The effectiveness of influenza vaccination has been studied in an old age colony during the winters 1977-78 and 1978-79. The subjects, vaccinated and non-vaccinated, were submitted to continuous clinical and virological controls for about 6 months during the cold season. Three blood samples were obtained from the vaccinated subjects: 2 in November (before and after vaccination) and 1 in May of the following year. From the non-vaccinated 2 blood samples were obtained in November and May, respectively. The hemagglutino-inhibition test (HAI) permitted studying the serological response after vaccination, and verifying changes of the antibody titer to be related to influenza virus infections, contracted during the winter, which had escaped the clinico-virological detection. The serological response to the bivalent vaccine employed in November 1977 (A/Victoria/3/75 - 600 I.U.; B/Hong Kong/8/73 - 300 I.U.) was good for type A virus, and fair for B type. The low circulation of influenza virus, both among the 61 non-vaccinated and the 126 vaccinated subjects, during the following winter 1977-78 did not allow to arrive at any information on the vaccine protective efficacy. In autumn 1978 a trivalent vaccine had been employed, containing 300 I.U. of A/Texas/1/77, A/USSR/90/77, B/Hong Kong/8/73 viruses, respectively. During the following winter 1978-79, cases of infection, caused by type B/Hong Kong/5/72 - like influenza viruses were detected. The morbidity turned out significantly lower (P less than 0.05) in the vaccinated subjects (4 out of 140 cases) versus the non-vaccinated (7 out of 56 cases). The protection index supplied by the vaccine on such occasion was 77%.