652 
PROFESSOR W. RAMSAY AND DR. J. SHIELDS ON THE 
results are given in tlie Table which follows; the method of experiment is described 
on p. 671. 
Table IL—Surface-energy at — 89°-8 C. 
Substance. 
7 (M^;)i 
Percentage 
deviation. 
Found. 
Calculated. 
Ethyl oxide 
611-2 
599-4 
' 
+ 2-0 
Methyl formate 
5.57-0 
608-7 
-8-5 i 
Ethyl acetate 
761-5 
744-0 
+ 2-4 
The agreement is good except in the case of methyl formate; here, however, the 
negative value, judging from experiments with acetic acid and with the alcohols, may 
be due to molecular association. The positive value of the other two substances may 
be due to the fact that measurements were made in a different tube of wider 
capillary bore, and that the arrangements were necessarily less perfect than in 
determinations at higher temperatures. 
We consider, therefore, that we are justified in stating that the linear relation 
betweeh molecular surface-energy and temperature holds for ether within a range 
of 240° ; apj)roximately for methyl formate, of 270°; and for etiiyl acetate, of 290°. 
4. Molecular Surface-energy of Acetic Acid and oj Methyl and Ethyl Alcohols. 
The behaviour of acetic acid, and of methyl and ethyl alcohols, is exceptional; the 
surface-energy no longer shows a linear relation to temperature. Surface-energy, as has 
been stated, is the product of # and y, where y* represents one-sixth of the surface of a 
cube containing one gramme of licjuid, and y, the surface tension. It has been assumed, 
with Eotvos, that the molecular volume raised to the two-thirds power (My)y gives 
a comparison of surfaces on which equal numbers of molecules lie. It is fair to 
assume that we have proved a linear relation to hold for “ normal” liquids between 
y (My)^ and temperature. The question now arises—if certain liquids fail to show a 
linear relation, to what factor must the variation be ascribed ? 
Two of these factors, viz., y and (yf, are deduced from actual measurements, viz., 
of the height of ascent in a capillary tube, and of the volume of one gramme of the 
substance. But M has been taken as the molecular weight of the simple molecules, 
C3 H,j^ 02, CH^O, and CgHeO. For acetic acid, at least, this assumption is ackuow- 
ledgedly wrong, for it is well-known that its molecular weight, as determined from its 
gaseous density, is not a constant, but decreases with rise of temperature and fall of 
pressure. It has been shown by Ramsay and Young (‘ Phil. Mag.,’ 1887, p. 206), 
