Introduction

In nature, animals of the same species tend to aggregate and form groups whose number of individuals depends both on population density and available resources (Parrish and Edelstein-Keshet 1999; Bonabeau et al. 1999; Lof et al. 2008). The size of the group changes the frequency of inter-individual interactions. This induces physiological and behavioral effects which are not always linearly correlated with the number of individuals in the group (Bednekoff and Lima 2004; van der Marel et al. 2019; Beauchamp 2019; Wiedenová et al. 2018; Verschut et al. 2023). Larger vertebrate groups show enhanced vigilance against predation and allow increased food uptake (Wirtz and Wawra 1986; Wilson and Richards 2000; Rieucau et al. 2010; Scott-Samuel et al. 2015). Aggregation also changes body temperature, immune response and gene expression (Andrews et al. 1987; Runcie et al. 2013; Wiedenová et al. 2018; Hamilton et al. 2022). In social insects, the size of the group strongly affects work organization, metabolism, hormone titer and water intake (Hewitt et al. 1971; Mao and Henderson 2010; Fewell and Harrison 2016). Aggregation also impacts development, sexual and aggressive behaviors, pheromonal communication and reproduction (Robson and Traniello 1998; Kikuchi et al. 2008; Yoder et al. 2010; Amsalem and Hefetz 2011; Fujioka et al. 2019; Orlova and Amsalem 2021; Nixon et al. 2022). The study of the sub-social species Drosophila melanogaster — a favourable model for genetic manipulation (The FlyBase 2003) — identified genes underlying the aggregation effect involved in sexual and feeding behaviors, learning and memory, gene expression and reproduction (Kent et al. 2008; Lof et al. 2008; Saltz 2011; Billeter et al. 2012; Lihoreau et al. 2016; Sehdev et al. 2019; Rooke et al. 2020; Churchill et al. 2020; Ebanks et al. 2021; Muria et al. 2021; Bailly et al. 2023).

Virtually all insects carry a mixture of long chain hydrocarbons on their cuticle (cuticular hydrocarbons; CHCs (Qiu et al. 2012; Kefi et al. 2019). CHCs are primarily used in desiccation resistance (Gibbs et al. 1997; Ferveur et al. 2018; Yang et al. 2020) and protection against xenobiotics penetration (Balabanidou et al. 2018; Wang et al. 2017, 2019). CHC-related bidirectional permeability diverges between Drosophila species (Wang et al. 2020a, 2021). CHCs are also involved in intra- and inter-specific communication in many insects including Drosophila species (Howard and Blomquist 2005; Ferveur 2005). In the Drosophila melanogaster subgroup of species, some CHCs produced in relatively large amounts are either involved in mate recognition, copulation, or in territory and aggressive male-male behavior (Hoffmann 1990; Greenspan and Ferveur 2000; Wang et al. 2011). D. melanogaster males of temperate geographic regions produce a high level of 7-tricosene (7-T; 23 carbons) and much less 7-pentacosene (7-P; 25 carbons) whereas males of tropical and equatorial regions show an opposite pattern (Jallon 1984; Ferveur and Jallon 1996). Differently, D. melanogaster females produce low levels of 7-T and 7-P and high amount of 7,11-heptacosadiene (7,11-HD; 27 carbons; (Antony and Jallon 1982). Moreover, in their ejaculatory bulb males produce the volatile pheromone 11-cis-Vaccenyl acetate (cVa). The terminal steps of CHC and cVa biosynthesis depend on distinct enzymatic pathways and tissues (Guiraudie-Capraz et al. 2007; Jallon 1984; Ferveur et al. 1997). In D. melanogaster males, some CHCs can quantitatively vary with circadian activity and social interaction (Kent et al. 2008). However, in these experiments, CHCs did not accumulate on the cuticle since their amounts returned to initial levels after a 24-hour long cycle.

Given our limited knowledge on the effect of aggregation in Drosophila pheromones, we investigated whether CHCs and cVa amounts changed in males competing in groups of different sizes during early adult life. We evaluated this effect on “superficial” and “whole-body” amounts of CHCs and cVa, using two extraction procedures. After testing 6 laboratory-acclimated lines (four wild-type and one CHC-mutant D. melanogaster lines plus one D. simulans line), we tested 21 D. melanogaster lines freshly collected in nature after 1, 12 and 25 generations acclimation in the laboratory.

