The haemagglutinin content of monovalent influenza whole virus and Tween-ether split vaccines derived therefrom, were assayed comparatively using single radial immunodiffusion (SRID, the only test recommended for influenza vaccines by the European Pharmacopoeia Commission), quantitative SDS-polyacrylamide gel electrophoresis and immunization of guinea pigs. If SRID was performed with split vaccines, reduced haemagglutinin values were consistently recorded which were 50-25% of values obtained before disruption of virions. If, however, disruption was conducted in the presence of excess detergent thus preventing aggregate formation of solubilized haemagglutinin, test values comparable to those of whole virus vaccines were obtained. In agreement with these results, immunization experiments revealed that whole virus and the corresponding split vaccines exhibited comparable immunogenicity in guinea pigs. From SDS-polyacrylamide gel electrophoresis and densitometer tracings obtained by scanning the gels after staining with either Coomassie Blue or fluorescein isothiocyanate-labelled concanavalin A it was calculated that about 90% of whole virus HA2 was recovered in Tween-ether split vaccines. From our experiments we conclude that precise quantification of solubilized haemagglutinin is not achievable by the single radial immunodiffusion test alone. Aggregate formation of solubilized haemagglutinin frequently occurs when the applied detergent is removed and, therefore, a physico-chemical method including an effective disaggregation procedure like SDS treatment in combination with PAGE is recommended.