marked the tips of all four wings with a unique mance, energy reserves and territorial behaviour of their combination of coloured spots using the Day-Glo powder hosts. described above. The powder rubbed gently onto the wing tips remained visible for up to 3 weeks. Long-lived males were occasionally recaptured to refresh their wing colours. METHODS Records were kept of arrival and departure times for each male, and we counted the number of times they copulated We studied L. pulchella dragonflies at two ponds in Centre with females. We collected a representative sample of County, Pennsylvania, U.S.A., during June and July in males from Darlington Pond near the end of their 2000 and 2001. Our main study site in 2000 was Ten Acre territorial tenure and tested them for muscle contractile Pond (40 480N, 77 560W). In 2001, we primarily studied performance. We also used data from this pond to dragonflies at a small unnamed pond (which we called examine the relationship between territorial effort and Darlington Pond; 40 510N, 77 500W). These two ponds longevity. were separated by approximately 12 km. Dragonflies from both ponds were transported to the Females visit territories defended by male L. pulchella laboratory to determine their muscle contractile perfor- dragonflies to copulate and deposit their eggs while being mance (see Marden et al. 2001 for detailed methodology). guarded by the male (Pezalla 1979). Owing to differences Briefly, we removed the head, wings and legs and in pond size and terrain, our ability to monitor these cemented the thorax and abdomen in a controlled tem- events differed at the two ponds, which required us to use perature chamber. A series of shallow incisions were made different sampling schemes. Ten Acre Pond was too large around the dorsal attachment point of the mesothoracic to monitor the entire pond, so we studied short-term basalar muscle to mechanically isolate the muscle and territorial and mating success of focal individuals or pairs attach it to the lever arm of an ergometer. The lever of individuals. Our focal observations were made on imposed a sinusoidal oscillation of muscle length in samples of highly aggressive territorial males and non- a manner that closely matched the in vivo length territorial satellite males that were active along the border oscillation that occurs in this muscle during flight. of one or more established territories. Whenever possible, Electrical stimulation of the muscle was supplied in we collected these samples in a paired fashion so that we a manner that corresponded to the in vivo phase relation- could compare a territory holder with a satellite simulta- ship between neural stimuli and muscle length. Muscle neously present at the same location. Territorial males temperature was regulated at 32–34 C. We recorded the were identified by their frequent patrolling of a well- time course of muscle length and tension at integer defined area, willingness to engage in escalated encoun- contraction frequencies ranging from 20 to 45 Hz (which ters with neighbours and interlopers, and successful slightly exceeds the range of wingbeat frequencies ob- copulation and guarding of females. Satellites were iden- served during free flight; Marden et al. 1999) and used tified by their submissive behaviour ( gliding and leaving those data to compute net work and power output. We the area) when confronted by a territorial male, the used the highest power output observed at any of those ARTICLE IN PRESS

