CHART 3-INFLUENZA, KELLEYS ISLAND, 1920.

This earlier occurrence of the epidemic among school attendants might be explained by assuming, first, a common source of infection for school and nonschool cases, and that the school children, being in a special age group, possess a shorter period of incubation; second, that the school offered a special opportunity for infection. If the latter assumption be true, it would seem that school attendants who spend the greater time at school might be affected somewhat more frequently and earlier than those spending less time, and it was thought that a comparison of the school children who carried their dinners to school, with those who went to their homes for lunch, might furnish a test of this hypothesis. Reference to Table IV will show that there were 46 school cases among 80 school attendants who carried their dinners, and 41 cases among 77 who did not carry their dinner, or attack rates of 57.5 and 53.2 per cent, respectively. Referring to Chart IV it will be seen that cases in those pupils who ate their dinners at school had a definite tendency to occur earlier than cases in those who went to their homes for lunch; for in this group 30 cases (65 per cent of the total) developed before, and 16 after, the closure of school, whereas in those who went home for dinner, 15 cases (37 per cent) occurred before, and 26 after, the closure of the school, notwithstanding the fact that those who carried their dinners lived for the most part in the more remote portions of the island where one would expect them to be less exposed to early infection through contact with the general island population. These two groups of school attendants, moreover, are fairly comparable in respect to age, although there is a somewhat larger proportion of younger children among those who carry their dinner than in the other group. As will be mentioned later, however, the primary room had the lightest attack rate of any room in the school. Again, if the school actually served as a center for the dissemination of influenza, one might expect that the nonschool-attending members of families having children in school, one or more of the latter developing the disease, would suffer earlier than the members of households where no school attendants resided; while if all were exposed equally through a generalized or common source of infection, and the earlier occurrence of cases among school children is due to a shorter period of incubation, there should, it would appear, be no great difference in the curves representing the chronology of the epidemic in the two groups of people, since the two groups are similar in all particulars excepting exposure to school attendants.

Referring to Table IV, and Chart 5, which represent the daily occurrence of new cases in the two groups, it may be seen that there is not only a marked difference in the character of the curves, to be considered later, but that the people associated with school cases reach the maximum number for any one day three days in advance

of those not associated with the school cases. It is difficult to explain these facts on other assumptions than that the school was a source of disseminating the infection. (A small group associated with school cases who were not ill is given in Table IV, but their number is too small to indicate whether children who were exposed but did not become ill may or may not have carried infection.) That this conclusion is correct is, we believe, also indicated by comparing the curve which represents the daily occurrence of new cases among the individuals attending school or in contact with school attendants, with the graph for the remainder of the population (Charts 2 and 5). It will be seen that the curves representing the chronology of the epidemic among school exposures have two peaks; whereas the curve for the nonschool exposures is a relatively smooth curve of one peak, and it may be that the remission in the epidemic among the former group is related to the closure of the school. On January 28 there were 18 pupils absent from school; on January 29, owing to the illness of a teacher, the second room was not in session, and 58 pupils were absent; and on January 30, owing to the illness of another teacher, another room was not in session, and on that day 88 pupils were absent. School did not open on January 31, nor thereafter until the epidemic had subsided.

It seems that the first wave of influenza among the school attendants (Chart 3) is definitely due in some way to school exposure as described above, and the remission of January 30, it would seem, may be due to the dismissal of a portion of the rooms and to a thinning out of the remainder of the pupils, which certainly would render the school a less efficient means of exposure. This conception, moreover, coincides with what we believe is the probable incubation period of the disease, as will be discussed later. The second increase in daily occurrence of cases among school exposures, as shown in Charts 2, 3, and 5, corresponds rather definitely with the increase among the associates who did not attend school, and is probably due to the same general causes.

If we admit that the school was instrumental in spreading the disease, it naturally becomes of interest to ascertain the circumstances conducive to the spread of the epidemic among the pupils. The seating arrangements and plan of the various schoolrooms, together with the dates of attack, etc., are shown graphically in Figure 1. The school building is a large, brick structure, centrally situated on the island. There are six rooms in the building, four of which are used as study and recitation rooms, and two for recitations only, various classes going to them for a few minutes each day. The building is heated by hot air furnaces located in the basement. The water supply is from a large cistern, which stores the water collected from the roof at time of rainfall. No water had been

permitted to enter the cistern for over two months. A large stone vessel, kept filled with this rain water and supplied with a vertical type of bubbler, which may be touched by the lips, holds the drinking water. There are no common drinking cups. The base

ment is also supplied with stationary wash bowls and bar soap. Paper towels are furnished. The school is supplied with open privies. A library is maintained, from which pupils may take books to their homes-a privilege more or less used by all classes. The

O

O

Primary Room.

Third Room.

FIG.1.

L.H.WILDER