'clerical errors.' Of these two, the data for the first were taken from the annual reports for five years of the chief of police of New York city. It was not there stated that sickness was the cause of absence from duty, but it is safe to assume that it was the usual one. It is, however, rather interesting to note that immediately following Christmas, New Years, and other holidays, an unusual number was laid off, but we may charitably suppose that the weather was excessively deadly in its effects at those times. The tabulation shows that 105%, or 5% in excess of the normal number, were off duty on calm days. This would hardly be more in accordance with our expectation than was the school attendance under such conditions. If perfectly calm days were the most agreeable of all kinds we might suppose that our stalwart guardians of the peace had chosen them for picnics, but gentle breezes are generally accepted as being more delectable than dead calms, and we must look for some other causes for the absence of bluecoats under the latter conditions.
The last class of data given, that having to do with 'clerical errors,' was studied as a make-shift. I wished to determine the influence of different weather conditions upon the intellectual as well as educational states of man, and to that end sought long and earnestly for school records which showed a daily marking of class work, but without success. If any teacher who may read this has such, I should be very grateful to him if they could be placed in my hands. While wondering what other records might be made to supply the lack, I came across the statement that in the Bank of England certain sets of books, an error in which would prove cumulative and produce disastrous results later on, were never worked upon during some kinds of weather, especially London fogs, as it had been proved that clerical errors were much more frequent at such times.
Following the clue here given I gained access to the books of some of the largest banks in the Wall Street district, with the result that in the records for two years, the number of errors stated were found and tabulated with reference to their daily occurrence. The results showed 104% of the normal, or an excess of 4% for the calm days.
To state in a sentence the occurrence of data of all these classes under the condition of calm: absence from school, death, policemen off duty, and clerical errors were all above the normal, while misdemeanors in school and penitentiary, arrests for assault and battery, drunkenness and insanity, and suicide, were below.
The facts so far given do not show whether the change with an increase in wind was a gradual one or not. As a matter of fact it was not. Had it been so, there would have been less excuse for this paper. The most striking thing about the curves upon which it is based is the sudden change which takes place in the occurrence of nearly all the