Psilorhynchus, , Ramaswami, 1952

Conway, Kevin W., 2011, Osteology of the South Asian Genus Psilorhynchus McClelland, 1839 (Teleostei: Ostariophysi: Psilorhynchidae), with investigation of its phylogenetic relationships within the order Cypriniformes, Zoological Journal of the Linnean Society 163 (5), pp. 50-154 : 138-139

PSILORHYNCHUS

Previous authors investigating the morphology of Psilorhynchus have largely refrained from commenting on the interrelationships of its species, which are relatively unknown. In both parsimony analyses, P. sucatio , a widespread species occurring throughout the Ganges–Brahmaputra drainage, was recovered as the sister group to all other members of the genus included for analysis ( Fig. 44A ). The monophyly of the sister group clade to P. sucatio (labelled clade 1 in Fig. 44A ) is supported by three derived characters: (1) the presence of a postepiphysial fontanelle (character 15, 0> 1); (2) the temporal commisure situated dorsal to the supraoccipital (character 19, 0> 10); and (3) the absence of basal radii along the anterior edge of scales (character 125, 0> 1). None of these characters represent unique support for this subgrouping of Psilorhynchus , and all are widespread amongst cypriniform fishes (see character descriptions; in addition see Kullander, 2008, for a discussion of cranial fontanelle in the Cyprinidae and Cavender & Coburn, 1992, for discussions of scale characteristics in the Cyprinidae ). Psilorhynchus pseudecheneis , an elongate species endemic to Nepal ( Shrestha, 1981), was recovered as the most basal member of clade 1, forming the sister group to all remaining species. The monophyly of the sister-group clade to P. pseudecheneis (clade 2 in Fig. 44A ) is supported by a single derived character, the presence of a large keel along the ventral surface of the parasphenoid (character 11, 0> 1). Relationships of clade 2 were poorly resolved, uniting P. tenura , the only species of the genus known to occur in Peninsular India ( Arunachalam & Muralidharan, 2008), in a polytomy with two smaller Psilorhynchus clades (clades 3 and 4 in Fig. 44A ).

The first of these smaller Psilorhynchus clades (clade 3) includes P. gracilis , P. robustus , and P. melissa , and is supported by three characters, one of which represents a reversal to the plesiomorphic condition, including: (1) presence of the mandibular sensory canal (character 58, 1> 0); (2) presence of a dorsally directed process on the dorsal surface of pectoral radial 3 (character 116, 0> 1); and (3) the presence of a hiatus between the tip of hypural 1 and the compound centrum (character 121, 0> 1). Conway & Kottelat (2010) hypothesized a close relationship between the members of this clade based on the presence of the hiatus between the tip of hypural 1 and the compound centrum, which is unique to these taxa amongst species of Psilorhynchus and rare amongst other cypriniforms (see character description).

The second of these smaller clades (clade 4) includes P. balitora , P. breviminor , P. nepalensis , P. rahmani , P. pavimentatus , and P. brachyrhynchus , and is supported by a single character, reduction of the anteriormost branchiostegal ray (character 89, 0> 1), a derived character that is also exhibited by P. sucatio . Conway & Kottelat (2010) hypothesized a close relationship between P. sucatio , P. balitora , P. breviminor , P. pavimentatus , and P. rahmani based on the reduction of the anteriormost brachiostegal in these five species. In the present analysis the distribution of this character is most parsimoniously explained as the result of homoplasy. Forcing monophyly upon these species results in a less parsimonious cladogram (two steps longer), and as such this hypothesis of close relationship is not supported by the parsimony analysis of the morphological data set used herein. Interestingly, Bayesian analysis of the data set recovered all Psilorhynchus species in possession of a reduced anteriormost branchiostegal ray as a monophyletic group ( Fig. 44C ). This latter grouping is, however, weakly supported in terms of posterior probability (pp 0.52).

