72' 



TEMPERATURE. 



This is obviously the key to the whole problem, the solution of which would be an important 

 addition to our knowledge of the physics of the atmosphere. Already I have spent more 

 time over this problem than over any other discussed in this work. As the investigation has 

 led only to negative results, this is not the place to go into details. It is to be hoped 

 that the problem will be taken up a^ain when more data become available in the future. 



Fourier Coefficients. 



The Fourier Coefficients for the daily variation of temperature are given in the following 

 table. The coefficients lead to no new results. The value of a— the amplitude of the whole- 

 day period — is greatest in January and it is slightly le.ss in December than in the two neigh- 

 bouring months. The phase of the whole-day period does not vary to any large extent from 

 September to March, when the daily variation is governed by the sun. During the period 

 April to August, when the solar energy is weak, both the amplitude and phase of the whole- 

 day period are extremely irregular, this is to be expected from the curves for this period 

 mven in figure 20, page 54. The amplitude of the half-day period is small throughout, and 

 the irregularity of the phases for the different months shows that this jjeriod has no real 

 physical basis. 



Table 30. 

 Fourier Coefficients of the Ditihj Variation of Temperature. 



