Abstract: Stepwise discriminant analysis was made from the data obtained from 9 traits of leaves (i.e. leaf length,leaf width,leaf thickness,width and thickness of vascular bundle,thickness of palisade tissue and mechanical tissue,thickness of epidermis and number of stomata),originating from 22 provenances of Cunninghamia lanceolata (Lamb.) Hook,Taiwania cryptomerioides Hayata and C. lanceolata (Lamb.) Hook.var.konisii (Hay.)Fujita. The optimized discriminant formula was developed by selecting thickness of palisade tissue (X₆),leaf length (X₁) and leaf thickness (X₃) as three important factors,i.e. F₁(Taiwania cryptomerioides Hayata)=-235.8714+17.3104X₆+26.3587X₁+5.9046X₃ F₂(Cunninghamia lanceolata (Lamb.) Hook)=-564.9259+78.4254X₆+93.1923X₁+0.6295X₃ F₃(C. lanceolata (Lamb.) Hook.var. konisii (Hay.) Fujita)=-349.4205+64.1926X₆+62. 6479X₁+0.8969X₃ and can be used to differentiate quantitively these three species, aiming at probing into the application of stepwise discriminant analysis to the classfication of tree species.

Key words: Cunniinghamia lanceolata (Lamb.) Hook, Taiwania cryptomerioides Hayata, C. lanceolate (Lamb.) Hook.var konisii (Hay.) Fujita, Stepwise discriminant analysis

摘要： 通过对杉木、台湾杉、香杉22个种源的叶长、叶宽、叶厚、维管束的宽与厚、栅栏组织厚、机械组织厚、表皮厚和气孔数等9个叶性状进行逐步判别分析,选取栅栏组织厚(X₆)、叶长(X₁)和叶厚(X₃)三个重要因子,建立优化判别方程:F₁(台湾杉)=-235.8714+17.3104X₆+26.3587X₁+5.9046X₃ F₂(杉木)=-564.9259+78.4254X₆+93.1923X₁+0.6295X₃ F₃(香杉)=-349.4205+64.1926X₆+62.6479X₁+0.8969X₃定量地区分这三个树种,旨在对逐步判别分析法在树种分类上的应用进行探讨。

关键词: 杉木, 台湾杉, 香杉, 逐步判别分析