94 On Friction between Surfaces at Low Speeds. [Mar. 22, 



of the material whose friction against steel was to be measured. The 

 disk was caused to revolve and then left to itself, when it came to rest 

 in consequence of the friction on the ends of the axle. The rate of 

 retardation was found as follows : — A strip of paper 2| inches broad was 

 stretched round the periphery of the disk, and a pendulum was caused to 

 swing across this paper in a plane perpendicular to that of the disk. On 

 the pendulum was fastened a fine glass siphon, one end of which dipped 

 into a box containing ink, whilst the other stood at a short distance from 

 the paper strip, across which it was carried as the pendulum oscillated. 

 By keeping the ink-box strongly electrified ink was deposited on the paper 

 by the point of the siphon in a rapid succession of fine spots. By this 

 means, without the introduction of any new source of friction, a per- 

 manent record was made of the resultant motion of the pendulum and 

 the revolving disk. This frictionless method of recording was designed 

 by Sir William Thomson for telegraphic purposes, and is employed in his 

 siphon recorder. From the curve drawn in this way it was easy to 

 determine the rate of retardation of the disk (and therefore the friction) 

 corresponding to various velocities of the rubbing surfaces. The lowest 

 velocity for which the determinations were definite was about 0*0002 foot 

 per second, and the highest velocity to which the experiments extended 

 was 0*01 foot per second. The surfaces examined were steel on steel, steel 

 on brass, steel on agate, steel on beech, and steel on greenheart — in each 

 case under the three conditions, dry, oiled, and wet with water. In the cases 

 steel on beech oiled or wet with water, and steel on greenheart oiled or 

 wet with water, the coefficient of friction increased as the velocity dimi- 

 nished between the two limits given above, the increase amounting to 

 about twenty per cent, of the lower value. It appeared that at the 

 higher limit of velocity there was little further tendency to change in 

 the coefficient ; but it is impossible to say how much additional change 

 might take place between the lower limit of the velocity and rest. In 

 the case of steel on agate wet with water there was a similar but much 

 less marked increase of friction as the velocity decreased ; and in the 

 case of steel on steel oiled there was a slight and somewhat uncertain 

 change of the opposite character — that is, a decrease o2 friction as the 

 velocity decreased. This case, however, would require further examina- 

 tion. In all other cases the friction seemed to be perfectly constant and 

 independent of the velocity. Out of all the sets of circumstances inves- 

 tigated, the only ones in which there was a large difference between the 

 static and kinetic values of the coefficient of friction were those in which 

 a decided increase was observed in the kinetic value as the speed 

 decreased. This result renders it exceedingly probable that there is con- 

 tinuity between the two kinds of friction. 



