PTERYGOPLICHTHINI, Armbruster, 2004

Armbruster, Jonathan W., 2004, Phylogenetic relationships of the suckermouth armoured catfishes (Loricariidae) with emphasis on the Hypostominae and the Ancistrinae, Zoological Journal of the Linnean Society 141 (1), pp. 1-80 : 53-54

PTERYGOPLICHTHINI

Several problems are evident in Weber’s (1991, 1992) treatment of the species of Pterygoplichthys s.l. The first is that the phylogeny given is not the single shortest tree. Given the data, a second most parsimonious tree could be constructed with the positions of Pterygoplichthys and Glyptoperichthys switched, and the tree should have more properly included a trichotomy of Pterygoplichthys , Glyptoperichthys , and Megalancistrus . Second, a close relationship of Pterygoplichthys and Megalancistrus would suggest that the distinct alteration to the opercle in the Ancistrinae (Schaefer, 1986, 1987) had to evolve twice (see character 75: 1; Fig. 19C, D ). Third, none of the genera are uniquely diagnosed. Glyptoperichthys with the inclusion of G. punctatus Natterer has no unique synapomorphies, although a particular pattern of plates between the dorsal-fin spine and the head putatively diagnoses both Glyptoperichthys and Liposarcus ( Weber, 1992) . The pattern is also present in many other loricariids and seems of little phylogenetic usefulness (pers. observ.). The other characteristic that is listed as synapomorphic for Liposarcus , thin anterolateral processes of the basipterygium, is polymorphic within the genus (pers. observ.). Finally, it appears that Weber lumps the species of his Pterygoplichthys together based on the putatively plesiomorphic condition of the postdorsal plates mentioned above. Two of the species retained in Pterygoplichthys ( P. undecimalis (Steindachner) and P. zuliaensis Weber ) occur to the west of the Andes in the Río Magdalena and Lago Maracaibo basins, respectively, and the other species ( P. etentaculatus ) occurs on the opposite side of South America in the Rio São Francisco. Although not impossible, the range suggested by Weber (1991, 1992) is unlikely.

The monophyly of Glyptoperichthys + Liposarcus + Pterygoplichthys ( Fig. 37 ) is supported by several characteristics: a diminutive or absent interhyal located posteriorly (26: 0, 27: 2), 8-11 postdorsal vertebrae (121: 2), and eight or more dorsal-fin rays (142: 0). The decay index is fairly high (DI = 4). Of the three genera recognized by Weber, only Liposarcus is monophyletic in this analysis, and support is weak (DI = 1, Fig. 36 ). If monophyly is forced upon both Glyptoperichthys and Pterygoplichthys sensu Weber (1991, 1992 ), there are no characteristics that would be synapomorphic for either genus. Thus, in order to retain Liposarcus and Glyptoperichthys as valid taxa, a new genus would have to be described for P. zuliaensis + P. undecimalis (these two taxa are most likely to be sisters; Weber, 1992), and a second would have to be described for G. punctatus . Given that the species of Pterygoplichthys , Glyptoperichthys , and Liposarcus form a well-diagnosed clade and the fact that they are all readily identifiable from most other species of the Loricariidae by their high number of dorsal-fin rays (ten or more, rarely 9 vs. 7), it is more practical to refer the species to a single genus ( Pterygoplichthys ) and to place Liposarcus and Glyptoperichthys in the synonymy of Pterygoplichthys .

Weber (1992) contends that Megalancistrus is most closely related to some Pterygoplichthys ( Fig. 5 ). In this study, based on specimens examined, the type species M. gigas (Boulenger) shares numerous synapomorphies with Acanthicus (most notably an enlarged swimbladder capsule), does not possess a modified stomach, and is clearly not related to Pterygoplichthys . At MNRJ I examined some uncatalogued specimens from the Rio São Francisco basin identified as M. barrae and which conform to the original description of the species by Steindachner (1910). These are clearly different from, and almost certainly not related to, M. gigas , although they are very similar to Pterygoplichthys . Because I have not examined the type of M. barrae or examined the stomachs of the MNRJ specimens, I defer transferring M. barrae to Pterygoplichthys .

Based on a unique modification of the stomach (213: 1), I have already suggested ( Armbruster, 1998c) that Pterygoplichthys , Glyptoperichthys , and Liposarcus represent a monophyletic group sister to the Hemiancistrus annectens group, which consists of several species formerly placed in Hemiancistrus (represented in this analysis by H. holostictus , H. maracaiboensis , and H. panamensis ). The results of this study support the conclusion of the earlier one, and the Pterygoplichthini is supported both by the modified stomach and the presence of 2–3 plates between the suprapreopercle and the exposed opercle (81: 2).

The support for the H. annectens group is provided by three characters, all of which are also found in some Hypostomus : an invagination in the fifth ceratobranchial (11: 1), presence of an interoperculo-mandibular ligament (74: 0), and the anterior process of the pterotic-supracleithrum extending at least midway through the eye (112: 1). However, the H. annectens group does not form a monophyletic group with Hypostomus in the phylogeny because of the shared presence of a modified stomach with Pterygoplichthys and the presence of evertible cheek plates. The connective tissue sheet found in the Pterygoplichthini is a complex characteristic found in no other loricariids. Despite the fact that many Hypostomus live in waters at least as hypoxic as those where Pterygoplichthys and the H. annectens group occur, they never develop a connective tissue sheet. Because there is support for the monophyly of the H. annectens group and because it is not closely related to Hypostomus in the phylogenetic analyses, a new genus for the H. annectens group should be described. This genus will be described in a future manuscript detailing the Pterygoplichthini .

Figures

Figure 5. Phylogeny of Pterygoplichthys and related genera from Weber (1992) based on osteology and external features.

Figure 19. Opercle, right side, mesial view. A, Hemiancistrus maracaiboensis, EBRG 2855. B, Hypostomus taphorni (Lilyestrom), ANSP 16195. C, Peckoltia sp., FMNH 70863. D, Ancistrus pirareta Müller, UMMZ 206085. Scale bars = 1 mm.

Figure 36. Relationships of Astroblepus, Lithogenes, hypostomine tribes, and the nonhypostomine loricariids based on the ordered analysis. This is part of the strict consensus of 5098 most parsimonious trees of trees of 1328 steps, CI = 0.203, remainder of the consensus tree is in Figs 37 and 38. Numbers above the branches are clade numbers, numbers below are decay indices.

Figure 37. Relationships of the hypostomine tribes and taxa within the Corymbophanini, the Rhinelepini, the Hypostomini, and the Pterygoplichthini based on the ordered analysis. This is part of the strict consensus of 5098 most parsimonious trees of trees of 1328 steps, CI = 0.203, remainder of the strict consensus tree is in Figs 36 and 38. Numbers above the branches are clade numbers, numbers below are decay indices. 1Aphanotorulus, Cochliodon, and Isorineloricaria are placed in Hypostomus. 2The Hemiancistrus annectens group represents an undescribed genus, Hypostomus panamensis is now placed in Hemiancistrus. 3Glyptoperichthys and Liposarcus are placed in Pterygoplichthys.