14 A. ÏANAKADATE. 



As these effects vary with the squares of amplitudes they will 

 be entirely negligible when the deviation from the normal ])Osi- 

 tion is one or two minutes. 



llie Deßectloii Krjteriment was done wdth two distances which 

 y re fixed in each l^ar {i\ V2 of § 4). The order of various posi- 

 tions of the magnet was Fier. 3. 

 as indicated in Fig. 3, ^^^ ^^^ ^^^ ^^^ 



(1) (2) (3) (4), ^'' -= = ^ = -^ ^ 



( — 4) ( — 3) ( — 2) (~1) and again (1) for check, where — sign 

 shows that the magnet is reversed end for end. This order 

 seems to be the usual one : it is found though too late that 

 better order is (1) (-1), (2) (-2), (3) (-3), (4) (-4), reversing 

 at each particular position ; series of experiments were made for 

 the express purpose of comparing the methods, and it was found 

 that the latter order gave more consistent values than the 

 former. The reason is obvious, in the former method the slight 

 change in declination during the experiment, which takes about 



15 minutes in our case, produces error in the angle of deflec- 

 tion, while in the latter method this error is minimized. 



The temperature coefficient of the magnetic moment of each 

 bar magnet, was found directly from the experiment ; sufficient 

 differences of temperature being usually obtained from the morn- 

 ing and afternoon observations as is seen in the complete list 

 under the headings " M." and " Temp." The fractional decre- 

 ment of moment per degree centigrade varied between .00070 

 and .00082 and the correction is ajiplied as shown in the speci- 

 men i)age where tv and ta are the mean temperatures in vibra- 

 tion and deflection experiment respectively. 



The Dip Circle was set into the magnetic meridian by the 

 usual process of observing its position in which the needle stands 