It is interesting to note that two of the Psilorhynchus clades recovered in the first parsimony analysis (clades 3 and 4) include taxa from drainages on both sides of the Indoburman Range. Previous molecular phylogenetic investigations of Indoburman fishes have recovered similar biogeographic patterns ( Rüber et al., 2004; Britz et al., 2009). In a molecular phylogenetic analysis of the Badidae, Rüber et al. (2004) found geographic splits within two clades of badid genera ( Badis and Dario ) inhabiting drainages on opposite sides of the Indoburman range: the Ganga– Brahmaputra drainages (North Eastern India) and the upper Irrawaddy drainages (northern Myanmar and adjacent China). Using a calibration derived from mitochondrial cytochrome b sequence divergence, Rüber et al. (2004) dated the split between the Ganga–Brahmaputra clades and Irrawaddy clades of Badis and Dario at around 19–23 Ma. This split was consistent with geological evidence for the separation of these drainages being caused by the tectonic uplifts of eastern Tibet ( Clark et al., 2004). Rüber et al. (2004) concluded that early cladogenesis within the Badidae was probably prompted by a vicariant event that separated the Ganga–Brahmaputra and Irrawaddy drainages, and speculated that the Miocene uplift of the Indoburman range might also have played an important role in shaping the current distribution patterns of badid fishes. As the relationships of clades 3 and 4 remain unresolved, it would be premature to speculate at this time whether the same vicariant events have played an important role in the evolutionary history of Psilorhynchus . Resolving the relationships of these terminal clades of Psilorhynchus may provide further insight into the biogeographic history of Indo-Burma.

Based mostly on osteological characters and a small number of external morphological characters, Yazdani et al. (1991) suggested a close relationship between P. homaloptera and P. pseudecheneis , for which they erected a new genus, Psilorhynchoides (type species P. homaloptera ). Conway & Kottelat (2007) showed that the characters used by Yazdani et al. (1991) to diagnose Psilorhynchoides had a mosaic distribution within Psilorhynchus s.l., and returned P. homaloptera and P. pseudecheneis back to Psilorhynchus , pending the results of the present phylogenetic investigation. Unfortunately, osteological material of P. homaloptera was not available for examination, and it was only possible to include this species in the phylogenetic analysis by scoring characters based on the osteological illustrations provided by Chen (1981) and Nebeshwar et al. (2007). Using this approach only 54 of the 127 characters used in the phylogenetic analysis could be scored. The inclusion of taxa with large quantities of missing data in a phylogenetic analysis frequently results in an increase in the number of equally parsimonious cladograms, and a decrease in the resolution of the consensus resulting from such cladograms ( Kitching et al., 1998; Wiens, 2003a, b). Although including taxa with large quantities of missing data in phylogenetic analyses can be illuminating (for a recent example, see Egge & Simons, 2009), in this particular case, little new insight could be gained on the phylogenetic placement of P. homaloptera , the inclusion of which resulted in a significant increase in the number of equally parsimonious cladograms (from 14 to 98), and a decrease in the resolution between the members of the clade to which it was recovered ( Fig. 44B ). It should be noted, however, that despite the large volume of missing data, P. homaloptera is recovered inside Psilorhynchus , and not in a sister-group relationship to all other species included for analysis. Although Psilorhynchoides , if recognized as valid, would render Psilorhynchus paraphyletic in the phylogenetic hypothesis presented in Figures 43 and 44 , definitive conclusions regarding the validity of this generic name should be reserved until the relations of P. homaloptera with the other species of Psilorhynchus are more satisfactorily resolved. Until then the generic name Psilorhynchus is preferred for this species (following Conway & Kottelat, 2007).

Figures

Figure 44. Intrarelationships of Psilorhynchus. A, Psilorhynchus clade from the strict consensus of the 14 equally parsimonious cladograms obtained from the first parsimony analysis (analysis 1; clade highlighted in grey in Fig. 43). Numbers above branches indicate clades discussed in text. B, the Psilorhynchus clade obtained in the strict consensus of 98 equally parsimonious cladograms obtained from the second parsimony analysis (analysis 2) with the inclusion of Psilorhynchus homaloptera. C, the Psilorhynchus grouping obtained in the Bayesian analysis. Numbers below branches indicate Bremer/bootstrap support for particular clades in A and B, and Bayesian posterior probabilities for nodes (derived from 9500 Markov chain Monte Carlo sampled trees) in C.