Introduction

Life history traits, such as development time, fertility, survival, and reproductive success, are key phenotypic components that directly influence individual fitness (Braendle and Paaby, 2024; Roff, 1992; Stearns, 1992). These traits often face trade-offs: investing in one can compromise another (Fisher, 1930; Stearns, 1989; Williams, 1966). For instance, high reproductive effort may reduce survival due to the energetic costs of gamete production or increased mortality associated with mating behaviors or competition (Flatt, 2011; 2020). Trade-offs also emerge between pre- and post-copulatory strategies (Simmons, 2001; Warner et al., 1995). Males prioritizing pre-copulatory investment, such as aggressive competition for mates, may have limited resources for post-copulatory mechanisms, including sperm competition or mate guarding (Andersson, 1994; Brennan and Orbach, 2021). Understanding how such trade-offs are shaped by natural selection offers insight into how organisms optimize fitness in competitive social environments (Roff, 1992; Stearns, 1976).

The fruit fly Drosophila melanogaster (D. mel.), provides a powerful model for studying life-history trade-offs, with well-characterized developmental, behavioral, and reproductive traits (Flatt, 2020). These traits are shaped by both genetic and environmental factors and can evolve under ecological pressures. Faster development, earlier reproduction and high mating rate, for instance, are advantageous in changing environments but come at the cost of reduced longevity and fecundity (Partridge and Farquhar, 1981; Stearns, 2000). Experimental manipulations of environmental conditions such as temperature and food availability reveal how external factors shape trade-offs between survival and reproduction (Chippindale et al., 1996; Marshall and Sinclair, 2010).

Beyond development and fecundity, D. mel. serves as a model to study a wide range of social behaviors, including aggregation, aggression, and courtship, that are largely mediated by chemical communication (Kim et al., 2017; Kohl et al., 2015). Cuticular hydrocarbons (CHCs) and pheromones play diverse roles in social interaction: 7-Tricosene (7-T), a male-specific compound, increases female receptivity and modulates aggression (Grillet et al., 2006; Wang et al., 2011); cis-vaccenyl acetate (cVA), a male-transferred pheromone, attracts females, repels other males, and promotes aggregation (Kurtovic et al., 2007; Laturney and Billeter, 2016); and the female-specific hydrocarbon 7,11-heptacosadiene (7,11-HD) strongly elicits male courtship and supports species-specific mate recognition (Toda et al., 2012).

Among these interactions, male-male competition plays a central role in shaping reproductive strategies. Pre-copulatory competition involves resource defense, courtship displays, and aggressive interactions to secure mating opportunities. Courtship in Drosophila includes stereotyped behaviors such as orientation, wing extensions (producing courtship song), tapping, licking, and attempted copulation (Yamamoto and Koganezawa, 2013). Successful courtship displays lead to female acceptance, outcompeting other males that fail to meet the female’s preferences (Baxter et al., 2018; Hindmarsh Sten et al., 2025). Aggression intensifies this competition: more aggressive or dominant males often monopolize access to females (Dow and von Schilcher, 1975; Filice and Dukas, 2019; Gao et al., 2024; Nandy et al., 2016; Prunier and Trannoy, 2024). Aggressive behaviors range from wing threats and lunging to high-intensity physical confrontations like tussling and boxing (Chen et al., 2002). These behaviors promote territorial dominance and enhance mating success (Gao et al., 2024), although subordinate males may also achieve reproductive success through first access to females and prolonged copulation (Filice and Dukas, 2019).

Post-copulatory competition further shapes male reproductive success. After mating, males use multiple strategies to ensure paternity, including the transfer of seminal fluid proteins (SFPs) including Sex Peptide (SP) and Accessory gland proteins (Acps), which promote sperm storage, stimulate oviposition, and reduce female receptivity to further mating (Chapman et al., 2003; Fricke et al., 2009). These factors also contribute to the formation of a mating plug that retains sperm, acting as a passive mate-guarding strategy (Avila et al., 2011; Dunham and Rudolf, 2009). In parallel, chemical mate guarding via pheromone transfer (e.g. cVA and 7-T) reduce female attractiveness and their likelihood of re-mating (Billeter et al., 2009; Gillott, 2003; Laturney and Billeter, 2016; Miyamoto and Amrein, 2008; Scott, 1986; Zawistowski and Richmond, 1986). Male reproductive behavior is plastic: in the presence of rivals, males extend their mating duration (Bretman et al., 2009; Bretman et al., 2013; Kim et al., 2013), increase the number of sperm in their ejaculate (Garbaczewska et al., 2013), adapt seminal fuids content in their ejaculate(Hopkins et al., 2019), and present higher levels of aggression toward opponents to guard their mates (Baxter et al., 2015; Bretman et al., 2011; Dore et al., 2021; Mazzi et al., 2009). Together, this demonstrate that male-male aggression influences access to mates and reproductive outcomes through both pre- and post-copulatory competition.