MARDEN & COBB: DRAGONFLY MUSCLES AND PARASITES 659 contraction frequencies as our measure of maximal power gregarine gut parasites (protists; Fig. 1). Only three of the output. The basalar muscle was removed from the thorax 13 territory holders were infected, whereas seven of the 2 and weighed after each experiment; its fresh mass was eight satellites were infected (chi-square test: c1 Z 9.0, used to determine mass specific power output. P Z 0.003). All infections appeared to be a single species of To assess gregarine parasite burdens, we made a dorsal gregarine (Hoplorhynchus sp., family Actinocephalidae: incision along the entire length of the abdomen and subfamily Menosporinae; Clopton 2002). The infected midgut of all collected dragonflies. Lateral margins of the territorial males had an average of 11 trophozoites abdomen were pinned flat and the midgut contents were (SE Z 2; N Z 3) compared with an average of 30 (SE Z 11; examined under a stereomicroscope with fibre optic illu- N Z 7) for satellite males. A crude measure of age, a four- mination. We removed partially digested food (typically level ranking based on wing wear, was not significantly contained within a peritrophic membrane) and counted associated with territorial/satellite status (chi-square test: 2 Z Z gregarine trophozoites throughout the midgut, including c2 4.7, P 0.09), presence or absence of gregarine 2 the recesses surrounding the proventriculus. parasites (c2 Z 3.8, P Z 0.15) or the mean number of We determined levels of energy reserves by extracting parasites (ANOVA: F2,24 Z 1.0, P Z 0.40). Because wing lipids from a subsample of dragonflies that were tested for wear is dependent in part on behaviour and perhaps also muscle performance. The thorax and abdomen were dried on the ability of the insect to maintain itself, the apparent to a constant mass, weighed, placed in refluxing chloro- lack of age effects must be interpreted with some scepti- form in a Soxhlet apparatus for 8 h, then dried and cism. Note, however, that L. pulchella frequently contain weighed again. The difference between dry mass and lean large numbers of gregarines even when they are newly dry mass was used as our measure of lipid content. mature, and none of the satellite males in our sample showed extensive wing wear. Of the 174 mature males that we marked at Darlington RESULTS Pond, 46 spent time defending territories over a time interval as great as 25 days after their initial arrival at the We measured muscle power output for 15 of the 21 pond (Fig. 2). We collected a representative sample of individuals collected from Ten Acre Pond for which we these dragonflies to determine how muscle power and had determined territorial versus satellite behavioural parasitism were related to long-term mating success. status. There was no difference in the mean or variance Collections were timed in such a way that the dragonflies of power output of territorial and satellite males. This was had already accomplished most of the mating that they true for both our entire sample (X G SE; territorial males: were likely to have achieved had we not collected them. 104 G 8 W/kg, N Z 9; satellite males: 107 G 11 W/kg, Specifically, we collected two dragonflies (Males 4 and 21) N Z 6; t13 Z 0.2, P Z 0.86) and for 10 males that were after 15 days of observation when their early-emerging collected in a paired fashion (mean difference Z cohort was disappearing (Fig. 2) and they showed consid- 3 G 21 W/kg; one-tailed t test of the hypothesis that erable wing wear. We collected another group (Males 37, territory holders have greater muscle power output than 41, 59, 63, 127, 135, 161) 43 days after the start of the satellites: t4 Z 0.2, P Z 0.44). Thus, in the population at experiment, by which time the arrival rate of ovipositing Ten Acre Pond, there was no support for the hypothesis females had dropped to such a low level that little mating that a high muscle power output enhanced success in was occurring. A final group of freshly mature males short-term competition for mating territories. (Males 180, 181 and 182) were collected late in the season Territorial and satellite males at Ten Acre Pond differed shortly after our behavioural observations had ended. Two markedly in their incidence and level of infection by additional males (Males 112 and 136) were collected and

Figure 1. (a) Distribution of parasitized and unparasitized males among territory holders and satellites. (b) A L. pulchella midgut containing parasitic gregarine trophozoites (oblong white objects) and a small number of newly formed gamontocysts (spherical white objects). (c) Close- up of a single gregarine trophozoite (Hoplorhynchus sp., family Actinocephalidae: subfamily Menosporinae). ARTICLE IN PRESS

660 ANIMAL BEHAVIOUR, 68, -

170 165 160 C 155 150 145 140 135 C 130 C 125 120 115 110 105 100 95 90 85 80 75

70 65 Dragonfly identification number C 60 C 55

40 C C 35 30 25 C 20

15 10

5 C

5 10 15 20 25 30 35 40 Day Figure 2. Territorial and mating activity of 174 individual male L. pulchella dragonﬂies at Darlington Pond during the summer of 2001. The experiment began on 12 June, which is Day 1 on the horizontal axis. Bar width represents time spent defending a territory relative to the total observation time that day. Bar height above the minimum height, which represents only presence, is proportional to the number of copulations obtained that day. The smallest bars include males that were present at the pond only momentarily; their width and height do not represent relative measures of territoriality or mating. The letter ‘C’ indicates the day of capture for dragonﬂies collected for muscle performance analyses. Vertical grey bars represent rainy days (single-day width) or weekends (2-day width) when we did not collect data. ARTICLE IN PRESS

MARDEN & COBB: DRAGONFLY MUSCLES AND PARASITES 661

3.5 15

3 10

5 2.5

0 2 1000 2000 Total time defending territory (min) total time on territory (min) 10 1.5 15 Log

