对极小种群野生植物贵州地宝兰(Geodorum eulophioides Schltr.)17个群落进行样方调查，采用生态位宽度和生态位重叠对群落草本层种类间的生态位进行分析，并运用方差比率(VR)、χ2检验、Pearson相关系数和Spearman秩相关系数分析草本层种类的种间关系。结果表明：贵州地宝兰群落中共有维管植物41科78属94种（含亚种、变种）。在草本层中，贵州地宝兰的重要值为15.10%，位列第4；贵州地宝兰的Levins生态位宽度和Shannon-Wiener生态位宽度均位列第1，分别为14.85和2.76。贵州地宝兰群落草本层有23个种类形成生态位重叠，共计100对，Pianka生态位重叠值在0.00~0.85之间；贵州地宝兰与其他22个种类均存在生态位重叠，其中与仙茅(Curculigo orchioides Gaertn.)、求米草〔Oplismenus undulatifolius (Arduino) Beauv.〕、藿香蓟(Ageratum conyzoides Linn.)和海金沙〔Lygodium japonicum (Thunb.) Sw.〕的Pianka生态位重叠值较大，分别为0.79、0.66、0.61和0.57。贵州地宝兰群落草本层种类间总体联结为不显著负联结。对贵州地宝兰群落草本层中形成生态位重叠的23个种类进行进一步分析，χ2检验结果显示这23个种类间呈不显著负联结的种对数占总种对数（253）的65.2%；Pearson相关性分析和Spearman秩相关性分析结果也均表明大多数种对间呈不显著相关，其中不显著负相关的种对数分别占总种对数的67.6%和53.0%，贵州地宝兰仅与白茅〔Imperata cylindrica (Linn.) Beauv.〕存在极显著正相关，与爵床(Justicia procumbens Linn.)、丝叶薹草(Carex capilliformis Franch.)和求米草存在显著正相关。综上所述，贵州地宝兰群落草本层种间关系总体上较为松散，种间竞争也不明显，各种类较为独立，具有一定的随机性，但贵州地宝兰与少数种类生态位重叠较大，在未来群落物种共存及演替过程中可能会存在竞争关系，建议对贵州地宝兰居群建立保护小区(点)进行就地保护，同时进行动态监测，加强对放牧活动的管控力度。

 Quadrat investigation was conducted for the 17 communities of wild plant Geodorum eulophioides Schltr. with extremely small population, the niches among species in herb layers of communities were analyzed by using niche breadth and niche overlap, and the interspecific relationships of herb layer species were analyzed by using variance ratio (VR), χ2 test, Pearson correlation coefficient, and Spearman rank correlation coefficient. The results show that there are 94 species (including subspecies and varieties) of vascular plants belonging to 78 genera of 41 families in G. eulophioides communities. In the herb layer, the importance value of G. eulophioides is 15.10%, ranking fourth; the Levins niche breadth and ShannonWiener niche breadth of G. eulophioides rank first, with the values of 14.85 and 2.76 respectively. There are 23 species forming 100 pairs of niche overlaps in herb layers of G. eulophioides communities, and the Pianka niche overlap values are within 0.00-0.85; there are niche overlaps between G. eulophioides and the other 22 species, in which the Pianka niche overlap values with Curculigo orchioides Gaertn., Oplismenus undulatifolius (Arduino) Beauv., Ageratum conyzoides Linn., and Lygodium japonicum (Thunb.) Sw. are relatively large, which are 0.79, 0.66, 0.61, and 0.57 respectively. The overall connection among species in herb layers of G. eulophioides communities is nonsignificant negative connection. Further analyses are made on 23 species forming niche overlaps in herb layers of G. eulophioides communities, and the χ2 test result shows that number of species pairs showing nonsignificant negative connection among these 23 species accounts for 65.2% of the total species pair number (253); the results of Pearson correlation analysis and Spearman rank correlation analysis both show that the correlations among most species pairs are not significant, in which numbers of  species pairs with non-significant negative correlation account for 67.6% and 53.0% of the total species pair number respectively, and G. eulophioides only shows an extremely significant positive correlation with Imperata cylindrica (Linn.) Beauv. and shows significant positive correlations with Justicia procumbens Linn., Carex capilliformis Franch., and O. undulatifolius. In summary, the interspecific relationship in herb layers of G. eulophioides communities is relatively loose in general, the interspecific competition is not obvious, and each species is relatively independent and has a certain randomness. However, the niche overlaps between G. eulophioides and a few species are relatively large, and there may be a competitive relationship in the process of community species coexistence and succession in the future. It is suggested to establish a protecting area (site) for in situ conservation of G. eulophioides populations, conduct dynamic monitoring, and strengthen the control of grazing activities.