Abstract Chong and Stohlgren [Chong, G.W., Stohlgren, T.J., 2007. Species–area curves indicate the importance of habitats’ contributions to regional biodiversity. Ecol. Indic. 7, 387–395] presented a combined ranking index to measure the habitats’ contribution to regional biodiversity, in which the species–area curves perform poorly in the prediction of the species richness in multiple plots. After re-examination of the dataset (17 vegetation types of the Rocky Mountain National Park, CO, USA), I present the reason for this poor performance: (1) species–area curve is not identical with species accumulation curve; (2) the latter is steeper than the former due to: (a) rare species normally appearing at coarser spatial scales and (b) beta diversity being positively correlated to the distance between samples. This result implies that Chong and Stohlgren's ranking index is scale dependent. The total number of species in different habitats could serve as an intuitive indicator that is highly correlated with the Chong–Stohlgren index.