1.9 2 2.1 2.2 2.3 10

Log10 muscle power output (W/kg) Figure 3. The relationship between a male’s mass specific muscle 5 power output and the total time he spent defending a territory. Data Number of males were log transformed to achieve normality. 0 10 20 30 40 50 Number of copulations checked for parasites after they were injured when we netted them at the time of their initial arrival at the pond. 2 Among the nine males at Darlington Pond for which we obtained data for both long-term mating success and 1 muscle contractile performance, there was a significant positive relationship between muscle power output and Z 80 100 120 140 160 180 total time spent defending a territory (Fig. 3; F1,7 4.1, one-tailed P Z 0.04) and total number of copulations Muscle power output (W/kg) (Fig. 4; ANOVA: F1,7 Z 8.8, one-tailed P Z 0.01; all of these Figure 5. Frequency distributions of time spent defending a territory, variables were log transformed to achieve normality). In number of copulations and mass specific muscle power output for a multivariate model, only the total time defending dragonflies at Darlington Pond. In each plot the darkened bar a territory had a significant effect on the number of represents Male 41, which was by far the most successful male in that population. copulations (F1,6 Z 15.4, one-tailed P Z 0.004), although the effect of muscle power approached significance Z Z (F1,6 3.0, one-tailed P 0.07). There was no effect of significant additional effect on the ability to obtain body size (measured by thorax mass) in any of these copulations. analyses. These results suggest that the primary effect of One of the males at Darlington Pond (Male 41) obtained high muscle performance was on the ability to establish approximately twice the number of copulations and spent and maintain a territory, after which there was no twice as much time at the pond as the next best competitor (Fig. 5). He also had the highest muscle power output (184 W/kg) among all of the males that we have sampled from this species (Fig. 6; N Z 51, which includes data from Marden et al. 2001 and represents dragonflies 1.5

1 15

10 0.5

total number of copulations 5 10 Number of males 0 Log 0 60 80 100 120 140 160 180 200 1.9 2 2.1 2.2 2.3 Muscle power output (W/kg) Log muscle power output (W/kg) 10 Figure 6. Frequency distribution of mass speciﬁc muscle power Figure 4. The relationship between a male’s mass speciﬁc muscle output from 51 mature L. pulchella males. The darkened bar power output and the total number of copulations he obtained. represents Male 41, which was by far the most successful male in Data were log transformed to achieve normality. our study of long-term mating success at Darlington Pond. ARTICLE IN PRESS

