Аннотация:Abstract: In the present study, we studied the population genetic structure and demographic history of Gansu pika (O. cansus) by using multiple mitochondrial loci (COI, Cyt b, tRNA-Thr, tRNA-Pro, control region, totally 2417 base pairs). Bayesian analysis of the concatenated genes revealed two major lineages: Lineage 1 included samples from Baishuijiang and Tangjiahe; Lineage 2 contained samples from the other 12 locations. The phylogenetic structure of haplotype network built based on the same genes was almost the same as that of the Bayesian tree. Bayesian analysis of population structure (BAPS) divided O. cansus into four geographic populations. Population A contained samples from Baishuijiang and Tangjiahe;B included WL (Wolong, Sichuan)only; C included North Qinghai population (GC, MY, QL), South Qinghai-North Sichuan population (JW, JZ, RE, ZK), KD (Kangding, Sichuan), TB (Taibai Mountain, Shaanxi); and D included LF (Loufan, Shanxi) and WQ (WuQi, Shaanxi). Species tree constructed on mtDNA showed that the major divergence time in four geographic populations were: 0.72 Ma, 0.32 Ma, 0.12 Ma, which generally coincided with the Kunhuang Movement (1.10-0.60Ma)and Gonghe Movement (0.15 Ma), the two dramatic geological events during the uplift of Qinghai-Plateau. Moreover, we estimated the demographic history of O. cansus by using three methods: neutrality tests, mismatch distributions and extended Bayesian skyline plots. The results indicated that O. cansus might have not experienced population expansion, but it was significantly decreased since the Last Glacial Maximum. The Ecological Niche Modeling indicated that the potential distribution of O. cansus in the Last Glacial Maximum (LGM) is smaller than Current and LIG, and the precipitation played a more important role than temperature in the population dynamics of Gansu pika.This is possibly because precipitation strongly influences the food resources and habitats of Gansu Pika by affecting the vegetation succession.