While aggression can enhance male mating success in natural settings, the relationship between aggression and reproductive fitness is not always straightforward. In a previous study, Penn et al.( Penn et al., 2010) developed “Bully” males through 37 generations of selection for aggression originating from a Canton-S (Cs) background. These hyper-aggressive males frequently escalated to boxing, the most intense aggression phenotype (Chowdhury et al., 2017), but showed reduced reproductive success with virgin females compared to unselected control males (Penn et al., 2010). This suggests that although aggression may enhance competitive ability, it can also incur fitness costs. This raises important questions about the role of aggression in shaping life-history trade-offs, such as: to what extent does aggression influence life history traits related to reproduction, and does it affect the balance reproduction vs survival?

To address these questions, we investigated whether D. mel. males from the Bully line exhibit changes in life-history trade-offs compared to the unselected starting wild-type flies (Cs lines). We found that hyper-aggressive Bully males show reduced reproductive success and shorter mating duration across their lifespan, indicating lower pre-copulatory success. They also display less effective mate-guarding behavior, indicating a weaker post-copulatory strategy. However, this poor pre- and post-copulatory success was traded-off with an increased survival, which is accompanied by late-life reproductive success. We then provided a mechanistic basis for the reduced copulatory success of Bully males. They showed distinct CHCs profiles compared to Cs males, which may contribute to their lower mating success. Moreover, they transferred lower levels of cVA, a mate-guarding pheromone, to females, potentially decreasing efficiency of their post-mating strategy. Our findings underscore the complex interplay between aggression, reproductive behavior, survival, and chemical signaling, providing insights into the evolutionary dynamics of life-history trade-offs in competitive social contexts.

Results

Highly aggressive males exhibit reduced mating success and have shorter mating durations

To test the hypothesis that selection for male aggressive traits may confer mating advantages, we performed courtship competition assays in which Cs and Bully males competed for mating with virgin Cs females. No significant mating advantage was observed for Bully males (Figure 1A). We then assessed the reproductive capacity of Cs and Bully males in no-choice assays by pairing each male with a virgin Cs female for 30 minutes. Mating success was significantly lower in Bully males compared to what was observed with Cs males (Figure 1B), despite no differences in the percentage of Unilateral Wing Extension (UWE), courtship latency or mating latency, when they succeed mating (Figure 1C-E), suggesting intact courtship and mating motivation. These results were replicated using an independently selected Bully line (referred to as Bully B (Penn et al., 2010)), which also showed reduced mating success and increased aggression compared to Cs (Figure 1-figure supplement 1A, C-E). This indicates that two independent selections for male aggressive traits had simultaneously reduced males mating success.

Figure 1.

Differences between Cs and Bully extended to mating duration (MD) where Bully males mated 25% shorter than Cs males (Figure 1F), a difference also observed in Bully B males (Figure 1-figure supplement 1B). This MD reduction persisted when using decapitated females (Figure 1G), ruling out female-driven termination and implicating male-intrinsic factors.

As males raised with other males have lengthened MD (Bretman et al., 2013; Kim et al., 2013), we compared singly and group-housed males for their MD. We found that both Cs and Bully group-housed males exhibited significantly extended mating durations compared to single-housed males (Figure 1H). However, the MD in Bully males remained shorter compared to Cs males in group-housed condition (Figure 1H). This indicates that Cs and Bully males respond equally to early social experience and hence do not display gene-by-environment interactions.

Together, these findings show that selection for male aggression reduces male mating success and mating duration, key traits of pre-copulatory reproductive performance. This supports our main hypothesis that selection for aggressive traits influences either or both pre- and post-copulatory strategies.

Hyper-aggressive males display reduced mate-guarding efficiency, without compromising female fertility

