sure at which distillation of turpentine and pene- 

 tration of the wood could be carried on without 

 unduly lengthening the time of treatment. A 

 pressure of 15-20 pounds was found to be effective 

 in achieving the most favorable results. 



The time consumed in extraction varied from 

 3.5 to 4 hours and must in most cases be considered 

 a maximum, because of the attempt to recover all 

 the turpentine. The greater proportion of the run 

 was spent in distilling the last small portions of 

 oil. An important feature in the design of a 

 digester for practical operation would be the pro- 

 vision for rapid -steam distillation, so that the final 

 portions of volatile oil could be carried over as 

 soon as liberated by disintegration of the rosin. 

 This would shorten the time of preliminary treat- 

 ment and decrease the effect of the alkali on the 

 rosin and the wood. A large dome or distilling 

 head set with closed steam coils might successfully 

 break the froth and allow proper control of distilla- 

 tion from the digester. 



There was considerable attack of the ligno- 

 cellulose, as may be seen from the amounts of 

 humus, etc., dissolved (11.4 24.8 per cent of the 

 actual wood). This decomposition of lignin seemed 

 to depend as much on the time of contact with 

 the alkali as on the ratio of caustic used. This 

 action does not signify a waste of alkali or an 

 injury to the wood, but rather represents the first 

 stages of the ordinary pulping operation. The dis- 

 solving of the humus is objectionable only because 

 of its contamination of the rosin soap and because of 

 the extra steps involved in removing it from the 

 same. 



Rate of Turpentine Distillation 



Representative curves showing the rates of turp- 

 entine distillation are given in Figure 2. In order 

 to compare the results on a convenient basis the 

 yields from Expts. 5 and 6 have been taken as 

 standard and assumed to be 100 per cent of the 

 volatile oils in the wood. All turpentine fractions 

 in the four experiments considered have been calcu- 

 lated as percentages of this maximum value. The 



39 



