Coryphella Gray, 1850

Genus Coryphella Gray, 1850 View in CoL , restricted

( Figs 1, 2, 13–16; Table 5)

Korshunova et al. 2017a: 28–29, Coryphella restricted.

Type species: Eolidia verrucosa M. Sars, 1829 .

Diagnosis: Body moderately narrow. Notal edge almost completely reduced. Cerata in several groups. Rhinophores smooth with small tubercles. Anterior foot corners present. Central teeth with non-compressed moderately wide cusp and distinct denticles.Lateral teeth denticulated without attenuated process basally. Distal and proximal receptaculum seminis. Moderately short vas deferens expands to a broad penial sheath, prostate distinct, S-shaped. Penis disk-shaped with numerous small triangular processes at the disk edge.

Species included: Coryphella pseudoverrucosa Martynov et al., 2015 , reinstated, Coryphella longicaudata O’Donoghue, 1922 , reinstated, and Coryphella verrucosa (M. Sars, 1829) , restricted in Korshunova et al. 2017a and herein ( Figs 13–16). See detailed details. P, jaw, SEM. Q, details of masticatory process of jaw, SEM. R, details of masticatory process of jaw, SEM. S, radular teeth, posterior part of radula, SEM. КM 565 (T–W), Rich Passage, Washington, USA, 12.5 mm length (live). T, dorso-lateral view. U, living animal on substrate. V, details of cerata. W, radular teeth, posterior part of radula, SEM. X, scheme of reproductive system. Scale bars: D, 100 μm; E, 50 μm; F, 50 μm; H, 100 μm; I, 30 μm; J, 10 μm; K, 30 μm; L,10 μm; P, 200 μm; Q, 100 μm; R, 10 μm; S, 100 μm; W, 50 μm; X. 0.5 mm. Photos: Karin Fletcher (A–C, G, T–V), Tatiana Korshunova (M–O). SEM images: Alexander Martynov. Abbreviations: a, ampulla; an, anus; fgm, female gland mass; go, genital opening; p, penis; pr, prostate, psh, penial sheath; rsd, receptaculum seminis distal; rsp, receptaculum seminis proximal; vd, vas deferens.

morphological data in Korshunova et al. (2017a), and in the present study.

Remarks: The genus Coryphella significantly differs from all genera of the family Coryphellidae by the combination of an almost completely reduced notal edge (the last remnants may be detected, after a very scrupulous study mostly below the anterior ceratal clusters, and thus there is no considerable ambiguity to consider that state also as ‘completely reduced’), smooth to slightly tuberculate rhinophores, non-compressed cusp of central teeth, lateral teeth without an attenuated process basally, a disk-shaped penis with small triangular processes, and a distinct S-shaped prostate (previosly considered as a part of prostatic vas deferens). Importantly, this complex of characters is stable and morphologically recognizable in no less than three separate species of true Coryphella : C. longicaudata , C. pseudoverrucosa , and C. verrucosa (see detailed explanation and discussion below; Figs 13–16). Instead of a putatively single species’ distribution over both the North Atlantic and the North Pacific, at least three evidently separate species exist within the proper genus Coryphella , each with its own complex evolutionary history ( Korshunova et al. 2017a, present study; Figs 1, 2, 13–16).

