In the first phase of a two part WHO Collaborative study, fourteen laboratories from ten countries estimated the antigenic content of six trivalent inactivated poliovirus vaccine preparations using in vitro methods. All laboratories used a candidate standard method for D antigen assay (method A) and eight contributed results from established 'in-house' methods (method B). All methods assayed D antigen in an antigen capture ELISA format. Monoclonal antibodies were used as detector reagents in method A and in some laboratories for method B. The average difference in potency estimates for duplicate preparations A and C was used to assess within assay variation. Overall this was found to be 22% and 19% for methods A and B respectively. Within laboratory variation was measured as the geometric coefficient of variation for between assay repeatability. Results for methods A and B, 28% and 26% respectively were again very similar. Variation in potency estimates between laboratories was in the range 2- to 5-fold for most samples and most laboratories irrespective of the method used. However, a maximum 24-fold difference occurred when all results were taken into account. Method A gave significantly enhanced potency estimates for the type 3 component of preparation B, a vaccine shown to be immunogenic in humans in clinical trials, compared to method B. Method A also failed to assay the type 3 component of preparation F which was prepared by inactivation of the Sabin 3 strain of poliovirus. Further work is required to identify monoclonal antibodies, or combinations of monoclonal antibodies, suitable for universal application in D antigen assays of inactivated poliovirus vaccines. Further work is also required to improve control of the antigen-capture ELISA in some laboratories. The second phase of this WHO Collaborative Study evaluated the proficiency of in vivo potency assays. These results together with an evaluation of the correlation of immunogenicity and antigenic assay plus assessment of candidate reference materials will be reported separately.