Abstract: In order to explore the correlation between yield and phenotypic traits of Leymus chinensis germplasm resources and screen out excellent Leymus chinensis germplasm,28 Leymus chinensis germplasm resources were used as experimental materials in this study,and the correlations between 9 phenotypic traits were analyzed using two comprehensive evaluation methods. The results showed that the correlation orders between field phenotypic characteristics of Leymus chinensis and hay yield were obviously different between the two methods. Through the linear fitting analysis between the two methods and the actual hay production ranking results,it was found that the comprehensive evaluation by DTOPSIS entropy weight method and entropy method was the closest to the actual hay production results,and the correlation coefficients were 0.994 5 and 0.955 7,respectively.Through comprehensive evaluation,it was found that in the process of breeding new Leymus chinensis varieties with high yield,attention should be paid on increasing the proportion of spikelets,considering the proportion of leaf width,and other plant characters. After comprehensive evaluation,the three germplasm materials H8,H7 and H27 in the two evaluation methods had good high-yield potential,which could be further used as breeding materials for breeding high-yield varieties of Leymus chinensis.

Key words: Leymus chinensis, Grey correlation degree method, DTOPSIS method, Comprehensive evaluation

摘要： 为探讨羊草（Leymus chinensis）种质资源产量与表型性状之间的相关性，筛选出优良的羊草种质。本研究以28份羊草种质资源为试验材料，基于9个田间表型性状，采用2种综合评价方法，进行了羊草种质资源表型性状间相关性分析和优良种质筛选。结果表明，灰色关联度法和DTOPSIS法在羊草田间表型性状与干草产量的关联次序有明显的差异。通过2种方法分别与实际干草产量排序结果进行线性拟合分析，发现相较于灰色关联度法，DTOPSIS熵权值法和熵值法的综合评价最接近干草实际真实测产结果，相关系数分别达到了0.994 5和0.955 7。因此，在羊草表型性状与干草产量相关性田间综合性状分析中应采用DTOPSIS法，该方法更适合羊草多性状评价。同时，通过综合评价发现在羊草高产新品种选育过程中应优先注重提升小穗数量比例，并要考虑叶宽比例，也要注意兼顾其它性状。经过综合评价，在2种评价方法中前三位的种质材料H8、H7和H27具有良好的高产潜力，可进一步作为育种材料进行羊草高产品种的选育。

关键词: 羊草, 灰色关联度法, DTOPSIS法, 综合评价