In the present study, three species within the Coryphella clade were revealed. C. pseudoverrucosa clustered in a distinct and separate clade (PP = 0.99, BS = 91), sister to the separate clade C. longicaudata (PP = 1, BS = 95), except for specimen WS14404. Sister-clades C. pseudoverrucosa and C. longicaudata branch directly from the clade containing C. verrucosa ( Fig. 13). The haplotype network based on the COI gene marker was calculated in the present study to evaluate the genetic distribution of the different haplotypes within the Coryphella complex. The results showed a network of haplotypes ( Fig. 14A), that clearly cluster into three groups coincident with Coryphella verrucosa (M. Sars, 1829) , original description in Sars (1829), C. pseudoverrucosa Martynov et al., 2015 , original description in Martynov et al. (2015), and a species classified here as C. longicaudata O’Donoghue, 1922 (original description in: O’Donoghue 1922a), reinstated. To provide robust taxonomic assessment, neotype KM 569, 17 mm length (live), for C. longicaudata is selected here from Tacoma, Washington, USA, collected on 9 November 2014 by Karin Fletcher at the depth 13.4 m, from the geographic region close to the type locality, which included both the southern part of British Columbia ( Canada) and San Juan Islands, Washigton, USA ( O’Donoghue 1922a). The external, jaws, and radular characters of the neotype are presented in Figure 15A–F. Intragroup and intergroup genetic distances for the COI marker are shown in Table 6. Uncorrected COI p -distances within the Coryphella verrucosa group are 0–2.0% and within the C. pseudoverrucosa group are 0–1.2%. Uncorrected COI p -distances between C. verrucosa and C. pseudoverrucosa are 2%–3.7%. The distances within the C. longicaudata group are 0–2.0% and within the C. longicaudata dataset (except specimen WS14404) are 0–1.7%. The distances between C. longicaudata and C. pseudoverrucosa are 1.7/1.8–3.5% (see details in Table 6). Thus, the distances between C. pseudoverrucosa and C. verrucosa , as well as between C. pseudoverrucosa and C. longicaudata , exceed the intraspecies’ genetic distances. The distances between C. longicaudata and C. verrucosa are 0.8/1.3–3.5% (see details in Table 6), therefore, C. longicaudata is genetically closer to C. verrucosa than to C. pseudoverrucosa , despite the fact that C. verrucosa and C. longicaudata live in different oceans, while the distribution of C. longicaudata and C. pseudoverrucosa overlaps ( Fig. 14B). This case is similar to the recently discovered, very interesting phylogeographic and taxonomic patterns within the family Trinchesiidae from another superfamily ( Fig. 2), where the NW Pacific Diaphoreolis zvezda Korshunova et al., 2023 is sister to the geographically highly remote North Atlantic Diaphoreolis stipata (Alder and Hancock, 1843) and more distantly related to the geographically closer NW Pacific species Diaphoreolis midori Martynov et al., 2015 (see: Korshunova et al. 2023).

The distribution of C. verrucosa and C. pseudoverrucosa also shows distinct differences ( Fig. 14B). All 111 analysed C. verrucosa specimens were distributed in the North Atlantic and only touched the south-eastern border of the Arctic areas (at the Vaigatch island location on the East and at the Nunavut location on the West; Fig. 14B). Whereas 21 analysed C. pseudoverrucosa specimens were collected from the north-west Pacific (Sea of Japan, Kurile Islands, and Kamchatka), but never in the north-east Pacific. Twenty-three specimens of the third haplogroup ( C. longicaudata ) were collected from both the north-west and north-east Pacific Ocean and their habitats in the north-west partially overlap with C. pseudoverrucosa (the Kurile Islands area and Kamchatka), Fig. 14B.

Thereby, the molecular-phylogenetic analyses, the haplotype network, and phylogeographical results clearly indicate three different haplogroups, related to three different species: C. verrucosa , C. pseudoverrucosa , and C. longicaudata ( Figs 13, 14A, B). Results published in Ekimova et al. (2022) also revealed three haplogroups, named Haplogroups A, B, and C. However, the genetic distances were calculated only between haplogroup A and haplogroup B + haplogroup C ( Ekimova et al. 2022, table 1). Besides, the haplogroup names B and C were mistakenly mixed up in the text and in figure 3A (see Supporting Information, Fig. S1A–C, in the present paper, all errors marked in red). The data for the specimen listed under number MIMB14404, published in Ekimova et al. (2022), needs to be discussed separately here. This specimen according to our analysis belongs to C. longicaudata . In Ekimova et al. (2022), the specimen MIMB14404 belongs to haplogroup A in the text of the paper, but haplogroup B in figure 6, which according to fig. 3A is haplogroup C (Supporting Information, Fig. S1D). Thus, the minimum interspecific uncorrected p -distances in Ekimova et al. (2022) calculated between C. verrucosa specimens erroneously included data for specimen MIMB14404 and combined specimens from haplogroup B and haplogroup C ( C. pseudoverrucosa and C. longicaudata ). Which is why in Ekimova et al. (2022) the distances within ‘ C. verrucosa ’ is 2.3%, within ‘ C. pseudoverrucosa ’ it is 3.2%, and between ‘ C. verrucosa ’, and ‘ C. pseudoverrucosa ’ is 0.9% ( Fig. S1 C). We analysed the COI data for specimen MIMB14404 and concluded that these more variable data may be the result of a technical sequencing error or may appear for some other reason, but this specimen is closer to the C. longicaudata group. We also calculated the genetic distances excluding the molecular data of MIMB14404 and found that these data do not significantly affect the results. Nevertheless, these data could have had a bearing on the credibility of the species’ delimitation results in Ekimova et al. (2022).

