A flightless beetle's head shape is diverging inside Honshu, not just across the sea

A team led by Junji Konuma of Toho University, with Nobuaki Nagata of the National Museum of Nature and Science and Teiji Sota of Kyoto University, combined genetic data from 170 individuals at 70 sites with morphological data from 50 populations of the snail-eating ground beetle Carabus blaptoides ^[8][9]. Their analyses suggest its body shape has diverged adaptively not only among islands, including Hokkaido and Sado, but also between regions within Honshu, Japan's largest island ^[10][15].

The interesting part is the mechanism, which is unusually concrete for a divergence study. Shape variation in this beetle sits on a functional trade-off between two ways of eating a snail ^[6]. Slender beetles have long, narrow heads and thoraxes and push their heads deep into the shells of large snails ^[4]. Stout beetles have short, wide heads and strong mandibles, and crush the shells of relatively small snails instead ^[5]. Earlier work suggested that regional differences in the size of available snails contributed to this diversification ^[7]. So the phenotype is not an abstract morphometric axis; it is a feeding tool whose optimum changes with what the local snails look like.

Flightlessness is what makes the within-island result plausible rather than surprising. The beetle's hind wings are reduced, so it cannot fly and disperses on foot ^[3]. Distance on land therefore behaves somewhat like water does for a winged species: it limits gene flow, and limited gene flow lets local selection accumulate ^[14]. Konuma put it directly, saying that adaptive radiation has usually been discussed in terms of populations isolated on oceanic islands, and that on large continental islands geographic distance and environmental variation within the same landmass may promote adaptive divergence, a process he expects to matter particularly for flightless animals ^[12]. The species has already been split into eight subspecies across the Japanese archipelago, which is a taxonomic hint that this pattern was visible long before the sequencing ^[1][2].

Two limits are worth holding onto. The study infers history from present-day sampling, and the phys.org summary consistently reports that the analyses "suggested" divergence rather than demonstrated selection experimentally ^[10]. And the sampling is not dense: 170 individuals across 70 sites averages roughly 2.4 beetles per site ^[16], and shape was measured in 50 populations, twenty fewer than the number of sites genotyped ^[17]. Fine-scale conclusions about where breaks fall inside Honshu will depend on how those samples are distributed.

What to watch: whether the genetic breaks within Honshu line up with the existing eight-subspecies boundaries or cut across them ^[1]; whether other flightless taxa on the same island show congruent breaks, which would point to shared geography rather than beetle-specific history ^[14]; and whether snail-size data from the same sites are published alongside the beetle shapes, since that is the step from correlation to a tested cause ^[7]. The paper is in the Biological Journal of the Linnean Society, DOI 10.1093/biolinnean/blag065 ^[11]. Snail-feeding carabids have radiated across continents and continental islands throughout the Northern Hemisphere, so there is plenty of comparative material to test the claim against ^[13].