The histone code hypothesis proposes that covalently modified histone tails are binding sites for specific proteins. In vitro evidence suggests that factors containing bromodomains read the code by binding acetylated histone tails. Bromodomain Factor 1 (Bdf1), a protein that associates with TFIID, binds histone H4 with preference for multiply acetylated forms. In contrast, the closely related protein Bdf2 shows no preference for acetylated forms. A deletion of BDF1 but not BDF2 is lethal when combined with a mutant allele of ESA1 (a histone H4 acetyltransferase) or with nonacetylatable histone H4 variants. Bromodomain point mutations that block Bdf1 binding to histones disrupt transcription and reduce Bdf1 association with chromatin in vivo. Therefore, bromodomains with different specificity generate further complexity of the histone code.