The pre-embryos of a number of species produce platelet-activating factor (PAF). This study compared the sensitivity, reproducibility and cross-reactivity of a radioimmunoassay (RIA) for PAF with a quantitative PAF bioassay and assessed the use of both assays for the measurement of murine and human embryo-derived PAF. PAF was extracted and partially purified by thin-layer chromatography (TLC) from medium in which mouse (30 x 2-cell) or individual human embryos [produced by in-vitro fertilization (IVF)] had been cultured for 24 h. Dose-response curves were generated with synthetic PAF standards over a concentration range of 0.3-30 ng/ml. The RIA was slightly more sensitive than the bioassay since the concentration which gave 50% response in the RIA was 3 ng/ml compared to 6 ng/ml in the bioassay. The RIA showed greater reproducibility than the bioassay, since at these concentrations the coefficients of variation were 1.97 and 6.77% intra-assay and 4.96 and 10.93% inter-assay, respectively. Both assays had a detection limit of 0.3 ng/ml. C-PAF, a non-metabolizable bioactive analogue of PAF, showed 5.16 and 0.85% cross-reactivity in the bioassay and RIA, respectively, while lyso-PAF and phosphatidylcholine had no effect. Following extraction and TLC, the average PAF concentration was 14.52 ng/ml (range 0.10-141.09) and 20.06 ng/ml (range 0.07-106.15) for mouse embryos, and 16.15 ng/ml (range 0.40-78.31) and 23.70 ng/ml (range 2.00-119.10) for human embryos, as measured by bioassay and RIA respectively. Therefore, both the bioassay and RIA were capable of the sensitive quantitative measurement of embryo-derived PAF.