Pertussis toxin (PT) in its detoxified form is an important antigenic component of both acellular and whole cell pertussis vaccines. Limits on the content of active PT in acellular vaccines are set in official monographs (EP, WHO, USP) and evidence of compliance is therefore, required by regulatory authorities. The two assay methods which are currently used by most manufacturers and official national control laboratories to monitor residual PT activity in acellular pertussis vaccines (and also in whole cell vaccines) are histamine sensitising (HIST) assays and Chinese hamster ovary (CHO) cell assays. Currently, different reference preparations of PT are used by individual laboratories for these tests. We therefore organised an international collaborative study to examine, by these two assay methods, two freeze-dried purified preparations of PT, one preparation in ampoules coded JNIH-5 and one preparation in ampoules coded 90/518, together with in-house reference (IHR) preparations in current use. Data from this study confirm that both JNIH-5 and 90/518 show biological activity both in HIST assays and in CHO-cell assays. Both HSD50 and ED50 values obtained in this study differ significantly between laboratories and thus show that biological activity is not determined by the nominal masses of preparations. Estimates of relative potency of 90/518 in terms of JNIH-5 per ampoule for the HIST assays do not differ significantly between laboratories. The overall mean estimates of relative potency of 90/518 in terms of JNIH-5 do not differ significantly between the two methods. Data from this study further indicate that the biological activity of different preparations was not directly related to their stated protein content. The use of protein content to indicate the level of PT activity in different preparations would give misleading results. Thus, use of a common standard is shown to greatly improve between laboratory agreement of estimates.