384 



SCIENCE PROGRESS 



every tenth of an inch. This gave 200 figures. These were 

 then arranged in 15 sets, in rows of tens, elevens, twelves, 

 etc., up to twenty-fours, the rows of each set being arranged 

 above one another ; for example, the elevens gave 1 8 rows of 

 eleven and one row of two, and the set was arranged so that 

 they formed two columns of 1 9 and nine columns of 1 8. Each 

 column was then added. In the case of the elevens we had 

 eleven results ; these eleven results were multiplied in suc- 

 cession by cosO, cos — , cos—, etc., the results added and 

 the sum squared ; they were then multiplied in succession 

 by sinO, sin—, sin—, etc., the results added, and the sum 

 squared. The two squares were then added. The result 



Fig. 6. 



is proportional to the energy belonging to the wave-length 

 i*i in. The other sets were treated similarly. In the case of 

 the 8-in. length the procedure was the same, but only the 

 first 80 ordinates were required. 



The results are shown in fig. 6. The dotted curve gives 

 the energy distribution for the 20 in., and the smooth curve 

 the energy distribution for the 8 in. The vertical scale is ten 

 times as great for the smooth curve as for the dotted curve. 

 As mentioned above, the wave-length of the curve seemed 

 from inspection to be about \"j in. If the period during 

 which the amplitude is attaining its full value is taken by 

 itself, however, most energy is associated with the wave-length 

 1 '5 in. The dotted curve has two maxima, one at 1*5 in. and 

 the other at 178 in. 



