68 



ME. GEORGE C. SIMPSON ON THE 



rising at exactly the same time as the other rises and falls, and both remaining 

 constant during the summer. These curves suggest that there is some relation 

 between the two phenomena ; this relation will be discussed later in the paper. 



The ratio of the negative dissipation to the positive (q) does not appear from these 

 results to have a regular yearly course, but when they are considered in connection 

 with the ionization it will be seen that it is very likely there is a yearly variation 

 with a maximum in the winter and a minimum in the summer. 



Ionization. Table III. gives the monthly mean values of the ionization. 



TABLE III. Ionization. 



These results, together with the weekly means, have been plotted in Curve III. 

 Here we have quite a different curve from either of the two previous ones. Instead 

 of the rapid fall and rise in the winter followed by a constant period during the 

 summer we have a six months' linear fall from August to February followed by a 

 similar six months' linear rise from February to August. 



That there should be such a great difference between the curves for the dissipation 

 and the ionization was not to be expected, and at first one would be inclined to doubt 

 the correctness of one or other of them. But this can be tested by the following 

 considerations. The dissipation depends practically only on two factors : ionization 

 and wind strength. If the effect of the latter could be eliminated, the course of the 

 dissipation should then be the same as that of the ionization. In order to see if this 

 were so, I took all my measurements of the dissipation and separated them according 

 to the strength of the wind as estimated at the time of observation, then, using only 

 one definite wind strength, took the means for each month and plotted them. ' The 

 result is shown in fig. 2. Each curve represents one wind strength, and it will at once 