Given that Bully males have shorter MD and reduced mating success, we hypothesized they would compensate this poor pre-copulatory success with more efficient post-mating strategies. After mating, males typically enter a refractory period during which they are no longer aroused and do not court the females they have just mated with (Lynn et al., 2024). We first tested the hypothesis that Bully males would have a shorter refractory period than Cs males, allowing them to potentially mate with more females. To investigate this, we paired Cs and Bully males with virgin Cs females, allowed them to mate, and measured the percentage of UWE directed by these males at the female they just finished mating with over a 5-minutes post-mating period. While Cs males performed minimal UWE towards just-mated females (Figure 2A’), Bully males displayed a significantly higher percentage of UWE in this context compared to Cs males (Figure 2A’), suggesting that Bully males have a shorter refractory period than Cs males to court mated females. This effect persisted when males were swapped after mating, with Bully males showing increased UWE towards Cs-mated females (Figure 2B–B’). This indicates that Bully males present a lower refractory period to court any mated females, allowing them to potentially have higher remating rate compared to Cs males. To test this, we conducted behavioral assays in which a single Cs or Bully male was consecutively paired with 6 novel virgin females over a 3h period and scored the number of mating. Contrary to our hypothesis, Bully males did not show higher remating rates, despite consistently shorter MD (Figures 2C–D). This suggests that, despite shorter refractory period and reduced MD, Bully males did not achieve a higher remating success rate. As mating accompanies ejaculate transfer and incurs a reduction in sperm and seminal fluid reserve in D. mel. males (Hihara, 1981; Sirot et al., 2009), we hypothesized that Bully males might retain more sperm, potentially enabling them to fertilize more females sequentially within a short period. We thus analyzed female fertility after each consecutive mating and found no significant difference between Bully and Cs males (Figure 2E). Females that mated with Bully males remained as fertile as those mated with Cs males (Figure 2E). Together, our results indicate that while Bully males present shorter refractory period after mating, but their remating rate and ability to fertilize females remain comparable to those of Cs males.

Figure 2.

Due to a shorter MD in Bully males, we reasoned that a lower quantity of CHCs, acting as mate-guarding pheromones, may be transferred to the females’ body, making Cs females that mated with Bully males more attractive to other males compared to those mated with Cs males. To assess female attractiveness independently of any confounding effects of female receptivity, we performed preference assays for decapitated females, who lack receptivity. Cs females were first mated with either Cs or Bully males, then decapitated and simultaneously presented to naïve males in a binary choice assay (Figure 2F). Both Cs and Bully males preferentially courted decapitated females that had previously mated with Bully males (Figure 2F’, green side), showing that females mated with Bully males remain more attractive than those mated with Cs males. In addition to mate guarding pheromones, males transfer seminal fluids that affect long term female receptivity, which may differ between Bully and Cs males. We therefore conducted mating assays in which virgin Cs females were initially mated with either Cs or Bully males. Several days later, these mated females were paired with naïve virgin Cs males, and the proportion of females that remated was analyzed (Figure 2G’). Females previously mated with Cs males, remated significantly less than females that had mated with Bully males (Figure 2G’, green), further supporting weaker mate-guarding by Bully males.

We conclude that Bully males fail to offset their poor pre-copulatory success with more efficient post-mating strategies. Their mates present a higher remating rate, likely due to reduced transfer of mate-guarding pheromones and seminal fluid. Thus, selection for aggressive traits compromises both pre- and post-copulatory male performance.

Hyper-aggressive males present differences in their CHCs profiles and transfer lower quantities of mate-guarding pheromones to females

To identify mechanisms underlying reduced mating success (Figure 1) and mate-guarding (Figure 2F-G) in Bully males, we reasoned that these could be attributable to differences in males and females CHCs profiles. To test this, we quantified the CHCs profiles of naïve Cs and Bully males. PCA analysis revealed distinct CHCs profiles between naïve Cs and Bully males (Figure 3A). Comparing individual compounds of naïve Bully males relatively to Cs ones, we found that Bully males exhibited significantly higher levels of alkanes, monoenes, and methyl-alkanes (Figure 3B, Table 1). These differences may contribute to modulating female attractiveness toward males.

Figure 3.

Table 1:

