958 
in that manner the very approximately rectilinear character of a 
large part of the diameter shows better and also because by 
that method the critical density ozq is obtained in the most rational 
way. Leaving out the deviations in the neighbourhood of the critical 
temperature and confining ourselves to those which refer to the 
direction of the diameter from (3) below the critical temperature 
downwards, it may be remarked, that these deviations although 
somewhat larger than those found with many other substances still 
in their general character do not differ from those shown by say 
carbon-dioxide and argon; the latter substances also possess at the 
lower temperatures a diameter which is convex on the side of the 
temperature-axis. 
As regards the deviations near the critical point, although appa- 
rently they increase systematically, we cannot help being surprised 
at the exceptionally high amount of them (nearly 2°/,). A syste- 
matie fault of the apparatus or the method of working, although in 
the region of the higher temperatures our apparatus does not gua- 
rantee the same degree of accuracy which we may consider as 
assured in the region of the lower temperatures, may be considered 
as excluded by the results obtained with oxygen and argon. Again 
although the manner in which in the absence of all experimental 
material concerning the isotherms of nitrogen at low temperatures 
we were obliged to apply the corrections for the dead spaces in the 
glass capillary in the cryostat (and therefore at the low temperature) 
is open to some objection *), still deviations of the said magnitude 
cannot thereby be explained. 
In fact these corrections need not be known with more than a 
very small relative accuracy. It would be less improbable to look 
for the explanation in the presence of some admixture whose influence 
was too small to make itself felt in the test referred to by means 
of evaporation under constant pressure. Taking everything into account 
we consider it probable, that the large deviations do not find their 
explanation in the uncertainties of observation or calculation, but 
are connected with the deviation of the critical point from the extra- 
polated vapour-pressure curve. In order to find a conclusive answer 
to the questions raised here, a special investigation will be needed 
of the diameter of nitrogen in the immediate neighbourhood of the 
critical point with apparatus specially constructed for the purpose, 
us also in general a thermodynamical investigation in that region in 
which amongst other things the mixing of the phases by stirring 
will have to be provided for. 
a) Comp. above § 4. 
