Paracanthobdella livanowi

NOTES ON OUR TAXONOMIC SCHEME Here, we present a taxonomic scheme for hookfaced fish worms (Acanthobdellida). Despite its historical significance, we do not retain the family Paracanthobdellidae for P. livanowi, because both morphological and molecular analyses suggest that the two acanthobdellidan species share many characteristics. Initially, Acanthobdellidae and Paracanthobdellidae were differentiated based on a few morphological differences that were viewed as taxonomically important (Epstein, 1987). Specifically, members of Paracanthobdellidae were said to possess a primitive prostomium and well-developed anterior sucker. Although the two-family system seemed well supported at the time, we have found that the prostomium of P. livanowi is not more conspicuous or developed than the anterior region of A. peledina. Moreover, the area between rows of chaetae in the cephalic extremity of the latter can be deepened such that it resembles a shallow sucker. It should also be noted that juvenile individuals of P. livanowi do not bear well-developed anterior suckers; this character is only common to large-bodied individuals of the species. The sum of this evidence suggests that the morphological differences between the two acanthobdellidan species, although pronounced, are not sufficient to warrant two families. We therefore classify both Acanthobdella and Paracanthobdella under the single family Acanthobdellidae. We establish a new suborder Acanthobdelliformes to match better the taxonomy erected for Hirudinea in prior work (Tessler et al., 2018 a), which divided Hirudinida (leeches) into five suborders (Americobdelliformes, Erpobdelliformes, Hirudiniformes, Glossiphoniiformes and Oceanobdelliformes). Acanthobdelliformes is defined by the presence of chaetae on each of five contiguous segments in the anterior body region, and 31 segments (mid-body ones are quadrannulate with annulus a 3 being subdivided) (Sawyer, 1986; Purschke et al., 1993; Bielecki et al., 2014). Although previous studies have reported different numbers of segments for each acanthobdellidan species (e. g. 29 for Acanthobdella and 30 for Paracanthobdella; Bielecki et al., 2014), this discrepancy is attributable to presumed differences in the number of segments that comprise the posterior sucker, which should be substantiated by a careful morphological analysis in the future. The higher taxonomy (class through order) follows our prior classification scheme (Tessler et al., 2018 a). Others have constructed alternative schemes, but we feel ours to be internally more consistent and phylogenetically appropriate. Alternative classifications include subclass Acanthobdellidea Livanow, 1905 (Archihirudinea Lukin, 1956 is an equivalent synonym) and subclass Acanthobdelliones (Epstein, 1987). It is important to note that the two acanthobdellidan species were originally in the same genus (Epstein, 1966). However, in the late 1980 s, P. livanowi was given its own genus and even family (Epstein, 1987). Ultimately, the decision to classify a monophyletic lineage of two species into one vs. two genera and / or families is subjective. We have decided to retain the genus-level classification proposed by Dr Epstein (rather than lump them) to honour his contributions to the study of Hirudinea and accentuate the differentiation of the anterior sucker, chaetae and internal anatomy that separates these species. This is also in concordance with the suggestions in the most recent, broad morphological comparison paper (Bielecki et al., 2014); see Table 1 for some of the differences between these species.