Our results suggest that the higher remating rate in females mated with Bully males (Figure 2G’) may be due to increased attractiveness compared to females mated with Cs males (Figure 2F’), a process influenced by pheromones transferred from males to females such as the blend of cVA/7-T (Laturney and Billeter, 2016). To test this hypothesis, we extracted CHCs from the cuticles of Cs virgin and Cs females mated to either a Cs or Bully males and compared their CHCs profiles. As a first step, PCA analysis showed that Cs virgin females exhibited distinct CHCs profiles from those of both Cs and Bully-mated females (Figure 3C). The first (PC1) and the second (PC2) dimensions explain 80.7% and 11.1% of the variance in CHCs profiles, respectively. High cos² values indicate a strong correlation between CHCs compounds and the principal components, suggesting that the PCA captures the major part of the variance in the data (Figure 3-figure supplement 1A). In addition, some compounds such as 2MeC24, C23:1(6), cVA and C25:1(6) strongly influence PC1 and explain the difference between the clusters, i.e. female mated status (Figure 3-figure supplement 1B-D). Individual compound analysis revealed that this difference was primarily due to the transfer of male-specific CHCs during mating not found in virgin females (Figure 3D and E, Table 2). Relatively to virgin females, C23:1(9) (9-Tricosene, 9-T), C23:1(7) (7-Tricosene, 7-T), C23:1(6), C25:1(6) and cVA were significantly elevated in mated females (Figure 3D and E, Table 2). These results are consistent with previous studies showing that 7-T and 9-T are acquired by females during mating (Everaerts et al., 2010; Laturney and Billeter, 2016). When comparing the two groups of mated females, individual compound analysis showed that C23:1(5) (5-Tricosene, 5-T) and C25:1(5) (5-Pentacosene, 5-P) were significantly elevated in Bully-mated females compared to Cs-mated females (Table 2, surrounded in red, Figure 3D). These findings suggest that 5-P and 5-T might contribute to promoting remating in females that have previously mated with Bully males (Figure 2F’ and G’). Given that Bully males also showed higher levels of both 5-P and 5-T compared to naïve Cs males (Figure 3B), it is likely that the elevated levels of these compounds observed in females result from their transfer during mating.

Table 2:

cVA, a male-specific compound that acts as an anti-aphrodisiac and a mate-guarding molecule, is primarily transferred to the female’s genital tract during mating, with only trace amounts deposited on the cuticle (Laturney and Billeter, 2016). Females with higher cVA levels (Kurtovic et al., 2007), combined with 7-T, are less attractive to males (Laturney and Billeter, 2016). Given that remating rate of Bully-mated females was higher than that of Cs-mated females (Figure 2G’), we hypothesized that Bully-mated females would receive less cVA in their genital tract during mating, thus decreasing their attractiveness than when mating with Cs males. To test this, we conducted mating assays with Cs females paired with either Cs or Bully males, followed by dissection of the females’ genital tracts to quantify the amount of cVA transferred. We found that females mated with Bully males had about three times lower cVA quantities in their genital tracts compared to those mated with Cs males (Figure 3F’). Combined with higher quantities of 5-T and 5-P, reduced cVA level in genital tract likely contribute to the increased attractiveness and higher remating rates observed in females that previously mated with Bully males.

Overall, our findings suggest a mechanistic basis for the reduced pre- and post-copulatory behaviors observed in Bully males. Differences in CHCs profiles between naïve Cs and Bully may underlie a reduced attractiveness to females, thereby reducing their mating success. Additionally, increased levels of 5-T and 5-P, combined with reduced cVA transfer to the female genital tract, may contribute to the lower mate-guarding efficiency observed in these hyper-aggressive males.

The lower copulatory strategy of aggressive males is offset by a longer lifespan and maintenance of reproductive success later in life

Based on the main life-history trade-off between reproduction and survival, we hypothesized that the reduced pre- and post-copulatory success in Bully males would increase their survival. To test this, we conducted longitudinal courtship assays in which individual Cs and Bully males were paired weekly with a new virgin Cs female until their natural death. Throughout their lifespan, Bully males consistently displayed shorter MD than Cs males (Figure 4A). While Cs males increased their MD with age or experience, a pattern reported in another Drosophila species (Dhole and Pfennig, 2014), this effect was absent in Bully males, whose MD remained stable across mating and throughout their lives (Figure 4A). Analysis of individual male slopes confirmed this divergence: most Cs males showed increasing MD across mating events, while Bully males did not (Figure 4—figure supplement 1). Mating success followed a similar pattern. Although Bully males mated less frequently than Cs males during the first 70 days, they maintained reproductive activity after all Cs males had ceased reproductive activity. All Cs males stopped mating by day 78 and died by day 97, whereas Bully males continued to mate until day 98 and survived up to 119 days (Figure 4B). Remarkably, these late-life matings produced viable offspring, demonstrating that Bully males retain fertility at older ages.

Figure 4.

Bully males survived significantly longer than Cs males, suggesting a shift in the reproduction-survival trade-off. To explore this further, Cs virgin females were mated with either Cs and Bully males, and progeny survival was monitored. Progeny from Bully mating showed a median survival (EC50) of 79 days, with the last male dying at day 106 (Figure 4C, green), while progeny from Cs males showed an EC50 of 57 days, with the last death occurring at day 90 (Figure 4C, yellow). This showed that the median survival was extended by ∼22 days (79 vs. 57 days) in Bully males. We then monitored male survival in mixed-sex groups composed of ten males and twelve females. Bully males still showed increased longevity, with an EC50 of 56 days and the last death on day 71, compared to Cs males, that had an EC50 of 43 days and all died by day 67 (Figure 4D). These findings are consistent with our previous results (Figure 4B and C), confirming that Bully males have a longer lifespan and can transmit this trait to their offspring.

