To clarify the periods and conditions of spontaneous infection with Japanese encephalitis virus (JEV) in Fukuoka Prefecture, slaughtered pigs, as well as sentinel pigs, were examined for the appearance of viremia and hemagglutination-inhibiting (HI) antibody in the prefecture over a four-year period of 1964 to 1967. In addition, studies were made on the occurrence of mosquitoes bearing JEV and the relationship between the prevalence of Japanese encephalitis among swine and that among human beings. The results obtained are summarized as follows.(1) The appearance of HI antibody in sentinel pigs was demonstrated in such period as varing with the year. Moreover, there were remarkable difference in conditions under which this antibody made its appearance among the years examined. In 1964, 1965 and 1967, HI antibody began to appear on July 3, 4 and 5, respectively. It took two week every year for the antibody to appear in all the experimental pigs. In 1066, however, the antibody appeared simultaneously in all the experimental pigs on July 25, or about three weeks later than in any of the three years mentioned above. These fluctuations in the appearance of HI antibody by year seem to be due to the period when infection took place among swine and frequency of chance of contracting infection, that is, the period of appearance of mosquitoes bearing JEV and the number of such insects which appeared. Furthermore, the period when HI antibody appeared in the sentinel pigs almost agreed with the time when the antibody-positive rate in the slaughtered pigs exceeded 50 per cent (the time of simultaneous rising) and coincided with the time of appearance of mosquitoes bearing JEV. Accordingly, it is possible to estimate the period of spreading of JEV from than of appearance of HI antibody in pigs.(2) In the case of slaughtered pigs, it is also possible to estimate almost correctly the period when they contacted infection by determining the 2-mercaptoethanol (2ME) sensitivity of HI antibody. Particularly in the early period of an outbreak, HI antibody susceptible to 2ME shows a high frequency of appearance. Therefore, it is possible to presume the state of spread of JEV from the occurrence of 2ME-sensitive HI antibody.(3) In the early period after the appearance of HI antibody, antigen of the JaGAr 01 strain frequently gave rise to four to eight times as high an antibody titer as antigen of the Nakayama (NIH) strain. Consequently, it seems reasonable to use antigen of the JaGAr 01 strain for an epizotiological survey.(4) Viremia was demonstrated in the sentinel pigs about one week prior to the appearance of HI antibody. Therefore, it was assumed that viremia might last about one week or less. In the case of the slaughtered pigs, the rate of detection of viremia seemed to be high about two weeks prior to the time of simultaneous rising of the antibody-positive rate. The occurrence of viremia was demonstrated in a total of 16 cases; that is, three cases in 1965, eight in 1966 and five in 1967. When the titer of virus contained in the blood was estimated, it was less than 101.0 BLU (50 per cent baby mouse lethal units) in nine cases, 101, 0 BLU in one case, 102.0 BLU in three cases, and 103.0 BLU or more in three cases.(5) Close correlations were observedd among the time of appearance of HI antibody in swine, the time of appearance of mosquitoes bearing, JEV, and the period of prevalence of Japanese encephalitis in man.