|Interspecific interactions: constructing a general, neutral model for interaction type
|Year of Publication
Experiments have not been devised to provide unbiased estimates of the relative co-occurrence of all different interaction types in communities. Thus, we have no idea what types of interspecific interactions to expect if we randomly sample species from an environment. The probability that a unidirectional, non-zero interaction is negative (P-) or positive (P+) has a direct effect on the expected types of reciprocal interspecific interactions. Analysis of field studies where strengths and signs of interactions between seven or more species were documented (large studies were selected to avoid bias from choosing species known to interact) suggests that P+=P-=0.5. Data also suggest that the proportion of non-zero interactions (C) is generally less than 0.6 when 10 or more interactions are considered. The proportion of expected reciprocal interspecific interactions can be calculated given the proportion of direct interactions that are not significantly different from zero and assuming that non-zero interactions are equally likely to be positive or negative. In general, this neutral model predicts that if species are chosen from an environment randomly, non-interaction (0/0), commensalism (+/0), or amensalism (-/0) will be most common when C < 0.6. Predation and functionally similar interactions (+/-), competition (-/-), and mutualism (+/+) should be relatively rare when there are less than 10 species with C < 0.6. Predation should be twice as common as competition or mutualism, and mutualism as common as competition. The neutral model suggests that community ecologists place undue emphasis on competition and predation.