[A few words are necessary to explain why the present paper has hitherto been printed in Abstract only, and to show what modifications it has undergone since it was read more than two years ago.

In the paper, as it was first presented to the Society, I contented myself with the usual practice of extracting from the virial a negative term (— βp) to represent at least a portion of the part due to the molecular repulsion at impact. But, as will be seen by the Abstract printed at the time (Proc. Roy. Soc. Edin., 21/1/89), I stated that though this procedure is correct when molecular attraction is not taken into account, it requires considerable modification when such attractions are introduced. I also stated that its main effect would be to alter one of the disposable quantities (A) in my equation. I have since seen that the definition, of what we are now to understand by “temperature,” which I then introduced, leads naturally and directly to the writing of a part of A in the form

−e(E+C/(v+γ)),

where E is proportional to the absolute temperature and to the average energy of a free particle. This remark really substitutes the new undetermined quantity e for the β which occurred in my former expression.