Quantifying spatial patterns of species richness and determining the processes that give rise to these patterns are core problems In blodlveralty theory. The aim of the present paper was to more accurately detect patterns of vascular species richness at different scales along altitudinal gradients in order to further our understanding of biodlverslty patterns and to facilitate studies on relationships between biodiversity and environmental factors. Species richness patterns of total vascular plants species, including trees, shrubs, and herbs, were measured along an altitudinal gradient on one transect on a shady slope in the Dongling Mountains, near Beijing,China. Direct gradient analysis, regression analysis, and geostatistics were applied to describe the spatial patterns of species richness. We found that total vascular species richness did not exhibit a linear pattern of change with altitude, although species groups with different ecological features showed strong elevational patterns different from total species richness. In addition to total vascular plants, analysis of trees, shrubs, and herbs demonstrated remarkable hierarchical structures of species richness with altitude (i.e. patchy structures at small scales and gradients at large scales). Species richness for trees and shrubs had similar spatial characteristics at different scales, but differed from herbs. These results indicated that species groups with similar ecological features exhibit similar biodlveraity patterns with altitude, and studies of biodiversity based on species groups with similar ecological properties or life forms would advance our understanding of variations in species diversity. Furthermore, the gradients or trends appeared to be due mainly to local variations in species richness means with altitude. We also found that the range of spatial scale dependencies of species richness for total vascular plants, trees, shrubs, and herbs was relatively large. Thus, to detect the relationships betweenspecies richness with environmental factors along altitudinal gradients, it was necessary to quantify the scale dependencies of environmental factors in the sampling design or when establishing non-linear models.