566 
CHEMISTRY: W. D. HARKINS 
stant of attraction {a) is additive, and therefore comes to the conclusion 
that all such forces are physical. The calculations of Einstein,^ Klee- 
mann/ and of Harkins and Clark/ have also given coefficients of atomic 
attraction which are moderately exact constants. Since all of these facts 
when considered together make it probable that cohesional forces are 
often less specific than those involved in ordinary chemical reactions, 
while in many cases they are the same valence forces, it seems to me pre- 
ferable to use neither of the two words, physical or chemical, and to 
consider that cohesion is due to electrical and magnetic, or electromag- 
netic forces, which are probably largely electrical. In a paper ^° on 
"An Electromagnetic Hypothesis of the Kinetics of Heterogeneous 
Equilibrium, the Structure of Liquids, and Cohesion" I have already 
traced a connection between cohesion and the completeness of the 
outer or valence shell, of electrons in the atom or the molecule. The 
cohesion decreases as the comxpleteness of the outer shell of electrons 
in the molecule increases. 
On the relation between cohesion and cohesional and adhesional work 
and energy. — A number of attempts have been made to calculate the 
cohesion, which is the cohesional force per unit area of the cohesional 
pressure, from the cohesional surface vjork, and this attempt has been 
more or less justified by the fact that the values thus obtained, while 
not good in numerical agreement with those given by a/v^, lie on the 
whole in the same relative order. Such calculations have been made 
by Mathews,^ and by Hildebrand,^^ but neither of them have shown to 
what extent the cohesion and the cohesional surface work are related. 
The term a/v^ may be said to represent, more or less imperfectly it is 
true, the total effect of the molecular attraction in decreasing the 
external pressure, which decrease is the cohesion. The cohesional sur- 
face work, on the other hand does not represent the total effect of the 
molecular attraction, even as it acts in a surface, since the molecules 
move into the surface not only by means of the energy which is con- 
tributed in the form of work, but also by means of the utilization of the 
kinetic energy of molecular motion, or the latent heat of the surface. 
Thus the formation of the surface of a pure liquid, with the exception 
of a few liquids in which liquid crystal formation is involved, is always 
accompanied by cooling. It is, therefore, the total cohesional surface 
energy and not the related work, which represents the total effect of the 
molecular attraction, or the cohesional effect. 
It may seem remarkable, from the point of viev/ of the last paragraph, 
that the calculation of even the relative cohesion of liquids^^ from a 
