Effect of juvenile hormone on senescence in males with terminal investment

Authors

  • D. González-Tokman,

    1. Departamento de Ecología Evolutiva, Instituto de Ecología, Universidad Nacional Autónoma de México, México, D. F. México
    Search for more papers by this author
  • I. González-Santoyo,

    1. Departamento de Ecología Evolutiva, Instituto de Ecología, Universidad Nacional Autónoma de México, México, D. F. México
    Search for more papers by this author
  • R. Munguía-Steyer,

    1. Departamento de Ecología Evolutiva, Instituto de Ecología, Universidad Nacional Autónoma de México, México, D. F. México
    Search for more papers by this author
  • A. Córdoba-Aguilar

    Corresponding author
    1. Departamento de Ecología Evolutiva, Instituto de Ecología, Universidad Nacional Autónoma de México, México, D. F. México
    • Correspondence: Alex Córdoba-Aguilar, Departamento de Ecología Evolutiva, Instituto de Ecología, Universidad Nacional Autónoma de México, Apdo. Postal 70-275, Ciudad Universitaria, 04510, México, D. F., México. Tel.: +52 55 56 22 9003; fax: +52 55 56 22 8995;

      e-mail: acordoba@ecologia.unam.mx

    Search for more papers by this author

Abstract

Senescence, a decline in survival and reproductive prospects with age, is controlled by hormones. In insects, juvenile hormone (JH) is involved in senescence with captive individuals, but its effect under natural conditions is unknown. We have addressed this gap by increasing JH levels in young and old wild males of the damselfly Hetaerina americana. We assessed survival in males that were treated with a JH analogue (methoprene), which is known to promote sexual activity, and an immune challenge, which is known to promote terminal investment in reproduction in the studied species. We replicated the same procedure in captivity (to control for environmental variation), where males were deprived of any activity or food. We expected old males to show the lowest survival after being treated with JH and immune-challenged, because the effect of terminal investment on senescence would be exacerbated by JH. However, this should be the case for wild animals, but not for captive animals, as the effects of JH and immune challenge should lead to an increase in high energetic-demanding activities only occurring in the wild. Old animals died sooner compared with young animals in both the wild and captivity, confirming that males are subject to senescence. In wild but not captive animals, JH decreased survival in young males and increased it in old males, confirming that JH is sensitive to the environment when shaping animal senescence. Immune challenge had no effect on survival, suggesting no effect of terminal investment on senescence. Additionally, contrary to the expected effects of terminal investment, with an immune challenge, recapture rates increased in young males and decreased in old males. Our results show that male senescence in the wild is mediated by JH and that terminal investment does not cause senescence. One explanation is that animals undergoing senescence and terminal investment modify their feeding behaviour to compensate for their physiological state.

Introduction

Senescence, or ageing, is an inevitable decline in survival and reproductive prospects with age, which is caused by a progressive reduction in physiological function (Partridge & Gems, 2006; Flatt & Schmidt, 2009). Every living organism is subject to senescence (Ackermann et al., 2007; Książek, 2010). Antagonistic pleiotropy, one of the main explanations for the origin of senescence, poses that beneficial mutations with pleiotropic effects early in life come to have deleterious effects in old age (Williams, 1957; Flatt & Promislow, 2007), implying that senescence is an inevitable part of trade-offs.

Trade-offs can be genetic and physiological, occurring in individuals when demands for limited resources require an organism to choose between different functions. One example is the need to assign limited resources to survival and reproduction (Zera & Harshman, 2001). Given that physiological trade-offs underlie senescence, animals should balance the assignment of resources to various reproductive events (Stearns, 1992; Roff, 2002). This balance occurs even at the risk of increased mortality, an idea otherwise known as terminal investment. The interaction between senescence and terminal investment has been examined mostly in terms of the effects on investment in reproduction (Hoffman et al., 2010; Massot et al., 2011), whereas the effects on survival still need further experimental evaluation (but see Part et al., 1992; Festa-Bianchet & King, 2007).