Materials and Methods

Flies

Laboratory Lines

We used five Drosophila melanogaster stocks which have been already intensively studied — however not for group-size effect — : the laboratory strains Canton-S (Cs, caught in the 1930’s USA), Dijon2000 (Di2, caught in 2000 in France), Oregon-R (Or-R, caught in the 1950’s in the USA; provided by Professor Jean-Christophe Billeter) and Zimbabwe30 (Z30, caught in the 1980’s in Zimbawbe; provided by Professor Jerry Coyne) and the desat11573-Gal4 homozygous mutant strain (desat1; introgressed with the genetic background of Di2 strain). These mutant flies produce a reduced quantity of 7-T, 7-P and 7,11-HD (Marcillac et al. 2005a, b); desat1 males also show reduced cVa emission (Cortot et al. 2022a). We also tested a D. simulans stock (#K509; provided by Professor Daisuke Yamamoto). The reason why we compared D. simulans males to D. melanogaster Cs and Di2 males, is based on the fact that males of both species produce high level of 7-T, low level of 7-P and cVa (Ferveur 1991; Schaner et al. 1987). We also compared females of the Cs, Di2 and desat1 strains, but given that they did not show any group-size effect in our preliminary experiments, we did not pursue further investigation with females. Stocks were maintained on alcohol-free standard cornmeal medium mixed with killed yeast in 30 ml glass vials, at 24 ± 0.5 °C and 65 ± 5% humidity on a 12:12 dark: light cycle. To reduce larval competition, only 8–10 pairs of progenitor flies were kept in each vial and transferred to fresh food vial every 2–3 days to generate enough tester flies. 1- to 2-hour old flies were sexed under light carbon dioxide anaesthesia 2–4 h after lights on and were kept in fresh-food glass vials randomly distributed either in small groups of 5 flies or in larger groups of 25 or 50 flies. In the first experiment, we tested groups of 50 flies only in series2 after finding a clear group-size effect in groups of 25 flies (compared to groups of 5 flies) in series1. Given that no major difference was found between groups of 25 flies and or 50 flies, we only tested groups of 25 flies in the other experiments. All flies tested were 4 day-old, at the age at which they showed a mature CHC profile and a relatively limited lethality when kept in larger groups (Cortot et al. 2022b; this study).

Freshly Caught Lines

In September 2021, flies were caught in traps containing different types of fresh fruits (fig, peach, mirabelle, banana, melon). All groups of founder flies (F0) caught in the same spot (an suburban orchard in Dijon, France), and trapped in each vial were used to initiate a distinct line in the laboratory (Ferveur et al. 2024a). At the next generation, some F1 males were kept for chemical analysis. In each F1 line, species identity was visually checked under binoculars. Then, we performed reciprocal crosses between F2 flies and control Cs flies to observe the presence of F3 progenies of both sexes. To avoid bottleneck effects, each line was maintained in a minimum of 6 vials (and in 12–18 vials at F11 and F24 generations preceding F12 and F25 extractions). For the sake of clarity, each line was identified with a number.

Pheromones

We used two extraction procedures on different insects to measure both the superficial and whole-body amounts of pheromones. The superficial amounts of pheromones are the quantities immediately available by other flies during the sexual or social interactions. Whole-body amounts correspond to the total of internal and superficial pheromones at a given time and provide a good indication of pheromone biosynthesis. Frozen flies were individually immersed either for 5 min into vials containing 30 µl hexane, or for 24 h in 30 µl dichloromethane at room temperature. This allowed us to respectively extract the superficial or the whole-body amount of cVa and of CHCs present on (in) each fly (Cortot et al. 2022a, b; Ferveur et al. 2024b). Each solvent contained 3.33 ng/µl of C26 (n-hexacosane) and 3.33 ng/µl of C30 (n-triacontane) used as internal standards (ISs). Amounts of cVa and CHCs were quantified by gas chromatography using a Varian CP3380 gas chromatograph fitted with a flame ionization detector, a CP Sil 5CB column (25 m x 0.25 mm internal diameter; 0.1 μm film thickness; Agilent), and a split–splitless injector (60 ml/min split-flow; valve opening 30 s after injection) with helium as the carrier gas (50 cm/sec at 120 °C). The temperature program began at 120 °C, ramping at 10 °C/min to 140 °C, then ramping at 2 °C/min to 290 °C, and holding for 10 min. The chemical identity of each peak was determined according to (Everaerts et al. 2010). The amount (ng/insect) of each compound was calculated based on the readings obtained from the ISs.

