ON COMPAKING AND REDUCING MAGNETIC OBSERVATIONS. 31 



Appendix II. Magnetic Dishbrbances at Fort Bae in 1882-83. 



At the International Polar Conference in Vienna, in the spring of 

 1884, the subject of the treatment of magnetic disturbances was much 

 discussed, but among the numerous schemes proposed none was univer- 

 sally accepted, and the matter was left undecided. 



I have thought it worth while to make trial, on the observations of 

 declination and horizontal intensity made at Fort Rae in 1882-83, of a 

 method founded on that proposed by Dr. Wild, and the results obtaiaed 

 will not, I hope, be without interest. 



From a comparison of lists of magnetically undisturbed days, supplied 

 by most of the circumpolar stations, Dr. Wild selected from four to six 

 days in each month, when the diurnal variation appeared to follow n 

 normal course ; and the hourly means obtained from these days are given 

 in the Fort Rae observations. When, however, these values are plotteii 

 down they do not afford a very regular curve, because, though these day s 

 are free from disturbance as a whole, a good many decidedly distui'bed 

 observations are included. The readings for hours of magnetic distur- 

 bances were therefore struck out, the question of disturbance beinLf 

 decided by reference to the original observation-book, for, as nine obser- 

 vations were taken at each hour, it was easy to see whether the instru- 

 ments were steady or not. Rather under 4 per cent, of the readings 

 were so struck out, and from the remainder a fairly satisfactory set of 

 hourly mean values was obtained for each month. 



The mean of these values at any hour for any two adjacent months 

 was then assumed to be the normal value at that hour on the middle dav 

 of the two months. Thas, for example, the mean horizontal intensitv 

 from the selected undisturbed days being, for January, at 11 a.m., '07653 ; 

 for February, at the same hour, •07661 ; and for March, -07669 ; then the 

 normal value for that hour was taken as "07657 on January 30, and •07665 

 on March 2. 



Having thus obtained a set of hourly values at intervals of about a 

 month, the values for the intermediate days were easily interpolated, an'l 

 in this manner a normal value of declination and horizontal intensity was 

 obtained for every hour of the year. 



By subtracting each of these values from the corresponding observed 

 value the ' disturbance ' at each hour was obtained, sometimes with a + , 

 sometimes with a — sign. These positive and negative disturbances weie 

 then entered in separate sheets, and their means are given in the appended 

 tables (I. — IV.) These means are obtained by dividing the sums of dis- 

 turbance by the number of observations ; not by the number of -)- or — 

 disturbances. 



In order to determine whether the larger and smaller disturbances 

 foUow different laws the disturbances were classified according to their 

 magnitude. Table V. shows the number of each class occurring at each 

 hour. 



H. P. Dawson, Maj. B.A. 

 April 20, 1887. 