662 ANIMAL BEHAVIOUR, 68, -

from a number of ponds over 4 years). The sample sizes in midgut from dragonflies collected at Ten Acre Pond to our study of mating success (174 dragonflies) and our four of the marked males at Darlington Pond (Males 70, survey of muscle performance (51 dragonflies) makes it 81, 86 and 105) on day 16 of our territorial observations highly unlikely that, based on chance, an individual during the summer of 2001. Preliminary experiments would be an outlier for both of these measures. showed that this technique results in many intact troph- Only three of the 14 dragonflies collected from Darling- ozoites becoming established in the midgut. As a control, ton Pond were infected with gregarines (Males 127, 136 three other males (Males 50, 80 and 98) were fed un- and 180), and all three had relatively low levels of infected midgut. The sample size of this pilot experiment infection (N Z 1, 10 and 3 trophozoites, respectively). proved too small to clearly determine the effect of This contrasts sharply with what we found at Ten Acre experimental gregarine infection on territorial behaviour, Pond, where 18 of the 36 dragonflies collected during and it would not be mentioned here except for an 2000 and 2001 were infected with gregarines. The mean incident that involved one of the experimentally infected number of gregarines among all individuals (i.e. both males (Male 86). Although he continued to be one of the infected and uninfected) at the two ponds differed by more successful males during the 2 days immediately a factor of 30 (1.0 G 0.7, N Z 14 at Darlington Pond versus following infection with trophozoites, at 9 days post- 30.0 G 9.9 at Ten Acre Pond; Wilcoxon chi-square ap- infection he was observed acting in a highly submissive 2 proximation: c1 Z 5.7, P Z 0.02). None of the males manner at the pond, so much so that he was apparently at Darlington Pond was observed to consistently behave mistaken for a female and grasped in a precopulatory as a satellite, nor did any of them consistently retreat from position by a territorial male. In 15 years of experience territorial encounters by gliding. This is quite different with this species, the senior author (J.H.M.) has never from what we observed at Ten Acre pond, where sub- previously observed a homosexual pairing of L. pulchella missive satellite behaviour was common, especially males, so it is noteworthy that this event corresponded among males that were infected with gregarines. with experimental gregarine infection and a sharp change Among the three infected dragonflies at Darlington from the earlier aggressive behaviour of this individual. Pond, only Male 127 was represented in our long-term Gregarine parasites may affect the behaviour of their observational data (Male 136 was injured at initial capture dragonfly hosts by a mechanism that involves some and Male 180 arrived after we had concluded the behav- combination of changes in neural, hormonal, muscle ioural survey). During the 20-day span from his first contractile and metabolic traits. We initiated an explora- appearance at the pond until he was collected, Male 127 tion of these possibilities by comparing muscle contractile defended a territory for a total of only 35 min. This is the performance and lipid content between parasitized and second lowest total among the 46 dragonflies in our unparasitized dragonflies. We found no significant effect sample that spent any time defending a territory, and he of parasitism on power output (N Z 32 males sampled for never copulated. both power output and parasitism; F1,30 Z 0.3, P Z 0.6), In the 2001 study at Darlington Pond we sought to but there was a significant decrease in the fat content of Z Z determine whether there was a negative relationship parasitized dragonflies (F1,38 4.2, one-tailed P 0.024; between male longevity and territorial intensity. If true, Fig. 7). There was also a marked change in the relationship this might provide an incentive for gregarine parasites to between fat content and muscle power output (Fig. 8). manipulate their hosts in order to make them less Dragonflies that were free of parasites showed a strong aggressive and thereby extend host survival, allowing positive relationship between muscle power output and 2 more time for trophozoites to mature and reach the fat content (r Z 0.63, F1,9 Z 15.1, P Z 0.004), whereas gamontocyst stage of their life cycle (Fig. 1). We tested there was no such relationship in parasitized dragonflies 2 for such a trade-off by comparing the number of copula- (r Z 0.25, F1,9 Z 0.7, P Z 0.17). tions per minute of territorial defence to the total time spent defending a territory, and the total time spent defending a territory with the maximum known age (days DISCUSSION between initial marking and last sighting; this is an underestimate of chronological age because it does not To our knowledge this is the first study that relates include the teneral period of early adult maturation). a detailed measure of muscle contractile performance with Contrary to our expectation of a trade-off, both of these competitive ability in nonhuman animals. The result is relationships showed significant positive slopes (ANOVA: quite clear: in the absence of parasites, long-term territo- Z Z Z Z F1,44 6.7, P 0.01 and F1,44 12.9, P 0.001, res- rial and mating success has a significant positive associa- pectively). Thus, there is either no trade-off between these tion with muscle power output in L. pulchella (Figs 3–5). life history traits, or more likely, a large amount of Given such strong sexual selection favouring high muscle between-individual variation in male performance and performance, why do all dragonfly males not possess longevity obscures the presence or absence of trade-offs high-power flight muscles? The tight positive relationship within individuals (van Noordwijk & deJong 1986; Stearns between muscle power and fat content in healthy drag- 1992). Thus, it remains an open question whether there onflies (Fig. 8) provides a probable answer. Males with exists an opportunity for gregarine parasites to beneficially different levels of fat appear to make a phenotypic manipulate the behaviour of their dragonfly hosts. adjustment in their muscle contractile performance that To determine whether gregarine parasitism causes sat- presumably allows the maximum energy consumption ellite behaviour, we fed portions of gregarine-infected rate of their muscles to match the rate at which energy can ARTICLE IN PRESS

MARDEN & COBB: DRAGONFLY MUSCLES AND PARASITES 663

25 Unparasitized

160

140 20

120

100 15

60 10 Fat content (% dry mass) 510152025

Parasitized

5 160

Unparasitized Parasitized 140 Muscle power output (W/kg)

120

Figure 7. Fat content (% of dry mass of the abdomen and thorax) of 100 dragonﬂies with and without gregarine gut parasites.

