335 
25 
thanks to the laboratory and to Mr. Cunningham, whose active interest made 
the decisive tests possible. 
Three experimental tests of the kind described above, which were 
made at Aberdeen, will be discussed in §§16-20 which follow. 
The precise formulation of these problems is this: 
The equations of state which correspond to (19) are assumed to be 
as simple as possible without impairing the significance of the results, in 
order to facilitate the computations. Specifically 
(40) p = p, Cv) 
is assumed to be a polynomial. It is well known from hydrodynamics that in 
order to be realistic, the curve (40) must be convex from below. According- 
ly 
(41) Pr Pet) = T= Na Ry 
was chosen. This "substance" "collapses" (i.e. y=O ) for p =|, and it 
"cavitates" (i.e, p = © ) for v=2 3 however it behaves reasonably in 
the intervals 
(42) Ox We, SS pis Oo. 
and 
ts) Ne et pos! by =. Somers 
was used as the "normal", initial state of the substance. By (30) the 
sound velocity is 
(44) (ee We is Vie 
hence the "normal", initial value is 
(45) Cc, = Ji va). ee Ohl 
The boundary conditions correspond to the collision of this sub- 
stance with a wall, as described at the beginning of §11. It is preferable, 
however, to have a definite finite number of particles, i.e. of values ofa, 
