The bee family Andrenidae includes nearly 3000 species of pollen-collecting species, most of which are specialists of particular host flowers. Andrenidae is a widely distributed family comprising approximately 40 genera, two of which are exceptionally diverse: Andrena, the most species-rich bee genus in the Holarctic, and Perdita, the most species-rich bee genus in North American deserts. This talk will summarize my dissertation research and subsequent research on the systematics of the Andrenidae, focusing on new hypotheses of phylogenetic relationships among andrenid subfamilies, tribes, and genera based on combined cladistic analysis of morphological data and DNA sequences for the nuclear gene EF-1alpha. Phylogenetic signal from adult morphological, larval morphogical, and DNA sequence data are generally congruent and together support monophyly of family Andrenidae and the following novel hypothesis of subfamilial relationships: (Andreninae (Oxaeinae Panurginae)). New tribal and generic classifications proposed for each subfamily will be summarized, as will evidence for expansion of subfamily Panurginae to include an unusual bee recently discovered in the Atacama Desert of Chile. Biogeographic implications will also be discussed, with emphasis on amphitropical disjunctions between southern South America and western North America, and disjunctions between the Eastern and Western Hemispheres. In addition, revised evolutionary hypotheses for morphological structures such as facial foveae will be presented. Finally, I will briefly summarize preliminary results from ongoing collaborative phylogenetic studies of Andrena and Bombus (Apidae), and from my current research on the evolution of pollen host specialization in Panurginae.