192 ANNUAL REPORT SMITHSONIAN INSTITUTION, 1908. 



the strength of the signals received at near-by stations was the same 

 during the day as during the night, while there was great variation 

 in the strength of signals received at stations farther away. 



It was also found that the absorption at a given instant was a func- 

 tion of the direction as well as of the distance, since on a given night 

 the signals received by stations in one direction w^ould be greatly 

 weakened, while there Avould be less weakening of the signals received 

 by stations lying in another direction, while a few hours or a few 

 minutes later the reverse would be the case. 



This was thought to be connected with the coming weather condi- 

 tions, but before this fact is proved a much larger amount of data 

 must be collected. Through the kindness of the United States 

 Weather Bureau I was enabled to obtain a chart of the magnetic 

 variations, and on comparison of these with the absorption between 

 the Massachusetts and Scotland stations there appeared to be a quite 

 definite relation, i. e., the greater the absorption the greater the mag- 

 netic variation. Here also, however, much more data is needed before 

 arriving at a definite conclusion. The fact that the absorption did 

 not take place in the neighborhood of the sending station having thus 

 been definitely settled the next point to be investigated was whether 

 or not there was any way of overcoming it. 



The fact that variations in the absorption occurred with extreme 

 rapidity, the absorption increasing sometimes a hundred fold in a 

 single minute, and at night, when the effect could not be due to the 

 sun directly, seemed to indicate that the body producing the absorp- 

 tion, whatever it was, was not in a state of continuity, but was 

 broken up into masses like clouds." This also was in accordance with 

 some experiments made in Brazil in 1905. 



From optical theories it is known that where the absorption is 

 produced by conducting masses of a more or less definite size the 

 absorption is to a certain extent selective. The next point in the 

 investigation was, therefore, to determine whether there was any 

 possibility of this being the fact in the case of the absorption of 

 wireless signals. 



Comparative tests were therefore made of the absorption at night 

 and during the day between Brant Rock and Washington, with wave 

 lengths varying from a fraction of a mile up to four or five miles. 

 It was found that the absorption did not increase continuously with 

 the wave length, but reached a maximum and then fell off with great 

 suddenness. 



Figure 8 shows the general character of the curve, the ordinates 

 referring to the amount" of the absorption and the abscissas to the 

 wave frequency. 



" Electrical Review, May 18, 190G. 



