An HPLC method was developed to determine simultaneously in a single analysis prostaglandin E2, prostaglandin E3, tetranor-prostaglandin E1 and delta 17-tetranor-prostaglandin E1 in rat urine. As internal standard omega-nor-prostaglandin E2 was added to the samples at the beginning of the analysis. The assay was applied in feeding experiments in which rats were fed diets with mixtures of (n-6) and (n-3) fatty acids (linoleate, arachidonate, alpha-linolenate and eicosapentaenoate). The level of urinary prostaglandin E2 was not very much affected by the diet and prostaglandin E3 could never be detected in significant amounts. These primary prostaglandins are assumed to reflect prostaglandin biosynthesis in the kidney medulla. Tetranor-prostaglandin E1 is a characteristic urinary metabolite of prostaglandin E2 in the rat; its level increased after feeding arachidonic acid. delta 17-Tetranor-prostaglandin E1 became a major metabolite when the rats received eicosapentaenoic acid. However, we found that the ratio of urinary tetranor-prostaglandin E1/delta 17-tetranor-prostaglandin E1 is not a very reliable measure of the ratio of prostaglandin E2/prostaglandin E3 formed in the body, because prostaglandin E3 is converted to a much greater extent into delta 17-tetranor-prostaglandin E1 than is prostaglandin E2 into tetranor-prostaglandin E1. As a matter of fact, incubations of tissue homogenates of rats resulted always in predominant formation of prostaglandins of the 2-series, even after high eicosapentaenoate diets. We conclude, in agreement with work carried out earlier, that biosynthetic pathways leading to prostaglandins of the 3-series are of minor importance.