A key step in retrieving the information stored in the complex genomes of eukaryotes involves the identification of transcription units and, more specifically, the recognition of promoter sequences by RNA polymerase. In eukaryotes, the task of recognizing nuclear gene promoters and then transcribing the genes is divided among three highly related enzymes, RNA polymerases I, II, and III. Each of these RNA polymerases is dedicated to the transcription of specific sets of genes, and each depends on accessory factors, the so-called transcription factors, to recognize its cognate promoter sequences.RNA polymerase I is unique among the nuclear RNA polymerases in transcribing only one set of genes, the large, tandemly repeated, ribosomal RNA genes, and thus in having to recognize a single promoter structure. RNA polymerase II transcribes the protein-coding genes (mRNA genes) as well as some small nuclear RNA (snRNA) genes. The RNA polymerase II promoters can be divided into a core region, defined as the minimal region capable of directing transcription in vitro, and a regulatory region. The regulatory regions are highly varied in structure, reflecting the highly varied synthesis patterns of cellular proteins and the need for exquisite and complex regulation of these patterns. The core promoters themselves come in different types that, in mRNA-encoding genes, can contain a TATA box, an initiator, a downstream promoter element, or various combinations thereof. The assembly of a functional RNA polymerase II transcription complex on a promoter consisting of just a TATA box has been extensively studied. All the factors involved in the process have been identified, and much is known about how these factors interact with DNA and with each other to recruit, eventually, RNA polymerase II (for reviews, see Orphanides et al. 1996;Woychik and Hampsey 2002). How RNA polymerase II transcription complexes assemble on TATA-less promoters is, however, not as well understood.RNA polymerase III is dedicated to the transcription of an eclectic collection of genes whose main common features are that they encode structural or catalytic RNAs and that they are, as a rule, shorter than 400 base pairs (bp). This length limit is consistent with the elongation properties of RNA polymerase III, which recognizes a simple run of T residues as a termination signal. The genes transcribed by RNA polymerase III encode RNA molecules involved in fundamental metabolic processes, specifically components of the protein synthesis apparatus and components of the splicing and tRNA processing apparatus, as well as RNAs of unknown function. The RNA polymerase III promoters are more varied in structure than the uniform RNA polymerase I promoters, and yet not as diverse as the RNA polymerase II promoters. They have been divided into three main types, two of which are gene-internal and generally TATA-less, and one of which is gene-external and contains a TATA box. Remarkably, we have a good, and in some cases a very detailed, understanding of how RNA polymerase III ...