During recent years, some observations respecting vaginal infections in children attending the Vanderbilt Clinic suggested that there was a decreasing incidence of the gonococcus as an etiologic agent. A number of recorded studies in other localities supported this belief. Some investigators, however, have expressed the belief that the gonococcus is still the usual cause of vaginal discharge.1 Therefore the present study was designed to determine what agent, or agents, is responsible for the production of this symptom in girls and at the same time to investigate the possible relation of vaginal discharge and the occurrence of infections of the upper respiratory tract. It also seemed desirable to obtain some information concerning the bacterial inhabitants of the healthy vaginal tract. This report consists, then, for the most part, in a record of data on smears and cultures and the clinical status for 16 premature infants studied in the nursery of Sloane Hospital for Women, for an unselected group of 81 children observed in the wards of the Babies Hospital and for 93 children with vaginal discharge who during an interrupted period of eighteen months, from May 1939 to October 1940, were referred to my associates and me from the pediatric section of the Vanderbilt Clinic. METHODS Premature Infants.-Premature infants rather than full term babies were chosen to represent the newborn group because of their longer period of care in the hos¬ pital nursery. Material for smears and cultures was obtained from the subjects at irregular intervals, commencing eight hours after birth and continuing up to the age of 3 months. Vaginal secretions were collected by means of a sterile glass catheter dipped in solution of sodium chloride. Each examination was planned to include the study of (1) a Gram-stained smear, (2) an iodine-stained smear and (3) a culture on at least one plain blood agar plate incubated aerobically. These preparations were supplemented on occasion by smears stained with the Giemsa and the nigrosin methods and by cultures made on other mediums and with other methods of incubation. The approximate hydrogen ion concentration of the vaginal secretion was determined in a