Predicting sex bias in mobility from functional traits in flying insects

Understanding the movement patterns of organisms is crucial for effective biodiversity conservation in the increasingly dynamic and fragmented landscapes. Since the colonization of habitat patches relies largely on females, sex differences in movement capability must also be considered. However, obtaining direct measurements of mobility and dispersal, and sex biases in these traits, is often challenging. This underscores the importance of predicting sex-specific estimates of mobility based on species’ functional traits. Our phylogenetic comparative study aims at identifying species traits that could serve as proxies for sex bias in mobility in flying insects. We rely on a comprehensive dataset on the sex ratios of 454 moth species (Lepidoptera: Macroheterocera), captured by light traps of the Finnish national moth monitoring scheme. We first demonstrate that, in the vast majority of species, males outnumber females among the captured individuals. Our phylogenetic regression models reveal that species-specific sex ratios correlate with traits typically predicted to be associated with sex differences in mobility. Female proportions decrease as sexual dimorphism in wingspan becomes more male-biased and female wing loading relative to males increases. Proportions of females are also lower in larger species. Females are particularly scarce in trap samples of species in which the reproductive output of females is primarily determined by larval-derived resources (i.e., capital breeders). These associations suggest that the observed variations in sex ratios do indeed mirror the variation in sex bias in mobility across species. Our findings highlight the potential of trait-based approaches to identify meaningful indicators of insect mobility, including sex biases in mobility. The availability of such proxies facilitates predictions about how different species might respond to contemporary challenges, such as light pollution and habitat loss and fragmentation. The detected associations also advance ordination schemes of insect life histories by integrating mobility measures into relevant analyses.

Methods
Sex ratio data

We took advantage of a vast dataset generated by the Finnish Moth Monitoring Scheme 'Nocturna', which has been operating an extensive nationwide network of light traps for over three decades (Huikkonen et al. 2024). The essence of using light trapping in moth monitoring lies in the well-known habit of these nocturnal insects to be attracted to artificial light sources. The light traps deployed by the monitoring scheme cover the entire flight period of moths in the region; they are installed in early spring (early April to early May, depending on latitude) and dismantled in autumn after moths cease flying (early October to early November). The traps are typically serviced and emptied once a week, with all captured macroheteroceran moths identified to the species level by voluntary lepidopterists, and the number of individuals per species is reported in an online database.

Conveniently for our study, many contributors of the scheme systematically report the numbers of trapped individuals separately for males and females (53 % of the total number of individuals in the database), thus providing data on the relative proportions of the two sexes among the trapped individuals. Sex-identified catches from 34 light traps, each operated for at least ten years between 1993 and 2012, provided a total of 358 trap-year samples, comprising 1,383,297 moth individuals. Data used in our analyses were extracted from the moth monitoring database (Yöpeti) on 22 October 2014. In our analyses, we included only species with more than 20 individuals captured over the trapping years. This criterion was met by 483 out of the 635 species in the total dataset. We applied two further inclusion criteria. Firstly, we excluded nine species with short-winged or wingless females, as light traps capture only males from these species. Secondly, we did not consider twenty species known to exhibit significant day-flying activity. In such species, there are often sex-related differences in circadian rhythms, potentially introducing unpredictable effects on sex ratios derived from night-operated light traps. For instance, in the lasiocampid Macrothylacia rubi, males primarily fly in the afternoon, whereas females fly only at night, causing a strong female bias in light traps. Applying these additional criteria left us with sex ratio data on 454 moth species, representing ten moth families, forming the basis of our study. 

Species trait data

Data on species' ecological traits (host plant specialization, habitat preference, presence/absence of adult feeding) and sex-specific data on body-size-related (wingspan, wing loading, body mass, abdomen mass, relative abdomen size of females) traits were obtained from previously published studies. Data on host plant specialization were retrieved from Teder (2020), where species were categorized as monophagous, oligophagous, or polyphagous. Habitat preference was rated on a scale from 1 (strictly open-habitat species) to 5 (strictly forest species) (see Ude et al. 2024 for details). Species classification by the presence or absence of adult feeding was based on data from regional taxonomic handbooks.

Male and female wingspan measurements were obtained from Teder (2020), where wingspan was defined as the maximum distance between the forewing tips measured in collection specimens mounted in the traditional way. Sex-specific wingspan data were available for all species for which we had sex ratio data. For a subset of 38 species within the Geometridae family, we could use additional sex-specific size-related traits, empirically recorded for a comparative study by Davis et al. (2016). These variables included sex-specific measurements of dry body mass and dry abdomen mass, as well as relative abdomen size in females (RAF), calculated as abdomen mass divided by body mass at adult eclosion. Finally, the wing loading index was computed for each sex by dividing body mass by the squared wingspan, with the squared wingspan serving as a proxy for wing area. Higher values of the wing loading index indicate a larger body mass relative to wing size, suggesting a higher energetic cost of flying and potentially reduced mobility.

For each trait with sex-specific values available, we calculated an index of sexual dimorphism (hereafter SSDwingspan, SSDmass, SSDabdomen, SSDwingloading). Using the algorithm proposed by Lovich and Gibbons (1992), SSD was calculated as follows: SSD = [(trait value of the larger sex) / (trait value of the smaller sex)] – 1. The resulting SSD values were assigned a positive sign if females were larger and a negative sign if males were larger. The SSD values calculated this way are symmetrically distributed around zero, where SSD = 0 indicates no size difference between sexes, while SSD > 0 and SSD < 0 denote female- and male-biased dimorphism, respectively. 

Phylogeny of the species

We took advantage of the fully resolved phylogenetic relationships of northern European Geometridae (Õunap et al. 2024), which were based on 11 molecular markers commonly used for phylogeny reconstruction in this group. In total, the data matrix encompassed 458 species of geometrids and 37 outgroup taxa. The maximum likelihood (ML) tree was constructed using IQ-TREE v2.1.2 (Nguyen et al. 2015), and subsequently pruned to include only species relevant to our study (see Holm et al. 2018 for further details).