recovery may explain this discrepancy. Previous 

 tests had shown that a trace of humus was dis- 

 solved by the ether, but this amount was not 

 enough to appreciably affect the rosin determina- 

 tion. 



The alkali-extracted chips, after thorough wash- 

 ing with water, were dried, weighed and extracted 

 with ether. Only a faint trace of resinous material 

 was obtained on evaporation of the ether, showing 

 that the alkali had completely dissolved the rosin. 

 The weight of the extracted chips varied from 45 to 

 50 per cent of the weight of the original wood, 

 whereas the actual wood content was found to be 

 over 54 per cent, as determined after direct ether 

 extraction of the wood. The discrepancy indicates 

 the extent of lignin decomposition by the alkali. 



The percentage of volatile oils could not be 

 accurately determined on a small scale, so the 

 highest yield obtained in the subsequent steaming 

 experiments has been taken as a close approxima- 

 tion of the true " turpentine" content of the wood. 

 In Experiments 5 and 6, 1,500 grams of wood 

 yielded 145 cubic centimeters of turps., weighing 

 approximately 126 grams on the basis of 0.87 

 specific gravity. (Actual specific gravity 0.8743 at 

 15C.) This shows that the wood contained close 

 to 8.4 per cent by weight of volatile oils. 



The average composition of the wood supply 

 was found to be as follows: 



Actual wood 54 . 3% 



Rosin (by ether) 32. 1% 



Volatile oils 8.4% 



Moisture (by diff.) 5.2% 



100.0% 



Rosin (by NaOH) '. . . 31.8% 



Thus the total oleoresin amounted to 32.1 + 8.4 = 

 40.5 per cent made up of 79.2 per cent rosin and 

 20.8 per cent volatile oils. 



Apparatus 



The digester used in carrying out the extraction 

 and cooking of the wood was made from an extra- 

 heavy cast iron tee, 6 in. x 6 in. x 4 in. The 



28 