Senescence is associated with changes in the immune system (DeVeale et al., 2004; Stanley, 2012). The immune response involves the release of reactive oxygen species, leading to the accumulation of oxidative damage (Nappi & Christensen, 2005; González-Santoyo & Córdoba-Aguilar, 2012). The latter is associated with senescence (Finkel & Holbrook, 2000). In general, the immune response declines with age (Doums et al., 2002; Moret & Schmid-Hempel, 2009), and the great demand for the immune system in youth inevitably shortens individual lifespan, especially when resources are scarce (Moret & Schmid-Hempel, 2000). Despite the fact that a decrease in the immune response during senescence has been documented in both vertebrates and invertebrates (Miller, 1991; Doums et al., 2002), the effect on senescence of activating and using the immune system has received much less attention (but see Moret & Schmid-Hempel, 2009).

In insects, juvenile hormone (JH) is a key regulator of ageing (Tu et al., 2006) and terminal investment (i.e. Fronstin & Hatle, 2008). Juvenile hormone is naturally synthesized in the corpora allata, an endocrine gland that is attached to the insect brain, in response to stimulation from the central nervous system, via insulin signalling (reviewed by Flatt et al., 2005). JH regulates many developmental and life history traits, such as resource allocation to reproductive activities or physiological maintenance of the organism, including immunity (Flatt et al., 2005; González-Tokman et al., 2012b). In general, older animals have higher titres of JH than younger individuals (Herman et al., 1981; Fluri et al., 1982; Jesudason et al., 1990). Whereas providing insects with JH, or JH analogues such as methoprene, accelerates ageing (i.e. increasing the probability of physiological deterioration and mortality), down-regulating JH slows ageing (Herman & Tatar, 2001; Tatar et al., 2001, 2001).

Synthesis of JH is strongly sensitive to the environment and can be affected by nutrient intake and mating activity (Tobe & Chapman, 1979; Lessman & Herman, 1983; Lee & Horodyski, 2006; Nouzova et al., 2012). However, studies of JH effects on ageing have been carried out under captivity, where feeding and mating activities are controlled (Keller & Jemielity, 2006). This limitation is especially serious if we consider that ageing is also sensitive to resource availability, which is highly variable in nature (Pletcher et al., 2002).

Recent evidence in rubyspot damselflies, Hetaerina americana, indicates that males show terminal investment as a result of changes in reproduction that come with age, being strongly dependent on health status (González-Tokman et al., 2013). Specifically, old males exhibit high territorial activity when their risk of mortality is experimentally increased with an immune challenge. Contrarily, young males reduce their territorial activity when immune-challenged (González-Tokman et al., 2013), presumably as a cautious behaviour to save resources for future prospects of reproduction (McNamara et al., 2009). Given that terminal investment has negative repercussions on an animal's energetic condition (González-Tokman et al., 2013), it necessarily impacts individual survival and must therefore be regulated by a mechanism involved in the allocation of resources to different life history traits. Juvenile hormone is an evident candidate for such regulatory activity, given that it is involved in trade-offs between immunity and reproduction in calopterygid damselflies and other insects (Rolff & Siva-Jothy, 2002; Rantala et al., 2003; Contreras-Garduño et al., 2009, 2011; González-Tokman et al., 2012b).

Here, we have investigated the role of JH on senescence and survival using males of the damselfly H. americana and employing an experimental approach that involves tests in the wild and in the laboratory. The main goal of the present study is to understand the consequences of JH treatment and male terminal investment on individual survival. Although it is known that infected older males of H. americana increase their mating activity at the cost of an impaired energetic condition (lower fat reserves) (González-Tokman et al., 2013), the possible effect of this terminal investment on survival is still not clear.

First, we tested the effect of JH supplementation on survival probabilities of young and old males under natural conditions. Second, we added an immune challenge to test for the possible effect of JH on terminal investment and the effect of the latter on senescence. These two tests were carried out in the wild, specifically in the mating sites of males. A similar manipulation was produced in captivity to control for varying environmental conditions and energetic expenditure that animals would incur in the wild (Adamo et al., 2010; González-Tokman et al., 2011). We proposed four theoretical suppositions (i) that older animals would die sooner than young animals, both in the wild and in captivity (confirming senescence), (ii) that JH and immune challenge would interact to explain senescence in the wild, but not in captivity (as feeding and sexual activities, which are major drivers of energetic exhaustion, are prevented in the latter situation), (iii) that mortality in the wild would be caused by the synergistic effect of JH and immune challenge on senescence and terminal investment and (iv) that terminal investment and senescence would determine fidelity to reproductive sites and, consequently, the recapture rates of our analyses in the wild.

