甘草野生种群果实与种子形态特征的地理变异研究

Study of the Geographic Variation of Fruit and Seed Morphological Characteristics of Glycyrrhiza Uralensis Fisch.Wild Population

目的:研究甘草荚果和种子形态特征地理变异规律并分析其形成的生态学机制。方法:采用样方调查的方法测定全国4省区8个旗县甘草野生种群的地上植株形态特征,采用双重筛选逐步回归分析方法探讨甘草荚果和种子形态特征的地理变异规律及其形成的生态学机制。结果:(1)不同甘草种群的荚果数、荚果腺毛长度、种室数目、每荚种子粒数和结实率5项指标的差异达到了极显著水平(P<0.01);荚果长度、荚果厚度、荚果腺毛密度和种子长度4项指标的差异达到了显著水平(P<0.05)。(2)荚果厚度、种室数目等5项形态特征与经度呈显著相关,只有荚果腺毛长度1项指标与纬度显著相关。结论:甘草野生种群荚果与种子形态特征存在显著的地理变异,其变异趋势以经向变异为主,光照因子差异是导致地理变异的主要原因,其次为温度因子和水分因子,其相关关系可用多元回归模型解释。

Objective： To investigate the geographic variation of pod and seed morphological characteristics of Glycyrrhiza uralensis Fisch. wild population and analyse it＇s ecology mechanism. Methods： Quadrat survey method was used to determine the pod and seed morphological characteristics of Glycyrrhiza uralensis Fisch. wild population from 8 counties of 4 provinces in China. Double screening stepwise regression analysis method was used to investigate geographic variation regular patterns of pod and seed morphological characteristics of Glycyrrhiza uralensis Fisch. wild population and it＇s ecology mechanism. Results： （1） There was extremely significant differences （P〈0.01）among different Glycyrrhiza uralensis Fisch. wild populations in pod number, pod gandular hairs length, seed room number, seed number of per pod and seed setting rate. There was significant differences （P〈0.05）in pod length, pod thickness, pod gandular hairs density and seed length. （2） There was tight correlation between longitude and five morphological characteristics factors as pod thickness, seed room numbe, etc. Just pod gandular hairs length has tight correlation with latitude. Conclusion： There were significant geographic variations in the pod and seed morphological characteristics of Glycyrrhiza uralensis Fisch. wild population. The geographic variations were mainly show as longitude variation. The geographic variations were mainly due to the light factors, as the temperature and moisture factors could also cause geographic variation, the correlation between them could be explained by multivariate regression models.