circumstances it might be expected that the weight of the men, taken as a whole, would remain sensibly the same; and that the numbers losing or gaining, as well as the quantities lost or gained, would vary little month by month; or that, if any marked variation occurred, it would be of an accidental character, depending on the greater or less amount of sickness during any particular month. The results, however, showed that a marked periodicity existed, and that, taking an average of years, there were two distinct series of months, during the one of which there was a constant loss of weight, and during the other a constant gain, so that, if the year were divided into quarters, there was a loss during the first and fourth quarters, and a gain during the second and third.
The two series of gaining and losing months were unbroken, except in one instance. On reference to the results it was found that in November, which was in the losing series, a gain occurred. The amount gained was very small, and the discrepancy was caused by the arrival of large numbers of prisoners in September and October, who usually gained weight for a short time after they were received, so that probably this break in the series resulted from the influence of the stage of imprisonment, which rather more than balanced the influence of season. On estimating carefully the facts which showed the average gain or loss per prisoner weighed, it was seen that, beginning at December, the amount lost per man increased rapidly and very steadily till March, but that between March and April there was a very abrupt transition from loss to gain. The gains then continued till August, the amount gained increasing on the whole, by a series of jerks, each alternate month presenting a larger and a smaller gain respectively: so that, to obtain a steadily increasing series, it was necessary to couple the summer months in pairs. Between August and September a change of weight occurred, about equal in amount, but in the opposite direction to that which took place between March and April. The changes between March-April and August-September were far greater in amount than the changes which took place between any other pairs of consecutive months; and this remark applied with greater force to the percentages of men gaining or losing, and to the net gains and losses per man.
The inferences which may be fairly drawn from these observations were: 1. The body becomes heavier during the summer months, and the gain varies in an increasing ratio. 2. The body becomes lighter during the winter months, and the loss varies in an increasing ratio. 3. The changes from gain to loss, and the reverse, are abrupt, and take place about the end of March and the beginning of September.
The results, which were thus gathered from the study of a large number of periodical weighings, presented a remarkable relation to the facts obtained by Dr. Edward Smith from a series of most valu-
