Save

Adult sex ratios affect mating behaviour in the common housefly Musca domestica L. (Diptera; Muscidae)

In: Behaviour
Authors:
Stefan ter Haar Groningen Institute for Evolutionary Life Sciences; University of Groningen; Groningen, The Netherlands

Search for other papers by Stefan ter Haar in
Current site
Google Scholar
PubMed
Close
https://orcid.org/0000-0002-3215-4755
,
Ljubinka Francuski Groningen Institute for Evolutionary Life Sciences; University of Groningen; Groningen, The Netherlands
Protix BV, Industriestraat 3, 5107 NC Dongen, The Netherlands

Search for other papers by Ljubinka Francuski in
Current site
Google Scholar
PubMed
Close
https://orcid.org/0000-0002-6549-3315
,
Jean-Christophe Billeter Groningen Institute for Evolutionary Life Sciences; University of Groningen; Groningen, The Netherlands

Search for other papers by Jean-Christophe Billeter in
Current site
Google Scholar
PubMed
Close
https://orcid.org/0000-0001-7796-4068
,
Martijn A. Schenkel Groningen Institute for Evolutionary Life Sciences; University of Groningen; Groningen, The Netherlands
Department of Biology, Georgetown University, Washington, DC, USA

Search for other papers by Martijn A. Schenkel in
Current site
Google Scholar
PubMed
Close
https://orcid.org/0000-0002-7721-319X
, and
Leo W. Beukeboom Groningen Institute for Evolutionary Life Sciences; University of Groningen; Groningen, The Netherlands

Search for other papers by Leo W. Beukeboom in
Current site
Google Scholar
PubMed
Close
https://orcid.org/0000-0001-9838-9314
Open Access

Abstract

Adult sex ratio determines the level of mate availability and intrasexual competition for each sex. Sex ratio biases have been proposed to enhance the productivity of animal rearing procedures. However, behaviour may change in response to sex ratio manipulations that may counteract potential benefits. We investigated how sex ratios affected mating behaviour of the housefly Musca domestica, a species used in the animal feed industry. We hypothesized a reduced courtship effort and mating latency and increased ejaculate allocation (copulation duration) under male-biased sex ratios, whereas female-biased sex ratios would lead to the opposite effects. However, courtship effort was reduced in female-biased groups, implying reduced male harassment. Mating latency was lower and copulation lasted longer in female-biased groups, which may reduce reproduction time and increase female fecundity and lifespan. Our results indicate that in houseflies, female-biased sex ratios cause behavioural changes in both sexes that could positively contribute to reproductive output.

1. Introduction

Sexual selection is a major factor in the evolution of mating systems and in shaping mating behaviour (Shuster, 2009). Males produce smaller gametes than females (Parker et al., 1972; Parker & Pizzari, 2010) and therefore invest less in individual offspring in terms of allocating resources to each gamete (Arnqvist & Nilsson, 2000). This could allow males to have considerably more offspring than females, provided that both sexes dedicate the same amount of energy to gamete production (Bateman, 1948). Consequently, males and females can evolve different reproductive strategies, where males typically compete with each other for access to females and females usually display mate choice (Lindsay et al., 2019). Both sexes can also invest in other factors related to offspring production. Males typically invest into securing mating and fertilization (Arnqvist & Nilsson, 2000) via mate searching, competition for mates and courtship performance (Leftwich et al., 2012), as well as producing large quantities of sperm in response to the risk of sperm competition, which reduces the likelihood of successful fertilization (Parker & Pizzari, 2010). Females invest in offspring production and parental care (Arnqvist & Nilsson, 2000) and female choice is costly through delayed commencement of reproduction, and the risk of remaining unmated (Andersson & Simmons, 2006; Lindsay et al., 2019; Kohlmeier et al., 2021). Investment in specific aspects of reproduction is usually plastic and their payoff in terms of costs and benefits depends on conditions such as the social environment, level of competition and mate availability (Billeter et al., 2012; Leftwich et al., 2012; Dore et al., 2020).

