CAVENDISH. 437 



it was plainly a corollary by that discovery it was 

 manifestly suggested. 



The former experiments, both those of Cavendish 

 and those on which Watt reasoned, were all syn- 

 thetical and decisive that of Lavoisier was analytical 

 and radically defective. It proved nothing conclusively : 

 it was well enough after the experimentum crucis had 

 demonstrated the proposition ; to that proposition it was 

 a corollary it was nothing like a critical experiment. 

 He placed water in a retort exposed to heat ; the vapour 

 of the retort, when the water boiled, was passed through 

 a tube (a gun barrel with the breech-pin knocked out 

 was generally used) ; the tube, if made of earthenware, 

 had iron filings placed in its course ; it was placed 

 in a fire ; its further extremity was connected with a 

 receiver, in which cold water or mercury rose to fill it 

 entirely. As the water slowly boiled there came 

 through the tube, and into the glass receiver, a current 

 of gas, which, upon examination, was found to be hy- 

 drogen gas, while the iron filings were converted into 

 calx or oxide. The weight of the gas produced, added 

 to the weight acquired by the gun barrel or by the 

 filings during the process, was found to be nearly equal 

 to the weight lost by the water in the retort. Hence 

 the inference was, that the lost portion of water had 

 been decomposed into its two elements, the oxygen gas 

 forming the calx of the iron and the hydrogen gas being 

 received in the glass vessel. But the adversaries of the 

 new doctrine had an answer to this inference far more 

 formidable than any that they could urge against the 

 conclusion drawn from the synthetical experiment. The 

 analytical experiment was liable to all the uncertainty 



