The study of small-scale species density has attracted considerable attention over the past two decades (Grime, ). This attention has been driven in large part by a persistent interest in the processes believed to regulate species density; namely, competition, colonization, disturbance, stress, and facilitation. Studies of species density have examined a number of the processes thought to be involved, Abstract: Here we propose that an important cause of variation in species density may be prior environmental conditions that continue to influence current patterns. In this paper we investigated the degree to which species density varies with location within the landscape, independent of contemporaneous environmental conditions. The area studied was a coastal marsh landscape subject to periodic storm events. To evaluate the impact of historical effects, it was assumed that the landscape position of a plot relative to the river's mouth ("distance from sea") and to the edge of a stream channel ("distance from shore") would correlate with the impact of prior storm events, an assumption supported by previous studies. To evaluate the importance of spatial location on species density, data were collected from five sites located at increasing distances from the river's mouth along the Middle Pearl River in Louisiana. At each site, plots were established systematically along transects perpendicular to the shoreline. For each of the 175 plots, we measured elevation, soil salinity, percent of plot recently disturbed, percent of sunlight captured by the plant canopy (as a measure of plant abundance), and plant species density. Structural equation analysis ascertained the degree to which landscape position variables explained variation in species density that could not be explained by current environmental indicators. Without considering landscape variables, 54% of the variation in species density could be explained by the effects of salinity, flooding, and plant abundance. When landscape variables were included, distance from shore was unimportant but distance from sea explained an additional 12% of the variance in species density (R 2 of final model = 66%). Based on these results it appears that at least some of the otherwise unexplained variation in species density can be attributed to landscape position, and presumably previous storm events. We suggest that future studies may gain additional insight into the factors controlling current patterns of species density by examining the effects of position within the landscape.Résumé : Les auteurs estiment qu'une cause importante des variations de la densité des espèces a trait aux conditions environnementales passées qui continuent d'influencer les patrons actuels. Dans cet article, nous examinons le degré avec lequel la densité des espèces varie en fonction de la localisation, indépendamment des conditions environnementales actuelles. Le site à l'étude est un marais côtier de la Louisiane sujet à des tempêtes périodiques. Afin d'évaluer l'impact des facteurs historiques, nous a...