在山桐子（Idesia polycarpa Maxim.）自然分布区南界（广东省乳源县）、北界（河北省易县）及主要集中分布区（四川省的芦山县和夹江县以及贵州省道真县）选择16个山桐子样本，对其完整果穗和成熟果实的性状（包括14个表型性状和果实含油率）进行变异分析、遗传多样性分析和相关性分析，并基于表型性状的主成分分析结果对供试样本进行聚类分析。结果表明：不同山桐子样本间果穗和果实性状差异极显著（P<0.001），各性状的变异系数为2.75%~58.23%，Simpson指数为0.75~0.88，Shannon-Weaver指数为2.30~3.13。相关性分析结果表明：果穗长、果穗宽、单果穗质量、单果穗果实质量和果穗小枝长间基本呈显著（P<0.05）或极显著（P<0.01）正相关，果实占比、单果质量、果实纵径、果实横径、单果种子质量和单果果肉质量间均呈极显著正相关，但果穗和果实表型性状与果实含油率的相关性均不显著。主成分分析结果表明：前4个主成分的累计贡献率为92.281%，能够反映山桐子果穗和果实表型性状的大部分信息；第1、第2、第3和第4主成分的贡献率分别为40.863%、31.979%、10.477%和8.962%，分别主要反映山桐子的果实大小、果穗长度和疏密度、果实形状及果柄长特征。聚类分析结果表明：供试样本可分为中果中疏穗中柄型（Ⅰ）、小果中疏穗长柄型（Ⅱ）、大果长疏穗中柄型（Ⅲ）及中果短密穗中柄型（Ⅳ）4个类群。其中，Ⅲ类群的果实含油率最高（31.07%），Ⅳ类群的果实含油率最低（21.28%）。研究结果显示：山桐子果穗和果实性状变异丰富，遗传多样性较高；果实大小和形状、果穗长度和疏密度以及果柄长度是影响山桐子果穗和果实表型多样性的关键性状；大果型和（或）长穗型山桐子具有一定的油用开发优势。

 Sixteen samples of Idesia polycarpa Maxim. were selected from the southern boundary （Ruyuan County of Guangdong Province）， northern boundary （Yixian County of Hebei Province）， and main concentrated distribution area （Lushan County and Jiajiang County of Sichuan Province， and Daozhen County of Guizhou Province） of its natural distribution areas. The variation analysis， genetic diversity analysis， and correlation analysis were conducted on characters （containing fourteen phenotypic characters and fruit oil rate） of their full spikes and ripe fruits， meanwhile， basing on the result of principal component analysis on phenotypic characters， the cluster analysis on test samples was performed. The results show that there are extremely significant （P<0.001） differences in characters of spikes and fruits of different samples of I. polycarpa， and the coefficient of variation of each trait is 2.75%-58.23%， the Simpson index is 0.75-0.88， and the ShannonWeaver index is 2.30-3.13. The result of correlation analysis shows that there are generally significant （P<0.05） or extremely significant （P<0.01） positive correlations among spike length， spike width， single spike mass， fruit mass of single spike， and offshoot length of spike， and there are extremely significant positive correlations among fruit proportion， single fruit mass， fruit vertical diameter， fruit horizontal diameter， seed mass of single fruit， and flesh mass of single fruit， but the correlations of phenotypic characters of spikes and fruits with fruit oil rate are all not significant. The result of principal component analysis shows that the cumulative contribution rate of the first four principal components is 92.281%， which can reflect most information of phenotypic characters of spikes and fruits of I. polycarpa； the contribution rates of the first， second， third， and fourth principal components are 40.863%， 31.979%， 10.477%， and 8.962%， respectively， which mainly represent the characteristics of fruit size， spike length and density， fruit shape， and petiole length of I. polycarpa. The result of cluster analysis shows that the test samples can be divided into four groups， including middle fruit-middle and sparse spike-middle petiole type （Ⅰ）， small fruit-middle and sparse spike-long petiole type （Ⅱ）， big fruit-long and sparse spike-middle petiole type （Ⅲ）， and middle fruit-short and dense spike-middle petiole type （Ⅳ）. Among them， the fruit oil rate of group Ⅲ is the highest （31.07%）， and that of group Ⅳ is the lowest （21.28%）. It is suggested that the variations of characters of spikes and fruits of I. polycarpa are abundant， and their genetic diversity is relatively high； fruit size and shape， spike length and density， and petiole length are the key characters affecting the phenotypic diversity of spikes and fruits of I. polycarpa； I. polycarpa with big fruits and （or） long spikes has certain oil development advantages.