ENTROPY AND PROBABILITY 59where the mean free path of the molecules is small compared with the linear
dimensions of the container [E38]. At that time it was believed by some that the
motion of foils in a radiometer was somehow induced by radiation pressure. Ein-
stein's paper, which complements earlier work by Knudsen, was a contribution
toward the elimination of this incorrect idea.
1924-5. Three papers on the quantum theory of a molecular gas; discovery
of the condensation phenomenon named after Einstein and also after Rose; Ein-
stein's last application of fluctuation theory, which leads him to particle-wave
duality for matter by a route independent of the one taken earlier by de Broglie
[E39, E40, E41].
Reviews. In 1911 Einstein summarized the status of the specific heat problem
before the first Solvay conference [E42]. In 1915 he wrote a semipopular review
on kinetic problems [E43].This concludes the introductory summary of Einstein's work on statistical phys-
ics and related subjects. I shall, of course, return in more detail to the main topics
mentioned in this chronology. Sections 4c and 4d deal with the 1902-4 papers and
with Einstein's subsequent involvement with Boltzmann's principle. Chapter 5,
which opens with introductory remarks on the highly complex subject of molec-
ular reality in the nineteenth century, is devoted mainly to Einstein's doctoral
thesis, Brownian motion, and critical opalescence. All the principal papers men-
tioned above that belong to the area of quantum physics will be discussed in Chap-
ters 19 to 24.
At the beginning of this section, I remarked that Einstein devoted some but not
much attention to his contributions to statistical physics when, at age seventy, he
looked back on his work. At that time, he had much more to say about his rela-
tivity theories and devoted more space to his critique of quantum mechanics than
to all the work summarized above [El]. It is an additional purpose of the foregoing
chronology to make clear that in doing so he did not fully convey the breadth of
his life's work.
Einstein's position regarding questions of principle in statistical mechanics is
best explained by first reviewing briefly the contributions of Maxwell and, espe-
cially, of Boltzmann. Gibbs will not enter into this review because he did not
influence Einstein and also because, as Lorentz noted in Einstein's presence, the
Einstein and Gibbs approaches are different [L2]. Einstein did not disagree.
Indeed, in responding to Lorentz's remark, he observed, '[My] point of view is
characterized by the fact that one introduces the probability of a specific state in
a phenomenological manner. In that way one has the advantage of not interposing
any particular theory, for example, any statistical mechanics' [E44]. His critical
attitude to Boltzmann's approach, implied by this statement, will be discussed in
Section 4d. One of the aims of this chapter is to explain what Einstein had in
mind with his phenomenological approach.
In concluding this introduction, I note that the period of Einstein's activities