as the milk supplies, mosquitoes, flies, and other disease-spreading insects which thrive during the warmer months. It is worthy of note that the cities of this diagram come from the southern latitudes, where the effect of climate is more marked, especially as to the duration of the fly period.

SEASONAL DISTRIBUTION OF TYPHOID FEVER.

It has been noted for many years that typhoid fever shows a marked difference in the cases found during the various seasons of the year. At certain times, notably the colder months, typhoid in some cities may be, and usually is, relatively low; while during the summer months, from July to October, very high typhoid fever rates are often noted. Such variations of typhoid fever incidence do not always, if ever, correspond to variations in the quality of water. The water supply may indeed vary in quality and be materially worse at some seasons of the year than others. It may often happen, however, that the times of bad quality of the water supply follow the times of high incidence of typhoid fever by an appreciable period. This may be taken to show rather definitely that the initial cause of the high typhoid fever rate at certain seasons of the year does not depend upon the water supply, but more likely the bad water supply may be the result of earlier typhoid fever conditions.

Fig. 5 shows in a little more detail the seasonal distribution of typhoid fever for the cities of Birmingham, Ala.; Washington, D. C.; Richmond, Va., and Baltimore, Md. The rates of typhoid fever indicated for the months from January to June average about twenty deaths per 100,000 population, and this figure is fairly uniform. Starting with the month of June and culminating usually with the month of August, but sometimes later, the typhoid fever deaths rapidly rise to a high figure of somewhere from seventy to eighty per 100,000. The real cause of such typhoid fever death-rates is to be found in the prevalence of flies under unfavorable sanitary conditions, such as unscreened privies and other unprotected sources of possible pollution, the prevalence of mosquitoes during the summer months, the possibility of bad milk during this season, and other similar sources of infection.

VOL. CLXXX, No. 1075-5

Fig. 6 illustrates the seasonal distribution of typhoid fever in the city of Jacksonville, Fla. The three curves given are for the three years, 1910, 1911, and 1912, 1910 being the solid line showing the highest rate, 1911 the dashed line showing the intermediate rate, and 1912 the dot-and-dash line showing the lowest rate. As in the previous diagram, the typhoid fever death-rate in the winter months was relatively low, around twenty

JAN. FEB. MAR. APR. MAY JUNE JULY AUG. SEPT. OCT. NOV. DEC.

DIAGRAM SHOWING SEASONAL DISTRIBUTION OF TYPHOID FEVER

per 100,000. During the warm summer months the typhoid fever cases and rates were very high. On the same reasoning as before noted, it is to be expected that these high summer rates were attributable to flies or other similar causes.

In this actual case, that of Jacksonville, it may be said that the proof of the pudding was in the eating, for the improvement shown by the three curves was effected on the theory above noted.

The heavy line shows the 1910 typhoid fever case rate, when all privies and water-closets were left in their ordinary loose, unscreened, and unprotected condition. In this year the rate ran up to 110 in the month of July. After this a start was made in cleaning up and making fly- and mosquito-proof all these

ANNUAL DEATH RATE PER 100 000 WITH RATIO OF 5 CASES TO I DEATH

200

150

100

75

50

25

JAN. FEB. MAR. APR. MAY JUNE JULY AUG, SEP. OCT. NOV. DEC.

DIAGRAM SHOWING SEASONAL DISTRIBUTION OF TYPHOID FEVER

JACKSONVILLE - FLA.

water-closets and privies. The curve in 1911 shows the results of this work. The highest rate reported for the year 1911 was thirty in the month of June. During this summer, 1911, seventyfive per cent. of the closets were screened. In the year 1912 all the closets were screened and the highest rate recorded was eighteen in the month of June.

Taking the whole year, we find that the typhoid cases were, for the years 1910, 1911, and 1912, 329, 158, and 94, respectively, and that the deaths for these years were 62, 40, and 18, respectively.

There exists very little doubt that all the improvements which have been made were the result only of the cleaning of the closets. It is not likely that the water supply in the city of Jacksonville, taken from very deep wells, was the cause of any typhoid fever.

THE BIRMINGHAM, ALA., WATER SUPPLY.

The relative significance of the bacterial content of water, of the coli content, the effect of water supplies on typhoid fever, as compared with the effect of flies, mosquitoes, milk, and other possible sources of typhoid fever infection, was under consideration in the case of the Birmingham Water Works Company, in which the speaker was interested early this year. This case depended on a complaint against the water supply received from the water works company on the ground that the water did not meet with the provisions of the franchise compact, which specified

"that the water so furnished shall be clear, wholesome,
and suitable for all domestic purposes, and ordinary
manufacturing purposes.”

From the sanitary standpoint this controversy did not involve the question of purity of the water in the raw condition. Essentially, it was a proposition of whether the opportunity for pollution of the raw water might have caused the raw water to become so polluted as to increase the load of work to be done by the filters beyond what might be reasonably caused by safe filtration practice, and so put a burden on the filters greater than that found in filter plants elsewhere, such as are producing what might be fairly termed pure and wholesome water, as that expression is used in the practical art of water supply and water purification.

The first step in the investigation of this problem was a visit to the watershed and an examination of local conditions. On going over the watershed of Five Mile Creek no seriously objectionable conditions were found. There were occasionally a few farmhouses. No villages existed on the watershed, and

there was no sign of any direct sewage pollution. There remained, of course, the opportunity for indirect contamination, owing to soil wash, and the possibilities of the flushing of polluted matters by heavy rains into the feeders of the various streams above the head of the canal.

Considering the total population tributary to the water supply, and a comparison to see if an unduly large number of people discharge their waste into this water supply, Table VI shows a comparison of the Birmingham water supply with a number of other cities to show what is the density of population tributary to the watershed from which the city supply is derived.

TABLE VI.

POPULATION RESIDENT ON CERTAIN WATERSHEDS.

Note-Populations per square mile, except Birmingham, refer to 1900 population.

It appears that the Five Mile Creek watershed contained a population of thirty-five to the square mile; the main Cahaba watershed a population of ten to the square mile; and the east Cahaba watershed a population of forty-four to the square mile. As compared with this, the old Boston watershed, Cochituate supply, has a population as high as 775 to the square mile; the Saugus River watershed, for Lynn, Mass., a population