and on the Ball-ended Magnets of Robison. 



33 



sheet of paper, fixed to a drawing-board held firm by lead weights, 

 the magnetic north and south line OA is drawn and distances of 

 10, 12, 14... cm. from are marked off along OA. The mag- 

 netometer is then adjusted so that the point of its bob is accurately- 

 above 0, and the time of vibration under the earth's force is 

 determined. The Robison magnet is then placed with its lower 

 pole on one of the points marked off on OA, and is held, with its 

 axis vertical, by means of a clip. If the magnet be placed to the 

 south of the magnetometer, its north pole should rest on the 

 paper, so that the magnetic force due to the magnet may be in 

 the same direction as the earth's force. Two or three observations 

 of the time of 30 vibrations are then made for each position of the 

 magnet. If F be the horizontal component of the magnetic force 

 due to the magnet, the total horizontal force is F + H E , and this 

 is compared with H E by the method of § 4. If the centre of the 

 lower sphere be on the same level as the magnetometer magnet, 

 and if the distance between the centres of the spheres of the 

 Robison magnet be h, then, when the lower pole is at a distance r 

 from the magnetometer, the pole strength is given by 



m = r*F\l 



(r 2 + hff- 



For the following detailed results of an experiment of this 

 kind I am indebted to Miss J. M. W. Slater. 



H F = -176 gauss. T E = 6-127 sees, h = 30"5 cm. 



The close agreement among these values of m shews that with 

 careful observations this form of Coulomb's experiment leads to a 

 very satisfactory confirmation of the law of the inverse square. 



The magnet used in these experiments had a rod of " silver 

 steel " -23 cm. in diameter and was fitted with steel spheres J inch 

 (l , 27cm.) in diameter. I find that magnets of this size are con- 

 venient for this experiment. 



In conclusion I have to thank Mr W. G. Pye and Mr F. Lincoln, 

 the instrument makers at the Cavendish Laboratory, for their care 

 in the construction of the vibration magnetometers and of the 

 Robison magnets, as well as for some useful suggestions. 



VOL. XII. pt. I. 3 