The plasticity of reproductive behaviour, and the influence of controllable factors thereon, provides both opportunities and threats in the context of applications in which reproductive capacity is of major importance. Examples of this include the mass rearing of insects, e.g. for feed production and as a more sustainable alternative to traditional livestock (van Huis & Oonincx, 2017), or for biological control and sterile insect technique (SIT) programmes (Sørensen et al., 2012). One controllable factor that influences reproductive behaviour is manipulation of the sex ratio, as biased sex ratios alter the level of intrasexual competition and availability of mates. Sex ratio manipulation could be used to select for desirable reproductive behaviour that contributes to optimization of mass rearing of insects (Fitz-Earle & Barclay, 1989; Desa et al., 2018; González-López et al., 2019; Gou et al., 2019), and it is therefore important to study how population sex ratios affect mating behaviour.

Different sex ratio types have been defined, including the primary sex ratio (number of male and female offspring; Székely et al., 2014), adult sex ratio (number of sexually mature males and females; Székely et al., 2014) and operational sex ratio (ratio of males and females available for mating/reproduction; Emlen & Oring, 1977). Although the adult and operational sex ratio can be similar and are sometimes used interchangeably, it is important to distinguish between the two (Kokko & Jennions, 2008). The adult sex ratio determines the initial operational sex ratio but is not dependent on reproductive behaviour, whereas the operational sex ratio is altered whenever an individual’s mating status changes, which can cause the adult and operational sex ratios to diverge over time. Adult and operational sex ratios are therefore not necessarily closely linked and can depict different temporal patterns (Carmona-Isunza et al., 2017) and are predicted to differently affect the evolution of sex roles and mating behaviour (Kokko & Jennions, 2008; Székely et al., 2014; Jennions & Fromhage, 2017). As such it may be important to consider which sex ratio to manipulate.

Sex ratio biases may have different effects on female and male behaviour depending on the sex that is overrepresented. Under male-biased sex ratios, males experience higher levels of intrasexual competition, and may therefore invest in pre-copulatory traits that enhance their mating success (Leftwich et al., 2012). Males increased courtship performance in the presence of other males in the fruit fly Ceratitis capitata and housefly Musca domestica (Carrillo et al., 2012; Leftwich et al., 2012). The presence of males prior to copulation can also decrease the mating latency (Bretman et al., 2009; Abraham et al., 2015; Dore et al., 2020), possibly speeding up reproduction time. Moreover, female remating exposes both sexes to post-copulatory sexual selection, e.g. through sperm competition and (cryptic) female choice (Birkhead & Pizzari, 2002). Males can increase sperm and seminal fluid protein production (Bretman et al., 2009; Fedorka et al., 2011) and allocation (Wigby et al., 2009; Garbaczewska et al., 2013; Abraham et al., 2015) under increased risk and intensity of sperm competition, which may influence female fecundity. For example, female houseflies M. domestica had larger first clutch size and higher offspring survival under male-biased sex ratios, presumably resulting from increased accessory gland protein transmission to females by males (Carrillo et al., 2012). Male-biased sex ratios could therefore positively affect productivity, depending on the species.

Female-biased sex ratios are however more commonly expected to enhance reproductive output compared to equal or male-biased sex ratios (Fitz-Earle & Barclay, 1989). This is because only females lay eggs and owing to their vastly larger output in terms of gametes, one male ought to suffice to inseminate many females. Resources can then be used more efficiently as fewer resources are expended on maintaining a surplus of males, effectively bypassing the twofold cost of sex (Maynard Smith, 1971, 1978). Indeed, female fecundity was increased in female-biased sex ratios in the Tsetse flies Glossina fuscipes fuscipes and Glossina pallidipes (Desa et al., 2018) and the Mexican fruit fly Anastrepha ludens (González-López et al., 2019). Moreover, female Drosophila melanogaster initiate oviposition sooner when in a group than when alone, and sooner in higher densities (Bailly et al., 2021), which may increase production in the short term but lower it in the long term. However, female-biased sex ratios may only be profitable to a certain extent, as large biases can lead to a shortage of males and unmated females (Desa et al., 2018). Moreover, males may alter their reproductive behaviour as well, which may confound the expected effect of a female-biased sex ratio on productivity. Given that both male- and female-biased sex ratios can alter reproductive behaviour in a manner that may benefit the reproductive output, it is important to study how exactly biased sex ratios influence the mating behaviour in the species of interest.

