134 
OX THE DISTILLATION OF MIXTURES. 
the liquid would acquire a bronze, a blue, or a pink colour; the kind of colour being 
dependent on the amount of aniline present,—bronze being the result of much aniline, 
and pink of a very little. In this way at least the ^-g^th part of a grain of aniline was 
easily recognized. 
To another portion of the acid liquid placed upon a white porcelain plate, a little 
peroxide of lead or red prussiate of potash was added, and a blue or purple reaction 
followed. This test is not so delicate as the last, for it fails when the amount of aniline 
is less than the y^oyth of a grain. 
Other tests may be resorted to if necessary, as when the quantity of aniline is large. 
Thus peroxide of manganese or bichromate of potash may be used in the same way as 
the red prussiate of potash in the last experiment; but these tests will not answer with 
less than the -^L^th of a grain of aniline. Lastly, a drop of a solution of chloride of lime 
may be added to the acid liquid, and if the quantity of aniline exceeds the Tijyth part of 
a grain it will cause a purple reaction. 
9th. If the quantity of saline residue from the last operation is large, and there is 
reason to believe that much ammonia is present, this alkali must be got rid of, for it 
greatly interferes with the success of the colour-experiments. The residue, therefore, is 
made moist with water, and rubbed down with about twice its bulk of neutral carbonate 
of soda. It is then exposed to the air for a short time until the odour of ammonia has 
passed away. It is then treated with strong alcohol, filtered, acidulated with dilute sul¬ 
phuric acid, and again evaporated. The aniline is now fit for the colour-experiments. 
There are no fallacies to these experiments ; for although, as I have elsewhere show r n, 
strychnia will give nearly the same colour-reactions, yet in the first place this alkali is 
not volatile like aniline, and will not therefore distil over as the latter does; and in the 
next place, while the best effects, in respect of colour, are developed with dilute acid and 
aniline, strychnia requires the concentrated acid. These differences are sufficient to pre¬ 
vent any embarrassment as regards the two alkaloids. 
ON THE DISTILLATION OF MIXTURES: A CONTRIBUTION TO THE 
THEORY OF FRACTIONAL DISTILLATION. 
BY J. A. WANKLYN, ESQ. 
There are many points in the boiling of mixtures which are obscure. The tension of 
the vapours at the temperature whereat the mixture boils, and the proportions in which 
the constituents of the mixture are present, are not the only factors which determine the 
relative rates at which the constituents distil. There have, for instance, to be taken into 
account the adhesion of the liquids to one another, and the vapour-densities of these 
liquids. On the present occasion I have to call attention .to the influence of this latter 
element, w T hich influence seems to have been lost sight of by most of those who have 
applied themselves to this subject. 
Leaving out of account for a moment the influence of adhesion, and simplifying the 
influence of the proportion in which the ingredients are present by taking equal weights 
of two liquids of different boiling-points, v r e may set down the rates at which these in¬ 
gredients will distil as determined by the tensions of the liquids and the densities of the 
vapours. In the first instant of time the quantity of each ingredient which distils will 
be found by multiplying its tension at the boiling-point of the mixture by its vapour- 
density. It thus appears that the liquid with the highest tension will not of necessity 
distil the quickest, for what the other liquids w r ant in tension they may make up by the 
greater density of the vapours which they give off. And so when w r e mix a more volatile 
with a less volatile liquid and proceed to distil the mixture, we shall now r and then find 
that the less volatile liquid distils faster than the more volatile one. I will here bring 
forward an experiment to illustrate this point. 
Vapour-density. 
Methyl-alcohol boils at 06° C. 1*107 
Iodide of ethyl boils at 72° C. 5*397 
I took 18 grammes of methyl-alcohol and 17 grammes of iodide of ethyl, mixed them, 
and distilled off rather more than one-third of the mixture. The distillate consisted of 
