Comparative analysis of multivariate morphology of four species of Schizothoracinae in the Yarkand River basin of Xinjiang

The Yarkand River is located on the western edge of the Tarim Basin in southwestern Xinjiang, and the longest source of the Tarim River. Affected by the uplift of the Qinghai-Tibet Plateau at the end of the Tertiary Period, this waters gradually formed a unique fish population structure, with Schizothoracinae as its typical representative species. Typical representatives of Schizothoracinae included Schizothorax irregularis, S. eurystomus, S. biddulphi and S. barbatus. The four common Schizothoracinae are highly similar in appearance, and the countable and measurable traits are mostly crossed. Affected by work experience and cognition, the four species are prone to misjudgment in the identification, and many scholars often question this classification system. This paper intends to establish a discrimination system for Schizothoracinae identification through comparing traditional morphometric and Box-truss data system, and provide reference and information for subsequent research. The differences among the species of Schizothoracinae (4 species, 494 individuals, 48 groups of parameters) were detected through ANOVA, PCA, DFA and Cluster Analysis. The results of ANOVA showed that 9 parameters were significantly different in the 4 species of Schizothoracinae at the same time (P<0.05), and the difference was mainly located in the head, front trunk and rear trunk of the fish. PCA results showed that there was significant difference in PC1 axis between S. barbatus and S. irregularis based on the morphological data, which is mainly represented the head characteristics. However, the results of the Box-truss data showed that the difference between S. biddulphi and S. eurystomus mainly reflected in the rear trunk. The 8 and 6 main parameters were selected to establish the discriminant formulas for two data structures, and the initial discrimination success rates reached 85.7% and 75.7%, respectively. The cluster analysis results showed that the two data systems are divided into two branches, among which S. barbatus is a separate group, the S. biddulphi and S. irregularis are closer in morphology, followed by S. eurystomus. The using of multivariate statistical methods can better distinguish the four Schizothoracinae species in the Yarkand River Basin, and the morphological data performed better than the Box-truss data by contrasting two data systems. However, there is still confused distinguish characteristic in species identification. These findings can be used to quickly verify the Schizothoracinae that are difficult to accurately identify based on the lip characteristics, and would accumulate basic data to provide a reference for follow-up research.