The author reports on the effect of a previous attack of influenza caused by virus A1 on susceptibility to virus A2 in the light of analyses made of the rate of sickness from influenza and acute catarrhs of the respiratory tract in a ball-bearing factory in the USSR in 1957. There were spring and summer rises in sickness and an autumn epidemic.In the summer rise, caused by A2 virus, the sickness rate for those who had been ill in the spring (with A1 influenza) was about half (4.7%) that for those who had not been affected in the spring (10.5%), but this was clearly an unstable and short-lived immunity as by the time of the autumn A2 epidemic the rate for those sick in the spring rose to 21.1% although this was still 1.6 times lower than that for those who had not been ill either in the spring or in the summer.Vaccination with a polyvalent vaccine (A, A1 and B strains) was partially effective in the summer epidemic, but slightly less so than seroprophylaxis with the same strains conducted in October. This also seems to indicate the working of the same phenomenon of partial immunity derived from an old variety of influenza virus of the same serological type as that causing an outbreak not long beforehand.