RESEARCHES ON THE VOLATILE HYDROCARBONS. 159 
in many cases so considerable, as to make it appear almost useless to attempt further 
generalizations upon those unreliable data. It may be hoped, however, that the supe- 
rior means which my process furnishes for separating mixtures of liquids, will lead to 
the accumulation of reliable facts of sufficient number and variety for a profitable 
review of this question in its different bearings, which, from its importance, it clearly 
merits. 
The frequent inaccuracy of the determinations of boiling-points, upon which Kopp 
has justly laid so much stress, may, I think, be more reasonably attributed, at least in 
a majority of cases, to a want of purity in the substances themselves, rather than to a 
neglect of the precautions and corrections which he recommends to be observed in 
such determinations; although errors as great as those mentioned by Kopp* may 
doubtless occur, and in the particular instances which he had in mind may have 
occurred from the cause which he assigned for them. It should be borne in mind, 
however, that these errors, in the case of an impure substance, may be compensating 
errors ; or, on the other hand, they may go to increase that which would arise from 
impurity. 
That the conditions under which my results have been obtained may be clearly 
understood, and hence the value of these results fairly estimated, in comparison with 
those of others, I shall endeavor, as I proceed with these researches, to specify, in 
sufficient detail, the processes which I have employed. Having, in the memoir pre- 
viously referred to, described the process by which the hydrocarbons were separated, 
the special object of this paper only requires, in this regard, that I should add a 
description of the method. employed in determining the boiling-points of these bodies, 
which has already been partially given in the foregoing memoir, when treating of the 
boiling-point of benzole. 
Of the Method of determining Boiling-points. — 1 use for this purpose a small tubulated 
glass retort, and usually operate on about 150 ¢.c. of the liquid. The thermometer 
extends into the liquid, even nearly to the bottom of the retort, taking care that the 
bulb shall not come in contact with the glass, but remain free in the liquid. To pre- 
vent abnormal elevation of temperature from adhesion to the glass, — which I have 
observed in some instances, when operating on impure hydrocarbons, to amount to 
several degrees, —I introduce pieces of sodium, instead of platinum, as it seems to 
serve at least as well for this purpose, and at the same time tends to preserve the 
purity of the material. Sodium has also this advantage over platinum for hydrocar- 
* « Bestimmung des Siedepunkts.” Poggendorff’s Annalen, 1847, LXXII. 38. 