Materials and methods

Study subject

The damselfly H. americana shows intense male–male competition for riverine territories where females arrive to mate (Grether, 1996; Córdoba-Aguilar et al., 2009). Males that can acquire and defend territories achieve much higher mating success than nonterritorial males (Grether, 1996; Serrano-Meneses et al., 2007). Territorial status is related to physiological condition – compared with nonterritorial males, territorial males have more energetic reserves and are more immunocompetent, as measured by encapsulation, phenoloxidase and lytic activity, and survival following bacterial infection (Contreras-Garduño et al., 2006, 2007).

Field work

The present study was carried out along the Apatlaco River, Morelos, Mexico (18°45′55′'N, 99°14′45′'W), in October and November of 2012. Between 1100 and 1500, the period of the greatest territorial activity, we captured males with a butterfly net on the river shore. To classify male age, we used criteria described by Plaistow & Siva-Jothy (1996), according to which adult damselflies are separated into four age classifications based on appearance. Age 1 comprises the youngest teneral animals that have just emerged, with their bodies still soft and their wings flexible and undamaged, and without the pigmentation typical of adults. Age 2 encompasses young and sexually mature animals that already fight for territories, with their wings flexible from the nodus to the tip and with typical adult pigmentation already fixed. Age 3 animals have fully matured, with harder bodies and less flexible wings and with some signs of pruinescence in the abdomen and thorax. Age 4 animals are the oldest, with their wings unflexible, papery-like and frequently damaged, and with signs of pruinescence on their abdomen and thorax. We only used males of ages 2 (young) and 4 (old), as at these ages animals cope well with marking effects (D. González-Tokman, I. González-Santoyo, R. Munguía-Steyer and A. Córdoba-Aguilar, unpublished) and should differ more between groups than between 2 and 3 or 3 and 4.

Experimental protocol

Immediately after capture, each male was kept in a vial stored on ice for 10 min (0 °C) prior to manipulation with a combination of two treatments: one of two hormonal treatments (JH) and one of two immune treatments (Imm). Individuals were randomly allocated to those treatments.

Hormonal treatment (Met) consisted of a topical application of methoprene acid, an analogue to juvenile hormone. From a dilution of 5 mg of methoprene acid in 1 mL of distilled water, we took 1 μL and diluted it in 1 mL of acetone. Using a micropipette, 3 μL of the solution (15 ng of methoprene acid) was placed on the male dorsal part of the head so that it would rapidly penetrate the cuticle near the corpora allata (Flatt et al., 2005). This dose has been calculated for damselflies based on JH titres found in other insects (Contreras-Garduño et al., 2011). This treatment has been successfully used to increase reproductive behaviour and affects immunity and survival in damselflies (including H. americana) and other insects (Teal et al., 2000; Rantala et al., 2003; Contreras-Garduño et al., 2009; González-Tokman et al., 2012b). As a control for the hormonal treatment, we used 3 μL of acetone (Ac).

Immediately after the application of the JH treatment, all males received an immune treatment (Imm). In the experimental group, a 2-mm piece of rubbed nylon (0.18 mm thick) was implanted in the ventral part of the fourth abdominal segment (Rantala et al., 2000). When implanted, insects respond immunologically by depositing a melanin layer around the nylon (Rantala & Roff, 2007; González-Tokman et al., 2012a; Moreno-García et al., 2013). As a control for the immune treatment (Control), a similar piece of nylon was inserted into the abdomen and immediately removed, avoiding melanization. To estimate body size, the left anterior wing was measured from the point of insertion to the tip with a digital calliper (± 0.01 mm).

