2 BULLETIN" 607, IT. S. DEPARTMENT OF AGRICULTURE. 
The most important variables in the treatment of wood affecting 
absorption and penetration are: (1) The species, character, and con- 
dition of the wood; (2) the properties of the oil; and (3) the methods 
of treatment. To eliminate the variability dne to species, and re- 
duce that due to the character and condition of the wood, th© ex- 
periments were confined to wood of one species (longleaf pine), and 
so far as possible specimens of uniform quality were selected. This 
made more readily determinable the relative importance of varia- 
tions in the oil and in methods of treatment. 
The relative absorptions and penetrations were determined by 
two forms of tests : 
1. Penetrance tests, in which the preservative was applied to a 
small area in the specimen and measurements made of the penetra- 
tions secured. 1 
2. Impregnation tests, in which the specimens were treated with 
the preservative under pressure in a cylinder. 
The bulletin describes (a) tests made with mixtures of coal-tar 
creosote and coal tar from which the free carbon had been removed 
and similar mixtures containing varying amounts of free carbon; 
and (b) tests made with commercial coal-tar creosotes. The object 
in both cases was to determine the effect on penetration of differences 
in the preservative. 
The bulletin also gives the results obtained in experiments to deter- 
mine the effect of varying pressure, time of treatment, and tempera- 
ture. The object of these experiments was to determine whether, 
by increasing the time of treatment, intensity of pressure, and tem- 
perature of preservative, absorptions and penetrations could be 
obtained with tar and creosote mixtures equivalent to those ob- 
tained with creosote, and to determine which of these factors was the 
most important. 
MATERIALS USED. 
WOOD. 
Selected air-dried heartwood of longleaf pine was employed in all 
of the experiments. The specimens used in the penetrance tests 
were approximately 2 inches by 4 inches in cross section and of 
varying lengths from 18 inches to 24 inches. Pieces cut from the same 
stick were regarded as matched and were designated as a set of 
specimens. Test pieces selected in this manner were more uniform 
in structure than unmatched material, and the variables introduced 
by differences in structure were thus reduced. From two to four 
sets of matched specimens were used in each series of tests, the results 
being averaged for each test. Where these results are represented 
graphically, each figure includes only tests with such matched speci- 
1 A description of the apparatus used (fig. 14) is given in the Appendix. 
