The work of Shwartzman (1) on toxic filtrates of organisms of which heretofore no specific toxic filtrate had been definitely shown, suggested testing B. pertussis for toxin production. It occurred to us that the failure of all previous attempts to clearly demonstrate a specific pertussis toxin may have been due to a combination of two factors: (a) the method of growing the organisms for toxin production, (b) the method of testing for toxicity. The present investigation was undertaken with the above two points in view. We selected two strains of B. pertussis for the experiment. One of them was recently obtained from Bordet and was designated “Bordet,” and the other was one of our old stock strains; both of them were type B. In our preliminary tests we grew the organisms in the same way as for a vaccine.