We investigated the effect of both male- and female-biased adult sex ratios on mating behaviour in the common housefly M. domestica. Houseflies are increasingly used for feed production, are easily reared, and the ability of its larvae to grow on waste products makes them a useful resource in establishing a more circular economy (Francuski et al., 2020). Female egg production is a limiting factor in rearing efficiency of houseflies (Pastor et al., 2015). If reproductive behaviour is influenced by the adult sex ratio, biasing the sex ratio could improve the efficiency of rearing, e.g. through shorter courtship display or reduced mating latencies. Moreover, male seminal fluid proteins are associated with increased female fecundity and lifespan (Arnqvist & Andrés, 2006). Adult sex ratios may vary in natural housefly populations (Feldman-Muhsam, 1944) and interpopulation crosses can cause strong sex ratio biases as populations differ in the number and composition of male and female sex determiners (Feldmeyer et al., 2008; Kozielska et al., 2008; Pastor et al., 2014; Hamm et al., 2015; Li et al., 2022). Houseflies also display protandry, the faster development and earlier emergence of males compared to females (L. Francuski, data not shown). Houseflies may therefore experience biased sex ratios spatially and temporally (Carrillo et al., 2012) and be able to adjust their mating behaviour accordingly.

Here, we test the effects of adult sex ratio bias on the courtship and mating behaviour of houseflies. Male-biased sex ratios are generally expected to lead to increased male intra-sexual competition and therefore increased courtship effort, for example in the butterfly Bicyclus anynana (Holveck et al., 2015) and med fly Ceratitis capitata (Leftwich et al., 2012). However, male-biased sex ratios can also lead to a lower courtship rate through increased courtship interruptions or lower female encounters (Weir et al., 2011). Male houseflies spent less effort on courtship performance in high compared to low density groups (Hicks, 2004). Moreover male D. melanogaster evolving under male-biased sex ratios adapted reduced and different courtship effort (Dore et al., 2020). Therefore, we first tested the hypothesis that males will reduce courtship effort in male-biased groups but increase their effort in female-biased groups. Females depicted reduced mating latencies in male-biased groups in the fruit fly Anastrepha fraterculus (Abraham et al., 2015) and D. melanogaster (Bretman et al., 2009), but D. melanogaster females evolving in female-biased sex ratios exhibited the opposite in another study (Dore et al., 2020). Second, we hypothesized that houseflies will exhibit lower mating latencies in male-biased sex ratios, and increased mating latencies in female-biased sex ratios.

Housefly matings last about an hour, during which sperm is transferred in the first approx. 15 min, whereafter only seminal fluid proteins are transferred (Riemann et al., 1967; Leopold et al., 1971a). Male seminal fluid proteins generally inhibit female remating at least until oviposition, and the inhibitory effect is dependent on the quantity of seminal fluids transferred, as indicated by increased female remating frequencies when copulations are interrupted. (Riemann et al., 1967). Seminal fluid proteins are costly to produce (Chapman, 2008) and males are depleted of their inhibitory seminal fluid proteins after multiple (3–4) matings (Riemann et al., 1967; Leopold et al., 1971a). Therefore, males are hypothesized to allocate seminal fluid proteins prudently to mitigate the costs of sperm competition. Using copulation duration as a proxy for resource allocation we hypothesized males to increase the copulation duration in male-biased groups to avoid sperm competition and decrease allocation in female-biased groups where sperm competition risk is low. Copulation duration is positively correlated with increased body size difference between sexes in houseflies (Baldwin & Bryant, 1981). Body size is generally correlated with fitness in insects (Beukeboom, 2018). Considering that male houseflies prefer larger females (Shin et al., 2003), we hypothesized that copulations last longer between mating partners with increased body size differences as males may invest more ejaculate. As males are expected to alter mating behaviour such as courtship performance and ejaculate investment in response to changes in sex ratio, female productivity is likely also influenced under mass-rearing conditions.

