36 Mr. W. B. Hardy and Lieut. J. K. Hardy 



The appearances suggest that it is owing to minute and very- 

 rapid vibrations or oscillations of the watch-glass at right 

 angles to the direction of pull. When the glass tears the 

 '•■give" is not directly in the line of the pull, the watch- 

 glass is thus thrown out of its balance to one side or the 

 other with the result that it chatters forward vibrating 

 rapidly from side to side. The breadth of the track is 

 determined by the amplitude of these vibrations. When a 

 watch-glass is moved by hand, owing to the unsteadiness of 

 the motion the track commonly begins in more than one 

 (five or six) fine lines which keep separate for say 40 /jl and; 

 then blend into one wide track. 



The appearance under the microscope is consistent with; 

 the view that the faces are seized before movement starts,, 

 and the thesis developed in these pages is that static friction 

 both of clean and lubricated faces is due to cohesion which,, 

 in the case of clean faces, causes the glass itself to seize and 

 in the case of lubricated faces causes the film of lubricant on 

 the one face to seize to that on the other face. 



We can find no justification for the view commonly held 

 that static friction is due to inequalities of the surfaces. 

 Actual measurements proved that the threshold lvalue for 

 glass with an " optical^ face is the same as that for ordinary 

 plate glass, whilst the value for ground glass is lower pro- 

 bably because it is impossible to clean it. It static friction 

 be due to inequalities these must be of insensible magnitude,, 

 and "inequalities " of this order are indistinguishable from 

 the attractions of individual molecules* across the interface. 



The weight of the watch-glass will form a depression in- 

 the plate, and it might be supposed that resistance to slip is 

 due to this fact, the surface of the plate being heaped up 

 into a crest which ruptures and so starts the tear when the 

 stress exceeds a certain limit. This amounts to supposing 

 that the watch-glass acts as a cutting tool. Deformation 

 of both surfaces must occur, but the action cannot be of the 

 kind just described since both faces are torn, and always in 

 such a way as to detach very thin flakes. The microscope 

 reveals no sign of a burst through. 



The resistance offered to slip is large. If it be due to 

 cohesion, why should there be no sensible resistance to 

 displacement of the surfaces along the normal ? 



Two answers may tentatively be jgiven to this very difficult 

 question. The first is that, in movement along the normal,. 



* This word is used merely to denote the units out of which the 

 matter is built. 



