Mr. H. A. Rowland on Magnetic Distribution. 267 



seems to me that this method is the only one capable of giving 

 exact results directly. 



The coils of wire which I used consisted of from twenty 

 to one hundred turns of fine wire wound on thin paper tubes 

 which just fitted the bar and extended considerably beyond 

 the coils. The coils were mostly from *1 to *25 of an inch 

 wide and from '1 to \2 inch thick. A measure being laid by 

 the side of the given bar under experiment, the coil was 

 moved from one division of the rule to the next very quickly, 

 and the deflection produced on an ordinary astatic galvanometer 

 noted. After experience this could be done with great accuracy. 

 It might be better in some cases to have the coil slide over a 

 limited distance on the tube, though for the use to which I 

 intend to put the results the other is best. 



Up to 35° Q e is nearly proportional to the deflection ; and 

 when any larger value is put down in the Tables, it is the sum 

 of two or more deflections. I have not the data in most cases 

 to reduce my results to absolute measure, but took pains to 

 ensure that certain series of experiments should be comparable 

 among themselves. 



Having measured Q 6 at all points of a rod, we may find Q by 

 adding up the values of Q e from the end of the rod. 



The magnetizing force to which the bar was subjected was in 

 all cases a helix placed at some part of the bar. The iron bars 

 were of course demagnetized thoroughly before use by placing 

 them in the proper position with reference to the magnetic me- 

 ridian and striking them. 



In the Tables L is the distance in inches from the zero-point, 

 Q e is the deflection of the galvanometer when the helix is passed 

 between the points indicated in the first column. Thus, in 

 Table II., 34*7 is the deflection on the galvanometer when the 

 helix was moved from the tenth to the eleventh inch from 

 the zero-point ; and so we may consider it as the value of Q e at 

 10^ inches ; so that the values of Q e refer to the half inches, 

 but Q to the even inches. 



In all the calculations the constants in the formulae were taken 

 to represent Q. most nearly, and then the corresponding formulae 

 for Qe taken w 7 ith the same constants. 



For ease in calculating by ordinary logarithmic Tables, we 

 may put ^sslO" 4 *** 



IV. 



Table I. is from a bar 17-g inches long with a magnetizing 

 helix H inch long at one end, the zero-point being at the other. 

 Table II. is from a bar 9 feet long with a helix 44 inches long 



