DO ALTERNATE STABLE COMMUNITY STATES EXIST IN THE GULF OF MAINE ROCKY INTERTIDAL ZONE? COMMENT

Petraitis and Latham (1999) hypothesized that rockweed stands and barnacle-mussel beds on sheltered rocky shores in New England may represent alternative states and outlined experimental tests suitable to address this question. They advanced this hypothesis because mussels and the seaweed Ascophyllum nodosum are often codominant in sheltered bays (Lewis 1964), and their co-occurrence is not adequately explained by the standard paradigm of consumer control (Menge 1976, Lubchenco and Menge 1978). Petraitis and Latham (1999) also suggested that the development of alternative states depends upon the timing and spatial scale of pulse perturbations. In Petraitis and Dudgeon (1999) and Dudgeon and Petraitis (2001), we reported the results of experiments that show patch size determines the rates of predation and recruitment. Our results provide evidence consistent with the formation of alternative communities on rocky shores of sheltered bays. Bertness et al. (2002) extended and tested our hypothesis about alternative states to other intertidal habitats in New England. They concluded that rockweed stands and barnacle-mussel beds are highly deterministic systems driven by the interaction of water flow, larval input, and consumer pressure. They suggested the type of habitat in which mussels and seaweeds exist as alternative states is likely to be rare. We are delighted that Bertness et al. have conducted experiments to test if mussel beds and rockweed stands are alternative states. We agree that it is crucial to duplicate experiments so that ecologists can move beyond debates of "different ecologies or different ecologists" (Underwood and Fairweather 1986). We applaud such efforts. Our applause, however, is muted because Bertness et al.'s (2002) experiments do not replicate our design as claimed and fall short of being valid tests of alternative stable states. We first compare their design with our design because the problems with their study as a test of alternative stable states arise from the design itself. We then discuss the criteria outlined by Petraitis and Latham (1999) for testing the hypothesis of alternative stable states and the limitations of Bertness et