The prevailing hypothesis concerning the ecology of Coccidioides immitis and C. posadasii is that these human pathogenic fungi are soil fungi endemic to hot, dry, salty regions of the New World and that humans and the local, small-mammal fauna are only accidental hosts. Here we advance an alternative hypothesis that Coccidioides spp. live in small mammals as endozoans, which are kept inactive but alive in host granulomas and which transform into spore-producing hyphae when the mammal dies. The endozoan hypothesis incorporates results from comparative genomic analyses of Coccidioides spp. and related taxa that have shown a reduction in gene families associated with deconstruction of plant cell walls and an increase in those associated with digestion of animal protein, consistent with an evolutionary shift in substrate from plants to animals. If true, the endozoan hypothesis requires that models of the prevalence of human coccidioidomycosis account not only for direct effects of climate and soil parameters on the growth and reproduction of Coccidioides spp. but also consider indirect effects on these fungi that come from the plants that support the growth and reproduction of the small mammals that, in turn, support these endozoic fungi.