Which traits of species predict population declines in experimental forest fragments?

Theory suggests that species with particular traits are at greater risk of extinction than others. We assumed that a decline in abundance in forest fragments, compared to continuous forest, equated to an increase in extinction risk. We then tested the relationships between five traits of species and decline in abundance for 69 beetle species in an experimentally fragmented forest landscape at Mt. Wog Wog in southeastern Australia. The experiment was controlled and replicated. Monitoring ran for two years before forest fragmentation; in this paper, we examine data for five years postfragmentation. We tested five hypotheses: (1) Species that occur naturally at low abundance are more likely to decline as a result of fragmentation than are abundant species. (2) Isolated species are more likely to decline than species that are not isolated. (3) Large species are more likely to decline than small species. (4) Species in trophic groups at the top end of food chains are more likely to decline than species in trophic groups lower in the food chain. (5) Because traits are often shared by related species, populations of more closely related species will respond in the same way. We found that: (1) rare species were more likely to decline than abundant species; (2) isolated species were more likely to decline than species that were not isolated; (3) body size was not correlated with response to fragmentation; (4) among species that declined, predators declined most; and (5) taxonomically related species did not respond in the same way to fragmentation. Thus, our results confirm theories predicting that isolated, rare, or predaceous species will be lost first from fragmented landscapes.