delay protein synthesis unduly if an empty adaptor (i.e. without its amino acid) could fit onto the template, and there was nothing in the original theory to prevent this. This difficulty cam be got round im @ mumber of ways. The obvious thing bo @o te 06 Gover Un vine speertic bace- sequence, needed for base pairing with the template, until the amino acid is attached. One rather neat way of doing this suggests itself. This is to make the soluble RNA fold back on itself to form a DNA-like structure. (We know from studies ot "Poly A plus Polly U" that a polynucleotide with a ribose backbone can take up a configuration like DNA). This requirement could well account for the length of the RNA being greater than expected. The idea that soluble RNA may have some secondary structure is not new. Die, Peart Berg thought at one time that he had some evidence for iy vend so, more recently, have others) and it had been suggested to us in discucsion by Dr. Boman and by Ue) Bach. The only novelty in our idea is that it suggests a reason for such a secondary SUrUCIUre, Ti 25 exmsus. The idea, then, is that without the amino acid the molecules of soluble RNA fold on themselves so that their "Adaptor sequence” is covered. At some time after the amino acid is attached this sequence is: uncovered, and can then base- paid with the template RNA of the ribosomes. Is there any evidence that soluble RNA has secondary structure? This is too recent and complex a subject to discuss here, but we may note that the base composition of soluble RNA, determined by three laboratories (Kendrick Smith, David Dunn and Jim Offengand, with their various collaborators - personal communications) have shown that the base ratios tend towards the DNA rule; that is a =: U. plus pseudo-U and GG == ¢. (Incidentally if soluble RNA resembles DNA it may act as a template for its own replication). There is one difficulty about this idea which is imaginary. This is: how does the activating enzyme recognise the correct molecules of S=RNA if the adaptor eo =