One of the basic truths to have emerged from the short history of eukaryotic molecular biology-and probably the main reason that the field is so excitingis that things frequently turn out very differently than anticipated. There are any number of examples of this, including genomic DNA rearrangements, 5’ capping, splicing and transsplicing, and non-template-encoded nucleotide incorporation. But almost nowhere has the truth,of “expect the unexpected” been more evident than with transcription by RNA polymerase III (pol Ill). Promoters for transcription by pol Ill were the first eukaryotic control regions to be carefully mapped, and based on prior results with prokaryotic RNA polymerases many investigators believed that they would also reside upstream of the initiation site. The first major surprise came when the control region of the Xenopus 5s gene was found to be within the coding region (Sakonju et al., Cell 79, 13-25, 1980); in this classic example, the sequences essential for transcription are actually closer to the 3’ than the 5’ end of the gene. Sequence elements necessary and sufficient for transcription of the adenovirus-encoded VAI gene and several tRNA genes were also shown to reside within the gene body (reviewed in Ciliberto et al., Curr. Topics Dev. Biol. 78, 59-88, 1983). Only after the “fact” of gene-internal pol Ill promoters was well established in textbooks did it become generally recognized that crucial elements of pol III promoters can also reside external to the coding region. First considered almost an oddity of the Bombyx system, the requirement for upstream sequences (in conjunction with geneinternal promoter domains) has now been documented in a variety of organisms and for numerous pol Ill-transcribed genes, yncluding those for several tRNAs, 5s RNA, and 7SL RNA (e.g., Larson et al., PNAS, 80, 3416-3420, 1983; Johnson and Raymond, JBC 259, 5990-5994, 1984; Morton and Sprague, PNAS 87, 5519-5922, 1984; Ullu and Weiner, Nature 378, 371-374, 1985). In at least some instances, the need for these external promoter domains (located within ~25 bp upstream of the initiation site) only becomes apparent under stringent reaction conditions. And now we have come full circle, with the again unexpected revelation that promoter domains essential for pol Ill transcription can also reside entirely upstream of the gene body. Deletion derivatives that lack the entire region downstream of the initiation site of the 7SK gene and the U6 gene can still be actively transcribed (Murphy et al., Cell 57,81-87, 1987; Das et al., EMBO J. 7, 1988, in press). These genes do contain a box A sequence-a geneinternal element that, in conjunction with box B (tRNA) or box C (5s) has been present in all pol Ill promoters (see Minireview