Hymenodoridae, Lunina & Kulagin & Vereshchaka, 2024

Lunina, Anastasiia, Kulagin, Dmitry & Vereshchaka, Alexander, 2024, The taxonomic status of Hymenodora (Crustacea: Oplophoroidea): morphological and molecular analyses suggest a new family and an undescribed diversity deep in the sea, Zoological Journal of the Linnean Society (Zool. J. Linn. Soc.) 200 (2), pp. 336-351 : 348-349

Evolutionary divergence within Hymenodoridae and cryptic speciation

Hymenodoridae show an unusually great evolutionary divergence. In fact, K2P distances between species of Hymenodora are similar to those between genera of the sister clade Acanthephyridae . Even within the clades H. glacialis (between H. glacialis A, H. glacialis B, and H. glacialis C) and H. gracilis (between H. gracilis A and H. gracilis B), COI genetic distances range from 0.095 to 0.145 and are comparable to those between species of Acanthephyra (0.024 –0.257; Vereshchaka et al. 2022). The K2P distances between all clades of H. glacialis and H. gracilis thus fit the ranges of interspecies K2P distances in other Oplophoroidea , and these taxa can be considered as cryptic species.

Cryptic speciation seems to be common in the pelagic eucarids and has been found recently in various groups, such as Oplophoridae and Euphausiidae : J. spinicauda and Meningodora mollis Smith, 1882 ( Wong et al. 2015); Oplophorus gracilirostris A. Milne-Edwards, 1881 and Systellaspis debilis A. Milne-Edwards, 1881 ( Lunina et al. 2019); and Hansarsia (former Nematoscelis ; Kulagin et al. 2021).

Although K2P distances and COI - and 16S-based trees show a hidden biodiversity, the morphological differences within H. frontalis + H. gracilis and H. glacialis clades are inconspicuous. Only H. frontalis can be identified reliably, owing to its distinctive elongated rostrum. The clade H. glacialis differs from the rest of Hymenodora only in the presence of an additional lateral sulcus on the carapace, but the three embedded molecular clades cannot be distinguished on the basis of qualitative characters. Likewise, no distinct morphological differences were found between H. gracilis A and H. gracilis B in the clade H. gracilis . The two remaining species of Hymenodora , H. acanthitelsonis and H. chacei , forming the ‘ Hymenodora acanthitelsonis ’ clade, were collected long ago and could not be sequenced, but their unique morphological characters suggest significant divergence on molecular trees.

Given that confident identification of molecular clades within H. glacialis and H. gracilis complexed is currently possible only through sequencing, available data on their geographical distribution are intriguing, albeit scant ( Fig. 8 ). In the Atlantic and adjacent areas, the H. gracilis complex occurs in the Caribbean ( H. gracilis A) and in the subtropical waters of the Southern Hemisphere (mainly H. gracilis B). Both clades are likely to have a panoceanic distribution and the spatial gaps may be explained by the lack of information. The H. glacialis complex was recorded in the Arctic ( H. glacialis A) and in the Central Atlantic ( H. glacialis B and H. gracilis C). Hymenodora glacialis A shows allopatric speciation, whereas H. glacialis B and H. glacialis C are sympatric.

Hymenodora glacialis A, H. glacialis B and H. glacialis C show distribution driven by the expected factors. The Arctic water mass (in the case of H. glacialis A) is known to delimit the distribution of cryptic lineages in copepods (Сornils et al. 2017) and mesopelagic chaetognaths ( Kulagin et al. 2014, 2017). Ocean-scale basin ecoregions or large-scale gyres (in the case of H. glacialis B and H. glacialis C) are also expected to delimit genetic lineages of plankton organisms ( Goetze 2005, 2011, Miyamoto et al. 2010, Blanco-Bercial et al. 2011, Burridge et al. 2015, Kulagin and Neretina 2017, Choo et al. 2021, Kulagin et al. 2021). In contrast, the distributions of H. gracilis A (Caribbean, South Atlantic and Southeast Pacific) and H. gracilis B (Indian Ocean and Southeast Pacific) do not fit any precise biogeographical model, and only new findings of these lineages might clarify the drivers of their geographical distribution.

The existence of complexes of cryptic lineages is usual for the benthic decapods. For example, at least five cryptic lineages were found in the decapod Thor amboinensis (De Man, 1888) ( Titus et al. 2018) , and even 10–15 putative cryptic species were suggested for the Saron marmoratus Olivier, 1811 and Saron neglectus De Man, 1902 species complexes ( Baeza et al. 2023). Decapod cryptic complexes have also been suggested for the pelagic lineages: widespread pelagic genera usually encompass at least one panoceanic species that likely represents a complex of cryptic lineages; Vereshchaka et al. 2014. Here, we confirm the cryptic speciation in the pelagic Hymenodora and show that even cryptic lineages can be identified morphologically, at least in our database. In fact, the simultaneous use of seven characters allows a successful visualization of morphological distances between all lineages (multidimensional scaling ordination; Fig. 6B ) and a confident identification of lineages (GLMMs; Supporting Information, Table S6).

Our dataset is too scant to provide a script for a model identification of cryptic lineages in new datasets; furthermore, we failed to find diagnostic morphological characters to separate cryptic lineages within H. glacialis and H. gracilis complexes from each other. Therefore, in the absence of clear distinct characters to identify cryptic lineages of Hymenodora , we adopt a conservative approach, considering them as genetic lineages, and do not erect new species. Instead, we state here the existence of the species complexes H. glacialis and H. gracilis and provide an amended key to identify them along with the other species of Hymenodoridae .

Figures

Figure 6. Principal component analysis of seven morphological characters of the clade Hymenodoridae. A, molecular clades within Hymenodora glacialis and Hymenodora gracilis combined (analysis 6). B, molecular clades within H. glacialis and H.gracilis analysed separately (analysis 7). The x- and y-axes are eigenvalue scales; the morphometric dataset is presented in the Supporting Information (Table S5); only four most significant factors are shown.

Figure 8. Geographical distribution of Hymenodora gracilis A, H. gracilis B, Hymenodora glacialis A, H. glacialis B and H. glacialis C.