80 be mobilized from storage pools. The ability to phenotyp- ically adjust muscle power output and ATP consumption 60 to match nutritional state has been suggested previously 510152025 for mammals (Russell et al. 1984) and is an intuitively Fat content (% of dry mass) appealing extension of other known forms of cellular and Figure 8. Relationship between fat content and muscle power organismal energy homeostasis. output for dragonflies with and without gregarine gut parasites. An alternative hypothesis for the correlation between muscle performance and fat content is that more powerful flight muscles allow for greater foraging success. We think der Houven van Oordt et al. 2000) is rudimentary; that this is unlikely because most L. pulchella foraging however, this is an important topic, because it is likely flights appear to involve a much lower level of exertion to be one of the primary mechanisms by which organisms than the flights used to obtain and defend territories and accomplish proteomic diversity and adaptive phenotypic mates. These large dragonflies appear to capture their plasticity. much smaller prey with ease, and therefore, there seems One of our study ponds had a much higher incidence little reason to believe that differences in muscle perfor- and intensity of gregarine parasitism. At that pond we mance could cause large differences in foraging success found no relationship between muscle performance and and accumulation of fat reserves. short-term territorial success. Instead, males assorted Our previous work with L. pulchella dragonflies has themselves into aggressive territory holders and submis- revealed the mechanism that allows for phenotypic sive satellites, with the large majority of territory holders adjustment of muscle power output (Marden et al. having no parasites and nearly all of the satellites 2001). Alternative splicing of mRNA that encodes the parasitized. Together with the changed relationship be- muscle regulatory protein troponin-t creates isoform di- tween fat and muscle power in parasitized dragonflies, versity that affects how the muscle responds to activation these results support the hypothesis that gregarine para- by calcium, which in turn affects force, work and power sites alter the physiology and behaviour of their dragonfly output during cyclical contraction. The relationship be- hosts, thereby changing the processes and mechanisms tween power and fat that we have found in the present that determine the outcome of intermale competition. study makes a significant advance in this line of work However, we cannot be confident that the effect of because it implies that there is a signalling pathway that gregarine parasites on the dynamics of male territorial communicates information about global energy reserves competition is generally true because our study involved to the intracellular mechanisms that control alternative just two ponds that differed greatly in size (i.e. we have no splicing. Knowledge of the mechanisms that control replication of the effect of different levels of parasitism alternative splicing (Maniatis & Tasic 2002) and their between ponds). Crowding of males into highly over- connection with extracellular signalling pathways (van lapping territories at the small pond also may have played ARTICLE IN PRESS

