Sex pheromones are among the most studied insect mating signals, with their extensive diversity underscoring their crucial role in promoting behavioural isolation during speciation. In Chrysomelidae, cuticular wax (CW), a hydrophobic layer covering the insect cuticle, is a potential barrier that facilitates behavioural isolation. Male leaf beetles use female CW as a mating signal, and their species-specific profiles prevent heterospecific mating between closely related species, implying that divergence in CW profiles promotes reproductive isolation, and, hence, contributes to speciation. However, the role of CW as an isolating barrier remains unclear owing to limited knowledge regarding intraspecific divergence in female CW and its coevolution with male preferences. Through chemical analysis and behavioural experiments, we demonstrated that intraspecific divergence in female CW profiles contributes to partial behavioural isolation between leaf beetle populations with different dispersal traits: the flight-capable macropterous and flightless brachypterous forms of Galerucella grisescens (Coleoptera: Chrysomelidae). To the best of our knowledge, this is the first study on the coevolution of female CW and male preferences at the intraspecific level in Coleoptera. Our results support the potential role of CW as an isolating barrier in Chrysomelidae and are consistent with previous findings that loss of flight enhances beetle diversification.

(a) Insects

The macropterous population originated from Ogata, Akita, Japan (40°0ʹ N, 139°56ʹ E), and the brachypterous population originated from Sendai, Miyagi, Japan (38°16ʹ N, 140°52ʹ E). Each population was reared under the same conditions to avoid the potential effects of extrinsic factors (humidity, temperature, and nutrition) on CW profiles and to investigate genetically based divergence in phenotypes. Both populations were maintained in a climate chamber at 25 ± 1 ℃ and 60% relative humidity under a 16L:8D photoperiod and reared on fresh leaves of Rumex obtusifolius. Newly emerged adults were sexed and kept separately from each other. Unmated adults at 7–8 days post-emergence were used in the experiments.

(b) Hindwing polymorphism

To describe hindwing polymorphism, we examined the development of hindwings in females and males from each macropterous and brachypterous population (n = 25). After the beetles were freeze-killed, their right forewings and hindwings were dissected. Images were captured using a camera (Axiocam 305 Colour; Zeiss Microscopy, Jena, Germany) on a stereomicroscope (SteREO Discovery.V12, Zeiss Microscopy). Each forewing and hindwing area was measured using ImageJ version 1.53t. To correct wing size variations caused by body size variations, the development of hindwings was evaluated by calculating the ratios of hindwing areas to forewing areas.

(c) CW extraction

Female and male CW were extracted from 100 individuals of each macropterous and brachypterous population per extraction. After beetles were freeze-killed at –20 ℃, they were thawed for 15 min at room temperature. They were then immersed in two sequential 20 mL aliquots of n-hexane for 30 min each. The two extracts were mixed, and the solvent was evaporated at 39 ℃ using a rotary vacuum evaporator (N1200A‐S, EYELA, Tokyo, Japan). The extracts were dried under a gentle stream of nitrogen gas. The obtained CW was used for chemical analysis and mating experiments.

(d) Chemical analysis

To investigate the divergence in the CW profiles, we conducted a chemical analysis. The CW samples were reconstituted in 500 µL of n-hexane containing 2 µg of n-pentadecane as an internal standard. Three samples were prepared from females and males from each macropterous and brachypterous population. An aliquot of 1 µL of each sample was injected manually into a gas chromatography-mass spectrometry (GC‐MS) (GS-MS-QP2010 Ultra, Shimadzu, Kyoto, Japan) equipped with a DB‐5ms column (30 m × 0.25 mm i.d., 0.25 µm film thickness, J&W Scientific, CA, USA). The analytical conditions were partly based on those used in previous studies (Sugeno et al. 2006; Chiba et al. 2023). Helium was used as the carrier gas. The inlet temperature was 280 ℃, and injection was performed in split mode (1:10). The oven temperature program was started at 50 ℃ and heated to 200 ℃ at a rate of 30 ℃/min, followed by an increase to 320 ℃ at a rate of 2 ℃/min, and finally maintained for 15 min. The mass spectrometer was operated in an electron impact mode at 70 eV with a source temperature of 230 ℃. Saturated hydrocarbons, the major components in the CW of G. grisescens, were identified by their mass spectra and Kovats indices. The Kovats indices of the detected compounds were calculated from the retention times of a standard mixture of n-alkanes (C9–C40) (GL Sciences, Tokyo, Japan).

(e) Mating experiment

To elucidate whether male preferences diverged between populations, we conducted mating experiments. All the experiments below were conducted at 25 ± 1 ℃ during the light period of the rearing condition.

(i) Male preference for live females in a two-choice design

First, we explored male mating preferences through two-choice experiments. Macropterous and brachypterous females were released together on a 48‐mm‐diameter glass Petri dish lined with filter paper, after which a male was released in the same test arena. We recorded which females were first chosen by males (n = 60). When males attempted aedeagal insertion, we regarded it as ‘chose’. The experiments were terminated when the males chose either the female or did not attempt aedeagal insertion within 3 h.

(ii) Male preference for female CW

We investigated male preferences for female CW to elucidate whether female CW mediated male mate preferences. A male was released on a 48‐mm‐diameter glass Petri dish lined with filter paper, on which a macropterous and brachypterous female was fixed with double-sided tape on the same centreline passing a circle at the centre of the Petri dish, 1 cm away from the wall. To prevent potential positional bias, the locations of the two females were exchanged for each replicate. The females were subjected to one of three treatments: 1) dead, 2) reapplied, or 3) exchanged. The procedure was based on a previous study (Chiba et al. 2023). Briefly, for Treatment 1, live females were freeze-killed at –20 ℃ for 30 min, following which they were thawed for 15 min at room temperature. For Treatment 2, the CW of the freeze-killed females was removed by immersing them in n-hexane twice for 30 min each, and then the CW was reapplied by pipetting one female equivalent of the n-hexane solution of CW that had been extracted from females of the same wing morph. For Treatment 3, n-hexane-washed females were covered with one female equivalent of CW extracted from females of different wing morph. We recorded which females were first chosen by males (n = 60). The experiments were terminated when the males chose either the female or did not attempt aedeagal insertion within 3 h.

(iii) Male preferences for live females under a no-choice design

When reproductive isolation is well established, mate preferences can be observed under both two- and no-choice designs. Thus, we evaluated male mate preferences in a no-choice situation. A macropterous or brachypterous female was released on a 48‐mm‐diameter glass Petri dish lined with filter paper, followed by a male in the same test arena (n = 30). We measured the time taken by males to copulate with paired females. In addition, because G. grisescens males inspect potential mates while mounting them and decide whether to attempt subsequent aedeagal insertion, we counted the number of times they dismounted from paired females without aedeagal insertion. The experiments were terminated when males copulated with paired females or when 3 h had elapsed since the start of the observations.

(f) Statistical analysis

The CW profiles between populations were compared by hierarchical clustering analysis using z-scores calculated for the peak areas of major CW components (peaks occurring at a mean relative abundance of over 0.5% in either or both populations). The analysis was based on the Euclidean distance measure and Ward’s clustering method. Male preferences for live females in the two-choice design were analysed using binomial tests. Mate preferences between treatments were compared using Fisher’s exact test with Holm correction to investigate male preferences for female CW. The latency to copulate with paired females was compared using log-rank tests when male mate preferences were investigated under a no-choice design. The number of times males dismounted from paired females without copulatory attempts was compared using the Mann–Whitney U test. All analyses were performed using R version 4.4.1.