Phylogeography and population genetics
Historically, glacial cycles probably had a strong influence on species distributions. Most of Iceland‘s shelf was covered by an ice sheet during the Last Glacial Maximum (~25 ka), which broke up around 15 ka due to a northward shift of the polar front and rising sea levels (reviewed by Geirsdóttir et al., 2009). At that time, today’s pattern of currents was established. The near-shelf populations ofH. bicuspis (represented by stations 1019, 1194 and 1219) were likely colonized following the break-up of the Icelandic ice shield. Deep-sea populations farther off the coast may have also been affected by changing temperatures, currents, etc. The observed demographic history for H. bicuspis I-III suggests a relatively recent decline in population sizes followed by a rapid recovery, though we do not have any age estimates (Figure 5). This overall pattern might reflect population declines and local extinctions during the Last Glacial Maximum. This was then followed by population expansions and recolonizations from unaffected deep-sea regions farther off the coast during the Holocene, which were scarcely sampled in our study. Colonization from less affected regions might also explain the rapid recovery of population sizes.
Our data suggests a complex colonization history of the near-shelf regions north of the GIF Ridge for all three lineages observed withinH. bicuspis I-III. The consistently observed differences between the mitochondrial COI and nuclear ddRAD data imply differing migration behavior between males and females. The maternally inherited COI exhibited consistently higher levels of genetic differentiation between populations, which suggests males migrate more actively and potentially over wider distances, while females appear to be rather stationary. This is surprising, as Haploniscus bicuspis does not exhibit a strong sexual dimorphism in swimming structures as observed in other isopod families (e.g., Brix et al., 2020; Bober et al., 2018; Hessler, 1970; Riehl, Wilson, & Hessler, 2012). However, adult males of H. bicuspis have stouter second antennae with many more sensory sensilla and aesthetascs than females or juvenile and preparatory males. This difference was not described by Brökeland & Wägele (2004), as they did not have adult males available in the type series. The aesthetascs are hypothesized to have a chemosensory function, possibly enabling males to detect females across large distances. This suggests that males have a more active lifestyle, roaming around in search for females, which would explain the inferred sex-specific migration patterns and the different distribution patterns of COI and ddRAD lineages. As stated in Brix et al. (2020), locomotion of the adult stages does influence migration patterns and distance for Pacific isopod families, but also a “male behavior” is reflected in the Pacific data for Haploniscidae.
The ‘central’ region north of the GIF Ridge around stations 1194 and 1219 was probably first colonized by lineage II, possibly originating from close-by populations in the Norwegian Sea to the east or north-east of Iceland. Subsequently, males from northern (lineage III) and southern (lineage I) populations migrated into the area around station 1219, largely replacing the local population. The result is a local population at station 1219 that comprises a hybrid nuclear genome of lineages I and III, but the mitochondrial genome of lineage II. These newly arriving males must have outnumbered the local (male) population or were better adapted to the local environmental conditions, or both. We propose a similar scenario for the north-eastern populations around stations 1159 and 1172. These were probably first colonized by lineage I, possibly from the central or southern Norwegian Sea, with subsequent male-biased migration and introgression from lineage II, resulting in populations with lineage I mitochondrial and lineage II nuclear genomes. The reverse, a female-biased migration and introgression, is highly unlikely in both cases due to the overall higher rates of population differentiation in COI (suggesting lower female migration rates) and because in such a scenario these females should have contributed to the nuclear genome as well.
Overall, our results show that genetic differentiation between populations is usually high, even at comparably low geographic distances, suggesting low dispersal and gene flow rates. This is also supported by elevated inbreeding coefficients in some populations, implying that reproduction occurs frequently among closely related individuals. The above-described hybridization scenarios appear to represent rare events, which probably occurred at a time when local population densities were still low in the early phase of recolonization. The strongly male-biased migration behavior is noteworthy, with the only apparent sexual dimorphism in sensory structures. If such patterns hold true for other asellote isopods, population differentiation inferred from mitochondrial markers like COI might systematically underestimate dispersal and gene flow rates.
Haploniscus bicuspis IV is the only putative species with consistently positive values in Tajima’s D and Fu’s Fs, suggesting that it is currently undergoing a genetic bottleneck. At least in COI, this putative species appears closer related H. bicuspis I-III, which occurs exclusively in colder arctic waters. It is possible that H. bicuspis IV is also better adapted to colder temperatures and is negatively affected by the northward movement of the Arctic front throughout the Holocene. The ranges of H. bicuspis IV and V meet along the Reykjanes Ridge, and we found H. bicuspis V at most stations, but with less genetic diversity among those found at the Reykjanes Ridge stations. Whether both species coexist due to the topological, structural, and environmental diversity of the Reykjanes Ridge (German et al., 1994; Taylor et al., in review) or whetherH. bicuspis V is currently replacing H. bicuspis IV is an unanswered and interesting question.