Anticipatory flexibility: larval population density in moths determines male investment in antennae, wings and testes

Abstract

Developmental plasticity provides individuals with a distinct advantage when the reproductive environment changes dramatically. Variation in population density, in particular, can have profound effects on male reproductive success. Females may be easier to locate in dense populations, but there may be a greater risk of sperm competition. Thus, males should invest in traits that enhance fertilization success over traits that enhance mate location. Conversely, males in less dense populations should invest more in structures that will facilitate mate location. In Lepidoptera, this may result in the development of larger antennae to increase the likelihood of detecting female sex pheromones, and larger wings to fly more efficiently. We explored the effects of larval density on adult morphology in the gum-leaf skeletonizer moth, Uraba lugens, by manipulating both the number of larvae and the size of the rearing container. This experimental arrangement allowed us to reveal the cues used by larvae to assess whether absolute number or density influences adult responses. Male investment in testes size depended on the number of individuals, while male investment in wings and antennae depended upon larval density. By contrast, the size of female antennae and wings were influenced by an interaction of larval number and container size. This study demonstrates that male larvae are sensitive to cues that may reveal adult population density, and adjust investment in traits associated with fertilization success and mate detection accordingly.

1. Introduction

Life-history theory predicts trade-offs in the allocation of finite resources to morphological structures or physiological processes that respond to different selection pressures [1–3]. Such trade-offs may be linked with particular behavioural strategies that influence the timing of selection: for example, Simmons & Emlen [4] document a trade-off between horns and testes in Onthophagus dung beetles, with horned (guarder) males having smaller testes than hornless (sneaker) males. These trade-offs are particularly evident when different environmental conditions or population ecologies differentially benefit different phenotypes [3,5]. More generally, developmental plasticity, which allows these trade-offs, will be especially advantageous when the reproductive environment is highly variable [6–12]. Population density has an important influence on the reproductive environment: at lower population densities, individuals may face a greater challenge of finding a mating partner, while at high population densities individuals may face greater competition for fertilization success, manifested through male–male competition [2,10,13–15]. In species where population density may fluctuate between generations, selection will favour individuals that can assess their potential reproductive environment during development, and facultatively adjust their relative investment towards those structures best suited to that particular environment [9,12].

Darwin recognized that male mating success can depend upon the speed with which they discover receptive females, and suggested that sexual selection could favour more elaborate receptor organs if they improve the capacity to detect females [16]. The antennae of male moths are crucial structures for mate location, as they are used to detect the minute amounts of pheromone released by the signalling female [17–20]. Although it is often assumed that larger, more elaborate antennae have greater sensitivity to sex pheromones [16–21], there is surprisingly little experimental evidence [22–24]. Recent field experiments support Darwin's largely overlooked prediction: males of the moth Uraba lugens with relatively longer antennae are better able to detect lower quantities of female sex pheromone [22]. Accordingly, males should benefit by investing in larger antennae when population densities are low (or the population is widely dispersed) as this may increase their likelihood of detecting a less abundant chemical signal. Indeed, males of moth species with typically lower population densities have larger and more elaborate antennae, suggesting an evolutionary link between population density and antennal size [23]. Other traits involved in mate searching that may also be favoured at low population densities include larger wings, more flight muscle and larger visual sensory organs [10,13,15]. Investment in these traits may be traded-off against other reproductive functions: for example, the risk of sperm competition may increase with population density, and males may respond with greater investment in sperm production [9,11,15,25–27].

Here, we explore the effects of larval density on investment in adult mate searching and mating effort traits in the gum-leaf skeletonizer moth, U. lugens. This common species is found across Australia, with invasive populations in New Zealand [28]. The antennae of females are simple filiform structures, while males have bipectinate antennae with side branches coming off the main stem, thereby increasing the surface area of the antennae. Larvae feed on a wide range of eucalypts and can be locally damaging when in high densities [29]. They are gregarious in early instars, becoming more solitary in the later instars when the larvae disperse, prior to searching for pupation sites ([28] and references therein). Following Gage [15], we predict that males reared in lower densities will invest more in structures involved in mate searching, such as antennae and wings, while males reared at higher densities will invest in larger testes. However, we do not predict such differential investment in females. Additionally, we explore whether larvae can infer adult population density using different cues such as tactile and chemical cues [8,11], by varying both the number of larvae and the volume of space in which they are reared.