Survival in the field

To estimate survival in the field in response to age and treatments, we used a capture–mark–recapture approach (for similar methods see Munguía-Steyer et al., 2010; Buzatto et al., 2010; González-Tokman et al., 2012b) and analysed data with the MARK 6.1 program (White & Burnham, 1999). This approach enables the researcher to distinguish survival probabilities (φ) from recapture probabilities (p) by calculating maximum likelihood estimates from encounter histories of marked animals (Lebreton et al., 1992). With this method, a group of models is constructed and the best model is selected based on Akaike's information criterion (AIC).

After receiving the combination of JH and Imm treatments, males were marked on the left anterior wing with a three-digit number made with a black permanent ink felt pen. From October 25th to 29th, we collected a total of 325 males and allocated them to the following treatments: young (Ac-Control, N = 41; Ac-Imm, N = 41; Met-Control, N = 41; Met-Imm, N = 41) and old (Ac-Control, N = 41; Ac-Imm, N = 40; Met-Control, N = 40; Met-Imm, N = 40). Each capture and marking day, we collected the following number of young and old males: day 1, 39 young and 43 old; day 2, 38 young and 38 old; day 3, 45 young and 31 old; and day 4, 42 young and 49 old. Marked animals were released to the same site where they were captured and the presence of marked animals was recorded daily for 22 days (from October 26th to November 16th). Detection of marked animals was based on daily surveys carried out from 1100 to 1400 h by two observers walking in the same areas of the river.

Survival in captivity

An additional set of 140 old and young males were collected and manipulated with the same combinations of treatments mentioned above. Sample sizes were as follows: young (Ac-Control, N = 17; Ac-Imm, N = 20; Met-Control, N = 20; Met-Imm, N = 20) and old (Ac-Control, N = 14; Ac-Imm, N = 16; Met-Control, N = 16; Met-Imm, N = 17). After manipulation, animals were kept in captivity, placed individually in 5 mL essay tubes, with a perch and a cap of humid cotton to keep humidity inside the containers, and kept in the shade. Water and food were not provided. The presence of dead animals was registered every 4 h during the first day and every 8 h for the following days. The experiment ended when the last animal died.

Statistics

Survival in the field was analysed with Cormack–Jolly–Seber models that estimate survival and recapture parameters from encounter histories (Lebreton et al., 1992). In this procedure, many explanatory models are tested and compared with alternative models based on AIC. A difference of two AIC units was considered enough to differentiate between two alternative models. Test models included three factors as explanatory variables of survival and recapture (age, hormone treatment (JH), and immune treatment (Imm)). There was an extremely large number of possible models to test in our mark–recapture analysis, given that the model selection process involves testing the effect of including or removing a given explanatory variable (or interaction between two or three variables) from an initial model. Therefore, we were not able to test all possible models and used a set of criteria for model selection.

Firstly, the predictors of survival (φ) were fixed in the global model φ(Age × JH × Imm). Then, recapture predictors (p) were varied, starting with the global model p(Age × JH × Imm) and reducing it by removing terms one by one until we got the simplest model, where recapture probabilities are constant (p(.)). At this point, we had tested 19 models. We selected three models that best explained (based on the ΔAIC < 2 units) recapture probability: p(Age + JH + Imm + Age:JH + Age:Imm), p(Age + Imm + Age:Imm + JH:Imm) and p(Age + JH + Imm + Age:Imm). These three models were considered in the second step of model selection, where we fixed recapture (p) in all three models and varied survival (φ) parameters, starting with the global model φ(Age × JH × Imm) and reducing it by removing terms one by one until we got the simplest model, where survival probabilities are constant (φ(.)). A total of 73 models were tested (see Table S1). Model goodness of fit was tested in the global model φ(Age × JH × Imm), p(Age × JH × Imm) by estimating overdispersion with a median c-hat approach (White & Burnham, 1999). Values higher than three units indicate high overdispersion. Our estimated overdispersion value was relatively low (median c-hat = 1.495) and was corrected in all models. Therefore, we used AIC for overdispersed data (QAIC) in our model selection process (Burnham & Anderson, 2002). All models used a logit link function. Given that there was no single model with considerably higher support than the others (i.e. ΔQAIC ≥ 2), we used the averaging of models with ΔQAIC ≤ 6 (Richards, 2008; Grueber et al., 2011) to estimate final survival and recapture parameters. Averaged models had an explanatory power of 92% (QAIC weight sum = 0.92; Burnham & Anderson, 2002).

