THE following results, which I have recently obtained by a discussion of temperatures obtained in kite and captive balloon ascents, may be of interest in connection with Prof. Clayton's letter (NATURE, February 4) and Mr. Dines's remark that at a height of 1 km. the daily temperature variation becomes insignificant (NATURE, June 17). The daily variation of temperature at a height of 1 km. over Berlin, deduced from 2232 observations made during the five years 1903–7, is given in degrees C. by T = T3 - (4.4.40 ± 0.08) + (0.87 ± 0.13) sin (nt+θ2) + (0.14 ± 0.10) sin (2nt+θ2), where T3 is the mean surface temperature, and the probable errors are deduced by the method of least squares. The most probable values for θ1, θ2, are 197° and 123° respectively, the time being measured from midnight.