g-points under the pressure
of one atmosphere differ greatly from the ratios of Tc, an approximate
confirmation of the law of Trouton may be compatible with an approximate
confirmation of the consequence of the law of corresponding states. If
we take the term boiling-point in a more general sense, and put T in the
law of Trouton to represent the boiling-point under an arbitrary equal
pressure, we may take the pressure equal to pc for a certain substance.
For this substance mr/T would be equal to zero, and the values of mr/T
would no longer show a trace of equality. At present direct trustworthy
investigations about the value of r for different substances are
wanting; hence the question whether as to the quantity mr/T the
substances are to be divided into normal and associating ones cannot be
answered. Let us divide the latent heat into heat necessary for internal
work and heat necessary for external work. Let r' represent the former
of these two quantities, then:--

r = r' + p(v_v - v_l).

Then the same remark holds good for mr'/T as has been made for mr/T. The
ratio between r and that part that is necessary for external work is
given in the formula,

r T dp
------------ = ----.
p(v_v - v_l) p dT

By making use of the approximate formula for the vapour tension:--

p /Tc - T\
log_[epsilon] --- = [int]' (--------), we find--
p_c \ T /

r Tc
------------ = [int]' --.
p(v_v - v_l) T

At T = Tc we find for this ratio [int]', a value which, for normal
substances is equal to 3/0.4343 = 7. At the critical temperature the
quantities r and vv-vl are both equal to 0, but they have a finite
ratio. As we may equate p(v_v - v_l) with pv_v = RT at very low
temperatures, we get, if we take into consideration that R expressed in
calories is nearly equal to 2/m, the value 2[int]'Tc = 14Tc as limiting
value for mr for normal substances. This value for mr has, however,
merely the character of a rough approximation--especially since the
factor f' is not perfectly constant.

Nature of a liquid.

All the phenomena which accompany the condensation of gases into liquids
may be explained by the supposition, that the condition of aggregation
which we call liquid differs only in quantity, and not in quality, from
that which we call gas. We imagine a gas to consist of separate
molecules of a certain mass [mu], havi