ABSORPTION" AND PENETRATION" OF COAL TAR AND CREOSOTE. 25 
placed in the apparatus previous to testing in order to heat them 
uniformly to the required temperature. By the aid of mirror D, 
placed at the back of the oven, both ends of the specimens were made 
visible. 
In making the impregnation treatments on the paving-block speci- 
mens, small treating cylinders were used. The tests of the effect on 
penetration of differences in the preservative were made in a cylinder 
6 inches in diameter and 12 inches deep which was substituted for 
the clamping device in the penetrance apparatus. Other impreg- 
nation tests were made in a 1^-foot by 4-foot treating cylinder. 
METHOD OF OBTAINING CARBON-FREE TAR. 
The free carbon was removed from a portion of tars Nos. 1, 2, and 
3, as follows: 
The tar was first dissolved in chloroform and the mixture then 
passed through a single thickness of fine quantitative filter paper. 
This process of filtration was repeated on the mixture using fresh 
filter papers each time, until no residue was deposited on the paper 
after the final filtering. The residue left on the papers was boiled 
in chloroform to dissolve any of the soluble tar constituents that 
might be deposited with the free carbon. This mixture was also 
filtered a number of times until no appreciable residue was left on 
the filter papers. The chloroform was then distilled off from the 
two mixtures which were combined and considered to be carbon-free 
tar. Analyses and coking tests were made which indicated that 
the material extracted was nearly all free carbon. 
DETERMINATION OF FREE CARBON IN TARS. 
The percentage of free carbon in each of the tars was determined 
to the nearest one-half of 1 per cent by careful quantitative filtering, 
employing a heavily matted Gooch crucible. After filtration was 
practically completed, the residue in the crucible was washed with 
chloroform until there was no further coloration of the liquid. 
SPECIFIC GRAVITIES. 
Specific gravities were determined by means of a Westphal bal- 
ance. These determinations were accurate to one in the third deci- 
mal place for creosote and most of the mixtures of tar and creosote, 
and to two or three in the third decimal place for the heaviest tars. 
VISCOSITIES. 
The viscosities were determined by means of an Engler viscosi- 
meter, which was standardized with water at 20° C. 
