418 



SCIENCE 



[N. S. Vol. XL. No. 1029 



not be directly observed, but may be obtained 

 by measuring the radiation toward a black 

 surface of cavernous shape at known tempera- 

 tures, using Stefan's law. This method has 

 been used in the present research. If no 

 radiation from the earth's surface can pene- 

 trate the atmosphere, as in the case when the 

 sky is obscured by thick clouds, then the sky 

 has the same effective temperature as the 

 ground; hut according as more or less surface 

 radiation escapes to space through the air, 

 the effective or apparent temperature of the 

 sky diminishes, although never reaching abso- 

 lute zero, for that would mean complete ab- 

 sence of absorption. 



The radiation to the sky, measured either 

 by a bolometer or a thermopile, falls off very 

 slightly as the pointing of the instrument de- 

 parts from the zenith, but more rapidly near 

 the horizon where, if the blue is of inferior 

 purity, the whitish sky has an effective tem- 

 perature almost identical with that of the 

 earth's surface. The dulling of the blue of 

 the sky near the horizon is due to the dust and 

 haze of the lower atmosphere. 



its purity much better and a larger transmis- 

 sion, such as that indicated by the upper curve 

 of Fig. 1 (which, however, is meant for a 

 winter curve at sea level) may be given for 

 terrestrial radiation emanating at inclina- 

 tions with the horizon which are taken as 

 abscissae, the ordinates being the transmission. 

 In cold winter weather when the ground is 

 covered with snow, atmospheric dust is 

 greatly diminished, and the conditions at sea 

 level approximate to those on mountains. 

 These transmission curves for variously in- 

 clined rays have been obtained by dividing the 

 observed sky radiation by the unobstructed 

 radiation to space appropriate to the observed 

 temperature, assuming that space is at the 

 absolute zero of temperature. The lower curve 

 is founded on excellent observations, but the 

 upper one is not so reliable and its shape is 

 partly inferred. The maximum ordinate, how- 

 ever, is sometimes exceeded with the purest 

 skies of winter, and at this season these uni- 

 form and deep blue skies are not rare, the 

 difficulty in observing them being usually an 

 instrumental one arising from the necessity 



The lower curve of Fig. 1 is derived from 

 observations of sky radiation on a summer's 

 day of good blue sky near sea level. From 

 mountain summits lifted above a large part of 

 the atmospheric dust, the cerulean blue retains 



?-i>r<ftu; 



of working with exposed instruments. The 

 adopted values of summer transmission, I's, 

 and of winter transmission, T ^, from which 

 the curves are drawn, are given in the next 

 table : 