Beside the absolute amounts of 7-T and 7-P, we measured the level of n-tricosane (23Lin), a linear saturated CHC structurally close to 7-T) in some laboratory lines. We also determined the 7-T/7-P ratio that provides a relative measure of these two principal CHCs independently of their absolute quantities. In females, we also measured the amounts of 7,11-heptacosadiene (7,11-HD) and 7,11-nonacosadiene (7,11-ND). In males of freshly caught lines, we calculated the rate of variation between the “x5” (small) and “x25” (large) group sizes with the [(“x25” – “x5”)/ “x5”] formula.

Statistics

The amounts (and ratios) of the various chemicals were compared using Mann-Whitney test or Kruskal-Wallis test (with Monte-Carlo simulations, followed by Conover-Iman multiple pairwise comparisons —p = 0.05— with Bonferroni’s correction), where applicable. The amounts of cVa, 7-T and 7-P in whole body or superficial extracts were used as variable to compute principal component analysis (PCA; Pearson’s correlation matrix type; standardized values) with the type of fly (line and group size) used as individuals. All statistical analyses were performed using XLSTAT Premium 2021.5.1.1220 (Addinsoft 2021).

Results

Group-size Effect in Laboratory Lines

Repeated Analysis in Three well-studied Lines

We measured whole-body and superficial amounts of cVa, 7-T, 7-P and the 7-T/7-P ratio in 4 day-old males of three well-studied laboratory lines: Canton-S (Cs), Dijon2000 (Di2) and the desat1 mutant. Two series of analysis, separated by 7 months, were performed: the series1 was performed with groups composed of 5 or 25 males (for Cs and Di2) and also of 50 individuals for desat1, while the series2 was carried out on groups of 5, 25 or 50 individuals (indicated as “x5”, “x25”, “x50”, respectively).

Whole-body Compounds

We first measured whole-body amounts of CHCs and cVa (Fig. 1).

Fig. 1
figure 1

Group size effect on whole body pheromones in males of three laboratory lines. The data shown (from top to bottom) correspond to the whole-body amounts (in ng) of cis-Vaccenyl acetate (cVa), 7-tricosene (7-T) and 7-pentacosene (7-P) extracted in single 4 day-old males (with dichloromethane during 24 h). At the bottom, we show the 7-T/7-P ratio. Using three laboratory-established lines: Canton-S (Cs), Dijon2000 (Di2) and the desaturase1 mutant (desat1), we compared males kept in groups of 5, 25 or 50 individuals (“x5”; “x25”; ”x50”). We performed two experimental series (series1; series2) separated by 7 months. Data are shown as box and whisker diagrams. The lower and upper box edges represent the first and third quartiles, while the median value is indicated by the inner small horizontal bar and the mean by the empty dot. The ends of the whiskers above and below each box represent the limits beyond which values were considered anomalous. The levels of the various chemicals or ratio were compared using Mann-Whitney or Kruskal-Wallis test, where applicable. The significance levels are p = 0.05 (★), p = 0.01 (★★), p = 0.001 (★★★). Different letters indicate significant differences between males of the different lines at each generation. For series 1, N = 10 except for desat1 × 25 and x50 (N = 5); for series 2, N = 15

Full size image

In Cs males, cVa slightly decreased in series2 (between “x5” and “x50” groups; p = 0.041). 7-T increased in series1 (between “x5” and “x25” males; p = 0.001) and in series2 (between “x5” and ”x50” males; p = 0.018). 7-P gradually increased with the group size in the two series (p < 0.0001). The 7-T/7-P ratio decreased in males of larger group-size in the two series (p < 0.0001).

In Di2 males, cVa slightly decreased in larger group-size in series1 and 2 (p = 0.015; p = 0.024, respectively). 7-T decreased in larger groups of series1 (p < 0.0001) and showed no change in series2 (p = ns). In males of larger groups, 7-P slightly decreased in series1 (p = 0.016) and slightly increased in series2 (p = 0.012). The 7-T/7-P ratio did not change in series1 whereas it gradually decreased in the larger groups of series2 (p < 0.0001).

