In a residential home for children, 367 cases of streptococcal illness were observed in a period of 30 months. The children lived in groups of about twelve in separate cottages. There were 194 occasions on which a streptococcus was thought to have been newly introduced into and produced illness in a cottage; on 132 of these 194 occasions there were no secondary cases of illness. The remaining 62 cottage introductions were followed by one or more secondary cases.

In 27% of the 194 introductions, the primary case of illness seemed to have been infected from a healthy person in the cottage. In all, 30% of introductions of a new streptococcus into a cottage could be attributed to recognized contacts with one or more known infected children.

The most important factor determining spread within the cottage seemed to be the carrier state of the primary case, spread following more often when the primary case had streptococci in the nose either on admission to hospital, or in convalescence.

There was no evidence that spread within cottage bedrooms was of great importance.

In about 35% of the incidents with spread, the initial spread to secondary cases seemed to be from the incubation-stage carriage of the introducer; in 42% it was from his or her convalescent carriage.

The carrier rate in the healthy cottage-contacts was generally higher in cottages experiencing clinical spread of infection than in those that had single-case introductions. There was a strong correlation between the carrier rate in the first week after an introduction and the final bacteriological attack rate, and a weaker correlation with the final clinical attack rate.

Continued spread of infection in a cottage was commonly due to the arrival of new children and was almost always associated with the presence of nasal carriers of streptococci.

The 194 cottage introductions could be grouped into sixty-three overlapping Village epidemics, each apparently derving from a new importation of the particular type into the Village, although the evidence for this was often merely the absence of known infections within the previous few months. Only 13% of the introductions resulted in more than 10 cases, and some 80% had 5 or fewer. Introductions were more frequent in the cottages receiving children new to the homes than in those for the more permanent residents.

The principal factor found as determining the spread from the first cottage to others was the attack rate in the first cottage. Introductions in cottages for school-age children, and especially those in which a child attending the school in the Village grounds was the first to be attacked, also seemed to lead to spread more often than others.

The interval between successive Village introductions of one type did not appear to affect the extent of spread at the second; but the number of cases occurring in the first of two introductions had a notable effect: in no case did two successive introductions both result in a large number of cases of illness.