Our findings demonstrate that males selected for high levels of aggression show reduced investment in mating success but exhibit extended lifespan and sustained reproductive output later in life. Together, these findings support our initial hypothesis that selection for aggressive traits influences a key life-history trade-off, by favoring survival at the expense of reproductive traits.

Discussion

Our study reveals that selection for aggressive traits in D. mel. males influences one key life-history trade-off, in which survival is extended while reproductive success is reduced. Our results indicate that selection for aggressive traits directly or indirectly affects both pre-and post-copulatory strategies in Bully males, challenging the hypothesis that aggression necessarily enhances mating success (Andersson, 1994; West-Eberhard, 1983). Aggressive males had lower pre-copulatory success with decreased mating success and shorter mating duration than their less aggressive counterparts. Despite reduced refractory courtship period after mating, Bully males fail to achieve higher remating rate and exhibit a less effective mate-guarding strategy, as evidenced by the increased remating rate and potentially attractiveness of their previous mates. Differences between Bully and Cs males also extend to chemical profiles, which may account for reduced mating success in Bully males. In addition, females mated to highly aggressive males exhibited quantitative differences in CHCs profiles compared those mated with Cs males, along with a dramatic reduction in cVA quantities in their genital tracts after mating, providing a mechanism underlying a less efficient mate-guarding. Our findings suggest that artificial selection applied to male aggressive traits to produce the Bully line, is associated with changes in reproductive traits, enhanced survival, and increased CHCs levels, suggesting the existence of pleiotropic genes whose variation links aggression with these other traits.

The increased survival rate observed in highly aggressive males raises questions about the evolutionary consequences of aggression. While Bully males exhibit reduced pre- and post-copulatory success compared to Cs males, their extended lifespan allows them to sustain mating activity and produce viable offspring at later ages, challenging the classical view that early reproductive success is the primary driver of male fitness (Flatt, 2011). This shift suggests that alternative reproductive strategies, even those associated with lower mating success, can persist if they confer long-term survival advantages. Moreover, the extended lifespan of both Bully males and their progeny suggests the transmission of pleiotropic genetic traits that influence both aggression and survival, or the transmission of independent genetic loci affecting these traits separately. A previous study identified a limited number of genes that are differentially expressed in the Bully line compared to Cs (Chowdhury et al., 2017). One of these genes, CG13646, a putative transmembrane transporter, is down-regulated in Bully males compared to Cs. Future research should investigate the genetic and neurobiological mechanisms underlying the trade-off reproduction/survival by examining the roles of candidate genes whose expression is up- or down-regulated in the Bully line.

Despite maintaining high courtship intensity, highly aggressive males exhibit reduced mating success and mating duration, suggesting that aggression is linked to physiological or behavioral traits that may be suboptimal for reproductive performance. For example, males might be less attractive for reasons unrelated to their courtship efforts. Rather than maximizing immediate reproductive output, aggressive males might allocate energy in a way that prioritizes long-term survival, potentially reflecting a life-history shift toward extended reproductive capacity at the cost of reduced mating success.

We showed that naïve highly aggressive individuals present distinct chemical profiles compared to unselected wild-type flies, indicating that selection for aggressive traits may have broader impacts, such as indirectly influencing mate choice through male attractiveness to females. Elevated levels of most CHCs in aggressive males suggest that they could be perceived as more attractive to females than wild-type males. However, despite these chemical traits potentially enhancing female attraction, aggressive males exhibit lower mating success, highlighting that a higher quantity of CHCs alone does not guarantee mating success. However, besides 7-T as CHCs fostering male attractiveness, other cuticular compounds, such as palmitoleic acid signaling through Or47b, modulate courtship behaviors in an age-dependent manner (Dweck et al., 2015; Kohlmeier et al., 2021; Lin et al., 2016). This mechanism enables older males to initiate courtship more rapidly and with greater intensity, potentially enhancing their mating success relative to younger males (Lin et al., 2016). In our experiments, we did not measure palmitoleic acid levels in Bully and Cs males. However, it remains possible that Bully males have lower levels of palmitoleic acid compared to Cs males. This could explain their decreased mating success by potentially reducing their attractiveness. Aggressive males may signal their fitness through chemical cues, but their high level of aggression could hinder mate choice by females, who may prioritize other factors, such as male courtship behavior or compatibility in post-mating interactions, over chemical signaling alone. This raises the possibility that aggression-driven sexual selection operates through both direct behavioral competition and indirect chemical signaling mechanisms, influencing overall reproductive success.

