(8,9) or DNA recombinants (10, 11). However, the task of assigning these polypeptides to specific functions has just begun. Indeed, until now, no poxvirus proteins with known functions have been mapped by molecular or classical genetic methods.Early mRNAs of vaccinia virus that have been examined in detail do not appear to be spliced or shortened at their 5' ends (12-15). For one early mRNA, the nucleotides within the 60-bp sequence immediately upstream ofthe initiation site are 89% adenylate and thymidylate, and no homologies to prokaryotic or eukaryotic sequences were found further upstream (16). This finding led to the proposal that the vaccinia virus RNA polymerase (17, 18) has its own novel DNA recognition sequences.Although additional early vaccinia virus genes are being examined to develop a consensus sequence, such information alone will be insufficient to establish functional significance.In vitro mutagenesis provides a direct way ofcorrelating the two parameters: nucleotide sequence and function. Such studies are facilitated by the use of selectable genetic markers. For vaccinia virus, the most suitable marker is thymidine kinase (TK). Dubbs and Kit (19) demonstrated that TK activity increases after infection of TKV or TK-mouse L cells with wildtype vaccinia virus but not with a TK-mutant, providing evidence that the enzyme is virus-coded. However, no information existed regarding the genetic locus of the vaccinia TK. For herpesvirus, which also encodes TK (20)