为了探究三江源地区野生小花棘豆（Oxytropis glabra DC.）表型多样性及其与环境因子的关系，对三江源地区6个主要且具代表性的小花棘豆天然居群的18个表型性状进行变异特征和主成分分析，基于这些表型性状对6个居群进行聚类分析，并对这些表型性状与经度、纬度、海拔、土壤pH值和土壤电导率进行相关性分析。结果表明：除单株花序数和种形指数外，其余表型性状的F值均达到显著（p＜0.05）水平，其中，花序长的F值最大，叶形指数和种子千粒质量的F值较大。供试居群18个表型性状的变异系数为0.30%～53.79%，均值为14.99%，其中，叶形指数、单株花序数、花序长和单荚果种子数变异系数的均值均在20%以上，而种子千粒质量变异系数的均值最小（0.47%）。相关性分析结果表明：仅单叶鲜质量与纬度呈显著负相关，开花后株高和冠幅与海拔呈极显著（p＜0.01）负相关，单叶鲜质量和种子千粒质量与土壤电导率分别呈极显著和显著负相关。主成分分析结果显示：前3个主成分的累计贡献率为92.960%，说明这3个主成分能够反映小花棘豆表型性状的绝大部分信息。聚类分析结果表明：供试居群被分成2组，其中，甘德县柯曲镇和天峻县木里镇居群为一组，其余居群为另一组。综上所述，三江源地区小花棘豆表型性状变异较为丰富，但仅个别表型性状受纬度、海拔和土壤电导率的影响显著。

To explore the phenotypic diversity of wild Oxytropis glabra DC. in Sanjiangyuan region and its relationships with environmental factors, variation characteristics and principal component analysis were conducted for 18 phenotypic traits of 6 major and representative natural populations of O. glabra in Sanjiangyuan region, cluster analysis was carried out for the 6 populations based on these phenotypic traits, and correlation analysis was performed between these phenotypic traits and longitude, latitude, altitude, soil pH value, and soil electrical conductivity. The results show that except for inflorescence number per plant and seed shape index, the F values of other phenotypic traits all reach significant (p＜0.05) levels, in which, the F value of inflorescence length is the largest, the F values of leaf shape index and 1 000-grain mass of seeds are relatively large. The coefficients of variation of 18 phenotypic traits of test populations are 0.30%-53.79%, with an average of 14.99%, in which, the averages of coefficients of variation of leaf shape index, inflorescence number per plant, inflorescence length, and seed number per pod are all greater than 20%, while the average of coefficient of variation of 1 000-grain mass of seeds is the smallest (0.47%). The correlation analysis result shows that only fresh mass per leaf shows a significant negative correlation with latitude, plant height after flowering and crown breadth show extremely significant (p＜0.01) negative correlations with altitude, and fresh mass per leaf and 1 000-grain mass of seeds show extremely significant and significant negative correlations with soil electrical conductivity, respectively. The principal component analysis result shows that the cumulative contribution rate of the first three principal components is 92.960%, indicating that these three principal components can reflect most information of the phenotypic traits of O. glabra. The cluster analysis result shows that the test populations are divided into 2 groups, in which, the population in Kequ Town of Gade County and that in Muli Town of Tianjun County are clustered into one group, and other populations are clustered into the other group. In conclusion, the phenotypic trait variation of O. glabra in Sanjiangyuan region is relatively abundant, but only several phenotypic traits are significantly affected by latitude, altitude, and soil electrical conductivity.