20 REPORT OF THE SECRETARY. 



knowledge of one of the most remarkable classes of phenomena pre- 

 sented in the whole course of physical science, the results of this 

 discussion cannot but be received with much interest by the scientific 

 world. 



As the magnetic needle, for example, may be considered as sub- 

 jected at the same time to different forces, each tending to produce 

 one of the variations we have mentioned, it becomes a subject of nice 

 inquiry to eliminate the several effects, and to obtain the magnitude 

 and period of each separately. In the case under consideration it 

 was necessary to separate more especially the large apparently fitful 

 variations from the regular daily ones. To effect this, the process 

 proposed by Professor Peirce, of Cambridge, and founded on the 

 -doctrine of probabilities, was employed as a criterion in judging as 

 to the magnitude of a disturbance which should be considered as be- 

 longing to the class under consideration, and it was finally concluded 

 that all disturbances which exceeded 3'. 64 of an arc were abnormal, 

 and accordingly all observations differing by that amount or more from 

 their mean monthly values were marked. Next, a new hourly mean 

 was taken, omitting the values so marked, and each observation again 

 examined in reference to deviations from this new mean, and so on — 

 the last mean thus obtained for each hour during each month gave 

 what was considered the normal daily curve. 



From this it appears that the north end of the needle reaches its 

 ■greatest eastern position between 7 and 8 o'clock in the forenoon, and 

 its greatest western deviation about 1 o'clock in the afternoon. 



The author next proceeds to discuss the large disturbances, and 

 from these lie deduces the fact that a principal maximum of disturb- 

 ances occurs in October, a smaller one in April, and the two minima, 

 .nearly equal to each other, occur in the months of February and 

 •June. 



The diurnal variation arising from the large disturbances presents 

 one maximum and one minimum; its most prominent feature is the 

 easterly deflection, which occurs about a quarter after 8 o'clock 

 p. in., at which hour the maximum deflection amounts to 32" of an 

 arc; the great westerly deflection takes place at a quarter past 6 

 a. m., and ©n an average amounts to 14". 



These variations are compared with deductions made from similar 

 observations at Toronto, and are found to be the same in kind, but 

 less in magnitude. 



