The polycyclic ethers represented by brevetoxins (BTXs) are unique dinoflagellate metabolites, and the mechanism of their biosynthesis has been of great interest. More than ten years ago, our group and Nakanishi’s group independently reported the unusual features of the biosynthesis of these dinoflagellate polyethers. In their reports, it was suggested that brevetoxins are aberrant polyketides, in which methyl side chains are derived from both methionine and acetate. This dual mechanism of methyl introduction has been recently questioned, since similar aberrant dinoflagellate polyketides, okadaic acid derivatives were found to receive their methyl groups exclusively from acetate units, and the pathway was considered to be characteristic of dinoflagellate polyketide biosynthesis. In fact, 13C-labelling studies of BTX biosynthesis accompanied considerable randomization, which left certtain ambiguity. In order to reinvestigate the biosynthesis, we made a large number of new isolates of Gymnodinium breve from the existing culture, and selected those isolates which caused the least randomization. Using those strains, we unequivocally confirmed our previous finding; four methyl groups attached to C-8, C-22, C-25 and C-36 in BTX-B are indeed derived from methionine methyl, and the rest from acetate methyl. We also confirmed the previously reported incorporation patterns of labelled acetates, and, in addition, were able to obtain new information regarding the C-C connectivity and incorporation rate. Based on these results, we arrived at a new interpretation of BTX biosynthesis, which also may apply to other dinoflagellate polyketides in general . Supported by NIH grant GM28754.