664 ANIMAL BEHAVIOUR, 68, -

a role in the importance of muscle power and in the males were unlikely to defend a territory and females absence of males acting as satellites. mated predominately with noninfected territory holders Although we detected no difference in the average of widely varying muscle performance. Thus, there was muscle power output of parasitized and unparasitized a marked difference between these two ponds in the male dragonflies, our data reveal subtle physiological changes characteristics that were sorted by territorial interactions, associated with parasitism. There was a small decrease in but the difference did not appear to involve any change in the average fat content of parasitized dragonflies (Fig. 7) female behaviour. Rather, it appears that the strong and a marked change in the relationship between fat tendency of females to favour certain oviposition sites content and muscle power output (Fig. 8). Gregarines served as a consistent filter for mating with males that have previously been shown to reduce the fat content of were in peak condition and/or were free of parasites. parasitized male odonates (Siva-Jothy & Plaistow 1999), Libellula pulchella females often make rigorous evasive and differences in fat content among competing pairs of flights when approached by males, which provokes a chase males have been shown to strongly affect the outcome of that may constitute an additional test of male flight territorial battles (Marden & Waage 1990; Marden & performance. By being choosy about oviposition sites Rollins 1994; Plaistow & Siva-Jothy 1996). However, it is and sometimes evasive, females consistently avoid mating unlikely that reduced fat content was the cause of the poor with physiologically or immunologically inferior males territorial success of parasitized males at Ten Acre Pond despite large between-pond differences in male behaviour because the distribution of fat content was broadly over- and the incidence and intensity of parasitic infection. This lapping between healthy and parasitized males (Fig. 7). mating system, in which males largely sort themselves For this reason, it is unlikely that parasite-induced differ- prior to female arrival, appears to be a quick and effective ences in fat reserves were the mechanism that led to the way for females to choose high-quality mates indirectly strong association between infection and territorial/satel- (Wiley & Poston 1996) in an environment where male lite status (we did not measure the fat content of most of condition varies quite widely between individuals and the paired males in the 2000 study, so we cannot evaluate between ponds. the relative fat content of territory holders and satellites). Rather, we are more impressed by the change in the relationship between lipid content and muscle power output (Fig. 8). In parasitized dragonflies, the apparent Acknowledgments loss of the ability to match muscle contractility to the size This material is based upon work supported by the of the energy storage pool suggests that gregarines may National Science Foundation under Grant No. IBN- cause physiological changes that affect signalling path- 0091040. We thank M. Darlington for access to his pond ways and energy homeostasis. Indeed, dragonflies with and K. Flaherty and N. Mahoney for assistance with data gregarine trophozoites in their midgut have chronically collection. Comments from T. Getty and an anonymous activated p38 map kinase in their flight muscles (R. J. referee improved the manuscript. Schilder & J. H. Marden, unpublished data). This molecule is involved in pathways that control a wide variety of cellular functions and gene expression (Cowan & Storey 2003), which in mammals includes insulin signalling, References glucose transport, and the function of fat storage cells (Carlson et al. 2003; Fujishiro et al. 2003). Chronic and Berwaerts, K., Van Dyck, H. & Aerts, P. 2002. Does flight morphology relate to flight performance? An experimental test systemic changes in p38 MAPK signalling are likely to with the butterfly Pararge aegeria. Functional Ecology, 16, 484– have strong effects on internal physiology that could 491. affect behaviour and flight performance in ways that we Carlson, C. J., Koterski, S., Sciotti, R. J., Poccard, G. B. & did not detect with our measure of short-duration muscle Rondinone, C. M. 2003. Enhanced basal activation of mitogen- power output. We are presently performing additional activated protein kinases in adipocytes from type 2 diabetes: analyses of cellular physiology and signalling in this potential role of p38 in the downregulation of GLUT4 expression. species to gain a more detailed understanding of this Diabetes, 52, 634–641. interesting host–parasite interaction. Clopton, R. E. 2002. Phylum Apicomplexa Levine, 1970: Order Dragonflies are a classic example of resource defence Eugregarinorida Le´ger, 1900. In: Illustrated Guide to the Protozoa. polygyny, with males defending territories that contain 2nd edn (Ed. by J. J. Lee, G. Leedale, D. Patterson & P. C. Bradbury), sites that females use for oviposition (Fincke 1997). Libel- pp. 205–288. Lawrence, Kansas: Society of Protozoologists. lulid females are clearly choosy about where they oviposit, Coelho, J. R. & Holliday, C. W. 2001. Effects of size and flight and males appear to assess which sites are favoured by performance on intermale mate competition in the cicada killer, Sphecius speciosus Drury (Hymenoptera: Sphecidae). Journal of females (Pezalla 1979; Koenig & Albano 1985). At Darling- Insect Behavior, 14, 345–351. ton Pond, where gregarine parasites were rare, the out- Cowan, K. J. & Storey, K. B. 2003. Mitogen-activated protein come of territorial competition between males was kinases: new signaling pathways functioning in cellular responses correlated with differences in muscle contractile perfor- to environmental stress. Journal of Experimental Biology, 206, mance and presumably flight ability, and females mated 1107–1115. most frequently with the most powerful males. At Ten Fincke, O. M. 1997. Conflict resolution in the Odonata: implications Acre Pond, which was larger, less crowded, and had for understanding female mating patterns and female choice. a much higher incidence of gregarine parasitism, infected Biological Journal of the Linnean Society, 60, 201–220. ARTICLE IN PRESS

