PAPERS OX CHEMISTRY AND PHYSICS 



353 



figure. This result is quite the opposite from what 

 would be expected. If a coal which has been completely 

 oxidized will not coke, and if the resinic matter carries 

 the coking principle, then complete saturation of the 

 resinic portion with oxygen should destroy its coking or 

 bonding property. But it does not, the resulting sample 

 being indistinguishable from the standard made for com- 

 parison, 



Xow we will reverse the process, saturating the non- 

 coking cellulosic residue with oxygen and leaving the 

 resinic material in the fresh unoxidized state. Upon re- 

 combining and carbonizing, we have the result as shown 

 in Fig. 10. Here the coking property has been com- 

 pletely destroyed. 



Fig. 11. Progressive Coking- Lines in By-Product Oven. 



Evidently our theory of the coking process will have 

 to go farther back than the simple idea of a binding 

 material, to a study of the factors that enter into the 

 properties of both constituents and the conditions that 

 affect their modification on the one hand or their preser- 

 vation and indeed their accentuation on the other. Here 

 it would seem from this illustration that we have worked 

 out an answer to the question how not to make coke 

 instead of the question "How is it made?" On the con- 

 trary, it may very well happen that the most direct in- 

 formation as to how to do a thing may come from posi- 

 tive evidence as to how it should not be done. 