No significant variation was found in desat1 males. In series1, the 7-T/7-P ratio slightly increased just below the limit of significance (p = 0.051) between “x5” and “x25” males.

Superficial Compounds

Simultaneously to whole-body compound analysis, we measured the amounts of compounds superficially present on the cuticle (Fig. 2). Basically, despite of the much lower quantities detected, compared to whole-body amounts, grouping induced very similar effects on 7-T and 7-P. Differently, cVa showed no group-size related variation whereas the 7-T/7-P ratio showed a slight divergence between the two extraction modes.

In Cs males, 7-T and 7-P increased with the size of the group in series1 (p < 0.0001) and in series2 between “x25” and “x50” males (p < 0.0001). The 7-T/7-P ratio was not affected. In Di2 males, 7-T decreased in series1 (between “x5” and “x25” males; p = 0.008) and did not change in series2 (p = ns). 7-P was not affected in series1 while it increased in series2 (between “x5” and “x25”/”x50” males; p < 0.0001). The 7-T/7-P ratio decreased in larger groups in series1 (p = 0.005), and between “x5”/”x25” and “x50” males in series2 (p < 0.0001). desat1 mutant males only showed increased 7-T/7-P ratio in series1 (between “x5” and “x25”/”x50” males; p < 0.0001).

Fig. 2
figure 2

Group size effect on superficial pheromones in males of three laboratory lines. The data shown (from top to bottom) correspond to the superfical amounts (in ng) of cVa, 7-T and 7-P extracted in single 4 day-old males (with hexane during 5 min). For genotypes, statistics, and experimental conditions, see the legend of Fig. 1. For series 1, N = 10 except for Di2 × 25 (N = 7); for series 2, N = 15

Full size image

The data shown (from top to bottom) correspond to the superfical amounts (in ng) of cVa, 7-T and 7-P extracted in single 4 day-old males (with hexane during 5 min). For genotypes, statistics, and experimental conditions, see the legend of Fig. 1. For series 1, N = 10 except for Di2 × 25 (N = 7); for series 2, N = 15.

To check the specificity of the compound(s) affected by the size of the group, we measured the level of n-tricosane (23Lin; a saturated linear CHC with 23 C structurally close to 7-T). In series1, 23Lin significantly increased between “x5” and “x25” Cs males (whole-body = 557 and 1058 ng, p < 10− 4; superficial = 210 and 584 ng, p < 10− 4) while it decreased in Di2 males (whole-body = 398 and 161ng, p < 10− 4; superficial = 207 and 101 ng, p = 0.001). We performed a similar analysis on desat1 males which showed no 23Lin difference either in series1 (whole body compounds between “x5” and “x50” males: 1379 and 1406 ng, p = 0.978; superficial “x5”, “x25” and “x50”: 943, 999 and 853 ng, respectively, p = 0.843) and in whole-body 23Lin of series2 between “x5” “x25” and “x50” males (1737, 1709 and 1757 ng, respectively, p = 0.717).

The levels of 7-T, 7-P, 7,11-HD and 7,11-ND, measured in 4 day-old Cs, Di2 and desat1 females showed no difference between “x5” and “x25” groups (Suppl. Figure 1).

In summary, Cs and Di2 males kept in groups of larger size showed frequent variations of 7-T and 7-P. These changes were more constant in Cs males while they sometimes diverged between the two series, in Di2 males.

Single Analysis of Three Other Laboratory Lines

Beside the Cs, Di2 and desat1 lines, we compared (in a single series) “x5” and “x25” males in two other D. melanogaster wild-type lines: Oregon-R (Or-R) and Zimbabwe30 (Z30) and in one D. simulans line (Suppl. Figure 2).

In larger groups of Or-R males, whole-body cVa (p = 0.002), 7-T (p < 0.001) and 7-P (p < 0.0001) decreased while the 7-T/7-P ratio increased (p = 0.004). Superficial cVa was not affected whereas superficial 7-T and 7-P decreased in larger groups (p < 0.0001). 7-T/7-P ratio increased in larger size groups (p < 0.0001).

