JOURNAL OF LINGUISTICS select two animals that appear in the narratives, such as a jaguar and a stag, and say they are transformations. The remaining papers are not entirely uninteresting. Turner, for instance, more likely has something. At least a certain amount of culture learning must have taken place during his ethnographic field work so that his interpretations are presumably closer to the mark than those someone might write who had never lived with the people. But Turner provides us with no procedures, such as those developed by Metzger and Williams, for dealing formally with the data or testing it either with informants, or by other means. Leach's writing rearranges the interpretations of Radcliffe-Brown on Andamanese myths. He himself admits that whether or not a 'structuralist' approach gains insight in the rearranging of the 'functionalist' study by Radcliffe is 'a matter of opinion'. He then goes on to say, 'The purpose of this essay has simply been to show that the two frames of reference are not all that different, though I have to admit that if we are to read the works of the classical functionalists with structuralist understanding, a great deal of verbal euphoria has to be shorn away.' The book ends with a pretentious display that underlines an attitude among many analysts of traditional narratives which is difficult for me to understand. Four variants of a Sherente myth were sent to the contributors of the volume with an invitation to provide analyses and comments. The pretension here is that myth analysis has reached a sufficiently advanced state of scientific sophistication as to enable one or more myths to be subjected to various analyses in some sort of 'blind' situation, as though some sort of scientific control were being applied. Such analyses would then be uncontaminated by the influence of collateral ethnographic (or even narrative) data which was not accessible to the analysts. While those who took the bait all seemed to agree on the usefulness of collateral information, they nevertheless proceed to rush in with an analysis-all except Dundes, who refuses to play the game. Maybury-Lewis, in all his ethnographic wisdom, proceeds to give us his own discussion of the stories, cluing us in on the way things really are for the Sherente. The pronouncements of the ethnographer are the final test: 'Thus myths collected by members of the Harvard-Central Brazil project are being formally analysed and the results of the analysis will then be put to the ethnographers for confirmation.' Reviewed by B.