The species composition of pools in the intertidal zone on the coast of Washington State varies greatly from pool to pool and from time to time. While assemblages change somewhat predictably from the low— to the high—intertidal zone (presumably owing to different stress tolerances of the species), the variance among pools at a given tidal height cannot be ascribed to such physical factors. Some pools at each height are dominated by one species that monopolizes space on the rock or in the water column and modifies the pool environment. Each dominant species, once established, can spread rapidly through a pool (either by vegetative growth or by enhanced recruitment of its conspecifics) and is thus potentially self—perpetuating. When abundant, most dominants appear to prevent potential competitors from settling and surviving by monopolization of resources, abrasion of the substratum, and/or collection of sediment. Six such dominants were identified for Washington tidepools: from low to high pools, these are (1) the surfgrass Phyllospadix scouleri, (2) articulated coralline algae, (3) the mussel Mytilus californianus (exposed shores), (4) the cloning anemone Anthopleura elegantissima (more protected shores), (5) the red alga Rhodomela larix, and (6) the green alga Cladophora sp. Colonial diatoms also appear capable of dominating low pools in the absence of wave disturbance. However, each dominant monopolizes only 20—50% of the pools at any height. Disturbances, defined here as a loss of biomass exceeding 10% cover of a sessile species within 6 mo and caused by extrinsic forces, were observed frequently in regularly censused tidepools. Disturbance agents included waves, excessive heat, wave—driven logs or rocks, and unusual influxes of predators and herbivores. Severe disturbances (those affecting a large proportion of the organisms in a pool) tended to occur in high pools in the summer (due to heat stress) and low pools in the winter (due to wave damage). Overall, a disturbance occurred in every pool studied an average of every 1.6 yr. About half of the 231 observed disturbances affected one of the six dominant species. The frequencies of these disturbances ranged from one every 2—5 yr, and recovery of the species to its original level required 3 mo to>2 yr. Some species (e.g., Rhodomela) were disturbed frequently bu recovered quickly because of rapid vegetative growth. However if asexual propagation was not possible, such as when the entire population of a species was removed from a pool, the slowness and irregularity of recruitment of sexual propagules greatly impeded recovery. Experimental manipulations involving the total removal of dominant species from pools showed that such large disturbances often require>3 yr for recovery. The irregularity of planktonic recruitment can be compounded by the presence of herbivores, which can remove most settling organisms from the substratum, or by the absence of other organisms that are necessary for the settlement of a dominant (e.g., seed—attachment sites for Phyllospadix). The combination of high disturbance frequency and slow rates of recovery makes it impossible for any dominant to occupy all the pools in its tidal range at any one time. Disturbance is viewed in these habitats as the stochastic factor overlying other, more predictable, community—structuring factors such as tidal height, pool size, wave exposure, and levels of herbivory, predation, and competition. Thus combined deterministic processes and random events operate to produce a complex mosaic of species assemblages in pools in one region. None of the tidepool assemblages is "stable" over many generations; rather, they seem to exist in a dynamic state where disturbances are an integral structuring factor.