In Z30 males kept in larger groups, whole-body 7-T and 7-P decreased (p = 0.001 and p < 0.0001, respectively) while cVa and 7-T/7-P ratio did not vary. Only superficial 7-P slightly decreased in larger groups (p = 0.032).

D. simulans males showed no effect either for whole-body or superficial compounds (p = ns).

In summary, group-size generally induced parallel effects on whole body and superficial compounds in the six lines tested. Each line showed a specific pheromone variation pattern. Also, the comparison of whole-body and superficial levels shows a relatively similar dispersion in flies of these six laboratory-acclimated lines between the two extraction modes (Fig. 3).

Fig. 3
figure 3

Principal Component Analysis (PCA) of pheromones in males of 6 laboratory lines. Data show the dispersion of single males according to their genotype (see the color code) and social experience (group sizes are represented by different symbols). The PCA shown on top represents the whole-body compounds analysis and the PCA at the bottom shows the superfical compound analysis

Full size image

Group-size Effect in Freshly Collected Lines

Since males in the four wild-type D. melanogaster laboratory-acclimated lines — but not the desat1 mutant and the D. simulans males — showed a significant group-size effect, we tested males in 21 freshly-caught lines, after 1, 12 and 25 generations acclimation in the laboratory (F1, F12, F25, respectively; Fig. 4; Suppl. Figure 3). This allowed us to test the significance of our finding from the field. Since each line resulted of flies trapped with different fruits, we also analysed lines grouped according to their fruit preference (Fig. 5; Suppl. Figure 4). Our study focused on whole-body compounds which showed more robust data than superficial compounds in laboratory lines (see above). For the sake of clarity, we show the rate of variation between “x5” and “x25” groups (see Material and Methods; The complete data set is shown in Suppl. Figure 3).

Fig. 4
figure 4

Group size effect on whole body pheromones in males of 21 natural lines after laboratory acclimation. Each dot corresponds to the pheromone difference measured between 4 day-old males kept in small (x5) and in larger groups (x25) based on the [(“x25” – “x5”)/ “x5”] formula, in each line (corresponding line # are shown below the graph). Whole body amounts (in ng) of cVa,7-T, 7-P, and the 7-T/7-P ratio were analysed after 1, 12 and 25 generations acclimation in the laboratory (F1, F12, F25; shown on the top). The dotted lines show the separation between lines caught on different types of fruits. N = 8 (Fig), 5 (Peach), 4 (Mirabelle), 1 or 3 (Banana) and 1 (Melon). Detailed data are shown on Supplemental Fig. 3

Full size image

Line by line Analysis

F1 Generation

After one generation in the lab (F1) on standard food, “x 25” males showed significantly increased 7-P in 7 lines (among 19 lines tested; lines #19 and 20 were not tested; Fig. 4). 7-T increased in “x25” males of four lines and decreased in one line (#5) whereas cVa slightly decreased in “x25” males of one line (#11). The 7-T/7-P ratio significantly decreased in “x25” males of three lines.

F12 Generation

F12 males showed a group-size effect in fewer lines, compared to F1 and the variation was not always found in similar lines. In particular, “x25” males showed lower cVa in five lines and lower 7-T in four lines. In “x25” males, 7-P increased in three lines and decreased in one line (#10). The 7-T/7-P ratio decreased in “x25” males of six lines. In summary, the three compounds decreased in “x25” males of the line #10 while cVa and 7-T decreased in two lines (#5, 6). 7-T and 7-P showed a reciprocal variation in the line #1. Please note that the absolute amounts of all compounds were particularly low at F12 generation.

F25 Generation

At F25, 7-P increased in “x25” males of eight lines and decreased in one line (#21). In x25” males, 7-T decreased in four lines and increased in one line (#7) while cVa decreased in three lines (#4, 10, 21). The 7-T/7-P ratio decreased in “x25” males of ten lines. The ratio decrease was frequently found in lines showing increased 7-P. In summary, the three compounds decreased in “x25” males of the line #21 while both 7-T and 7-P increased in the line #7. A reciprocal variation was observed in “x25” males of the line #2 (decreased 7-T and increased 7-P) and of the line #10 (decreased cVa and increased 7-P).