2. Material and methods

Clusters of eggs of U. lugens were collected from Royal Park (Parkville, Melbourne, 37°47′36.7″ S, 144°57′13.0″ E) in August 2011 and initially reared in 500 ml plastic containers (120 × 170 × 35 mm) with two to three clutches of eggs in each container. The eggs were stored in incubators at 22°C and in 15 L : 9 D cycle. Larvae that reached their second instar were randomly assigned to a treatment within an orthogonal experimental design of four treatment groups, comprising containers of different dimensions (small: 500 ml, 120 × 170 × 35 mm; large: 1000 ml, 120 × 170 × 68 mm) with different numbers of larvae (20 or 50). The volume per larva for each treatment, therefore, varied as follows: 50 ml per larva (20 larvae—large container); 25 ml per larva (20 larvae—small container); 20 ml per larva (50 larvae—large container) and 10 ml per larva (50 larvae—small container). The container size and number of larvae in each treatment optimized larval survival and created a similar larval density in the largest container with many larvae and the small container with a few larvae. Larvae were fed leaves of eucalyptus species ad libitum and containers were cleaned regularly to remove old leaf matter and frass. Larvae that had fully pupated were removed, weighed and stored in individual containers in the same incubators. We recorded when each adult eclosed, and the adult was then stored in 70% ethanol. The development time for each individual was recorded as the number of days elapsed from the day of egg hatching to the day of adult eclosion.

The wing length and the antennal length of both adult males and females were measured, as was the testes area of adult males. One forewing and one antenna were removed from each individual, cutting the wing as close as possible to the base (figure 1a) and cutting the antennae below the scape. The male antenna was flattened on a small piece of putty (Blu-Tack™), ensuring the flagellum was completely flat (figure 1b). The female antenna and both the male and female wings were laid out on a white tile. Photos of the antennae and wing were taken using a camera with a macro lens (Sony T77) with a scale bar included, and measurements of the structures (figure 1) were taken from the images using ImageJ software [30]. The males were then dissected to obtain measures of testes size. The abdomen was cut open and white abdominal tissue teased apart to expose the testes, which are clearly visible by their dark red/brown colour. Any adjoining tissue was removed from the testes and images of the largest side were taken under a Leica MZ12 microscope and an auto-montaging camera (figure 1c). The areas of the testes were measured from the images using the ImageJ software [30].

Figure 1.

Illustration of how measurements of morphological structures were obtained, shown with a white line: (a) wing length, same for both males and females, (b) male antennae length, female antennae were measured in the same way and (c) testes area. (Online version in colour.)

Linear mixed models predicting pupal weight, developmental time, antennal length, testes area and forewing length were constructed with experimental treatments (number of larvae, container size) and sex as independent variables together with interaction terms. Container ID was included as a random effect and, due to significant effects of treatments on pupal weight (electronic supplementary material, table S1), pupal weight was included as a fixed effect in the analyses of morphological variables to control for body size. Sex was initially included in the models to determine statistically whether males and females responded similarly to the treatments. Significant interaction terms with sex suggest that males and females respond differently, so we ran separate models for males and females. Testes area was positively skewed and, therefore, −1/x transformed to improve normality. The full model was initially examined, non-significant (p > 0.05) interaction terms were sequentially removed and the reduced model was reported. All analyses were conducted in R [31], using the lme4 and lmerTest packages [32,33].

3. Results

Models for pupal weight, developmental time, antennal length and wing length included a significant interaction between sex and one or both treatments (electronic supplementary material, table S1). Therefore, we created models testing the sexes separately. Models describing the variation in pupal weight and developmental time revealed significant effects of both larval numbers and container size for both males and females (electronic supplementary material, table S1). Strong treatment effects on antennal and wing lengths, after controlling for pupal weight were revealed for males but not females (electronic supplementary material, table S1). The relative antennal length of males was longer when they were reared with fewer larvae (t_39.4 = 3.019, p = 0.004, figure 2c; electronic supplementary material, table S2) and in larger containers (t_41.9 = 4.791, p < 0.001, figure 2c; electronic supplementary material, table S2). Similarly, the relative forewing length was longer when males were reared with fewer larvae (t_40.8 = 3.641, p < 0.001, figure 2a; electronic supplementary material, table S2) and in larger containers (t_43.3 = 4.233, p < 0.001, figure 2a; electronic supplementary material, table S2). Larval density had the opposite effects for testes, which were relatively larger when larvae were reared with more individuals than with fewer individuals (t_37.7 = 5.837, p = 0.021; figure 3; electronic supplementary material, table S2), but the size of the container did not affect relative testes size (t_40.9 = 1.04, p = 0.305, figure 3; electronic supplementary material, table S2).

Figure 2.

Least squares mean ± s.e.m. wing length (millimetres) and antennal length (millimetres) of male and female moths from the four treatment groups. Columns with different letters above them are significantly different from each other.

Figure 3.

