the Directive Power of Magnets, §c. 225 



Magnets upon the raised ~\ power of small magnet ( 



floor J power of horse-shoe magnet ~~ 1 oT ' 



Magnets 0*5 inch above \ power of small magnet i 



the raised floor . . J power of horse-shoe magnet ~~~ ' 36 ' 



Magnets 1*0 inch abovel power of small magnet l 



the raised floor . .J power of horse-shoe magnet ~~ 1 35 ' 



With so great inequality the results are necessarily irregular. 

 I use T -|-o as the proportion for comparison, without asserting 

 that it is accurate. All results obtained for the coil without 

 core ought therefore to be divided by 120, to make them 

 comparable with the other results. 



The results obtained for the direction of magnetical force now 

 consisted of lines drawn upon paper. Upon examining these, 

 some very small irregularities were found, generally of syste- 

 matic character — partly arising from minute failure in the 

 neutralization of terrestrial magnetism, partly from a difference 

 in the intensity of the poles of the great magnet; these were 

 eliminated by the following graphical process : — The paper was 

 bent upon its longitudinal axis, and exposed to a strong light 

 passing through the two folds of paper ; the lines drawn upon 

 both sides of the magnet or coil were visible, and a mean line 

 bisecting the small angle between each pair was easily drawn. 

 Then the paper was unfolded and was bent upon its transversal 

 axis, and a similar operation was performed upon the mean lines 

 mentioned above. Thus for one fourth part of the circum- 

 ference of the magnet a series of lines was obtained representing 

 the mean of the four parts; these mean lines, repeated for the 

 four divisions of the magnet's or coil's circumference, are alone 

 used in further graphical deductions and in the subjoined 

 figure (p. 226) . 



The results, however, for magnitude of force were obtained in 

 numbers. The means of these were taken in an analogous 

 order — first taking the sums of those on opposite sides of the 

 magnet or coil, then taking the sums of the last-found sums for 

 opposite ends. The division by 4 was omitted ; and thus the 

 numbers in the Table below give the value of 400 x cotangent 

 of deviation. At two stations the proximity of the coil-terminals 

 made it difficult to obtain actual observations ; but there was no 

 difficulty in supplying them conjecturally, with great confidence 

 in their accuracy. 



The diagram below was drawn carefully to represent the 

 positions taken by the small magnet when the edge of the large 

 magnet is presented to the small magnet. The same diagram 

 will serve, almost without perceptible error, for the case when 

 the flat side of the large magnet is presented to the small 



Phil Mag. S. 4. Vol. 46. No. 305. Sept. 1873. R 