The influence of aggression extends beyond accessing mates, affecting post-mating behaviors, mate-guarding and chemical communication, shaping male reproductive strategies and female remating dynamics. Hyper-aggressive males exhibit a shorten refractory period to court females showed by an elevated post-mating courtship yet reduced mate-guarding efficiency, likely contributing to increasing remating rate of both males and females respectively. This suggests that while aggression may provide advantages in direct competition, it can incur reproductive costs by diminishing mate-guarding efficiency. The chemical analysis further supports this assumption, showing that females mated with the hyper-aggressive Bully males exhibit elevated levels of specific CHCs (5-T and 5-P) and significantly lower cVA transfer into the female genital tract, which may fail to efficiently decrease their attractiveness, thus triggering higher remating rate. This is supported by other study showing that lower cVA levels can shorten the effectiveness of chemical mate-guarding, allowing quicker female re-attractiveness (Ejima, 2015). As 7-T and cVA act in concert to modulate behavior (Laturney and Billeter, 2016; Verschut et al., 2023), it is possible that the elevated levels of 5-T and 5-P in Bully-mated females interact synergistically too, or with potential other CHCs, to fail to reduce female attractiveness. This would suggest that aggression may indirectly influence post-copulatory sexual selection by changing female chemical signatures.

In summary, our findings emphasize the need to consider aggression within a broader life-history and evolutionary framework. The influence of aggression on the trade-off survival/reproduction, potentially through chemical signaling, illustrates how selection for competitive traits does not occur in isolation but is shaped by interrelated behavioral and physiological factors. This supports the notion that aggression benefits to individuals in competitive environment but may induce reproductive constraints. By framing aggression as part of a broader life-history strategy rather than an isolated trait, our study highlights the intricate balance between competition, survival, and reproductive success. These insights offer novel perspectives on the evolutionary consequences of aggressive behavior and its role in shaping reproductive strategies in animal populations.

Material and methods

Experimental model and subject details

Fly stocks

Drosophila melanogaster strains were raised at 25°C under a 12h:12h light/dark cycle (LD= 9:00a.m. – 9:00p.m.) on standard fly food (74% corn flour, 28% yeast, 40% sugar, 8% agar, 20% Moldex). Canton-S (Cs) flies (from Edward Kravitz’s laboratory, RRID:BDSC_64349) were used as wild-type. The Bully A and B lines were originally selected in Edward Kravitz’s laboratory from Cs during 37 generations, resulting in hyper-aggressive male flies (Garbaczewska et al., 2013; Hopkins et al., 2019).

Experimental arena

Experimental arenas were described in details in Trannoy et al., 2015 (Trannoy et al., 2015). Briefly, resin blocks containing three circular behavioral arenas (dimension: 2.3 cm diameter, 1.7 cm height) were used to conduct behavioral assays. Each arena was divided in two equal sizes by a plastic divider allowing flies to acclimate without physical interactions with the second fly. Then, dividers are removed from the arenas to start the behavioral assays. All behavioral assays were performed in empty arenas, expect for aggression assays and courtship assays with decapitated females, in which a food cup (1.5 cm diameter, 1 cm height) containing fresh standard fly food with a drop of fresh yeast paste on the surface was placed in the center of each arena.

Method details

Commonalities between all behavioral assays

Late-stage pupae were collected from stocks raised at 25°C. Male pupae were individually placed in 5ml vials (Dutscher, ref 390597) containing about 1ml of standard fly food while female pupae were placed in group of 15 in ≈47ml vials (Dutscher, ref 789001B) containing about 15ml of standard fly food. All vials were placed at 25°C for 7 days. Seven-day-old individuals were randomly inserted in arenas by negative geotaxis, and allow to acclimate for 5 min before experiments. All behavioral assays were performed at 25°C. All tests were recorded with Basler camera Ace1970 through the StreamPix8 multi video recording software (norpix.com) and analyzed with BORIS software (Friard and Gamba, 2016). After 15 min or 30 min, if pairs of flies had not yet lunged or mated respectively, we considered that no aggression or mating occurred and arenas were not considered for analysis. No power analyses were performed prior to the experiments, as we chose to use sample sizes of approximately 20 replicates, which is considered standard practice in the field.

Aggression assays

Two days prior to the aggression assays, males were anesthetized with CO₂, and half were marked with a small white dot on the thorax before being returned to their original isolation vials. Aggression tests were conducted between ZT0 and ZT3 using size-matched male pairs. Males competed for access to a central food cup containing yeast. The following behaviors were recorded for 10 minutes following the first lunge: latency to lunge (i.e., the time between the initial encounter on the food cup and the first lunge), total number of lunges, and number of boxing events.