2. Material and methods

2.1. Strain collection and housing culturing

The laboratory strain SPA4-2020 (collected in 2020 in Barcelona, Spain, 41°30′19.2″N, 2°05′46.8″E) was used in all mating experiments. Adult flies had been kept in in plastic cages (3250 ml; Semadeni; Ostermundigen, Switzerland) in a climate room for 11 generations before use in experiments. Flies were kept at 25°C and a L12:D12 cycle, also during experiments. Adult flies were provided with water, sucrose solution (20%, w/v) and milk powder. After 4-6 days, egg substrate (wet food) was provided in 2 cups (35 ml each). Wet food was prepared by mixing 200 g dry food (1000 g wheat bran, 150 g flour, 120 g milk powder, 50 g inactivated yeast), 4 ml nipagin solution (10 g nipagin powder/100 ml ethanol 96%; Spruijt-Hillen; IJsselstein, The Netherlands), and 250 ml tap water in a beaker. After 3–5 days, egg-laying cups were removed from the cages (and if necessary, new cups were provided), the contents of the egg-laying cups were emptied into beakers (750 ml) and mixed with additional wet food. Beakers were closed off using lids with nylon meshes. After 7–10 days, emerging (virgin) flies were anaesthetized using CO2 (g), separated by sex and transferred to plastic cups (280 ml) with milk powder and 1 vial of sucrose solution (20%). Flies were kept in groups of 10 to prevent any potential differentiation in sperm/seminal fluid protein synthesis in males due to variation in perceived competition levels. Dead flies were removed and replaced using an aspirator to maintain equal group sizes. Flies used in the experiment were from the 11th to the 14th generation after initial collection.

2.2. Experimental setup

To study the effect of the adult sex ratio on mating behaviour, males and females were grouped in different adult sex ratios (male:female): M5:F1, M3:F1, M1:F1, M1:F3, M1:F5 and M3:F3. Adult sex ratio is the relative number of sexually mature males and females in a group. The M1:F1 and M3:F3 equal sex ratios served as control for the biased sex ratios, as well as for an effect of density. Flies were 4–8 days old when used in experiments. They were transferred from the storage cups to observation arenas with an aspirator. Each observation arena consisted of a 100 mm Petri dish. An observation area with a particular sex ratio was considered a replicate. A hole (approx. 1 × 0.5 cm) was created in the side of Petri dish using a soldering iron to allow for the addition of flies via an aspirator. Afterwards, tape was added to cover the hole to prevent fly escape. A 35 mm Petri dish with egg-laying substrate was attached to the centre of the observation arena with tape. Finally, a cotton plug dosed in sugar water (20%, w/v) was added to the side of the arena. Flies were handled briefly and with care to minimize handling stress. The only individual of the sex present in the arena of each sex ratio was added last to minimize unrecorded interaction between sexes, but for the M1:F1 and M3:F3 control males were added first, then females. Two replicates of each sex ratio (i.e., 12 observation arenas) were assessed per day. If one or more flies died during recording the replicate was excluded from data analysis.

Flies were colour coded under light CO2 anaesthesia at least 72 h prior to mating assays to minimize any potential influence of anaesthesia on mating behaviour. Where possible, dyes were applied simultaneously with sex separation to avoid repeatedly anaesthetizing flies. Dyes consisted of acryl model paint and were applied with a paint brush to the fly’s notum. Flies were dyed with white and green or red acrylic paint for identification under red light.

