NO, 3 RADIATION OF THE ATMOSPHERE ANGSTROM I 5 



(i) If the sky is clear, there will always be a positive radiation 

 from earth to sky, even in the middle of the day. 



(2) If the sky is cloudy, there will always, in the daytime, be a 

 radiation from sky to earth. 



(3) In the night-time the radiation for a clear as well as for a 

 cloudy sky always has the direction from earth to sky. 



Homen also made some measurements of the radiation to different 

 parts of the sky and found that this radiation decreases rapidly when 

 the zenith angle approaches the value 90 . His values of the noc- 

 turnal radiation vary between 0.13 and 0.22 for a clear sky. 



When relatively large quantities of heat are to be measured under 

 circumstances where the conduction and convection are subject to 

 considerable variation, it is favorable if one can apply a zero method, 

 where the instrument is kept the whole time at the temperature of its 

 surroundings. As the first attempt to discover such a method may be 

 regarded the experiment of Christiansen, who measured the thick- 

 ness of ice formed on metal disks that were placed on a water-surface 

 and exposed to the sky. In 1899 K. Angstrom published a descrip- 

 tion of the compensation pyrheliometer and shortly afterward (1903) 

 a modified type of this instrument was used by Exner * in order to 

 measure the nocturnal radiation on the top of Sonnblick. In agree- 

 ment with former investigations made by Maurer and Homen, Exner 

 found the radiation to be relatively constant during the night. He 

 points out that there are tendencies to a slight maximum of radiation 

 in the morning, one to two hours before sunrise. To the method 

 of Exner it can be objected that the radiation is only measured for 

 a part of the sky. In order to obtain the radiation to the whole sky, 

 Exner applied a correction with regard to the distribution of 

 radiation to the different zones given by Homen. It will be shown 

 in a later part of this paper that such a procedure is not entirely 

 reliable. 



In 1905 K. Angstrom" gave a description of an instrument 

 specially constructed for measuring the nocturnal radiation. The 

 instrument is founded upon the principle of electric compensation, 

 and, as it has been used in the work here published, I shall in a 

 following chapter give a more detailed consideration of it. With this 

 instrument Angstrom measured the nocturnal radiation during sev- 

 eral nights at Upsala and found values varying between 0.13 and 



1 Met. Zt., 1903, p. 409. 



2 Nova Acta Reg. Soc, Sc. Upsal., Ser. 4, Vol. 1, No. 2. 



