be distributed at dawn, by the milkman referred to above (F. D. K.), to the consumers in the McKnight district.

At the time of our first visit to the farm we discovered the cans of morning's milk quietly reposing entirely out of sight on the bottom of the well. This was in most respects an ordinary dug well, but of unusually large diameter. It was covered with a platform of old and badly worn planks loosely laid on, and with rounded edges bordering on wide cracks. To allow the cans to be sunk and lifted, the planks were loose and separate. The cans were lifted by worn ropes, each of which ran through a hole in a plank and ended above in a large knot. On one side of the well was a chain pump, the spout of which overhung the platform, so that careless pumping easily washed matters on the planks near by through the wide cracks into the well. On the planks lay clumps of manure evidently left there by the dirty boots of the men. A little pumping proved that these were easily washed into the well; and, indeed, ordinary stepping about upon the planks to handle the milk cans or to pump must have cleaned the boots of the men at the expense of the water in the well.

On lifting the cans, on one of our visits, we found, by drawing them up and inverting them, that out of nine four leaked milk around the wooden stoppers with which they seemed securely plugged. Moreover, not one of them was completely full. On shaking the cans, the splashing of the milk inside was plainly heard. It was clear that if milk would leak out when the cans were inverted in the air water would leak in when the same cans were submerged in water. Thus by submersion the milk was not only cooled in temperature but to some extent augmented in volume. The well water was plainly dirty, and bacterial analysis showed it to contain B. coli communis in abundance, a form which testifies to the presence of fecal contamination, at least of animal origin. There was no question in this case of pails or cans washed in infected water, for the former were washed in the house in mountain tap water, and the latter were steamed in Springfield.

This well, however, was at some distance from the house, and not used there for drinking. It was easy to see how, if the well were contaminated, the milk sunk in it might have become so. But how could the well have been connected with the cases in the house? A plausible hypothesis was the following: the excreta of the patients went into the privy, and there is reason to believe that they were not dis

MORTALITY

24-1 2-8 9-15 16-2223-29 30-5 6-12 13-1920-2627-2 3-9 10-16 17-2324-3031-6

27-2

JULY

AUGUST

SEPTEMBER

infected. The contents of the privy, shortly before the infection reached Springfield, were spread upon the tobacco field. From this field the workmen frequently passed to the well to get water and to work about the milk. It is not difficult to believe that in doing so they may have carried upon their boots masses of fæces, originally from the privy, from the field to the well, into which pieces fell through the cracks while the men trod upon the irregular planking. Whether or not this theory is true we shall probably never learn. An alternative hypothesis, moreover, remains. Upon most farms of this sort the "help" is constantly changing. Hired men only a few grades better than tramps come and go. There was a history of much of this kind in the present case, and it may easily have been that some such person acting for the time being as milkman was suffering from incipient or "walking" typhoid fever. In such a case a self-infection of the fingers, such as can only too easily. happen, may have led to a heavy infection of the warm milk, such as must have existed in the milk which was served by F. D. K. to the McKnight district in Springfield.

The milkman (F. D. K.) for a time maintained with bold assurance that the milk from this particular farm had never gone to the infected houses; and therefore, in order to have independent evidence, we undertook a minute and systematic study of the somewhat intricate system of milk supply in Springfield. The problem was complicated by the fact that this particular milkman delivered from two separate wagons, and loaded these with milk from several dairies at two distinct points. But after many vexatious researches, including a visit by one of us (W. H. C.) to all of the dairies. selling him milk, for the sake of obtaining documentary and statistical evidence of the dates and amounts of milk supplied by the several farms, delivered by the carriers and loaded upon each of his wagons, with the time of starting and the route of each, as well as the actual purchases of his wholesale customers, to whom he now affirmed that all of the Feeding Hills milk had always gone (thus contradicting flatly his original statement to Dr. H—); after interviewing his assistant, or "striker," who had delivered the milk; after having had the aid, also, of the able inspector of the Board, Mr. McCaffrey, we at last obtained positive evidence from wholly independent data that the milk which went to the infected houses must have come, beyond a peradventure, from the farm in Feeding Hills. In all of this we were very greatly aided by the officers of

the Springfield Co-operative Milk Association, who courteously gave us access to their books and in many ways afforded us invaluable assistance.

It now remains to describe somewhat more fully the epidemic itself. The location of the cases is indicated by the red circles upon the map. The houses (upon the hill) served by the milkman concerned are shown by the blue circles. It must not be forgotten that many other milkmen served other houses in this district, which is thickly settled, but in none of these was there any fever. The correspondence of the cases of typhoid fever with their milk supply is plain, although, as would be expected, some houses served by this milkman had no fever. There is good reason to believe that the infected milk was not the only kind that he carried, and that a different milk was sometimes delivered, e. g., in the Ingersoll Grove district.

It will be observed that there were a few cases of typhoid fever in other parts of the city. A very few of these were imported. Of the others, those on Tenth and Essex streets deserve special comment. After the milkman became alarmed, he is known to have refused for a time to take milk from the dairy in question. We have positive evidence that milk from this dairy afterwards went to Essex Street; and that here, also, the fever appeared, is one of the strongest links in our chain of intrinsic evidence. Similarly, seven cases of typhoid fever suddenly appeared about August 20, in an hotel in the heart of the city; and while no positive evidence was found that milk from the infected dairy went to this hotel, we did discover that at the time when this milk was under a cloud, and was held by the Association, one can of reserved milk (ten quarts) was sold by them to that hotel, that this was an unusual occurrence, and at the very time required to have conveyed the disease to the hotel. As for the rest of the city, it was throughout the epidemic period almost wholly free from typhoid fever. The annual autumnal wave had not yet begun to rise.

The whole number of cases discovered and investigated was one hundred and fifty, the whole number of deaths belonging to the epidemic was twenty-five. The course of the epidemic may be seen upon the diagram of weekly mortality and morbidity opposite p. 722. its intensity, upon the diagram on p. 725 of monthly mortality from typhoid fever, which shows well also the limits of the epidemic