Effect Across Generations

Several lines showed a similar group-size effect across generations (Suppl. Figure 3). In “x25” males, 7-P increased in the lines # 2, 3, 10, 17 (at F1 + F25) and in the lines #7 and 8 (F12 + F25). Note that 7-P decreased at F12 in the line #10. In “x25” males, 7-T decreased in the line #5 (F12 + F25) and showed a reciprocal variation in the lines #2 and 11 (F1 + F25) and in the line #10 (F1 + F12). All significant changes of the 7-T/7-P ratio led to decreased values in “x25” males. The frequency of this effect increased between F1, F12 and F25 (in 3, 6 and 10 lines, respectively). The 7-T/7-P ratio was affected in the lines #3, 8 and 18 (at F12 + F25), and in the lines #10, 11 and 17 (at F1 + F25 ). In “x25” males of the line #10, cVa decreased at F12 + F25. The lines #12, 14 and 15 showed no group-size effect at any generation (Line #20 was not considered).

“Fruit lines” group analysis

Fig. 5
figure 5

Group size effect on whole body pheromones in males of four “Fruit lines” groups after laboratory acclimation. We compared the pheromonal productions (amounts of cVa,7-T, 7-P and the 7-T/7-P ratio) in lines grouped according to the fruit preference of their founder flies (Fig; Peach; Mirabelle; Banana; Melon included only one line and was excluded). Data correspond to pheromone difference measured between 4 day-old males kept in small (x5) and in larger groups (x25; see the legend of Fig. 3). Each dot corresponds to a line and the group is shown with its median (small horizontal bar) and mean (diamond) values. The levels of variation were compared using Kruskal-Wallis test. Different letters indicate significant differences for cVa between males of the different lines at F1 generation. N = 8 (Fig), 5 (Peach), 4 (Mirabelle) and 1 or 3 (Banana). Detailed data are shown on Supplemental Fig. 4

Full size image

We also analysed the lines grouped according to their initial fruit preference (Fig. 5; Suppl. Figure 4). cVa showed no effect at F1, but a general decrease occurred in “x25” males at F12 and F25 (except in “Fig” males at F25). 7-T increased in “x25” males of “Fig” and “Peach” lines at F1 whereas it decreased at F12 in these two groups together with the “Mirabelle” group. No 7-T variation was noted at F25. Across generations males of “Banana” lines showed no effect. All 7-P variations in “x25” males corresponded to increased amounts: at F1 in the “Fig”, “Peach” and “Mirabelle” groups; at F12 in “Fig” males; at F25 in “Fig”, “Mirabelle” and “Banana” males. “x25” males of the “Fig” group showed a constant increase from F1 to F25.

All variations of the 7-T/7-P ratio corresponded to decreased values in “x25” males: at F1 in “Peach” and “Mirabelle” males; At F12 in “Mirabelle” males; At F25 in all groups. Across generations, “x25” males of the “Mirabelle” group showed a persistent decrease.

Discussion

Our study reveals that D. melanogaster males kept in large groups during early adult development frequently changed the production of their two principal 7-monoenes (7-T, 7-P) — less often that of cVa — when compared to males kept in smaller groups. Increasing the density of males in laboratory-acclimated lines and in freshly caught lines induced effects varying (i) between lines taken at a given generation and (ii) between generations in given lines.

Increasing the size of the group in laboratory-acclimated males induced parallel effects on whole-body and superficial compounds with clearer effects on whole-body compounds which showed a relatively lower intra-treatment variability than superficial compounds (Figs. 1, 2 and 3). The group-size effect on whole-body 7-T and 7-P diverged between lines. Cs males kept in larger groups showed the most robust effect with strongly increased 7-T and 7-P (up to + 220% and + 120%, respectively) compared to smaller groups. The parallel variation of 7-T and 23Lin indicates that increased Cs male density similarly affected desaturated and saturated CHCs with 23 C (23 carbon chain length). However, the decreased 7-T/7-P ratio (in larger groups of Cs males) indicates that 7-P increased relatively more than 7-T. In Di2 males, the group-size effect was not constant: 7-T and 7-P differently varied within and between generations. This likely explaining why their 7-T/7-P ratio differently varied between the two series. The different “reproductibility” of the variations noted between Cs and Di2 males could be due to the possible presence of transposable elements in Di2 males — but not in Cs males — this possibly affecting gene regulation (Anxolabéhère et al. 1988; Quesneville and Anxolabéhère 1998). Or-R males kept in larger groups showed a mirrored variation compared to Cs males: 7-T and 7-P decreased while the 7-T/7-P ratio increased. This suggests that 7-T decreased relatively less than 7-P in Or-R males kept in larger groups. Z30 males kept in larger groups showed a parallel decrease of the two 7-monoenes without changing their ratio. The whole-body cVa amount decreased in three of the four wild-type males (Cs, Di2, Or-R) kept in larger groups.

