The messenger RNAs present in HeLa cells late in adenovirus 2 infection have been characterized utilizing three complementary methods of analysis. In each case the mRNAs were defined functionally, by identification of the polypeptides they encode when translated in a reticulocyte cell-free system. The RNAs were examined by (1) electrophoretic fractionation in agarose gels containing methyl-mercury hydroxide to estimate the sizes of the functional RNAs; (2) selection by annealing to viral DNA fragments immobilized on nitrocellulose to define the genomic origin of their constituent sequences; (3) hybrid-arrested translation using viral DNA fragments to locate their protein coding sequence. The results allow the positioning of both the coding and non-coding regions of all the late mRNAs with respect to the DNA. They demonstrate that the mature mRNAs originating in the major late transcription unit fall into five families. Within each family the mRNAs overlap in a staggered fashion, sharing the same 3′ terminus but differing at their 5′ termini. The data also reveal the existence of two new late proteins of 52,000 and 55,000 Mr encoded between map co-ordinates 30 to 34 and permit the positioning of the late RNAs, according to the polypeptides they encode, as follows: 52, 55K · IIIa: III · pVII · V: pVI · II: 100 K/33 K · pVIII: IV. (The colons signify the ends of families; centre dots the junctions within families; the slash an overlap of coding sequences; and the comma, coding sequences which have not been separated in this analysis. A 55 K protein has Mr = 55,000.) Those mRNAs which are either the only transcripts from a specific region (such as those for polypeptides IX, IVa2 and IV) or are the 3′-terminal transcripts within a family (such as those coding for polypeptides IIIa, V, II and pVIII) migrate as discrete species upon electrophoresis in agarose gels containing methyl-mercury hydroxide. However, the 5′-proximal RNAs within a family (such as those encoding polypeptides 52, 55 K, III, pVII, pVI, 100 K and 33 K) exhibit a broad distribution of molecular weights.