Both our analysis (Kerr et al. 2006), and Lees and Colwell's (2007) reanalysis, of patterns of bird and mammal diversity on Madagascar show that the central peak of richness predicted by the Mid-Domain Hypothesis (MDH) is not observed. Lees and Colwell emphasize an observation consistent with MDH predictions: a latitudinal mid-domain richness peak in the rainforest biome. They find (but do not mention) that no analogous peak is observed in the other two main Madagascan biomes. MDH fails nearly all its tests in Madagascar.