This paper addresses questions of community and patch stability as defined by the population biology of dominant plants in the context of different areas within a large kelp forest. We ask (1) @'Do large-scale episodic events override biological mechanisms as major community structuring processes?@', (2) @'Are different local areas characterized by different processes?@', and (3) @'How persistent are the patches or biological structure over decal and local spatial scales?@' We evaluate these questions with regard to the effects of various types of disturbance for as much as three decades on the populations of several species of kelp in the large kelp forest off Point Loma, San Diego, California. The most sensitive population factors we studied include recruitment, density, and survivorship. Patch stability was evaluated with regard to the persistence of patches already well established in 1971-1972. The study sites offer a cross-shore transect through the central part of a large kelp forest at depths of 8, 12, 15, 18, and 21 m; two additional sites at the north and south ends of the forest offer a longshore transect along the 18-m contour. There were marked differences among the decades with regard to the intensity of the disturbances. Compared with the 1980s, the two preceding decades were relatively benign. The 1980s had two extreme disturbance events: the 1982-1984 El Gino-Southern Oscillation (ENSO) was the most severe El Nino event in the last century, which included very warm, nutrient-depleted water, and a short but intense storm in January 1988 appeared to have been the most severe in perhaps 200 yr. The storm changed age-specific kelp mortality patterns and caused the first large-scale understory mortality in several decades. By sweeping away drift algae it caused intense local urchin grazing. The storm was followed by a strong La Nina event marked by cool, nutrient-rich water in 1988-1989. Differences in kelp recruitment and survivorship in different areas of the kelp forest are influenced by gradients in longshore currents, temperature, light, wave energy, floc, planktonic propagules, and physical disturbance. The areas are characterized by different plant population patterns and the effects of several species of herbivores. The massive disturbances of the 1980s obliterated much of the structure in the kelp forest. Certainly the disturbances caused many lag effects including outbreaks of understory algae such as Desmarestia ligulata, intraspecific competition, changes in grazing patterns, etc., which in turn resulted in between-area variation in recovery rates. However, in all cases this variation was overshadowed by the overwhelming competitive dominance of Macrocystis pyrifera. Most of the understory patches on the transect lines, some of which had persisted for 7 yr, died out by the end of 1990. The population biology of Macrocystis was remarkably similar in most areas, as the cohort longevity and survivorship curves were very similar, and the plant and stipe densities tended to level off in only a few years. Thus large-scale episodic events such as El Ninos, La Ninas, and rare storms exert dramatic impacts, but small-scale responses such as density-vague recruitment (neither density dependent nor density independent) and survival allow prompt recovery, often to preexisting patterns. The one exception was the southern site, which was marked by sea urchin grazing and poor kelp recruitment through the latter half of the 1980s, but a recent sea urchin disease event has led to kelp recruitment in fall 1991. A seeming paradox to the observed Macrocystis dominance is that in almost all areas, some understory patches of old plants have persisted through the 1980s.