
of Cavitation in Lubrication. 335 
been explained by a theory in which the liquid sap is under 
tension. In general the effect has been studied in liquids at 
rest. It may be supposed that if the movement of the 
cylinder on the plane in the experiment described in § 6 
were sufficiently slow, the negative tension would not be 
great enough to break the liquid but would help in drawing 
the liquid forward to fill the space left behind the moving 
surface. Experiments in which the motion is slow do not 
show the formation of a cavity, and may be regarded as 
supporting this view of the action of the tensional strength 
of the liquid. 
§ 11. The slipperiness of ice * has been attributed to the 
presence of a layer of lubricating water under the body 
pressing on the ice. The water is produced by the lowering 
of the freezing-point where the pressure is experienced. 
On this view the object glides on a liquid layer and 
consequently viscous friction in water takes the place of the 
rubbing friction between the solids. Joly shows by calcula- 
tion that the weight of a man concentrated on the blade of 
a skate is sufficient to lower the freezing-point very con- 
siderably. Reynolds, arguing from the difficulty of slipping 
on very cold ice, comes toa similar conclusion. I wish to 
point out that sliding on a liquid layer is a condition under 
which cavitation will occur in the liquid, and that this will 
aid the slipping. I find that a cavity can be seen when 
a convex lens is pressed strongly on ice and pushed along. 
A cavity of this kind may be formed behind the sliding 
contact of a hog-back skate, and behind that of a curling- 
stone. 
§ 12. We may now point out the influence which the 
formation of a cavity has on lubrication. In ball-bearings 
completely immersed in oil the experimenis show that there 
must be a small cavity near the point of nearest approach of 
each ball to its neighbours and also to the surface on which 
it is running. As the friction of the bearing is the viscous 
friction of the oil, from the considerations in § 5 it follows 
that it must be reduced by the formation of these cavities 
which are filled with relatively non-viscous vapour. In fact, 
if a steel ball thoroughly oiled be rolled against an oiled glass 
plate the cavity may be directly observed. The high lubri- 
cating property of oils owes its origin not only to their superior 
viscosity but also possibly to the facility with which cavities 
may be formed in them. ; 
* Joly, Proc. Roy. Dublin Soc. p. 453 (1886) ; O. Reynolds, Mem. and 
Proc. of the Lit. and Phil. Soc. of Manchester, xliii, (1899). 
2A 2 
