No. 3.] ATOMS AND MOLECULES. 571 
thrown, no one thinks of saying that the house and ball repel 
each other ; instead, it is recognized as the result of the impact 
of elastic bodies. In like manner, the impact of elastic mole- 
cules upon the sides of the containing vessel produce a pressure 
upon it proportional to the momentum of the molecules; and 
when molecules collide, they have their direction of movement 
changed, and they separate. 
Chemical affinity is another term that seems to imply that 
atoms are endowed with essential selective properties by which 
they are enabled to combine in definite ratios. In late years 
chemists have adopted the term Chemzsm in place of chemical 
affinity, and have given to it a greater range of proclivities, 
finding no difference but one of degree between it and cohesion. 
It makes but little difference to one who uses it who is con- 
cerned solely with chemical reactions ; but it makes a good deal 
of difference to one who is to explain the relations of chemical 
changes to other physical changes, whether or not, as soon as 
he touches atomic phenomena, he has to import into his discus- 
sion a factor which has no physical antecedent that can move 
and arrange atoms and molecules thus and thus, and yet itself 
not be subject to the laws that obtain among larger masses. 
As chemists no longer contend for a chemical force that is not 
derived from antecedent physical conditions, it is alluded to 
here for the double purpose of calling attention to the positive 
mischief that often comes from inappropriate terminology, and 
to the fact that, so far as my knowledge goes, chemists have 
not attempted to give a physical explanation of the cohesion of 
atoms into molecules, but have stopped with chemism, as if it 
were an ultimate fact or property. 
To give such a physical explanation of chemism or atomic 
cohesion, and to extend it to the building up of geometrical 
crystalline forms, is the object of this paper. It is not claimed 
to be demonstrated, but as a hypothesis having a good degree 
of probability in its favor. But even if it be inadequate, at least 
it gives one a mechanical idea of how such forms necessarily 
result from mechanical conditions. 
Investigations into the relations of chemical reactions to heat 
have developed the fact that all chemical changes involve 
definite quantities of heat, either absorbed or evolved. That all 
such changes are absolutely dependent upon the exchanges in 
