Growth rates and population variables of the seaweeds Laminaria longicruris De la Pylaie and L. digitata (Hudson) Lamouroux were monitored from July 1980 until April 1983 within adjacent harvested and control plots in southwestern Nova Scotia, Canada. Faster growth rates within the harvested plot, relative to the control plot, for new recruits of both species were detected only during the first year following a September 1980 total harvest of both species. Since standing crop had nearly recovered to pre-harvest levels after 1 yr, the increased growth rate was attributed to reduced standing crop during the first year post-harvest. Analysis of the relationship between each species' summer canopy biomass and the density of the most recent year-class developing in the sub-canopy at the same time provided evidence for interspecific competition, and provided parameter estimates for the relationship. A stochastic model of this 2-species interaction using reliable data on growth, recruitment, and mortality predicts long-term persistence of both species in the absence of harvesting, or other disturbances. Following total harvest, L. longicruris is initially more abundant than L. digitata, but an increasing canopy of L. digitata diminishes the dominance of L. longicruris over a period of 3 to 5 yr.Annual harvesting of only the canopy of L. longicruris (a practical and perhaps technically preferable strategy) produces only 4 % of the yield for a total annual harvest. Annual harvesting of the canopy of L. longicruris plus L. digitata produces only 11 % of the yield for a total annual harvest. These lower yields result from a large decrease in L. longicruris growth rate with increased density. This feature diminishes the effectiveness of removing L. digitata to reduce competition with faster growing L.longicruris. The need for controlled and monitored experimental harvests to evaluate these predictions is discussed, as is the need for more functional analyses of seaweed systems.