804 Transactions of the American Institute. 



peratnre of the oil vapors takes place similarly to, and can be con- 

 trolled better than, the cooling in high stills, without this provision. 

 The arrangements for cooling the sides and top of a still must there- 

 fore be more complete the lower or smaller the still employed is. 

 This law also teaches us that stills, to be used for " cracking " oil, 

 should have a flat bottom, and should have the ilues arranged in such 

 a manner as to permit the restriction of the fire to the bottom only, 

 which is necessary in the process of " cracking." Where superheated 

 steam is used as the heating medium it ought to be applied to the 

 bottom only. "Where the dimension of stills become huge, as those, 

 for instance, in the refinery of Reese & Graff, Pittsburg, Pennsyl- 

 vania, which have a capacity of 40,000 gallons each, a flat bottom 

 would hardly be strong enough ; in such cases the boiler shape is 

 usually employed. In order to have a practically sufiiciently large 

 heating surface, the fire has to reach up to a certain distance on the 

 round boiler. If the cpiantityof oil present in the still is so small as 

 to be below the boiler surface exposed to the fire, the rising oil 

 vapors will be superheated on this surface, and the resulting distillate 

 will be of greater specific gravity, and of darker color than it nor- 

 mally would have been, the product resembling more the oils result- 

 ing from the distillation of coal-tar. For this reason the residuum in 

 sucli stills, after reduction to the quantity mentioned, is frequently 

 removed to the smaller stills ; this diminution of the temperature of 

 the oil vapors causes a partial condensation and redistillation of the 

 oil, which diminishes its color and gravity. 4. The intensity of the 

 process of " cracking " is proportionate to the suddenness with which 

 the oil vapors are condensed before leaving the still. The thorough 

 application of this principle produces more rapidly those results men- 

 tioned as necessary in the preceding paragraph. 5. The difference in 

 gravity between that of the oil distilled, and of the desired distillate,- 

 is in direct proportion to the quantity of water produced in the pro- 

 cess. If from a heavy oil, an exceedingly light distillate is to be pro- 

 duced, the proportion of water is so immense as to show occasionally 

 a distillate of water with but a minute percentage of oil floating on 

 its surface. In all such cases small black particles of carbon float on 

 the top of the water, forming an intermediate layer between the lat- 

 ter and the oil. The quantity of carbon separated in this manner is 

 also proportionate to the quantity of water distilled over respectively 

 and the intensity of the process of " cracking " employed. This car- 

 bon is mechanically carried over by particles of water which, in con- 



