The biphasic nature of the 1957-1958 influenza epidemic and its contribution to the mortality during those years are revealed by data obtained in New York City. Hemagglutination-inhibition (HI) tests for Asian influenza were carried out from Sept. 16, 1957, to Feb. 21, 1958, on 2,719 serums negative for syphilis. The percentage of persons with HI antibodies positive to the NYC/1/57 influenza virus increased abruptly from a low figure to about 35% during October; the increase coincided with a first wave of deaths in excess of the normal seasonal expectation. The percentage of serologically positive persons showed a second abrupt increase in January, again simultaneous with an excessive mortality for the time of year. Between the two waves there was a period of quiescence of approximately two months. In January, 1958, a strain of virus was isolated that differed markedly from all previously tested strains, and three more strains of the new type were isolated in February. These facts suggest the possibility that older strains may become altered, or re-emerge after the lapse of years, or somehow recombine the antigens of earlier with those of more recent strains.