Observation arenas were placed on a white background in an incubator, resulting in a set-up of 12 arenas per rack. Two cameras (c920 HD pro webcam, Logitech, Lausanne, Switzerland) were placed above the arenas. Flies were recorded using the recording software ‘Security Monitor Pro’ (Deskshare, Plainview, NY, USA). As flies were kept at a L12:D12, and have low sensitivity for red light (Goldsmith, 1965), a red LED light strip of 630 nm (Buyledstrip; Geldrop, The Netherlands) was attached to the roof of the incubator to allow for continued observation during the dark phase. Flies were given time to accommodate to the new environment; hence, the first 5 min of recording were not scored. Fly courtship and mating behaviour were first video recorded for 6 h. The observation period was then extended to 24 h to obtain more data on mating behaviour. As 24 h video recordings were not feasible for storage and analysis, flies were instead photographed once every minute for 24 h. The first 10 min of the 24 h observation periods were still video recorded to observe courtship behaviour in detail. Courtship data from 6 h and 24 h observation periods were pooled whereas mating behaviour data stems only from 24 h observations.

2.3. Housefly courtship behaviour

Male houseflies exhibit several distinct courtship behaviours, usually in a certain order of steps (Colwell & Shorey, 1975; Meffert & Regan, 2002; Figure 1). First, courtship is initiated by a male jumping on a female, referred to as a strike (STRIKE). In response, a female moves her wings out of the way by spreading them perpendicular to the substrate she is standing on (WING OUT). During the strike the male can buzz his wings at virtually any moment (BUZZ). While mounted, the male moves forward positioning his head directly above the female’s head (FORWARD). In this position, the male can lift the female’s forelegs using his own forelegs (LIFT). The male can also hold his wings over the female’s head in about 180° in this position (HOLD). After courtship a male moves backwards on the female to touch her genitalia with his own. Upon rejection a male either dismounts or performs courtship again.

Figure 1.
Figure 1.

Housefly courtship behaviour. Courtship is initiated after a males has spotted and/or touched a female with a male jumping on the female (STRIKE, A–H), which it is continued by steps such as the male moving forward to the female’s head (FORWARD, A), buzzing his wings (BUZZ, B), lifting the female’s forelegs with his own (LIFT, C and D) and the male holding his wings over the female’s head (HOLD, E). In response to a male initiating courtship a female spreads her wings (WING OUT, F). After successful courtship the male moves backwards on the female to initiate mating (G and H).

Citation: Behaviour 160, 15 (2023) ; 10.1163/1568539X-bja10241

To assess the effect of group sex ratio on housefly courtship and mating behaviour, we scored (1) the number and duration of male housefly courtship steps; (2) mating propensity, i.e., whether copulation occurred within the observation period; (3) mating latency (measured as time until the 1st copulation and time in between subsequent copulations); (4) remating of individual flies; and (5) the duration of each copulation. As males use about 40 min to transfer seminal fluid proteins during mating, copulation duration was taken as a proxy for the quantity of seminal fluid proteins allocated (Leopold et al., 1971b).

Despite using dyes to mark individuals it was often not feasible to identify individuals during courtship bouts, as they lasted only a few seconds. All courtship behaviour was therefore assessed on a group level rather than an individual level, i.e., courtship behaviours were not assigned to specific individuals but rather reflect how males in a sex ratio generally behaved. Please note that all other data is based on individual metrics. Determination of copulation duration was started if no other courtship behaviours were observed 60 s after a male attempted courtship. Copulation durations under 20 min were considered pseudo-copulations and excluded, as housefly copulations generally last about 70 min (Leopold et al., 1971a). Sporadically, genitalia of a mating pair failed to dislodge, presumably because they were stuck. This could be inferred from mating flies moving in other/opposite directions but remaining attached to each other. As it was impossible to observe when copulation was terminated in these instances, these pairs were excluded from data analysis. Courtship behaviour was observed for the first 10 min of recording (after the acclimation period), mating latency and duration were observed for 24 h. To analyse courtship behaviours, videos were played back at 20% of their original speed. Videos were analysed with ‘Behavioural Observation Research Interactive Software’ (BORIS) software (Friard & Gamba, 2016).

