1. Rapidly labelled RNA from Escherichia coli K 12 was characterized by hybridization to denatured E. coli DNA on cellulose nitrate membrane filters. The experiments were designed to show that, if sufficient denatured DNA is offered in a single challenge, practically all the rapidly labelled RNA will hybridize. With the technique employed, 75-80% hybridization efficiency could be obtained as a maximum. Even if an excess of DNA sites were offered, this value could not be improved upon in any single challenge of rapidly labelled RNA with denatured E. coli DNA. 2. It was confirmed that the hybridization technique can separate the rapidly labelled RNA into two fractions. One of these (30% of the total) was efficiently hybridized with the low DNA/RNA ratio (10:1, w/w) used in tests. The other fraction (70% of the total) was hybridized to DNA at low efficiencies with the DNA/RNA ratio 10:1, and was hybridized progressively more effectively as the amount of denatured DNA was increased. A practical maximum of 80% hybridization of all the rapidly labelled RNA was first achieved at a DNA/RNA ratio 210:1 (+/-10:1). This fraction was fully representative of the rapidly labelled RNA with regard to kind and relative amount of materials hybridized. 3. In competition experiments, where additions were made of unlabelled RNA prepared from E. coli DNA, DNA-dependent RNA polymerase (EC 2.7.7.6) and nucleoside 5'-triphosphates, the rapidly labelled RNA fraction hybridized at a low (10:1) DNA/RNA ratio was shown to be competitive with a product from genes other than those responsible for ribosomal RNA synthesis and thus was presumably messenger RNA. At higher DNA/rapidly labelled RNA ratios (200:1), competition with added unlabelled E. coli ribosomal RNA (without messenger RNA contaminants) lowered the hybridization of the rapidly labelled RNA from its 80% maximum to 23%. This proportion of rapidly labelled RNA was not competitive with E. coli ribosomal RNA even when the latter was in large excess. The ribosomal RNA would also not compete with the 23% rapidly labelled RNA bound to DNA at low DNA/RNA ratios. It was thus demonstrated that the major part of E. coli rapidly labelled RNA (70%) is ribosomal RNA, presumably a precursor to the RNA in mature ribosomes. 4. These studies have shown that, when earlier workers used low DNA/RNA ratios (about 10:1) in the assay of messenger RNA in bacterial rapidly labelled RNA, a reasonable estimate of this fraction was achieved. Criticisms that individual messenger RNA species may be synthesized from single DNA sites in E. coli at rates that lead to low efficiencies of messenger RNA binding at low DNA/RNA ratios are refuted. In accordance with earlier results, estimations of the messenger RNA content of E. coli in both rapidly labelled and randomly labelled RNA show that this fraction is 1.8-1.9% of the total RNA. This shows that, if any messenger RNA of relatively long life exists in E. coli, it does not contribute a measurable weight to that of rapidly labelled messenger RNA.