d for the
movements of matter taken as a whole, should not be applied outside
the domain which belongs to them. Atomism, in fact, tends more and
more, in modern theories, to imitate the principle of the conservation
of energy or that of entropy, to disengage itself from the artificial
bonds which attached it to mechanics, and to put itself forward as an
independent principle.

Atomistic ideas also have undergone evolution, and this slow evolution
has been considerably quickened under the influence of modern
discoveries. These reach back to the most remote antiquity, and to
follow their development we should have to write the history of human
thought which they have always accompanied since the time of
Leucippus, Democritus, Epicurus, and Lucretius. The first observers
who noticed that the volume of a body could be diminished by
compression or cold, or augmented by heat, and who saw a soluble solid
body mix completely with the water which dissolved it, must have been
compelled to suppose that matter was not dispersed continuously
throughout the space it seemed to occupy. They were thus brought to
consider it discontinuous, and to admit that a substance having the
same composition and the same properties in all its parts--in a word,
perfectly homogeneous--ceases to present this homogeneity when
considered within a sufficiently small volume.

Modern experimenters have succeeded by direct experiments in placing
in evidence this heterogeneous character of matter when taken in small
mass. Thus, for example, the superficial tension, which is constant
for the same liquid at a given temperature, no longer has the same
value when the thickness of the layer of liquid becomes extremely
small. Newton noticed even in his time that a dark zone is seen to
form on a soap bubble at the moment when it becomes so thin that it
must burst. Professor Reinold and Sir Arthur Ruecker have shown that
this zone is no longer exactly spherical; and from this we must
conclude that the superficial tension, constant for all thicknesses
above a certain limit, commences to vary when the thickness falls
below a critical value, which these authors estimate, on optical
grounds, at about fifty millionths of a millimetre.

From experiments on capillarity, Prof. Quincke has obtained similar
results with regard to layers of solids. But it is not only capillary
properties which allow this characteristic to be revealed. All the
properties of a body are modifi