In contrast to wild-type D. melanogaster males, increasing the group size in desat1 mutant and D. simulans males induced no or a very limited effect. This suggests that the aggregation effect (i) depends on gene(s) involved in the biosynthetic pathway leading to final CHC products and (ii) can vary between closely related species. Could the difference observed between the two sibling species be related to their different biosynthetic pathway leading to sexually dimorphic CHCs in D. melanogaster but not in D. simulans flies (Jallon 1984; Chertemps et al. 2007)? This is unlikely since females kept in small and large groups showed no significant CHC differences. Instead, we believe that species difference is due to the higher plasticity of CHC production in D. melanogaster as compared to its sibling species. This difference may be reflected by the fact that the variation of the 7-T/7-P ratio depends on several genes in D. melanogaster males (Scott and Richmond 1988; Ferveur and Jallon 1996) and on a single gene in D. simulans flies (Luyten 1982; Ferveur 1991). However, we cannot exclude that the absence of effect observed in D. simulans is due to the limited number of D. simulans line tested here.

Males of freshly caught lines kept in larger groups showed varied effects on CHCs and cVa, depending on the line considered (Figs. 4 and 5). One or two (rarely three) compounds were affected at one, sometimes at two of the examined generations (F1, F12 or F25). Indeed, these effects were rarely repeated across generations (our interpretation remains reserved for F12 flies which produced much less compounds than F1 and F25 flies). Overall, the three pheromones showed distinct variation patterns. 7-P showed the most coherent variation during laboratory acclimation. In “x25” males, 7-P increased in 35% F1 and F25 lines whereas it only decreased in one line (#21; “melon”) at F25. 7-T showed a more contrasted variation: it increased in four F1 lines and one F25 line while it decreased in one F1 line and four F25 lines. Therefore, the frequently decreased 7-T/7-P ratio in “x25” males during laboratory acclimation (in three F1, six F12, and ten F25 lines) was mostly due to increased 7-P (in 8 lines). This effect was eventually enhanced by decreased 7-T (in 4 lines). Whole-body cVa varied less frequently in freshly-caught lines than in laboratory-acclimated lines (3/4): “x25” males produced less cVa in only one F1, five F12 and three F25 lines. The fact that, during acclimation, increased group-size induced distinct effects on cVa and CHCs suggests that the social interaction, and male-male competition, differently affected their biosynthesis. Such dissociation maybe due to the distinct enzymatic pathways, biosynthesis sites and/or externalization canals involved in CHCs and cVa production and emission (Jallon 1984; Ferveur et al. 1997; Howard and Blomquist 2005; Guiraudie-Capraz et al. 2007; Wang et al. 2020b; Cortot et al. 2022a).

Group-size related variation of pheromones may be — at least partly — under hormonal control. The juvenile hormones (JHs) are pivotal gonadotropic hormones involved in several aspects of insect reproduction (Uzsák and Schal 2013). Their synthesis depends on brain associated glands: the corpora allata (CA; (Feyereisen 1999). The density and caste of social insects can modulate the hormone level affecting both the behavior and CHCs. For example, while the increased density of Formosan subterranean termites showed elevated JH titer in individual workers, the subsequent introduction of soldiers lessened this level (Mao and Henderson 2010). A specific JH hormone (JH III) changed task-specific CHC profile in the ant Myrmicaria eumenoides (Lengyel et al. 2007). Both increased JH titer and CA size were correlated with higher reproductive dominance in a eusocial polistine wasp (Sledge et al. 2004).

