Studies of the upper respiratory flora in inhabitants of New York City over a period of fifteen years have shown that the prevalence of hemolytic streptococcus has a seasonal trend, which varies little from year to year. If one begins an analysis of the throat flora during the third quarter of the year he finds that hemolytic streptococcus is inactive. Few of these organisms are recovered and they are not virulent. This low carrier rate usually persists into the fourth quarter when sporadic infections occur. Early in the new year there is a steadily rising carrier rate and this is followed or accompanied by a steady increase in acute streptococcal infections. Finally, in the second quarter, the carrier rate and the incidence of infection reach a peak. Infections at this time of year are caused by many serological types of group A organisms, and frequently occur as epidemic outbreaks; however, the presence of hemolytic streptococcus in the throat flora of the general population makes difficult a thorough investigation of these outbreaks. If an isolated epidemic occurs in a season when the carrier rate is low one can identify the sources and trace the spread of infection. We observed such an outbreak during the fall months of 1939 when the carrier rate was low and streptococcal infections rare. This epidemic began in the infants' ward of the Babies Hospital. Since both patients and contacts could be controlled, it was possible to trace the spread and identify the sources of infection. In so doing careful bacteriological studies were conducted. These indicated that certain apparently harmless convalescent babies had the capacity to communicate infection and that this capacity was associated with certain cultural characteristics of the epidemic strain. It is our purpose to describe the epidemic, indicate the sources of infection,