通过Hoagland营养液培养法，对苜蓿属（Medicago Linn.）15种植物132份样本幼苗进行NaCl〔0（对照）、3、5、7和9 g·L^-1NaCl〕胁迫处理，测定了幼苗存活率（SR）、幼苗生长速率（GR）、叶片叶绿素含量（Chl）、苗高（SH）、叶片相对电导率（REC）、单株干质量（DW）和单株绿叶数（NGL）的变化；以这7个指标的耐盐系数为基础，运用主成分分析和隶属函数分析，结合耐盐性综合评价值（D）及聚类分析对各样本的耐盐性进行了综合评价和归类；并采用逐步回归分析建立了耐盐性预测回归方程。结果表明：经NaCl胁迫处理后，供试样本的SR、GR、Chl、SH、DW和NGL值均低于对照，仅REC值高于对照。不同样本7个指标的耐盐系数均存在明显差异，且7个指标的耐盐系数间均存在不同程度的相关性；SR、GR、Chl、SH、REC、DW和NGL的耐盐系数分别为0.394～0.926、0.147～0.886、0.393～0.913、0.272～0.896、1.136～4.120、0.293～0.961和0.283～0.975，其中，各指标耐盐系数最高的样本分别为10173、4187、4587、4799、10173、3172和4570，耐盐系数最低的样本分别为3754、4818、3773、4600、3803、4607和4818。主成分分析结果表明：前4个主成分的贡献率分别为45.656％、26.024％、9.637％、8.236％，累计贡献率达89.553％；其中，第1主成分的主要作用因子为SH、DW和NGL，第2主成分的主要作用因子为SR和GR。根据D值由大至小排序，供试132份样本可分为强耐盐型（22份样本）、中耐盐型（39份样本）、弱耐盐型（59份样本）和敏盐型（12份样本）4类，与聚类分析结果总体一致。其中，样本10173的D值最大，耐盐性最强；样本3754的D值最小，耐盐性最弱。通过逐步回归分析，获得回归方程Y=-0.353+0.471XChl-0.367XNGL+0.197XSH+0.094XREC（R²=0.971），表明Chl、NGL、SH和REC 4个指标对供试样本的耐盐性有显著影响。研究结果显示：供试132个样本的耐盐性差异明显，同种植物不同样本间的耐盐性也存在差异；在相同的实验条件下，可采用Chl、NGL、SH和REC 4个指标对苜蓿属植物苗期的耐盐性进行预测。

Changes in seedling survival rate （SR）， seedling growth rate （GR）， chlorophyll content in leaf （Chl）， seedling height （SH）， leaf relative electric conductivity （REC）， dry weight per plant （DW）, and green leaf number per plant （NGL） of seedlings of 132 samples of 15 species in Medicago Linn. were detected under NaCl stress treatment 〔0 （the control）， 3， 5， 7, and 9 g·L^-1 NaCl〕 by Hoagland nutrition culture method. On the basis of salt tolerance coefficient of seven indexes， salt tolerance of each sample was comprehensively evaluated and classified by principal component analysis and subordinative function analysis combining with salt tolerance comprehensive evaluation value （D） and cluster analysis； and the regression equation of salt tolerance predicted was established by stepwise regression analysis. The results show that SR， GR， Chl， SH， DW, and NGL values of samples tested are all lower than those of the control after NaCl stress treatment， but only REC value is higher than that of the control. Salt tolerance coefficients of seven indexes of different samples are all obviously different and show correlations at different degrees. Salt tolerance coefficient of SR， GR， Chl， SH， REC， DW, and NGL is 0.394-0.926， 0.147-0.886， 0.393-0.913， 0.272-0.896， 1.136-4.120， 0.293-0.961， and 0.283-0.975， respectively， in which， sample with the highest salt tolerance coefficient of each index is 10173， 4187， 4587， 4799， 10173， 3172， and 4570， respectively， that with the lowest salt tolerace coefficient is 3754， 4818， 3773， 4600， 3803， 4607， and 4818， respectively. The result of principal component analysis shows that contribution rate of the first four principal components is 45.656％， 26.024％， 9.637％, and 8.236％， respectively, with the cumulative contribution rate of 89.553％； in which， the main action factors of the first principal component are SH, DW, and NGL， those of the second principal component are SR and GR. According to the order of D value from big to small， 132 samples tested can be divided into four classes of strong salt tolerance （22 samples）， middle salt tolerance （39 samples）， weak salt tolerance （59 samples）, and salt sensitive （12 samples） types， which is generally in agreement with the cluster analysis result. Among them， D value of sample 10173 is the largest with the strongest salt tolerance； D value of sample 3754 is the smallest with the weakest salt tolerance. By stepwise regression analysis， regression equation of Y=-0.353+0.471XChl-0.367XNGL +0.197XSH+0.094XREC （R²=0.971） is established， indicating that four indexes of Chl， NGL， SH， and REC have significant effects on salt tolerance of samples tested. It is suggested that there is obvious difference in salt tolerance among 132 samples tested， even in different samples of the same plant； under the same experimental condition， four indexes of Chl, NGL, SH, and REC can be used to predict salt tolerance of Medicago plants at seedling stage.