We avoid the gravest difficulties when, giving up the attempt to frame hypotheses concerning the constitution of matter, we pursue statistical inquiries as a branch of rational mechanics.

From the preface to Elementary Principles in Statististical Mechanics (1902), p. ix.