187 Ecological Monographs, 71(2), 2001, pp. 187–217  2001 by the Ecological Society of America BOTTOM-UP AND TOP-DOWN FORCES IN TIDE POOLS: TEST OF A FOOD CHAIN MODEL IN AN INTERTIDAL COMMUNITY KARINA J. NIELSEN 1 Department of Zoology, Oregon State University, Corvallis, Oregon USA 97331-2914 Abstract. A simple food chain model of community structure was used to evaluate the roles of bottom-up and top-down factors in a rocky intertidal community. Predictions of the model were modified to incorporate known variation in the strength of species inter- actions and nutrient delivery rates along a wave-exposure gradient. To test the predictions of the model, I manipulated nutrients and consumers in tide pools chiseled into mudstone benches at two sites that varied in degree of wave exposure. The pools were located in the mid-intertidal zone between 1 and 1.5 m above mean lower low water (MLLW), at Boiler Bay, Oregon, USA. The focal organisms were the benthic macroalgae and mobile inver- tebrate herbivores that dominate naturally occurring tide pools at this site. I manipulated nutrient levels and the abundance of herbivores in these tide pools in a fully factorial randomized block design replicated six times at a wave-exposed and a wave- protected site. The experiment was maintained for two years (1994–1996). The abundances of herbivores and macroalgae were monitored in the spring, summer, and fall of each year. I measured primary productivity in the tide pools during the summer. Herbivores had a negative impact on algal abundance. The total effect of herbivory, and the efficiency of herbivores per se, on algal abundance was lower at the wave-exposed site. Nutrient additions had a positive effect on algal abundance, but this effect was reduced at the wave-exposed site. Nutrients also appeared to increase algal productivity, but only where herbivore abundance was low. Algal abundance patterns were generally consistent with model predictions for bottom-up, top-down, and hydrodynamic effects. In contrast to model predictions, herbivore abundance did not respond to the nutrient treatment. The decoupling of consumers from resource dynamics is interpreted to be the result of an herbivore preference for noncalcified seaweeds with higher potential growth rates. In wave-protected pools, where nutrients were most limiting and consumers were most effi- cient, seaweeds with the potential to translate elevated nutrient levels into growth had no effective refuge from consumers. The difference in scale between resource patches (tide pools) and the foraging range of the dominant herbivore, Tegula funebralis, may have augmented the ability of this herbivore to virtually exclude fleshy seaweeds from wave- protected pools. Expanding the domain of applicability of food chain models requires the incorporation of consumer preferences, variation in plant growth rates, environmental gra- dients, and differences in the relative scales of resource patches and foraging ranges of consumers. Key words: algal productivity; bottom-up; food chain model; herbivory; macroalgae; nutrients; rocky intertidal; seaweeds; tide pool; top-down; wave exposure. INTRODUCTION Ecological theory has not always embraced the idea that top-down and bottom-up forces are inextricably linked to produce patterns in community structure. In 1960 Hairston, Smith, and Slobodkin’s top-down view of terrestrial community structure provided a simple but elegant theoretical resolution to the historical de- bate between the proponents of competition vs. pre- dation structured populations, but subsequently spawned debate over the prevalence of bottom-up vs. top-down factors in structuring communities (Fretwell 1977, White 1978, Oksanen et al. 1981, Hunter and Price 1992, Power 1992, Strong 1992, Polis 1994). For Manuscript received 6 July 1999; revised 5 July 2000; ac- cepted 21 July 2000. 1 E-mail: nielsenk@bcc.orst.edu example, Hunter and Price have argued (1992) argued that bottom-up factors necessarily set the stage upon which all biological interactions are carried out, be- cause in the extreme case, without primary producers, there is no community. The basic, Lotka-Volterra type top-down model, modified by Fretwell (1977) and Oks- anen et al. (1981) to include variation in primary pro- ductivity, predicts that increasing primary productivity will cause the top trophic level in a system to increase in abundance, as will alternate levels below it, but not the intervening levels. Thus, according to the models, primary productivity can influence both the number of trophic levels, and the absolute abundance of organisms on those levels that are controlled by their resources. However, removal of the top trophic level in the system continues to result in the classic trophic cascade,