Thus the species that are molecularly closer to each other, C. verrucosa and C. longicaudata , inhabit geographically distant regions of the Atlantic/south-east sub-Arctic and Pacific Oceans. Coryphella pseudoverrucosa and C. longicaudata show distinct differences according to their genetic distances, despite overlapping ranges. Three distinct haplogroups in the Coryphella verrucosa complex were discussed in Ekimova et al. (2022), but the genetic distances were analysed only between two groups, with mixed data for C. longicaudata , which led to the erroneous conclusion that C. verrucosa and C. pseudoverrucosa are the same species. In addition, a map showing the distribution of the Coryphella verrucosa complex is absent from that work and the names of the locations are misleading. The localities for C. verrucosa noted as Norway and NW Atlantic are attributed to the Atlantic Ocean and those noted for the White Sea, Barents Sea, Kara Sea, and ‘Canadian Arctic’ are attributed to the Arctic Ocean. In reality, all the ‘Arctic’ waters are not the true arctic waters but are close to boreal, temperate regions. The Barents Sea is in reality the continuation of the boreal, temperate waters of Norway. The Kara Sea (mentioned due to Vaigatch Island) is only the border between the Kara and Barents Seas, and thus still under the influence of the Barents Sea, not the actual Arctic Seas. The localities in the waters of Canadian Nunavut and Labrador are also close to boreal regions. When species are mentioned as ‘trans-Arctic’ it suggests that species should habitat the true cold Arctic seas, such as the Laptev, East Siberian, and Chukchi Seas. The distribution of C. longicaudata in Alaska is so far limited to Fishermen’s Bend, Juneau, which is also close to boreal temperatures and does not apply to all Alaskan waters, and surely does not include true Alaskan Arctic waters. It was previously concluded ( Martynov et al. 2015: 61) that ‘In the Arctic C. verrucosa is credibly absent ( Martynov 2006b), and is not distributed to the East, far than the Barents Sea, which practically excludes gene exchanges between C. verrucosa and Pacific C. pseudoverrucosa .’ Since that time there are no data to disprove that conclusion.

Therefore, describing the habitat of C. verrucosa as ‘Arctic’ in Ekimova et al. (2022) is not the real Arctic. The distribution map of C. verrucosa , C. pseudoverrucosa , and C. longicaudata (mixed together under the single name C. verrucosa ) is provided in Ekimova et al. (2024), fig. 2l. However, this figure is misleading: distributions of these species are presented to be more ‘Arctic’ than they really are and some habitats are incorrectly marked in red. Fishermen’s Bend (Alaska) is not very far from definitely non-Arctic British Columbia, but both the Bering Sea and Bering Strait are marked in red. Vaigatch Island is located between the Kara and Barents Seas, but the entire Kara Sea is also marked in red. Franz Josef Land is incorrectly marked in red. The localities Disko Island ( Greenland) and Durban Harbour (Nunavut) are both marked as Canadian Arctic. Although Nichols (1900) and Lemche (1929) listed ‘ Coryphella rufibranchialis ’ from the Bering Sea, these were lists without any detailed analysis of any particular material and as such may represent other genera of Coryphellidae or completely unrelated Aeolidacea , and cannot be relied upon, therefore, with any credibility, and cannot be used for any phylogeographic analysis. In a later book based on Lemche’s observations ( Just and Edmunds 1985) there is no mention of the Bering Sea for Coryphella verrucosa . Using only morphological data, we previously concluded that the North Pacific true Coryphella , including the more southern parts of the Bering Sea, may represent separate species, which at the level of lists would have been presented as Coryphella ‘ verrucosa ’ ( Martynov 2006b) at that time. Therefore, it is completely unsubstantiated to consider the distribution of Coryphella ‘ verrucosa / rufibranchialis ’ to be ‘pan-Arctic’ using species’ lists from well before the molecular era.