2.4. Body size estimation

To test for the effect of body size on copulation behaviour, flies were collected after mating assays. They were anaesthetised with CO2 and placed individually in Eppendorf tubes (1.5 ml) and stored at 18°C. Right wings of individuals were gently dissected at the point of articulation with the thorax and mounted between microscope slides and coverslips using Euparal medium (Carl Roth, Germany, Karlsruhe). Wing images were captured using a digital camera Moticam 2000 (Motic Europe, Spain, Barcelona) connected to a stereomicroscope (Stemi SV617, Zeiss, Jena, Germany). Landmarks positioned at 17 vein intersections and terminations (Ludoški et al., 2014) were collected using TpsDig (version 2.31) (Rohlf, 2017a) and expressed as x- and y-coordinates in a Cartesian space (Figure A1 in the Appendix). Coordinates were superimposed using a generalized Procrustes analysis in tpsRelw (version 1.5) (Rohlf, 2017b) and wing centroid size was extracted. Centroid size, an isometric estimator of size, was calculated from the square root of the sum of the squared distances between the centre of the object (i.e., wing centroid) and its landmarks (Bookstein, 1991; Zelditch et al., 2004).

2.5. Statistical analysis

All data analysis and visualisation was carried out in R (version 4.0.2) (R Development Core Team, 2022) and RStudio (version 1.2.5033) (RStudio Team, 2022) using the ‘betareg’ (Cribari-Neto & Zeileis, 2010), ‘cowplot’, ‘ggbeeswarm’, ‘ggfortify’ (Tang et al., 2016), ‘multcomp’ (Hothorn et al., 2008), ‘tidyverse’ (Wickham et al., 2019), ‘lmtest’ (Zeileis & Hothorn, 2002), ‘sjPlot’, ‘emmeans’, ‘factoextra’, ‘readxl’, ‘EnvStats’ (Millard, 2013) and ‘viridis’ packages.

Courtship indices (CI) were calculated by summing the total time any male spent courting (i.e., ‘STRIKE’), divided by the duration of the observation period (10 min). CI per male was additionally calculated by dividing CI by the number of males in the treatment. Mating behaviours (latencies and copulation durations) and body size were normalized by subtracting the mean and dividing by the standard deviation; body sizes were normalized separately for males and females. Additionally, for each mating we obtained the body size differences (female size–male size), which was subsequently normalized as well.

The courtship indices (CI) were analysed using generalised linear models with a beta distribution and treatment as sole predictor. The average duration of courtship behaviours were analysed using a Principal Component Analysis and a linear model with treatment as sole predictor variable. Linear models were also used to analyse male- and female mating latency and copulation duration, with treatment and mating status as predictor variables. The effect of body size on copulation duration was also assessed with a linear model, with the body size difference and male- and female mating status and their interactions as predictor variables. Details on the specific analyses performed are presented in more detail per subject in the results section. Where applicable, post-hoc testing was done by pairwise comparisons between the control treatments versus the male-biased treatment to discover the effect of male–male competition, the control treatments versus the female-biased treatments for multiple mating partner effects, and the M1:F1 control versus the M3:F3 control treatments to assess the effect of group size.

3. Results

3.1. Courtship performance

Over 1300 courtship bouts were analysed in detail in the 10-min observation period, with at least 100 bouts per sex ratio. The number of replicates per sex ratio were M5:F1 ( N = 26), M3:F1 ( N = 29), M1:F3 ( N = 23), M1:F5 ( N = 26), M1:F1 ( N = 20) and M3:F3 ( N = 26) (Figure 2). Males were expected to reduce courtship effort in male-biased groups but increase effort in female-biased groups. Generalized linear models with a beta distribution were fitted with the Courtship Index (CI) or CI per male as response variable and treatment as sole predictor variable. Full models were compared to reduced models to test for significant effects using likelihood ratio testing. The courtship index (CI) and CI per male were significantly affected by the social condition (group χ 2 = 15.288, p = 0.0092; per male: χ 2 = 14.634, p = 0.012</