甜瓜种质资源的判别分析

DISCRIMINANT ANALYSIS OF GERMPLASM RESOURCES OF CUCUMIS MELO L.

在对甜瓜种质资源数量分类的基础上,应用判别分析建立了甜瓜3亚种、4变种的二级分类函数,为甜瓜种质资源的自动鉴定打下了基础。用单株结果数、茎毛尖钝、果皮厚度和果皮木栓化4个性状建立的一级分类函数,对野生甜瓜、薄皮甜瓜和厚皮甜瓜3亚种的判别准确度为97.6％。用子叶距、病害级别、两性花萼片长、花痕有无和子蔓节数5个性状建立的分类函数,对网纹甜瓜、非网纹甜瓜的判别准确度为90％。由果实酸度、种子囊衣、单性结实、两性花雄蕊数和果粗5个性状建立的判别函数,判别中国甜瓜和菜瓜的准确度为100％。另外还对判别分析的应用价值及其局限性进行了讨论。

On the basis of numerical taxnomy of germplasm resources of Cucumis melo L. , The grades classification functions were set up for three sup-species and four varieties of Cucumis melo L. by step-wise discriminant analysis with 98 characters,which stands a fundement for auto-identification of germplasm resources of Cucumis melo L.. The first grade classification function being made of the number of fruit per plant, the shape of hair top on stem, the thickness of skin and suberization of skin has 97.5% correct rate on distinguishing the three sup-species. The second grade classification function depending upon the character of distance of two cotyledon top, the grade of diseases, the length of sepal of complete flower, have/no of cicatrice of flower on fruit and the number of nodes of second branch per plant has 90% correct rate on discriminating netted melon and non-netted melon; the second classification function including characters of the sour of sarcocarp, bursa of seed, parthenocarpy, the number of stamens of complete flower and the diameter of friut distinguishing oriental melon and oriental pickling melon at 100% correct. In addition, the value and the limit of discriminant analysis used in germplasm study were discussed.