Least squares mean ± s.e.m. testis area (square millimetres) of male moths from the four treatment groups. Males reared with 50 larvae had relatively larger testes than males reared with 20 larvae. There was no effect of the container size.

By contrast, female relative wing and relative antennal length were explained by a larval number and container size interaction term (wings: t_28.1 = 2.97, p = 0.006; antennae: t_42.1 = 2.131, p = 0.039), with no significant larval number or container size effects (figure 2b,d; electronic supplementary material, table S1 and S2). Post hoc analysis of the least square means revealed that females reared with 50 larvae in a small container had relatively shorter wings than those reared with 20 larvae in a small container (t_42.91 = 2.714, p = 0.045), but there were no other significant differences between treatment groups for either relative length of wing or antennae (figure 2b,d; electronic supplementary material, tables S1 and S2).

4. Discussion

Darwin recognized that a key component of male reproductive success in a competitive environment is his rapid detection of, and subsequent movement to, a receptive female, and suggested that sexual selection will act on their associated traits [16]. Thus, males at low population densities would benefit from investing in structures that assist with rapid detection of females [2,10,13,14]. We demonstrate that males of U. lugens adjust their investment in receptor organs associated with mate detection accordingly. Males reared in lower larval densities (larger containers, or with fewer other larvae present) had relatively longer antennae and larger wings. A similar effect was not evident for females: relative wing and antennae length were influenced by an interaction between the two experimental treatments and females in the highest density containers had relatively shorter wings than the other treatments. These results are broadly consistent with previous studies of male differential investment in adult reproductive traits in anticipation of different reproductive environments [8], and provide novel evidence that antennal size covaries with population density. This response is facultative, and similar effects of population density on male investment in antennae may not occur over evolutionary time. Long-term selection experiments at different population densities did not result in differences in male antennae size [34].

It is widely held that larger and more elaborate antennae assist male moths in rapidly detecting the typically small quantities (in the order of 3–100 ng h⁻¹) of pheromones that are released by females [17–21]. However, only a small amount of evidence exists to support this idea: artificially adjusting the antennal lengths of male diamondback moths, Plutella xylostella, revealed a positive correlation between antennal length and mating rate [24], and comparative analyses suggest a link between low densities, or more widely dispersed populations, and larger antennae in moth species [23]. Previous experiments on U. lugens revealed that males attracted to a field trap containing one female have longer antennae than males attracted to a trap containing two females that, combined, would release more pheromone [22]. The facultative adjustment of investment in antennal size, according to larval densities, provides compelling support for Darwin's idea: males, anticipating low densities of adults, invest in sensory structures associated with mate location. Additionally, these males also invest in larger wings (see also [15]), which are associated with greater mate-searching success [35].

Theory also predicts greater investment in sperm production under higher risk of sperm competition [2]. Males of U. lugens invest in relatively larger testes when raised in denser juvenile populations, presumably anticipating a high potential for polyandry, a life-history adjustment reported in other insects [9,11,15,25,27]. Larger testes allow more efficient sperm production, which can benefit males in at least two ways. A male with larger testes can produce a new spermatophore more quickly, thereby ensuring he is capable of mating when he encounters a new female (cf. [36]), or he can produce larger spermatophores that may inhibit female re-mating [37–39] or secure a higher paternity share (e.g. [40–42]). Investment in testes size was influenced by the number of larvae in the container only, which is not particularly surprising as sperm competition is determined primarily by the number rather than density of competitors [11,43,44].

Males of U. lugens apparently use different information to determine their investment in different adult morphological structures. While testis size appears to be determined by the number but not density of larvae, antennal and wing length are influenced by larval density: males that eclosed from treatments with 20 larvae/small container and 50 larvae/large container (25 ml per larva and 20 ml per larva, respectively) have similar antennal and wing lengths. Insects use a range of cues to obtain information about the environment, drawing on chemical, acoustic and tactile modalities [8]. Perhaps, male larvae use chemical cues to discern population density, responding to the concentration of volatile odours emanating from larvae, but use tactile cues to assess population size, determined by the number of encounters with different individuals [8,11,15,45]. Using different cues may allow male larvae to fine-tune their relative investment in structures associated with mate search as adults and/or mating effort, giving individuals the greatest reproductive advantage possible.

Data accessibility

The data supporting this article have been made available in Dryad: http://dx.doi.org/10.5061/dryad.qp5g7 [46].

Authors' contributions

T.L.J., M.R.E.S. and M.A.E. conceived the project. T.L.J. collected and analysed the data and prepared the first draft of the manuscript. M.R.E.S. and M.A.E. advised on data analysis and contributed to preparing the manuscript.

Competing interests

We declare we have no competing interests.

Funding

This work was supported by an Australian Research Council (DP0987360) grant to M.A.E.