MR. MACDOWALL'S letter in NATURE of October 12 (p. 485), in which he directs attention to certain features in the sequence of annual number of hot days at Greenwich, is interesting. Nevertheless, I think he himself will acknowledge that his example, viz. the summer of 1911, is a happy one. What his diagram would lead one to expect if one were making a forecast, and not a retrospect, is that the number of hot days in 1911 would lie between his lower limit of 90 days and an upper limit of about 90+130, or 220 days; the value half-way between the two, i.e. 155 days, being the “ost probable” Clearly in this case the upper limit, and also the most probable value, may be disregarded, and the lower limit is sufficiently high to be worthy of note.