Ten years ago, the late Dr John Brownlee published in The Lancet a preliminary note giving the reasons which led him to conclude, provisionally, that epidemics of influenza tended to occur at intervals of approximately 33 weeks. In the following year two other short notes were published in the same journal by Mr Spear and by Dr Stallybrass. Dr Stallybrass, from an analysis of Liverpool data (not very fully set out in his paper), confirmed Brownlee's conclusion, but Mr Spear, using the London data, criticised Brownlee's inferences and methods, pointing out—inter alia—that an interval between successive epidemics of the order of 52 weeks agreed better with the observations than an interval of 33 weeks or a multiple thereof. In an appendix to the Report on the Pandemic of Influenza 1918–19 issued by the Ministry of Health, Dr Brownlee reaffirmed his original contentions, published some further collateral evidence and replied to Mr Spear's methodological criticism. That was, I think, the end of Dr Brownlee's connection with the subject; to his death he retained, I think, a half-serious faith in the 33 weeks' rule but did not regard the matter as of sufficient epidemiological interest to justify further investigation.