We did not, however, observe an association between ENM contrasts and mean relative emergence for parasites of maize (Fig. 3I). Given that relative crop area harvested did predict emergence on maize (Fig. 3C; Table S2), it may be that ENM contrasts perform poorly if alternate hosts are common. Alternate hosts may be particularly common in areas with the highest quality environments, which are also those predicted to have highest habitat suitability for parasites of maize (Fig. 4). For example, ENM contrasts predicted high suitability for maize specialization in central Nigeria (Fig. 5). However, sorghum was the dominant crop in this area in 2000 and other hosts are commonly grown (Fig. S1); indeed, most tested parasite populations from Nigeria appear to be maize/sorghum generalists (Fig. 2).  In addition to increased diversity of hosts in higher quality environments (Thrall et al. 2007), the observed lack of parasite specialization on maize could result from coevolutionary history.  Both sorghum and pearl millet originated in Africa, with evidence for centers of domestication for sorghum in eastern Sudan (Winchell et al. 2017) and for pearl millet in the western Sahara (Burgarella et al. 2018). In contrast, maize was domesticated in Mexico about 9,000 years ago (Matsuoka et al. 2002, Piperno et al. 2009) and improved varieties bred specifically for African environments were not available until the 1980s and 1990s (Evenson and Gollin 2003).