The influence of hormones was often reported on reproduction and CHCs in non-social insects such as Diptera. For example, D. melanogaster females present in large groups laid egg faster; this effect is dependent on light and JH activity (Bailly et al. 2023). In the female house fly, Musca domestica, and likely in other Diptera species, ovarian-produced ecdysteroids regulate CHC synthesis by affecting specific elongases (Tillman et al. 1999). The surgical removal of ovaries or of CA in M. domestica induced a drastic change in the internal level of its principal CHC pheromone (9-tricosene; 9-T); a close-to-normal level of 9-T was restored after the injection of ecdysone (Adams et al. 1995). Similarly, the decapitation of immature D. melanogaster adults strongly affected the CHC profile of mature adults; this profile was partly restored by the application of a JH synthetic analog (Wicker and Jallon 1995). The manipulation of the bursicon neurohormone receptor (Rickets) affected the externalization of CHCs (Flaven-Pouchon et al. 2016). Also, the sex-ratio of D. serrata populations affected the CHC profile of their males (Gershman and Rundle 2017) while D. melanogaster flies tested in a social context showed inter-individual variation of their chemical communication depending on the genotype X environment interaction (Kent et al. 2008).

Similarly to desat1 and D. simulans males, males of three freshly caught lines (#12, 14 & 15) showed no pheromone variation at any generation. How can we interpret the stability (or absence of effect) observed in these lines? Since the amount of 23Lin was not affected in larger groups of desat1 males (differently to Cs males), the absence of effect may not be related to the bias caused by their very low levels of 7-T and 7-P. We rather hypothesize that beside its “canonical” effect on the carbon chain desaturation (Marcillac et al. 2005a), the defective expression of the desaturase1 gene also altered mechanisms underlying CHC plasticity. This novel effect of desat1 can be added to another unexpected effect induced by this mutation on the altered externalization of cVa (Cortot et al. 2022a).

The increased frequency of lower 7-T/7-P ratio during laboratory acclimation (from F1 to F25) observed in freshly-caught line males kept in larger groups seems to mostly result of increased 7-P levels. We recently found a frequently decreased 7-T/7-P ratio in “x5” males of freshly-caught lines during laboratory acclimation (until F40; (Ferveur et al. 2024a). In these lines, the “across-generations” acclimation effect on 7-T/7-P decrease was found in males of the “Fig”, “Mirabelle” and “Banana” lines but not of the “Peach” lines. Taken together these data indicate that the 7-T/7-P ratio can be affected by the (i) increased group-size, (ii) duration of laboratory acclimation, (iii) initial fruit preference, and (iv) genetic background.

This study leaves several questions open. Beside genes and hormones, which factors activate or repress CHCs and cVa production? Since the finely tuned CHC biosynthesis depends on many biotic and abiotic factors, we believe that any subtle change of their interaction could explain the frequent and somewhat fluctuating variation of pheromone levels within and between generations. The regulation of some of the biosynthesis genes could be affected by the presence of transposable P-elements as found for metabolic and reproductive functions (Clark et al. 1995; Serrato-Capuchina et al. 2021). Could the natural fruit preference impact the subsequent CHC evolution in the laboratory where the food and environment are more constant? Since the different microbiota associated with different fruits drive the fly preference (Becher et al. 2012), the varied microbes ingested in nature could differently affect CHCs plasticity depending on their transmission and conservation in laboratory lines. This hypothesis is supported by the relationship found between the cuticle chemistry and the composition of bacterial microbiota in D. melanogaster flies (Mokeev et al. 2021). Overall, the genetic background variation in the “fruit lines” tested in this study may be responsible for the observed differences (Ferveur et al. 2024a). Indeed, most flies caught in the field might have a more variable genetic background than “old established” laboratory lines. Thus, allele variation in genes directly or indirectly responsible for pheromone production and externalization may account for the different responses to male density.

In summary, changing the number of D. melanogaster males in a limited space affected the amount of their sex pheromones in laboratory-acclimated wild-type lines and in freshly caught lines, but not in a CHC mutant and in the D. simulans species. During laboratory acclimation, the production of the CHC with the longer carbon chain showed the highest plasticity while the internally-produced pheromone less frequently varied. In conclusion, our study stress the importance of (1) keeping flies in small groups for pheromone analysis and (2) not extrapolating natural CHC phenotype with insects acclimated for many generations in the laboratory.