1869.] Lieut. Rokeby on Magnetic Observations. 



399 



The angles of deflection given are each the mean of two determinations. 



In deducing from these observations the ratio and product of the mag- 

 netic moment m of the magnet, and the earth's horizontal magnetic 

 intensity X, the induction- and temperature -corrections have always been 

 applied. 



In calculation of the ratio ^r, the third and subsequent terms of the 

 P Q 



series 1+— +-r-f &c. have always been omitted. The value of the con- 



y y 



stant P was found to be —0*00291, by the mean of ten determinations ob- 

 tained each from six pairs of deflection-observations at distances 1*0 and 

 13 foot. 



The mean of the values of ^ derived from deflection-observations at the 



distances 1*0 foot and 1*3 foot has been used in calculating the measure of 

 horizontal force. 



Observations made with the Differential Declinometer. 



Observations were made of the scale-reading of the differential declino- 

 meter at 3 a.m., and hourly from 6 a.m. to midnight, daily, from December 

 1863 to June 1865, from which date to the conclusion of the series the 

 3 a.m. observations were discontinued. 



The observations from the commencement until May 1864 were found 

 on inspection to be valueless for the purposes of reduction, for the reason 

 assigned in the introduction. 



The two years' records, June 1864 to 1866, were treated according to the 

 method employed by General Sabine in the reduction of declination-ob- 

 servations. 



In the first place, hourly means were taken for each month. All ob- 

 servations which vary from these means to a greater extent than 4'*0 were 

 then rejected, and new means taken of the unrejected observations. 



The mean reading for each month was then computed ; finally, the 

 hourly means were subtracted from the monthly means. 



Table 2 shows these differences as derived from the mean of two years' 

 observations, excepting in the case of 15 hours, which is the result of one 

 year only. 



In the Table, the sign + represents the north pole of the magnet to the 

 east of the mean position, and the sign — that it was to the west of the 

 mean. 



At the bottom of the Table semiannual and annual means are given, and 

 these means are exhibited in the form of curves in the diagram accompany- 

 ing the paper. 



Figure 1 shows the differences of the semiannual and annual means 

 from the normal position, which is represented by the straight horizontal 

 line. 



In figure 3 the annual mean is represented as a straight line, and the 



