ment laboratories. (U. S. Dept. Agr., Forest Service, 

 Bui. 105, p. 13.) 



Specific gravity was in all cases measured at 

 15C. by the use of cooling baths. 



The important variable in the various runs was 

 the ratio of alkali to wood. The unit basis for 

 calculating the amount of NaOH or Na 2 CO 3 to 

 be used in the preliminary treatment was the 

 weight of Na 2 theoretically necessary to saponify 

 the rosin in the wood. The average combining 

 weight of American rosins has been placed by 

 Lewkowitsch (" Chemical Technology and Analysis 

 of Oils, Fats and Waxes," Vol. 1, p. 499) at 346, 

 and for convenience this figure has been taken as 

 an approximation of the combining weight of the 

 rosin in the chips under consideration. On this 

 basis rosin would require 31/346 =8.96 per 

 cent of its weight of Na 2 O to produce a neutral 

 sodium resinate. The wood under consideration 

 contained 32.1 per cent of rosin, so that 100 parts 

 of resinous wood required at least 2.88 parts of 

 Na 2 for complete saponification of the rosin 

 therein. The amounts of NaaO actually used have 

 been expressed as multiples of this unit value. 

 The attempt was to find the least amount of caus- 

 tic which would completely extract the rosin before 

 neutralization of alkali by the rosin and humic 

 acids, under conditions allowing efficient 

 turpentine distillation and still preventing undue 

 attack of the wood itself. With the preliminary 

 extraction continued for 3.5-4 hours at 15-30 pounds 

 pressure, the lower limit for NaOH was found to 

 lie somewhere between 2.17 (Expt. 3) and 3.0 

 (Expt. 6) times the theoretical amount of caustic. 

 Under the conditions employed, the minimum 

 amount of caustic necessary for extraction could 

 be safely placed at about 2.5 times the theoretical 

 value. The excess above that needed for saponi- 

 fication of the rosin was of course neutralized by 

 combining with the decomposition products from 

 the wood itself. 



The relieving pressure was varied somewhat, but 

 without appreciable change in the other steaming 

 factors. The aim was to discover the lowest pres- 



38 