Mating assays

Seven-day-old individuals were tested in arenas without any food source. Latency to court and to mate, defined as the time between the first encounter and either the first Unilateral Wing Extension (UWE) or the initiation of mating, respectively, were measured. Total UWE duration was recorded over a 10-minute period starting from the first encounter. Mating proportion and mating duration were also scored. In all assays, males were reared in social isolation, except for Figure 1H, where they were housed in groups of 10 males for 7 days.

Mating assays with decapitated females

Females were anesthetized with CO₂ and placed under the binocular for head removal. Experiments with decapitated females were performed in arenas with food cup filled with fresh fly food but without yeast.

Multiple consecutive mating events assays

Single naïve Cs and Bully males were sequentially paired with six virgin Cs females. For each male, mating duration and the number of successful mating were recorded. After each copulation, the mated female was transferred to a 5 ml vial containing fly food to assess progeny production.

Mating assays throughout the lifespan

24 Cs and Bully males were reared in isolation and paired with seven-day-old virgin Cs females for mating assays, conducted every seven days until all males had died. At each time point, mating duration and mating success (proportion of mating) were recorded. After all Cs males had died, females that had mated with Bully males were transferred to 5 ml vials containing fly food to assess progeny production.

Attractiveness assays

Measure of post-mating UWE

Mating assays were performed by pairing Cs and Bully males with Cs females. Right after mating end, we scored the total time that males displayed UWE for 10 minutes towards the just-mated females.

Measure of post-mating UWE with switched females

Mating assays were performed with naive Cs or Bully males and virgin Cs females. Immediately after mating ends, females were switched: Cs males interacted with Bully-mated females and Bully males interacted with Cs-mated females. Then total time of UWE towards the females was scored during 10 min.

Measure of post-mating UWE with decapitated females

Mating assays were conducted using naïve Cs or Bully males paired with virgin Cs females. Following copulation, mated females were decapitated. One decapitated Cs-mated and one Bully-mated female were then placed on the surface of the food cup. A new naïve Cs or Bully male was introduced and allowed to interact with both decapitated females. The duration of Unilateral Wing Extensions (UWE) directed toward each female was recorded over a 2-minute period following the first encounter. Arenas in which males interacted with only one of the two females were excluded from the analysis, as no choice could be determined.

Re-mating assays

Mating assays were performed with naïve Cs or Bully males and virgin Cs females. Mated-females were then individually placed in 5ml vials containing about 1ml of standard fly food and kept at 25°C. 12 or 14 days after the initial mating, mated-females were paired with a new naïve Cs male in empty arena and the proportion of re-mating during 30 min was scored.

Survival assays

Survival with isolated males

Late-stage pupae from the progeny of ♀Cs × ♂Cs and ♀Cs × ♂Bully crosses were individually isolated in 5 ml vials containing 1ml of standard fly food and kept at 25 °C until adult emergence. The number of dead individuals was recorded daily until all flies had died. To maintain food quality, flies were transferred weekly to fresh vials containing approximately 1 ml of new fly food.

Survival with group-raised males

Late-stage Cs and Bully male pupae and Cs female pupae were placed in group of 12♀Cs-10♂Cs or 12♀Cs-10♂Bully in ≈47ml vials containing 15mL standard fly food and kept at 25 °C until adult emergence. The number of dead males was recorded daily until they all had died. To maintain food quality, flies were transferred twice a week to fresh vials containing approximately 15 mL of new fly food.

CHCs analysis

Extraction and analysis of cuticular hydrocarbons (CHCs) of naive Cs and Bully males, as well as virgin Cs, Cs- and Bully-mated females were performed in Jean-Christophe Billeter’s lab, Groningen, Netherlands, with a Gas Chromatography coupled with Flame Ionization Detection (GC-FID) as already described (Laturney and Billeter, 2016). Briefly, single whole flies were put in a 2ml glass vial (Supelco certified vial kit, 2ml clear glass vial 12×32mm and Supelco inserts) containing 50µl of nC18+nC26 10 ng/µl standard solution (made in Billeter’s lab) diluted in n-Hexane 99%+ (Acrōs Organic®). Vials were then vortexed at the lowest speed for 2min, afterward flies’ body were gently removed from vials using a “J-shape” paper clip sliced at the end, taking care not to damage the body to prevent from hemolymph contamination. Finally, samples were placed in GC-FID for apolar compounds analysis only.

