Geometric morphometric analyses define riverine and lacustrine species flocks of Himalayan snowtrout (Cyprinidae: Schizothorax) in Nepal

Abstract

Freshwater fishes in the river and lake systems in the Himalayas and Tibetan Plateau are morphologically diverged but the evolutionary relationship of putative subspecies separated in these freshwater systems has not been explored. Snowtrout (Schizothorax spp.) are minnows (Cyprinidae) broadly distributed in Asia. Body shapes of 3 Lake Rara (northwest Nepal) endemics (S. macrophthalmus, S. nepalensis, S. raraensis) and 2 widely distributed riverine species (S. progastus, S. richardsonii) across 3 drainages in Nepal (i.e. Karnali, Gandaki, and Koshi Rivers) were studied using geometric morphometry. Data were derived from museum voucher specimens/tissues collected in 1984-1986 and 1996 (Lake Rara). Cartesian coordinates of 18 anatomical points (Type I landmarks) from 528 individuals were digitized; shape variation was then quantified with principal component analysis and visualized with thin-plate splines derived from a Procrustes analysis. Models of shape variation (i.e. taxonomy versus geography) were tested with a multivariate analysis of variance and a morphological distance matrix. Phylogeographic relationships were examined with a haplotype network (N = 115) derived from 1140 base pairs of the mitochondrial DNA cytochrome b gene, and selected GenBank sequences (N = 5). Koshi River snowtrout diverged morphologically from conspecifics, consistent with the phylogeographic data. In contrast, Gandaki and Karnali River snowtrout grouped by morphotype (up- versus downstream) irrespective of geographic origin, yet clustered separately within the haplotype network. Lake Rara snowtrout were morphologically but not genetically distinct, due to incomplete lineage sorting. Morphological and genetic variability in Schizothorax from Nepal represent a mosaic driven by isolation (= vicariance) and specialization (= adaptation), with taxonomy insufficiently reflecting diversity. Additional data are required to appropriately derive management and effective conservation plans.

Parallel introgression, not recurrent emergence, explains apparent elevational ecotypes of polyploid Himalayan snowtrout

“…We first examined patterns of morphological convergence and divergence using geometric morphometric data [51] derived from 528 images captured from museum specimens representing all five Nepali species (for full sample metadata and KU voucher numbers, see Regmi et al [51]; for those used for ddRAD, see table S2). Our analyses were restricted to the Nepali sub-tree in that voucher specimens for Bhutanese samples were not available.…”

“…Recent phylogenetic analyses based on mtDNA also failed to support the nominal species, with relationships occurring instead at the drainage level [43]. Despite this, morphologies are consistent across drainages [38], suggesting the independent emergence of 'blunt-nosed' phenotypes (i.e., those associated with S. richardsonii) within each highland drainage.…”

“…Yet, patterns of convergence in previous studies are also broadly apparent [36,37]. Additionally, the rapid formation of multiple independent 'species flocks' has occurred within unique and isolated alpine lake habitat [38][39][40][41][42], with an emerging consensus indicating recurrent ecotypic formation along riverine gradients [38,43]. Here, two Himalayan species are of particular interest: Schizothorax progastus and S. richardsonii.…”

“…Coordinates were then superimposed across samples using Generalized Procrustes Analysis implemented in GEOMORPH [68]. Full details on geometric morphometric data processing, including bias and sensitivity analysis, can be found in Regmi et al [38].…”

“…Recent phylogenetic analyses based on mtDNA failed to support the nominal species, with relationships occurring instead at the drainage level [56]. Despite this, morphologies are consistent across drainages [51], suggesting an independent emergence of ‘blunt-nosed’ phenotypes (i.e., those associated with S. richardsonii ) within each highland drainage. Morphologically convergent ecotypes are also replicated within elevational gradients of Bhutan [61], again suggesting the formation of parallel ecotypes.…”