The complexity of the question of obtaining the true mean temperature led the late A. Buchan to look favorably on the mean daily extremes (maximum plus minimum)/2 as an expression of the mean temperature. This view has become traditional among English-speaking races, but it has two very grave objections. and it is not adopted by continental or South American meteorologists. These objections are briefly : (1) Maximum and minimum thermometers. especially the latter, are far more likely to develop systematic errors than is the ordinary dry bulb. This remark applies with great force to hot countries, where in addition the observers are frequently untrained. In fact, it was this very difficulty which led to the adoption of a column of “True mean temperature” in Reseau Mondial. (2) The relation (maximum plus minimum)/2 to the true mean is not, even approximately fixed, but varies with the location, with the cloudiness, and with the time of year.

The means of observations taken at 7 a. m., 1 or 2 p. m., and 9 p. m., giving the evening observation double weight, is recommended for general use.—H. L.