Friction seizure of clean solid faces

It is usually found that dean surfaces cannot be made to slide on each other without considerable damage. Hardy1 found that a curved slider of clean glass, moving on a flat glass plate, tears a track which begins 

The area of real contact was, with a curved slider, moving on a flat plate, approximately proportional to the load pressing the two surfaces together. It was also nearly equal to the area which would be theoretically required to support the metals under the applied load, calculating from 1 Proc. Roy. Soc. A, 169, 391 (1939). 

The proportionality between the area of real contact and the applied load supplies at last a rational explanation of the well-known Amontons 1 or Coulomb s2 law of friction, that the frictional force F is directly proportional to the total load P pressing the surfaces together. The meaning of this law has long been mysterious. Under ordinary circumstances the law holds fairly accurately, i.e. there exists a nearly constant coefficient of friction, /x = F/P. The frictional force is naturally proportional to the total area of these bridges consequently the frictional force should be proportional to the load. 

With both surfaces flat, the area of real contact was more difficult to estimate absolutely and appeared to increase more rapidly than in direct proportion to the load but there is some uncertainty here as to the actual number of regions of real contact.3 

It should be noted that the area of real contact is, as a rule, the minimum area required to reduce the pressure produced by a given normal load to that at which the metal no longer flows plastically. It is therefore necessarily proportional to the applied load. If the deformation were elastic, not plastic, the area of contact would be proportional to the two-thirds power of the load, assuming that the projections at which real contact commenced were segments of spheres. 

---