SPATIAL DISPERSION AND DENSITY DEPENDENCE IN A PERENNIAL DESERT SHRUB ( CHRYSOTHAMNUS NAUSEOSUS : ASTERACEAE)

Dogma asserts that desert shrubs should become uniformly dispersed through time through a process of self‐thinning caused by competition for water. We present long‐term data (1985–2001) on two nearby populations of the perennial shrub, Chrysothamnus nauseosus ssp. consimilis , in sand‐dune ecosystems near Mono Lake, California. Because previous studies established that water limited the growth, survival, and reproduction of this shrub, we searched for population‐level patterns of competition. A seven‐year history of survival, growth, and flowering of marked individuals delineated five demographic stage classes corresponding roughly to age and plant size: seedlings, nonflowering juveniles, juveniles that later flowered, adults that flowered intermittently, and “immortal” adults that flowered annually. For all stages except nonflowering juveniles, initial size was a strong predictor of life expectancy, and for the reproductive classes, flowering was associated with larger size and higher survival. Survival of nonflowering juveniles was strongly related to depth to groundwater (consistent with the biology of phreatophytes, defined as plants that use deep soil moisture) corresponding to a critical demographic shift at this phreatophytic threshold necessary for long‐term survival and reproduction. We therefore predicted that only the pre‐phreatophytic stage classes dependent on depletable upper soil moisture should compete with close neighbors. Neither mixed‐age population of C. nauseosus nor any single stage class within them was uniformly dispersed. Younger classes (seedling and juveniles) were aggregated, and older classes (from newly reproductive through oldest adults) were randomly dispersed with respect to their nearest neighbor. The younger the stage, the more aggregated was the dispersion. All five seedling cohorts became significantly less aggregated with time during the study, consistent with a hypothesis of competitive thinning. However, mortality was statistically density independent at a landscape scale in all classes and all years and resulted in thinning toward reduced aggregation, just as expected under competition. Two distinct stage‐specific patterns occurred. Whereas the phreatophytic stage classes experienced only density‐independent mortality, the pre‐phreatophytic stage classes experienced opposing density dependence at different spatial scales. Seedling cohorts in particular exhibited strong direct density dependence within the first 1–3 years in a local neighborhood of 20 cm or less, in both survival (higher mortality in dense aggregations) and growth (seedlings in groups were smaller and grew more slowly than those not in groups). In contrast, these classes exhibited inversely density‐dependent survival, and often growth, on a scale of 1–4 m caused by spatial heterogeneity and not facilitation such as by nurse plants. Although competition occurred for only the first 1–3 years of the life history, it caused size differences among individuals that increased through time and that significantly predicted long‐term survival and eventual recruitment to the reproductive class. Despite this potential for long‐lasting effects on population dynamics, competitive thinning was too brief and stage‐structured spatial processes were too complex to result in a uniform dispersion of individuals in this mixed‐age population. Corresponding Editor: S. D. Wilson