Morphological descriptions in Ekimova et al. (2022) for C. verrucosa also contain much inaccurate data. One example is the radulae of C. verrucosa , which are given in figure 6A–H [here and in the remainder of this paragraph figures refer to figures in Ekimova et al. (2022)]. Particularly, for the specimens whose radulae are illustrated in fig. 6A, B there are no molecular data. Therefore, the radula of C. verrucosa is reliably presented only in fig. 6C. Further, fig. 6D–H is a mixture of C. pseudoverrucosa and C. longicaudata (for details see Supporting Information, Fig. S1D). In the present study, a Phyloperiodic Table was built ( Fig. 16, present study) for the comparison of radular patterns of C. verrucosa , C. pseudoverrucosa , and C. longicaudata . On the basis of the molecular phylogenetic tree for three species of the genus Coryphella (= three groups), periods were built for each of the groups, depending on the dimension between the central cusp of the central teeth of the radula and the border of the lateral denticles (marked in yellow on Fig. 16, present study). In total four periods (<30 μm, 30–50 μm, 50–100 μm,>100 μm) were built. All available data for each of the three groups were given at the same scale and arranged into these four periods. Data from Ekimova et al. (2022: fig. 6С–H) were accurately redrawn with the original scale and placed into the periods in relation to the dimension between cusp and lateral denticles ( Fig. 16, present study). After filling in the Table, it was revealed that all radulae (fig. 6С–H), which have been presented as comparative material in Ekimova et al. (2022), belong to juvenile and subadult specimens and belong in I and II periods. It is well known that in juveniles of different molluscan taxa, morphological differences are significantly weaker compared to adults ( Martynov and Korshunova 2015). For illustration of the morphological characters of the radula in Ekimova et al. (2022), the radulae of adult specimens were not given; however, having a huge sequencebased dataset doubtless includes adult specimens. Using only juvenile or subadult radulae produced an incorrect impression about the putative absence of morphological differences in ‘ C. verrucosa ’ and, therefore, led to false conclusions about the true number of species Ekimova et al. (2022).

Applying the Phyloperiodic Table, the following morphological patterns have been revealed: adult C. verrucosa possess a shorter cusp of the central teeth with longer lateral teeth, adult C. pseudoverrucosa possess a longer, massive cusp with shorter lateral teeth, and adult C. longicaudata possess a longer and wider cusp with longer lateral teeth ( Fig. 16, present study). More fine-scale differences along this spectrum are expected, but general trends in the adult radular patterns within these three species are evident. These fine-scale characters further support a multilevel organismal diversity system where application of both morphological and molecular data has recently been employed on a practical basis in other groups of nudibranchs ( Korshunova and Martynov 2024).

The genus Coryphella forms a distinct clade according to the molecular phylogenetic analysis ( Figs 1, 13). Merging all the taxa of the family Coryphellidae into a single genus ‘ Coryphella ’ would completely dismiss the extensive, considerable morphological diversity of the large-scale to the fine-scale differences ( Korshunova et al. 2017a, present study) within Coryphellidae , disregard the obvious extensive molecular phylogenetic diversity ( Figs 13, 14), and destroy taxonomy, per se, since all the significant and indisputable differences of the entire Coryphellidae family (see below for a detailed synopsis) would need to be presented under the single pan-lumping name ‘ Coryphella ’. Such an action would represent nothing more than the formation of a completely non-diagnosable pseudo-unit with a collection of almost all possible character states, overlapping not only between more related families such Paracoryphellidae and Flabellinidae , but also between very distantly related families such as Samlidae and Facelinidae (see Synopsis above and below of all Aeolidacean families, Tables 1–4, and molecular phylogenetic tree in Figs 1, 2). Furthermore, there is even finer-scale epigenetic diversity within the proper genus Coryphella , since a taxon/species C. rufibranchialis apparently does not show significant genetic differences from C. verrucosa ( Eriksson et al. 2006) , but readily differs from it morphologically by the presence of markedly longer cerata, to such degree that it may be treated as a separate species ( Picton and Morrow 2023, 2024). All these large-scale and fine-scale differences are undoubtedly the results of the evolutionary process at different levels, but they are all clearly ignored when taxonomically the name ‘ Coryphella ’ is incorrectly applied to cover all the significant morphological and molecular diversity of the family Coryphellidae ( Fig. 13). Comparison of the genus Coryphella with all the valid, currently included Coryphellidae genera is presented in Table 5. See also detailed considerations in Results and Discussion ( Figs 1, 2, 13).