354 History of Mechanical Inventions, Sfc. 



dened into an uniform mass of no disagreeable colour, and of 

 very tolerable consistency. After a certain number of trials, 

 experience taught me that the best proportion for mingling 

 the substances was eighty parts of oil to one of strong fuming 

 nitric acid, and having increased my apparatus, I continued 

 with this receipt to prepare a considerable number of candles, 

 which answered their purpose sufficiently well. 



By degrees, however, he began to experience unaccountable 

 variations in the process ; for in spite of all his pains, the oleo- 

 cere sometimes would not harden, but continue unalterably of 

 the consistence of butter. For a long time he concluded these 

 defects to proceed either from the entrance of watery vapour 

 into the jars whilst boiling, or from the increasing heat of the 

 weather at the time. To remedy this he took every precau- 

 tion in shutting the jars, and when the process was over, placed 

 them behind a tattie to cool. Still this was to no advantage, 

 and many trials showed that the hardening of the oleocere was 

 a matter of the greatest uncertainty. 



Some time afterwards, being placed in more favourable cir- 

 cumstances for conducting his experiments, he adopted another 

 plan, which we give in his own words : " I erected a furnace 

 about four feet from the ground ; on this was placed a large 

 iron boiler to serve as a reservoir ; immediately adjacent to 

 this first furnace, was built another furnace about half the 

 height of the former, on which was placed a round iron vessel 

 whose side was about ten inches high, and whose capacity was 

 such as to contain seven of the China jars already mentioned. 

 In the side, about two inches below the level of the top of the 

 jars, was fixed a pipe, so that the water might rise to this level 

 and no more, whatever should be superfluous being carried off 

 by the pipe. Having then a quantity of water to boil both in 

 the reservoir on the copper furnace, and in the vessel on the 

 lower, and having prepared a long copper syphon, I placed 

 its short leg in the reservoir, and directed its long leg to the 

 lower vessel, so that a perpetual stream of water should be 

 conveyed from the upper receptacle to the lower. By this 

 contrivance, the water was pei7)etually kept boiling, and the 

 quantity in the lower vessel was uniform, — its loss was perpe- 

 tually supplied by the syphon, and its excess carried off by 



