516 Prof. W. McF. Orr on Clausius’ 
its steady state at rest in contact with the same atmosphere 
at rest, so much so that meteorites moving through the atmo- 
sphere are frequently heated to incandescence, and that if 
bodies composed of different substances, or the same substance 
but of different sizes, are so moved through the air, their steady 
states would not in general be such as could be maintained in 
contact with each other. We usually express these facts by 
stating that the temperature of the moving body is raised 
above that of the atmosphere to such an extent that the rate 
at which heat is generated in it by friction against the atmo- 
sphere is equal to the rate at which it loses heat by radiation 
&c.; the amount of this excess depends on the velocity, size, 
and nature of the body. 
Yet occasionally observers attempt to measure the tempera- 
ture of the air by whirling a thermometer through it *, 
It is then, I think, doubtful whether in such an extreme 
case as that of a gas rushing into a vacuum the temperature 
of any portion of it, however small, has at all times a meaning 
which is definite and in accordance with the laws which 
experience shows to hold for bodies at rest. (Bertrand’s 
objection to the use of the word ‘temperature ” appears, 
however, not to be that in a state of extreme agitation the 
“temperature ” of a small part has not a clear meaning, but 
merely that its value differs from point to point.) In many 
cases of less turbulent motion Planck’s view might be justified, 
and his definition of the entropy of a perfect gas therefore 
be allowed as valid. 
Bertrand’s objection to the term ‘‘ Pressure.” The Nature 
of the Stresses in a Viscous Fluid. 
11. But in establishing the theorem that the entropy of 
a system continually increases, Planck supposes, in the argu- 
ment of Arts. 2-4 above, that the system can at any time 
be divided into a number of small particles, each of which 
can be restored to its original state by a series of infinitely 
slow adiabatic and isothermal changes, constituting in fact a 
succession of equilibrium states in each of which the portion 
considered is subject to pressures wholly normal. Bertrand’s 
objection to the validity of such an argument seems to have 
great force. It may be pardonable here to recall some of our 
fundamental ideas as to the nature of the stress in fluids, as the 
bodies discussed in thermodynamics are usually either fluids 
or, if solids, are subjected only to such stresses as could exist 
in fluids. Defining a fluid as a body having the property that 
* E. g. Nansen, ‘ Farthest North,’ 
