Patterns of broad-scale plant species richness are thought to be largely determined by (1) variation in energy and water availability among sampling units (species energy hypothesis), (2) habitat and topographic heterogeneity within sampling units (spatial heterogeneity hypothesis), and (3) regional differences in geographic configuration and history (regional effects hypothesis). However, lack of taxonomic and distribution data, particularly for tropical regions, has impeded assessments of the relative importance of these three hypotheses. We used a large botanical database to estimate the pattern of relative vascular plant richness across the western Neotropics and regression models to measure the extent to which this estimated pattern supported predictions from each of the above three hypotheses. Variation in plant richness across three major paleophysiographic regions (northwest South America, southern Central America, and northern Central America) was primarily predicted by the spatial heterogeneity hypothesis, with secondary contributions from the species energy hypothesis and, to a lesser extent, the regional effects hypothesis. Regression models that incorporated the relative contributions of all three hypotheses predicted peaks of relative species richness mostly in topographically complex areas (e.g., Sierra Madre de Chiapas, Cordillera de Tilarán, Cordillera de Talamanca, Panama's Cordillera Central, the Andes, and the Venezuelan Guayana); relatively low richness in central Mexico and Yucatán, Los Llanos of Venezuela, and in the Gran Chaco region of Bolivia, Paraguay, and Argentina; and a richness trough in lowland Amazonia relative to southern Central America, the Andes, and the Venezuelan Guayana. We discussed the contrast between our results and previous assessments that found plant richness to be primarily determined by the species energy hypothesis and predicted different patterns of plant richness across the western Neotropics.