Survival in captivity was analysed with a proportional hazard Cox regression model that included Age × JH × Imm × Body size. Again, the best model was selected based on AIC. Differences in body size were analysed with t-test and one-way anova. Analyses were carried out in R software 2.10.0 (R Development Core Team, 2009) and MARK 6.1 (White & Burnham, 1999).

Results

Survival in the field

Despite the fact that old males were significantly larger that young males (t-test t = 7.516, P < 0.001, N = 325), there were no differences in size between treatments within either age group, whether young (anova F3,160 = 0.107, P = 0.956) or old (anova F3,157 = 0.518, P = 0.671).

Most of the best supported models used to explain differences in survival of marked animals in the field suggest that survival depends on the interaction of age and hormonal treatment (Age × JH) (Table 1) and that there is no effect of immune treatment (Imm) on male survival (Table 1, Fig. 1). The general trend in the present study was that with Met, survival decreased in young males and increased it in old males, independently of immune treatment (Fig. 1). Whereas the negative effect of Met treatment on survival was not significant in young males (Table S2), Met treatment significantly increased survival in old males when immune-challenged (Table S2; Fig. 1). Old males that were only manipulated with control treatments (Ac-Control) survived less than young control males (Table S2; Fig. 1), which confirms that senescence existed in control animals.

Table 1. Summary of model selection process to test the effect of age, a hormonal (JH) and an immune treatment (Imp) on daily survival and recapture probabilities of marked Hetaerina americana males in the wild
Model descriptionQAICcΔ QAICcAICc weightModel likelihoodNo. of parametersQ Deviance
Survival componentsRecapture components
  1. AIC, Akaike's information criterion; JH, juvenile hormone.

Age + JH + Age:JHAge + JH + Imm + Age:Imm2589.2970.0000.1741.00092571.154
Age + JH + Imm + Age:JH+Age:ImmAge + JH + Imm + Age:Imm2590.5131.2160.0950.544112568.303
Age + JH + Age:JHAge+JH+Imm+Age:Imm+JH:Imm2590.5151.2180.0950.544102570.340
Age + JH + Age:JHAge + JH + Imm + Age:JH + Age:Imm2591.0731.7770.0720.411102570.898
Age + JH + Imm + Age:JHAge + JH + Imm + Age:Imm2591.2421.9450.0660.378102571.067
Age + JH + Imm + Age:JH + Age:Imm + JH:ImmAge + JH + Imm + Age:Imm2591.3282.0320.0630.362122567.080
Age + JH + Imm + Age:JH + Age:ImmAge + JH + Imm + Age:Imm + JH:Imm2591.8212.5240.0490.283122567.572
Age + JH + Imm + Age:JH + JH:ImmAge + JH + Imm + Age:Imm2591.8572.5600.0480.278112569.647
Age + JH + Imm + Age:JH + Age:ImmAge + JH + Imm + Age:JH + Age:Imm2592.2572.9600.0400.228122568.008
Age + JH + Imm + Age:JHAge + JH + Imm + Age:Imm + JH:Imm2592.4583.1610.0360.206112570.247
Age + JH + Imm + Age:JH + Age:Imm + JH:ImmAge + JH + Imm + Age:Imm + JH:Imm2592.8543.5570.0290.169132566.564
Age + JH + Imm + Age:JHAge + JH + Imm + Age:JH + Age:Imm2593.0203.7230.0270.155112570.810
Age + JH + Imm + Age:JH + Age:Imm + JH:ImmAge + JH + Imm + Age:JH + Age:Imm2593.1033.8060.0260.149132566.813
Age + JH + Imm + Age:JH + JH:ImmAge + JH + Imm + Age:Imm + JH:Imm2593.3194.0230.0230.134122569.071
Age*JH*ImmAge + JH + Imm + Age:Imm2593.3204.0230.0230.134132567.030
Age + JH + Imm + Age:JH + JH:ImmAge + JH + Imm + Age:JH + Age:Imm2593.6534.3560.0200.113122569.404
Constant (.)Age + JH + Imm + Age:Imm2594.3505.0530.0140.08062582.283
Age*JH*ImmAge + JH + Imm + Age:Imm + JH:Imm2594.8515.5540.0110.062142566.515
Age*JH*ImmAge + JH + Imm + Age:JH + Age:Imm2595.0995.8020.0100.055142566.763
Figure 1.