MARDEN & COBB: DRAGONFLY MUSCLES AND PARASITES 665

Fujishiro, M., Gotoh, Y., Katagiri, H., Sakoda, H., Ogihara, T., Anai, Marden, J. H., Fitzhugh, G. H., Girgenrath, M., Wolf, M. R. & M., Onishi, Y., Ono, H., Abe, M., Shojima, N., Fukushima, Y., Girgenrath, S. 2001. Alternative splicing, muscle contraction and Kikuchi, M., Oka, Y. & Asano, T. 2003. Three mitogen-activated intraspecific variation: associations between troponin T transcripts, protein kinases inhibit insulin signaling by different mechanisms in calcium sensitivity, and the force and power output of dragonfly 3T3-L1 adipocytes. Molecular Endocrinology, 17, 487–497. flight muscles during oscillatory contraction. Journal of Experimen- Harrison, J. F. & Hall, H. G. 1993. African–European honeybee tal Biology, 204, 3457–3470. hybrids have low nonintermediate metabolic capacities. Nature, Moore, J. & Gotelli, N. J. 1996. Evolutionary patterns of altered 363, 258–260. behavior and susceptibility in parasitized hosts. Evolution, 50, 807– Harrison, J. F., Camazine, S., Marden, J. H., Kirkton, S. D., Rozo, 819. A. & Yang, X. 2001. Mite not make it home: parasitic tracheal Munger, J. & Karasov, W. 1989. Sublethal parasites and host mites reduce the safety margin for oxygen delivery of flying honey energy budgets: tapeworm infection in white-footed mice. bees. Journal of Experimental Biology, 204, 805–814. Ecology, 70, 904–921. Harwood, C. L., Young, I. S., Lee, D. L. & Altringham, J. D. 1996. van Noordwijk, A. J. & deJong, G. 1986. Acquisition and allocation The effect of Trichinella spiralis infection on the mechanical of resources: their influence on variation in life history tactics. properties of the mammalian diaphragm. Parasitology, 113, American Naturalist, 128, 137–142. 535–543. Pezalla, V. M. 1979. Behavioral ecology of the dragonfly Libellula van der Houven van Oordt, W. , Diaz-Meco, M. T., Lozano, J., pulchella Drury (Odonata: Anisoptera). American Midland Natural- Krainer, A. R., Moscat, J. & Caceres, J. F. 2000. The MKK(3/6)- ist, 102, 1–22. p38-signaling cascade alters the subcellular distribution of hnRNP Plaistow, S. & Siva-Jothy, M. T. 1996. Energetic constraints and A1 and modulates alternative splicing regulation. Journal of Cell male mate-securing tactics in the damselfly Calopteryx splendens Biology, 149, 307–316. xanthostoma (Charpentier). Proceedings of the Royal Society of Howard, R. D. & Minchella, D. J. 1990. Parasitism and mate London, Series B, 263, 1233–1238. competition. Oikos, 58, 120–122. Russell, D. M., Atwood, H. L., Whittaker, J. S., Itakura, T., Walker, Koenig, W. D. & Albano, S. S. 1985. Patterns of territoriality and P. M., Mickle, D. A. & Jeejeebhoy, K. N. 1984. The effect of mating success in the white-tailed skimmer Plathemis lydia fasting and hypocaloric diets on the functional and metabolic (Odonata: Anisoptera). American Midland Naturalist, 114, 1–12. characteristics of rat gastrocnemius muscle. Clinical Science, 67, Maniatis, T. & Tasic, B. 2002. Alternative pre-mRNA splicing and 185–194. proteome expansion in metazoans. Nature, 418, 236–243. Schall, J. J. & Sarni, G. A. 1987. Malarial parasitism and the behavior Marden, J. H. 1989. Bodybuilding dragonflies: costs and benefits of of the lizard, Sceloporus occidentalis. Copeia, 1987, 84–93. maximizing flight muscle. Physiological Zoology, 62, 505–521. Schall, J. J., Bennett, A. F. & Putnam, R. W. 1982. Lizards infected Marden, J. H. & Rollins, R. A. 1994. Assessment of energy reserves with malaria: physiological and behavioral consequences. Science, by damselflies engaged in aerial ‘wars of attrition’ for mating 217, 1057–1059. territories. Animal Behaviour, 48, 1023–1030. Siva-Jothy, M. T. & Plaistow, S. 1999. A fitness cost of Marden, J. H. & Waage, J. K. 1990. Escalated damselfly territorial eugregarine parasitism in a damselfly. Ecological Entomology, contests are energetic wars of attrition. Animal Behaviour, 39, 954– 25, 465–470. 959. Stearns, S. C. 1992. The Evolution of Life Histories. Oxford: Oxford Marden, J. H., Fitzhugh, G. H., Wolf, M. R., Arnold, K. D. & University Press. Rowan, B. 1999. Alternative splicing, muscle calcium sensitivity, Wiley, R. H. & Poston, J. 1996. Perspective: indirect mate choice, and the modulation of dragonfly flight performance. Proceedings competition for mates, and coevolution of the sexes. Evolution, 50, of the National Academy of Sciences, U.S.A., 96, 15304–15309. 1371–1381.