(namely, 2,842 and 2,869 acres, respectively) and yet we notice that while

13,516 cases of scarlatina appeared during the 10 years in the West Side district among 415,694 inhabitants, equal to 3.25 in 1,000, but 17,318 cases appeared in the East Side dis

trict in the same time among 608,524 inhabitants, equal to 2.79 in 1,000.

In the twelfth ward, with 476,632 inhabitants, comprising an area of 5,504 acres of land (out of 12,576 acres of the entire island of Manhattan) we find 14,569 cases of scarlet fever in these ten years, an annual morbidity rate of 3.10 in 1,000 inhabitants, but slightly better than the middle West Side (below Eighty-sixth Street) with 3.25 in 1,000, but yet decidedly worse than the middle. East Side with but 2.79 cases in 1,000 inhabitants.

Comparing, lastly, the two wards of the Bronx, we find that the twenty-third, with a population exactly three times as large and an area of ground one fifth the size of the twenty-fourth ward, has but a scarlet fever rate of 2.75 in 1,000, against 3.50 in 1,000 inhabitants of its neighbor.

These comparisons prove that it is not the density of the population, but the actual number of people living in a district which determines the frequency of scarlet fever.

If, then, scarlet fever is a contagious disease, appearing in crowded and in sparsely populated districts alike often, in direct proportion to the number of inhabitants, what factor is there common to all that can carry this contagion from house to house? Where is this common carrier of infection for scarlatina, like the water supplies of cities for typhoid fever?

On Chart IV, I have arranged the average number of scarlet fever cases in a year (6,295) according to the weeks in which they were reported, so that each column represents the average scarlet fever frequency of each week of the

year.

Here we first notice that scarlatina is found in New York during every week of the year. The highest number of cases (208) were reported during the first week in May, and the lowest (34) during the second week of September.

During the first half of the year three times as many cases occur as during the second half. Beginning with January 1st, we see scarlet fever increasing steadily from week to week to the middle of May, when its frequency diminishes gradually until the end of June, then to drop down suddenly from 110 to 60 cases in a week. This sudden drop at the end of June is the most important point on this chart. It repeats itself regularly every year. In 1901 this drop showed the

figures of 228 to 82, in 1902 from 149 to 19, and in 1903 from 132 to 94.

What could possibly be the cause of this sudden decrease in frequency recurring annually with unerring regularity? A highly contagious, epidemic disease which is more prevalent during May and June of each year in New York than during January and February cannot possibly be influenced as to its frequency by variations in the atmospheric condition so as to show this phenomenon about the first of July of each year, resulting in the fact that during May and June on the average 150 new cases of scarlatina are reported weekly, and during July and August but 50 in each week.

There can be but one explanation: All public schools close during the last week of June and therewith abruptly ceases the almost daily congregation of infectious material on small areas. (school houses and play grounds) for so many hours of the day. According to the official report of the health department 450,000 children attended the public schools on the daily average during the first three months of 1903, and about 75,000 the private and parochial schools. Many private schools close during May, and many scholars leave the city at that time annually, resulting in the first falling off in scarlet fever since New Year, but the bulk of this one half million of young inhabitants suddenly cease coming together daily at the end of June, and, as a result, the frequency of the disease comes down to one half in one week, a phenomenon seen at no other time during the entire year.

From July 1st on, the frequency of scarlatina. diminishes steadily from week to week during the entire vacation period until the second week of September, one week after the opening of the public schools, when again the number of new cases steadily increases to the end of the school year, July 1st. This steady decline of scarlatinal frequency during vacation, and this steady increase during the school term, are of high importance. If all schools closed on August 1st instead of July 1st and if no scholars left the city till then, the accumulation of new cases would then continue up to that date, instead of to the middle of May. And if all schools opened on October Ist, the increase in the scarlet fever rate would not begin until then.

If the schools are the chief common localities for most scarlatinal infections, how is it that the proportion of contagions is not larger in the densely populated districts than in the less populous sections of the city?