Daily survival probabilities (estimates ± 95% C. I.) of Hetaerina americana males of different ages exposed to a hormonal (Met or Ac) and an immune (Imp or Control) treatment. Estimates and C. I. were calculated from model averaging of the best supported models. Sample sizes are shown above the bars. Confidence intervals can be seen in Table S2.

Regarding recapture probabilities, there was a clear interaction of age and immune treatment (Age × Imm) (Table 1; Fig. 2). Recapture probability was found to increase by immune activation (Imm) in young males and to decrease with Imm in old males. In both young and old males, Met treatment had a negative effect on recapture probabilities, which was statistically significant for old males (Table S2; Fig. 2).

Figure 2.

Daily recapture probabilities (estimates ± 95% C. I.) of Hetaerina americana males of different ages exposed to a hormonal (Met or Ac) and an immune (Imp or Control) treatment. Estimates and C. I. were calculated from model averaging of the best supported models. Sample sizes are shown above the bars. Confidence intervals can be seen in Table S2.

Survival in captivity

Survival of males in captivity was only explained by age, with young males surviving significantly more than old males (z = 6.17, P < 0.001, N = 140; Fig. 3). None of the other factors – hormonal treatment (JH), immune treatment (Imm), or body size – explained differences in male survival in captivity.

Figure 3.

Survival in captivity of young (N = 77) and old (N = 63) Hetaerina americana males.

Discussion

We have shown that males of the territorial damselfly H. americana are subject to senescence, consistent with our first prediction: survival probabilities are lower for old than for young control animals both in the wild and in captivity. This occurred in spite of the inevitable sampling bias resulting from our experimental design: (i) when sampling old males, we are biasing our sample towards long-lived individuals (only individuals that were able to survive until old age), whereas when sampling young males, we are including both short- and long-lived individuals; and (ii) seasonal effects on body size given that old males were larger than young males (which is coherent with what has been previously documented in this species; Córdoba-Aguilar et al., 2009). Partly in agreement with our second prediction, the JH analogue ameliorated senescence in wild animals and had no effect in captive animals (the latter of which did not feed or mate). This suggests that either foraging or territorial activity may have been altered by the JH analogue treatment in wild animals.

The present finding that Met increased survival in old, immune-challenged males was unexpected, as most evidence associates senescence with high levels of JH (Herman & Tatar, 2001; Tatar et al., 2001; Flatt & Kawecki, 2007). However, this previous evidence comes from laboratory studies. In natural conditions, on the other hand, there are many uncontrolled factors favouring senescence that can be affected by JH. One such factor is how animals can compensate for experimentally induced high JH levels in the field. JH has different physiological effects when resource availability varies (Trumbo & Robinson, 2004), as occurs in natural conditions. Nutritional state is a key regulator of the neuroendocrine system, because it is implied in the insulin signalling pathway leading to JH production. In particular, when food is scarce, insulin-like peptide production decreases and expression of insulin-like genes is repressed (Ikeya et al., 2002), leading to low JH synthesis and increased lifespan (Flatt et al., 2005). Animals can vary their feeding activity in response to environmental pressures, by either increasing it (Lee et al., 2004; González-Tokman et al., 2011), decreasing it (Adamo et al., 2010) or selecting their diets adaptively (Mayntz et al., 2005; Pekár et al., 2010; Ponton et al., 2011). Given that dietary restriction is associated with prolonged lifespan in insects (Grandison et al., 2009), one possibility is that JH in old males may have induced a reduction in their foraging activity, thereby explaining their observed increased lifespan. Unfortunately, we did not conduct any behavioural study along with our experiment. Thus, future studies should evaluate the effect of JH on feeding behaviour.

