Again, the ages at which pneumonia is both most common and most deadly, namely, under five and over sixty-five, are precisely those at which this feature of exposure to the weather plays the most insignificant part. Last and most conclusive of all, since definite statistics have begun to be kept upon a large scale, pneumonia has been found to be emphatically a disease of cities, instead of country districts. Even under the favorable conditions existing in the United States, for instance, the death-rate per hundred thousand living, according to the last census, was in the cities two hundred and thirty-three, and for the country districts one hundred and thirty-five,--in other words, nearly seventy per cent greater in city populations. How, then, did the impression become so widely spread and so firmly rooted that pneumonia is chiefly due to exposure? Two things, I think, will explain most of this. One is, that the disease is most common in the winter-time, the other, that like all febrile diseases it most frequently begins with sensations of chilliness, varying all the way from a light shiver to a violent chill, or _rigor_. The savage, bone-freezing, teeth-rattling chill which ushers in an attack of pneumonia is one of the most striking characteristics of the disease, and occurs in twenty-five to fifty per cent of all cases. Its chief occurrence in the winter-time is an equally well-known and undisputed fact, and it has been for centuries set down in medical works as one of the diseases chiefly due to changes in temperature, humidity, and directions of the wind. Years of research have been expended in order to trace the relations between the different factors in the weather and the occurrence of pneumonia, and volumes, yes, whole libraries, published, pointing out how each one of these factors, the temperature, humidity, direction of wind, barometric pressure, and electric tension, is in succession the principal cause of the spread of this plague. Many interesting coincidences were shown. But one thing always puzzled us, and that was, that the heaviest mortality usually occurred, not just at the beginning of winter, when the shock of the cold would be severest, nor even in the months of lowest temperature, like December or January, but in the late winter and the early spring. Throughout the greater part of the temperate zone the death-rate for pneumonia begins to rise in December, increases in January, goes higher still in February, reaching its climax in that month or in March. April is almost as bad, and the decline often doesn't fairly set in until May.