38 Mr. W. Sturgeon on the Distribution of 



the magnetic poles ; then the diameter n n' drawn at right 

 angles to the former line will be the line of demarcation which 

 separates the disc into the two required halves. 



Now when the disc revolves in the direction of the exterior 

 arrow, the quadrantal portion vi o n will advance towards the 

 pole m ; whilst the quadi'antal portion m a n' will recede from 

 it. 



Letc be any radius of the disc approaching the magnet m; 

 then, in order that any point c in that line may arrive at m, 

 it must necessarily partake of the direction c b, which would 

 bring it towards the side of the magnet; and also of the direc- 

 tion b m, which would carry it to within the magnetic poles : 

 and as the lines c b and b m are respectively the exact measures 

 of the spaces through which the point c would have to travel in 

 those directions, whilst approaching the magnet, and are both 

 performed in the same time, — they are also the faithful repre- 

 sentatives of the respective vican velocities with which the point 

 c is carried in each direction whilst advancing from c to m. 



Now as c 6 and b m are respectively the sine and versed 

 sine of the angle c a ?h, the mean velocity from c to m in each 

 direction will always be proportional to those lines, from what- 

 ever point of the quadrant the point c has to travel. If c tra- 

 vels through an arc of 90^, or from n to ?h, the mean velocity 

 in each direction will be equal, because 7i a = o?n; but if the 

 arc be less than 90°, the mean velocities will be unequal. If 

 the arc 7i c be 45°, the mean velocity from n to c will be in 

 favour of the direction a m ; but between c and m the predo- 

 minating velocity will be in the dii'ection of c b. 



Now as the excitation is more powerful in the neighbour- 

 hood of the magnetic poles than in any other part of the disc, 

 tlie vicinal area c o ?« of the quadrant n a m will constantly be 

 receivinc: stronger impressions than the remote area n a c. 

 And as the predommatmg mean velocity m the area c o in is 

 in the direction c b, the ascendant influence will consequently 

 be due to that direction of motion. 



With regard to the quadrantal area m o n', nothing more 

 appeared necessary to be understood than to resolve its curvi- 

 linear motion into rectilinear directions in the manner already 

 considered in the other part of the disc, supposing it to be 

 receding from the magnet, instead of advancing towards it. 



Under these considerations the experiments necessary for 

 the inquiry, which at first view had appeared to present con- 

 siderable difficulty, became very mucii simplified, being re- 

 duced to four rectilinear motions of the plate; — attending to 

 the velocity in each direction, and taking into calculation 

 the observed phyenoniena under each individual circumstance. 



The experinienlb were made with a rectangular plate of 



