Abstract Aim To use published distributional and taxonomic information on Neotropical birds in a cladistic analysis of distributions and endemism (CADE) to generate a testable hypothesis of area-relationships for the arid areas of endemism, particularly those of Central South
America (the ‘arid diagonal’), and to clarify the different methodologies commonly associated with parsimony analysis of endemicity (PAE). Location South America. Methods Cladistic analysis of distributions and endemism. Results We obtain
a tree where the relationships of most areas are resolved, and we find support for an exclusive clade of Central South American areas, with the Caatinga as sister to both the Chaco and Cerrado. Main conclusions There is a substantial amount of historical signal in avian distributions,
when large numbers of taxa and multiple taxonomic levels are considered. Ecological noise and historical information are more easily distinguished in CADE analyses than they would be in PAE analyses. Based on our results we predict that among aridland birds, the Cerrado and Chaco species will
be more closely related to each other than to Caatinga species.