The influence of loss of diversity on community dynamics and ecosystem functioning has recently received considerable attention. Although study of biodiversity has a long history within ecology, empirical investigations exploring consequences of loss have been rare. Because many factors confound diversity comparisons, experimental manipulations of diversity offer the most direct way of attributing cause to diversity loss. The effects of reduction in number of species will depend on the strength and sign of species interactions affected by loss of diversity. An experiment performed on a high zone, rocky intertidal community in which macroalgal diversity was manipulated demonstrated that effects of diversity loss will be highly dependent on which species are removed. However, effects of diversity reductions were strongest at the harsh end of a stress gradient where interactions were more positive. Thus, factors that affect the strength and sign of species interactions such as the degree of physical stress may serve as a rough guide to where the effects of diversity loss will be most severe. An assessment of the influence of diversity on community response to a strong physical perturbation was performed using an experimentally-induced thermal stress. Higher diversity treatments were most strongly affected directly by the stress because such treatments had higher abundance and therefore more biomass to lose. However, those same treatments recovered more quickly from the stress. Community recovery of initially low diversity treatments was slowed by persistence of non-typical states or slow recovery of dominant species. A simulation study was performed to assess the ability of different experimental designs to detect biodiversity effects. Our ability to predict consequences of changes in diversity will be dependent on our ability to distinguish the most influential biodiversity "components" within a system. This study uncovered a phenomenon that will be common in biodiversity studies: misidentification of one biodiversity component (e. g., an effect of a keystone species) as a different component (e. g., an effect of the number of species). I call this phenomenon "aliasing." Because of the complexity of biodiversity, experiments and observational studies will be highly susceptible to aliasing and, thus, results will require careful interpretation. / Graduation date: 1997