572 
CHEMISTRY: W. D, HARKINS 
mental value is 352. As a matter of fact, the deviation between the 
result given by rule and by experiment seldom falls below 90 ergs, and 
is often higher, which seems to disprove the rule. 
In spite of the above objections it must be admitted that Antonow's 
rule has been of considerable use in connection with phase boundaries 
between water and organic liquids. It should be realized, too, that 
the above objections to it involve a certain hypothesis, that is that the 
amount of the organic Hquid dissolved in the mercury is too small to 
reduce the surface tension by 90 ergs or more when the mercury phase 
is split apart in such a way as to form surfaces from the interior part 
of the mercury phase. 
Preliminary experiments on the adsorption of benzene and other vapors 
on a mercury surface show that their positive adsorption is very much 
greater than on a water surface at the same concentration of vapor. 
Water present either in the air or in such organic liquids as hexane, ben- 
zene, ether, etc., is also adsorbed positively, and the adsorption is very 
marked even when only a very small amount of water is present. Mer- 
cury salts, such as mecurous iodide or bromide, when dissolved in an 
organic iodide or bromide, are strongly adsorbed. The high adsorption 
of such substances on mercury as compared with that on water is in ac- 
cord with my theory that the adsorption increases in general with the 
rate of drop in the intensity of the electromagnetic intermolecular stray 
field at the surface of the liquid, since this drop is very much more rapid 
for mercury than for water. According to the results obtained by 
F. Schmidt^ caesium, rubidium, and potassium, when dissolved in mer- 
cury are strongly, and sodium, tin, lead, and gold, are slightly positively 
adsorbed. Negative adsorption is most marked with barium as a 
solute, and to a lesser extent with strontium, calcium, cadmium, zinc, 
thalhum, and Kthium, in decreasing order of effect. 
The experimental difficulties in connection with the investigation of 
the surface of a liquid metal are considerable, and they have been solved 
thus far by Dr. Grafton and Mr. W. W. Ewing. The result which is 
used in table 1 for the surface tension of mercury in a vacuum, is not our 
own, and may be too high, but this would not affect the value of the 
results, since it introduces only a small constant difference in all of the 
numbers which represent the work of adsorption. The liquids were 
purified with great care. Many supposedly pure organic substances 
which are not sulphur compounds, and which have the correct boiling 
points, contain enough organic sulphides to make determinations 
made upon them entirely useless. This was especially true of the xylenes, 
which were boiled with mercury for several days to remove the sulphur. 
