(481 ) 
representing graphically in the same way as in fig. 1 how the 
density of a substance is distributed in a tube which is filled with 
carbon dioxide, mixed with a small proportion of oxygen, if that 
admixture increases in concentration from the bottom upwards. I 
also suppose that the temperature first rises from the critical tempe- 
rature of 31° C. to 38° C., and then falls again to 31° C. Further 
I assume that the concentration of the oxygen which at first decreased 
regularly from the top downwards, so that the greatest difference 
of concentration was 0.001 mol., at last, owing to a more rapid 
diffusion in the upper space, varies there less rapidly with the height 
than in the lower space’). Fig. 2 thus obtained, may really be 
looked upon as a diagrammatical reproduction of fig. 1; in the 
falling period the density curves, as in fig. 1, show a point of in- 
flection; in the upper half no maximum has yet been reached by 
the curve 0.450, but by adopting a more rapid diffusion in that 
space I might have brought about also this circumstance. 
§ 4. Conclusion. On the strength of what precedes we can there- 
fore firmly deny that Tricunnr’s observations’), at least with respect 
to the nature of the phenomena, are incompatible with the theory of 
ANDREWS and vaN DER Waars. Down to details these phenomena can 
be explained by the presence of admixtures, which are slowly diffu- 
sing through the substance; and calculations based on existent data 
have shown that in order to reach a quantitative agreement, we 
must assume a proportion of the admixture of the same order as that 
which actually was present in other experiments with so-called pure 
substances. Whether in the carbon tetrachloride, used by Turcunmr, the 
required proportion of any admixture, of which neither the nature 
nor the @ and 3 are known with certainty, has occurred, is a ques- 
tion that cannot be answered. It does not seem impossible, however, 
because carbon tetrachloride is a substance which, owing to the manner 
in which it is prepared, might contain many foreign components, 
and the constancy of the boiling point (to within 0°.1 CP) is not 
deemed by us a guarantee for sufficient chemical purity. We are 
even inclined to consider the existence of the deviations as a proof 
to the contrary, and the non-existence of the deviations (other 
1) Starting from a given condition, [ might evidently have worked out this problem 
in perfect harmony with reality; it appeared to me, however, that this would have 
been useless trouble, and that the scheme, I have given of it, does at any rate 
represent the phenomena qualitatively. 
2) The same conclusion holds for similar observations (pe HeeN, Gaxirzine, etc.) 
about the so-called abnormal phenomena near the critical point. 
30 
Proceedings Royal Acad. Amsterdam. Vol. VII, 
