Abstract
Functional trait diversity is increasingly used to model future changes in community structure despite a poor understanding of community disassembly's effects on functional diversity. By tracking the functional diversity of the North American large mammal fauna through the End-Pleistocene megafaunal extinction and up to the present, I show that contrary to expectations, functionally unique species are no more likely to go extinct than functionally redundant species. This makes total functional richness loss no worse than expected given similar taxonomic richness declines. However, where current species sit in functional space relative to pre-anthropogenic baselines is not random and likely explains ecosystem functional changes better than total functional richness declines. Prehistoric extinctions have left many extant species functionally isolated and future extinctions will cause even more rapid drops in functional richness.