Ambigolimax nyctelius

Vendetti, Jann E., Burnett, Emily, Carlton, Lidia, Curran, Anne T., Lee, Cedric, Matsumoto, Ron, Donnell, Rory Mc, Reich, Inga & Willadsen, Ole, 2019, The introduced terrestrial slugs Ambigolimax nyctelius (Bourguignatı 1861) and Ambigolimax valentianus (Férussacı 1821) (Gastropoda: Limacidae) in Californiaı with a discussion of taxonomyı systematicsı and discovery by citizen science, Journal of Natural History 53 (25), pp. 1607-1632 : 1622-1623

Systematics of A. nyctelius and A. valentianus and selected limacids

Ambigolimax nyctelius and A. valentianus collected from Los Angeles County could not be reliably differentiated by mantle patterning or the morphology of the radula or jawı a finding that supports the conclusion of authors including Mc Donnell et al. (2009) and Rowsonı Turnerı et al. (2014b). These taxa could be differentiated by distal genital morphology and COI barcode sequences ( Nitz et al. 2010; Rowson et al. 2014a). Such sequence data are not without limitations ( Davison et al. 2009; Sauer and Hausdorf 2012)ı but are extremely useful for specimen identificationı especially of phenotypically similar taxa such as A. nycteliusı A. valentianus and L. marginata . Notablyı COI barcode analysis would also allow for distinguishing between these and other taxa non-lethally ( Morinha et al. 2014) or when distal male genitalia are under-developedı as in juvenile slugs.

Sequence analysis herein revealed distinct COI haplotypes for A. nyctelius and A. valentianus that were not shared between species. Ambigolimax nyctelius haplotypes were more divergent than those of A. valentianus ı corroborating findings of Rowson et al. (2014a) who examined specimens collected in Europe. An 8.6% mean K2P interspecific COI divergence between the two Ambigolimax species is relatively low compared to other slug species pairs ( Rowson et al. 2014a) and is the lowest between any two limacid taxa examined in this study ( Table 2 ). Howeverı Nitz et al. (2010) found a comparatively low K2P COI distance of 10.8% between Limax species.

Both maximum likelihood and Bayesian inference analyses of COI ( Figure 6 ) show all eight speciesı represented by sequence data from at least two specimensı as monophyletic. Notablyı in all analyses A. nyctelius and A. valentianus are resolved as sister taxa. However and importantlyı taxon sampling is too poor to conclude that A. nyctelius and A. valentianus diverged most recently from a common ancestor. There are 15 additional described species in the genus Lehmannia ( Welter-Schultes 2012) and three in Malacolimax ı all of which lack sequence data in BOLD and GenBank databases. No studyı including this oneı has examined sequence data of COI or any other molecular marker for species in other limacid generaı including six in Gigantomilax ı seven in Turcomilax ı three in Eumilax ı two in Metalimax ı and one each in Svanetia and Caspilimax (MolluscaBase 2018) . This lack of data makes sequence-based inference of phylogenetic relationships within and between most genera within the Limacidae (e.g. Ambigolimaxı Lehmanniaı Malacolimaxı Bielzia )ı and any conclusions about trait evolutionı presently untenable.

Finallyı in a literature search of the species name ‘ valentiana ’ or ‘ valentianus ’ referring to a limacid slugı a sample of 40 publications from 1999 to 2018 included 33 unique first authors who used the genus name Ambigolimax nine timesı Lehmannia 23 timesı and Limax eight times. This generic lability reveals the need for a revision of what we refer to here as Ambigolimax spp. For exampleı our results and those reported elsewhere ( Klee et al. 2005; Rowson et al. 2014a) indicate that the genus Lehmannia is not monophyletic when its name is applied to A. nyctelius and/or A. valentianus . Additionallyı Pollonera (1887) proposed Ambigolimax for species that have a globular phallus appendixı as in A. valentianus ı but not in A. nyctelius . It isı thereforeı problematic that the genus name Ambigolimax is applied here and elsewhere to A. nyctelius ı especially as A. valentianus is the type species of the genus. Consequentlyı and because of the poor taxon sampling within Lehmanniaı Malacolimax ı and most other limacid genera (excluding Limax )ı we recommend a revision of the genus Ambigolimax after a thorough examination of molecular and morphological data from a diversity of limacid species. A resulting taxonomy could name a new genus for A. nyctelius ı reject Ambigolimax altogether in favour of a new or different genusı or redescribe Ambigolimax without reference to the morphology of the phallus appendix.

Figures

Figure 6. Phylogram based on a SATé FastTree analysis of aligned, partial COI mtDNA sequences from selected limacid gastropod species. Specimens are represented by their GenBank accession number with the following indications: bold = COI sequence generated by this study, boxed = first sequence of A. nyctelius from a specimen collected in California, underlined = collected in New York City, asterisk = radular morphology shown in Figure 4, diamond = external phenotype shown in Figure 1, square = jaw morphology shown in Figure 2, a star within a circle = distal male genital morphology shown in Figure 5. RAxML bootstrap support values are shown above and next to nodes, and MrBayes posterior probabilities are below. Where support values are absent indicates a node that was not resolved by all analyses or a node with very low support values.