cVA quantification into genital tract

Mating assays were performed with naïve Cs or Bully males and virgin Cs female. Right after mating, females were transferred into a 5ml vial previously put in ice in -80°C fridge to make them fall asleep quickly and prevent from sperm ejection. Mated-females’ genital tract was dissected in PBS with precision forceps under a binocular loupe. Protocol was already employed in (Verschut et al., 2022). Briefly, abdomen was torn to make ovaries and genital tract out and visible, then ovaries, oviduct, spermathecae, parovaria and cuticula were removed to keep only the uterus, taking care not to injure it to prevent from sperm leak. Each uterus was inserted in a 2ml glass vial (Supelco certified vial kit, 2ml clear glass vial 12×32mm and Supelco inserts) containing 50µl of n-Hexane 99%+ (Acrōs Organic) for 1h at ambient temperature. Hexane solution was then evaporated under helium flow to prevent from cVA oxidation with dioxygen. Samples were sent to Jean-Christophe Billeter’s lab for GC-FID analysis of cVA, as describe above in CHCs analysis method part.

Quantification and statistical analysis

No data have been removed from the statistical analysis. Some experiments were independently replicated by different authors to reduce potential bias. All statistical analyses were performed on R software version 4.0. (Team, 2020) with packages (Fox and Weisberg, 2019), coin (Hothorn et al., 2006), dplyr (Wickham et al., 2023), factoextra (Kassambara and Mundt, 2020), FactoMineR (Lê et al., 2008), lme4 (Bates et al., 2015), lmerTest (Kuznetsova et al., 2017), MASS (Venables and Ripley, 2002), MuMIn (Bartoń, 2023), nlme (Pinheiro et al., 2022), performance (Lüdecke et al., 2021), pgirmess (Giraudoux. P, 2018), predictmeans (Luo et al., 2021), survival (Therneau, 2022), vegan (Oksanen J, 2022).

Comparisons of means (k =2) were performed with a T-test when possible (i.e. mating duration) otherwise a Fisher Pitman permutation test was used (i.e. UWE, latency to court and to mate). For comparisons with more than 2 conditions, a linear model (LM) was used and tested either with a type II ANOVA when possible (i.e. mating duration) or a permutation test instead (i.e. latency to lunge and CHC concentrations). For comparison of proportions and counting events, a type II analysis of deviance table with a Likelihood Ratio Test (i.e. mating success, proportion of female re-mating, number of lunge and boxing events) or a Wald Chi-square test (i.e. number of consecutive mating events) was performed. A Principal Component Analysis (PCA) was performed to see different CHC profiles and tested with a PerMANOVA and a SIMPER analysis. Survival data were analysed with a Weibull regression model and a Likelihood Ratio Test.

When paired data, dependency was considered thanks to linear mixed model (LMM) with the source of dependency as random factor. Data following a distribution other than a normal distribution (e.g. binomial or Poisson distribution) were computed in a generalized linear model (GLM) (i.e. mating success, number of consecutive mating events, number of mating giving progeny, proportion of females re-mating, number of lunges and boxing events). When necessary, interactions were considered within models and tested. All post-hoc comparisons were performed with a sequential Bonferroni correction on the alpha level using either Holm method (Holm, 1979) when few numbers of comparisons or Benjamini and Hochberg method (Benjamini and Hochberg, 1995) when more comparisons. All statistical details are given in Table S1.

Acknowledgements

We thank the members of EXPLAIN and IVEP team at the CRCA for their helpful discussion and support on this study. We thank Karvitz’s lab for sharing the Bully lines. We also thank all the member of Billeter’s lab to host A. Defert and train him to GC-FID methodology. We acknowledge support from the French Research National Agency (ANR) (ANR-19-CE37-0018-01 to S.T.), the Fondation Fyssen (190573 to S.T.), and the Doctoral Mobility Campaign of Toulouse III University in 2023 (to A.D.).

Funders had no role in study design, data collection and analysis, decision to publish or preparation of the manuscript.

Additional information

Resource availability

Further information and requests for resources and reagents are accessible from the lead contact, Séverine Trannoy (severine.trannoy@univ-tlse3.fr).

Author contributions

Conceptualization: S.T., J.C.B., A.D.

Data curation: A.D., R.G., G.P., F.J., A.C., A.H., T.G., S.T.

Formal analysis: A.D., R.G., G.P., F.J., A.C., A.H., T.G., S.T.

Funding acquisition: S.T., J.C.B., A.D.

Project administration: S.T.

Supervision: S.T., J.C.B.

Writing: S.T., J.C.B., A.D.

Supporting informations

Figure 1-figure supplement 1.

Figure 1-table supplement 2.

Figure 3-figure supplement 1.

Figure 4-figure supplement 1.