My work for this edition has been mainly a revising of the text in the light of recent contributions to the literature. Many new references have been added, and there are also certain changes of emphasis.

The difficulties in the way of establishing chemical thermodynamics in a fully rigorous manner have been described afresh by Munster in his Classical Thermodynamics (1970). As he has said, the ‘laws’ do not constitute a complete set of axioms, especially in the case of systems having variable composition.

As regards entropy, one way of dealing with these difficulties is simply to postulate its existence, rather than seeking to prove it. However this method seems to me not sufficiently satisfying for the student. Far better, in my view, to put forward the classical arguments as well as they can be put, and to develop simultaneously the statistical interpretation of the second law, so as to create a linkage of thermodynamics with the rest of physics and chemistry.

This leaves my previous scheme for Chapter 1 essentially unchanged. But I have become better aware than previously, especially from Popper, that there is a certain hazard in using the statistical argument, even at the elementary level of the present volume. If the argument is put forward in terms of ‘lack of information’ about micro-states, this may well create the impression, although quite unwarrantably, that thermodynamics contains very subjective elements. Some of my re-writing has been intended to correct that impression.