TWO features which play important roles in the epidemiology of rubella are the incidence of inapparent infection and the occurrence of reinfection. There is considerable evidence that an attack of rubella confers life-long immunity to the disease, yet data have now accumulated which suggest that reinfection at the subclinical level does occur.1,2 At present it is not known how frequently this happens, nor whether it poses a hazard to the fetus if the mother is reinfected during early pregnancy. We have examined the problem of reinfection in two ways: (1) experimentally in chimpanzees given wild virus and subsequently challenged with homologous or heterologous strains; and (2) in antibody positive children exposed during a natural epidemic of rubella in a closed population. Experimental Reinfection of Chimpanzees.—Five young susceptible chimpanzees were used. Two of these had been exposed to rubella virus previously by ingestion of 106,5 InD50 of