There was no clear support for our third prediction in relation to a synergistic effect of JH and immune challenge on senescence and terminal investment, which would affect the survival of old males more intensively than younger males. According to this supposition, Met would increase reproductive activity and thus to faster mortality, especially in immune-challenged males (González-Tokman et al., 2012b, 2013). This expectation is based on previous evidence in terms of resource allocation to sexual traits in damselflies (Rantala et al., 2003; Contreras-Garduño et al., 2009, 2011). This was possibly the case for young males in our study, who showed increased mortality when treated with Met. Nevertheless, old males supplemented with Met showed reduced mortality, especially when they were immune-challenged, suggesting that JH determines how organisms schedule their investment in reproduction during their lifetime (McNamara et al., 2009). It seems that JH mediates a switch towards an alternative reproductive trajectory by which animals save resources for better chances of reproduction in the future (McNamara et al., 2009). Besides modifying nutrient intake, JH could have modified reproductive behaviour in males treated with Met, altering survival probabilities in our field study. Again, these behavioural effects should be investigated in the future.

Regarding the fourth prediction, a high recapture rate should be expected for animals whose mortality risk is also high, as a consequence of terminal investment. This prediction was not found to hold true in the present study, would imply that high recapture rate is positively correlated with territorial tenure, which is not necessarily the case. Indeed, there are factors other than territorial tenure that determine recapture rates. As a consequence, territorial animals would not always show higher recapture rates than nonterritorial animals (Munguía-Steyer et al., 2010).

On the other hand, it seems that immune challenge, rather than hormone treatment, influenced faithfulness to mating sites. In particular, with immune challenge, the recapture probability increased in young males and decreased in old males. There are at least three factors associated to faithfulness to a mating site in damselflies: mating experience (acquisition of mating in the past leads to further faithfulness; Switzer, 1997; Nagy et al., 2008), infection (the higher the re-infection probability, the more likely to abandon a mating area; Suhonen et al., 2010; Rantala et al., 2010) and male aggression (the greater the male aggressiveness, the less likely it is for damselflies to stay in a mating place; Córdoba-Aguilar, 1994). We do not have any data to assess the effect of mating experience. As for infection, the fact that we used the immune challenge approach suggests that avoiding re-infection may explain our results for old males. This age-related effect of infection on recapture rate may actually make sense in terms of expected mating opportunities, as these are linked to survival in the mating site for males of different ages (see also Rantala et al., 2010).

In terms of avoiding aggression, it is possible that old males searched for mates outside the mating riverine territories, as they were unable to face the intense male–male competition in these areas. It is common in Hetaerina that relatively old males are chased away from mating territories (Córdoba-Aguilar, 1994; Guillermo-Ferreira & Del-Claro, 2011). These old males may look for mating opportunities elsewhere to save their already reduced energetic stores or to replenish energy reserves.

Environmental conditions are key determinants of animal senescence, meaning that an evaluation of senescence under controlled laboratory conditions can provide an incomplete picture of natural selection acting on senescent animals (Williams et al., 2006). On the other hand, detecting ageing rates under natural conditions has been challenging for two main reasons. Firstly, this requires periodical monitoring of animals across their lifetimes, which can be logistically demanding (Nussey et al., 2008), and secondly because many factors can mask senescence. Using a capture–mark–recapture model, we have detected senescence in a wild population of a short-lived territorial insect. Senescence is not the cause of mortality, but it increases its probability. Our unexpected findings concerning the effect of JH on senescence in the wild highlight the importance of using longitudinal studies in natural populations to understand the evolution of senescence (Williams et al., 2006; Nussey et al., 2008).

Acknowledgments

The authors thank F. Baena-Díaz for help during fieldwork, and R. Torres-Avilés, H. Drummond and three reviewers for valuable comments on the article. This research was funded by a PAPIIT (UNAM-DGAPA) Grant IN 222312. The present article constitutes partial fulfilment of the Programa de Doctorado en Ciencias Biológicas at the Universidad Nacional Autónoma de México. DG-T and IG-S are grateful for the scholarship provided by the Consejo Nacional de Ciencia y Tecnología (CONACYT, México). Authors declare that they do not have any conflict of interest.

Ancillary