Vol,. 6, 1920 CHEMISTRY: HARKINS AND EWING 
51 
When a tube containing the charcoal is outgassed, and mercury is ad- 
mitted, this Uquid fills what are technically known as the voids between 
the lumps, but we have found that it does not enter the very small pores 
in the lumps, so in this way the total volume of the lumps, and also the 
lump or "block" density, may be determined. For the charcoal given 
in table 1 this is 0.865. What is known as the apparent density of a char- 
coal is the mass of the charcoal contained in a 100 cu. cm. flask divided 
by 100, or it is the density of the carbon, pores, and voids, all taken to- 
gether, and is usually about 0.5 for charcoals used by the Chemical War- 
fare Service. What is known usually as the "real density," though it 
is not the actual density, is that which is obtained by allowing water to 
penetrate the pores and by calculating the "real volume" of the carbon 
on the basis of the assumptions that the water completely fills the pores 
and that the water is not compressed by the attraction of the charcoal. 
The data of column 2 show that both of these assumptions cannot be 
correct. The densities of cocoanut shell charcoals as used by us when 
determined in water are 1.843 (table 1), 1.863, 1.835, and 1.808; that of 
the similar charcoal used, by Titoff,i 1.86; by Baerwald,^ 1.92; and by 
Miss Homfray,^ 1.66. 
The compressibility data in table 1 are those of Bridgman,* those on 
surface tension were obtained in this laboratory, and the viscosity data, 
by different investigators. The compressibility increases in exactly the 
same order as either the pore volume or the density, except that the in- 
dividual order for chloroform, benzene, and paraxylene is not known, 
though a study of Bridgman's curves of compression taken in connec- 
tion with Richard's^ data on their compressibility up to 500 atmospheres, 
makes the conclusion inevitable that as a class they lie between propyl 
alcohol and carbon bisulphide in accordance with their position in table 1 . 
The curves plotted from Bridgman's data do not in general cross each 
other when all of the liquids are non-polar, while the curves for polar 
liquids not only cross each other, but also those for non-polar liquids. 
The data indicate that the differences in the volume of liquid absorbed 
in 1 cu. cm. of the charcoal, listed as the pore volume in table 1, may be 
due either to the compressibility or the viscosity of the liquid, but on the 
whole they are in better agreement with the idea that it is the compressi- 
bility which is effective. The difference between the volume of ether and 
water absorbed is about 10%, and to produce this difference of volume 
by pressure exerted on the whole volume of liquid would require, accord- 
ing to Bridgman's data, between four and twelve thousand atmospheres. 
It is evident, however, that only a small fraction of the liquid is in actual 
contact with the charcoal. Dr. A. M. Williams^ has calculated the volume 
of liquid in the primary film in the cocoanut charcoal used by Titoff as 
0.30 ccm., which would be of the order of three-fifths of the pore volume. 
This would show that the average pressure in the primary film due to the 
