Phyllidiella cf. pustulosa (Cuvier, 1804)

Papu, Adelfia, Bogdanov, AleXander, Bara, Robert, Kehraus, Stefan, König, Gabriele M., YonoW, Nathalie & Wägele, Heike, 2022, Phyllidiidae (Nudibranchia, Heterobranchia, Gastropoda): an integrative taxonomic approach including chemical analyses, Organisms Diversity & Evolution (New York, N. Y.) 22 (3), pp. 585-629 : 607-609

Phyllidiella cf. pustulosa

Twenty specimens group together in a separate clade (bootstrap value 98) that is sister taxon to P. pustulosa ( Cuvier, 1804) , and this relationship is supported by a bootstrap value of 100 ( Figs. 13 , 14 , S 1 ). Specimens of Phyllidiella cf. pustulosa (Fig. 9.2a–d) are characterised by a black mantle covered in short, conical, compound, pink tubercles which have a somewhat broader base than the tubercles in all other specimens assigned to P. pustulosa . This also renders the specimens paler because the black lines are narrower. These are arranged in longitudinal and diagonal lines, forming an X-shaped pattern in the first cluster on the medial dorsum. A narrow pink margin is present. The rhinophores are black with pale white–pink bases, their openings lying at the transition of a black line and a posterior pink tubercular field, and there are no rhinophoral sheaths. The anal opening lies in the connection of the two longitudinal black lines. Specimens of this clade have white or pale pink oral tentacles and foot sole. The colour pattern somewhat resembles the pattern of Phyllidiella lizae Brunckhorst, 1993 , but analysis of the holotype and the paratypes of P. lizae (Australian Museum: C159493, C162784, C159492, Fig. 2g-i) allows the clear differentiation of our material from this species. Examination of the type material of P. lizae results in the following external characters: a white to pale pink background of the notum with a narrow cross-like black pattern, the margin is white, and the foot and oral tentacles are also pale pink. Phyllidiella lizae is also characterised by low pink rhinophoral sheaths that are absent in specimens of P. cf. pustulosa . It also appears that the tubercular field between the rhinophores is continuous in P. lizae but not in P. cf. pustulosa . One of the available P. lizae sequences (AF430365) retrieved from GenBank and identified by Valdés (2003) is clearly misidentified and clusters with P. rudmani (see below); unfortunately, there is no description or photograph of this specimen available (Á. Valdés pers. comm.). Our P. cf. pustulosa seems to match with one syntype and one paralectotype of P. nobilis (NHMD 91773, Fig. 2a; NHMD-633658a, Fig. 2d).

The species delimitation tests of the concatenated data set as well as of CO1 clearly distinguish between Phyllidiella pustulosa and Phyllidiella cf. pustulosa , the specimens of the latter united with a support value of 98. However, the ABGD test using only the 16S data set considers these two clades as one species (bootstrap value 100). The maximum distance between P. pustulosa and P. cf. pustulosa using the

◂ Fig. 10 Phyllidiella species and specimens with identifiers. Scale bars: 10 mm. 1a–f Phyllidiella zeylanica auctt.: a Phan16Bu1; b Phli18Bu1; c Phze18Bu1; d Phpu15Bu41; e Phpu18Bl4; f Phpu15Bu25. 2a, b Phyllidiella rudmani : a; Phpi15Bu2; b Phpi15Bu5. 3a–d Phyllidiella sp. b: a Phpu16Sa25; b Phpu16Sa39; c Phpu16Sa27; d Phpu18Bu5. 4a–d Phyllidiella sp. c subclade 1: a Phli16Sa8; b Phpu15Bu35; c Phli18Ba3; d Phpu18Po4

CO1 data set is 9%; this genetic distance value is lower in comparison to other interspecific genetic distances of Phyllidiella species (lowest minimum values between P. pustulosa and P. nigra is 9.84%, but in most cases higher than 10%, see Table S6), thus not providing adequate evidence for separate entities. Nevertheless, we consider P. cf. pustulosa as a separate subspecies because of the consistent external colour pattern, despite its broad distribution around Sulawesi. Two chemically analysed specimens (Phpu16Sa76, Phpu16Sa79) each represent a different branch of P. cf. pustulosa . LCMS analysis shows very different extract compositions in these two specimens (Fig. S9d), but both are dominated by sesquiterpene isonitriles and their derivatives, which nevertheless differ from those of true P. pustulosa .

Using the haplotype network analyses of CO1 and plotting chemical results on this network ( Fig. 18 ) highlight discrepancies in the results between the various methods. In the network analysis, P. cf. pustulosa cannot be considered as an independent species from P. pustulosa because several specimens that form a subclade of P. pustulosa in the tree cluster with P. cf. pustulosa . This subclade is still unique as it shares at least 18 nucleotide states, thus distinguishing it from P. cf. pustulosa . However, it does not share unique nucleotide states only with P. pustulosa . Metabolomic analysis distinguishes P. cf. pustulosa from true P. pustulosa but also provides evidence that a particular subclade (Phpu15Bu8 and Phpu16Sa52) of P. cf. pustulosa is more similar to the P. pustulosa clade ( Fig. 18 ). All these data indicate an ongoing process of speciation within Phyllidiella pustulosa and the lack of clear species boundaries at this moment in time.

Figures

Fig.13 Maximum likelihood tree of Phyllidiidae based on the concatenated data set of partial CO1 and 16S sequences. Terminal taxa on species level collapsed. Numbers at nodes are bootstrap values (1000 replicates). The genera are marked with the same colours as in Fig. S1. Numbers behind the species names indicate results of the

Fig. 14 Maximum likelihood tree of Phyllidiidae based on partial CO1 sequences. Terminal taxa on species level collapsed. Numbers at nodes are bootstrap values. The genera are marked with the same colours as in Fig. S1

Fig. 1 Sampling locations in Sulawesi, Indonesia

Fig. 18 Haplotype network of Phyllidiella pustulosa (74 sequences, white circles) and P. cf. pustulosa (20 sequences, grey circles) based on the CO1 gene, with the four described chemotypes (coloured cir-