Each school girl has but a certain number of

personal acquaintances at school, with whom she comes in direct close daily contact, or, correctly speaking, five times each week, whether she resides in the Bronx, in a Rivington Street tenement, or in a private house west of the Park. With these acquaintances (limited in number) the scholar exchanges kisses, eatables, and pencils to write with and to bite on. The majority of the members of her class do not touch her, any more than the rest of the scholars in the building. The longer school is open, the more the term advances, the more acquaintances the pupils will make among themselves (like among the passengers of a steamer crossing the ocean), and from week to week the average scholar comes in direct contact with more and more new sources of infection, carried from infected homes to the school, resulting in the steady increase of scarlet fever cases during the school term. If the scarlatinal contagion could be spread by exhalation and inhalation, into, and from, the air of the schoolroom, then the majority of a class would be infected within a few days by one scholar and local school endemics would constantly appear in different schools and the rise in September would be abrupt and not gradual, as well as through the entire school year.

The direct contact infection proved for scarlatina, we can readily understand why the scholar living in the tenth ward is in no greater danger of acquiring this disease on his way through the crowded streets then the one living in a fine flat in Harlem. During lessons both children are endangered only by possible contact with their immediate neighbor next to whom they sit and with whom they exchange courtesies. The number of such neighbors is alike to each child in every school, public and private, uptown and down

town.

During recess and on the way to and from home, the possibilities of direct contact are also limited to a certain number, limited by time and space, which govern all alike. In fact, a child attending school in a sparsely settled district is apt to make more friends than one in a crowded section, because the school is at a greater distance from home (taking a longer time to walk) and because the children are less suspicious of one another. Large crowds do not induce conviviality, any more in children than in adults. Children make no exceptions to this law of "most being alone in a crowd." When epidemics of scarlet fever strike small country villages and towns in Europe (where isolated farms are unknown), the percentage of scholars stricken is always much higher than in large cities during like conditions.

At the close of this investigation I submit the following conclusions:

1. Scarlet fever, always present in New York, is evenly distributed each year among the different districts, in direct proportion to the number of inhabitants.

2. It is most common among school children. 3. The chief common centres for contagion .are the schools.

4. Direct contact is necessary for infection.

II.

As a reliable specific remedy against scarlatina is not yet on hand, the question arises whether we should limit our therapeutical efforts to symptomatic and dietetic treatment, or should attempt to destroy some of the pathogenic organisms which have entered the body wherever they come within reach.

In scarlatina the inflammation of the skin begins about the neck and invariably progresses downward to the toes. Why this order of march? Why, if scarlatina is primarily a general blood infection, does not the rash appear all over the body at once, or begin occasionally somewhere else than near the throat, say at the toes, and progress upward instead of always downward? Does this not lead us to the conclusion that the primary invasion takes place in the pharynx, from there reaches the nearest skin surface, and thence spreads locally by continuation of surface in the lymphatics of the skin?

In the scarlatinal throat we first find unusual redness, then swelling, and then exudate on the invaded mucosa, and then the adjoining lymphnodes swell more and more from day to day, as the infectious material reaches them from the pharynx. Again we have here primarily a local infection.

When is scarlet fever a systemic blood infection? Let others answer this question: Jochmann (Deutsch. Archiv. f. klin. Medizin, lxxviii) has recently examined the blood of 161 scarlet fever patients in the Eppendorf Hospital in Hamburg and found organisms (streptococci) during the disease in only 15 per cent. of the cases, all of which proved fatal. During the height of the rash in the first days of illness streptococci were never found in the blood, not even in malignant cases, while in 50 per cent. of the dying patients streptococci were found shortly before death, their number being relatively small in comparison to those found in the blood of the dead.

These important findings of Jochmann prove that the scarlatinal organism lodges in the throat, in some lymph nodes, and in the skin of the patient for days, even in severe cases, before it enters the blood and then to kill the patient. In fact, the