Abstract
Analysis of the observations of long‐wave radiation from clear skies, R, made by Dines at Benson, yields a correlation coefficient of 0·99 between R and the black‐body radiation at the corresponding screen temperature T. A new series of measurements over wider ranges of temperature and humidity confirms this, with the same value for the correlation between R and σT^4^, the regression equation being: R = −17·195 σT^4^ (milliwatt cm^−^, T °K).
An alterlative representation of equals accuracy is R = 5·31.10^−14^ T^−6^ (Milliwatt cm^−2^, T°K) The latter formulation is probably better founded physically, and brings out the temperature dependence of the ‘effective emissivity’ ϵ (i.e. R/σT^4^), which the atmosphere must exhibit. Either expression provides an estimate of R in terms of T with a probable error less than 0·5 mw cm^−2^.
The present analysis omits any explicit reference to the influence of vapour pressure e on R, and so differs essentially from those due to Brunt and Angström. Re‐appraisal of these latter suggests that the relationships established therein between * and e result basically from a correlation between temperature and humidity. Both the nature and the degree of the correlation between RσT^4^ and e for a given locality would then depend on the temperature‐humidity regime occurring there. The wide variations from place to place, both in the values of the coefficients occurring in the Brunt and Angström equations, and in the degree of correlation found between R/σT^4^ and the corresponding function of e, are thereby explained.