206 
FREDERICK BARRY 
It is a simple matter to define temperature in terms of these con- 
ceptions. According to the kinetic theory of gases, which is the most 
thoroughly developed application of the atomistic conception of 
matter — and which in its fundamental assumptions is supported not 
only by its consistence with fact, but recently by the actual observa- 
tion and measurement of ceaseless and independent movement among 
particles approximately of molecular magnitude in colloidal systems 
— the temperature of a gas is directly proportional to the square of 
the mean rectilinear velocity of its molecules. More generally, it is a 
necessary consequence of the fundamental assumptions of this theory 
that thermal equilibrium between two or more gases indicates the 
condition that the mean molecular kinetic energies of both or all is 
the same, and that this is measurable in each system by the product of 
the molecular mass and the square of its velocity. Since now it is 
possible for a substance to pass without any abrupt change of condi- 
tion from the gaseous to the isotropic liquid and amorphous solid 
state, it is inevitable that we believe this correlation between tempera- 
ture and mean molecular energy of translation to exist in every state 
aggregation, assuming only that in crystalline liquids and solids the 
motion is restricted to a certain definite periodicity (3). 
Now, since a change in the total thermal energy of a system is a 
function not only of the rise in temperature, but also of some specific 
character of the substance heated — since, in other words, equal 
increments of total energy do not occasion the same temperature 
rise in equal masses of different substances — it follows from what has 
been said that this specific character, the "mass factor" of heat energy, 
or heat capacity, is a value dependent on the particular constitution 
of the molecule itself. Consequently, specific heats, which measure 
the relative heat capacities of equal masses of different substances, 
become of great theoretical interest. 
All are familiar with the success which has attended the attempt 
to explain chemical relationships and to account for the existence of 
isomeric substances, especially among the compounds of carbon, by 
assigning to each substance a definite, though conventional, molecular 
configuration; and it is equally well known how, in order to explain 
optical activity and optical isomerism, this representation has taken 
on a less conventional character, by assuming the existence of geom- 
etrical arrangement among the atoms in space (4). No one believes 
that such formulae represent actual static conditions : too many 
