472 JAMES WATT. 
134 and 135 were not there at first; they are added, and an 
indication is given as to where they belong; the writing is no 
longer that of Cavendish ; these additions are in the hand- 
writing of Blagden. And it must have been he who gave all 
the relative details to Lavoisier, for it is not said that Caven- 
dish held any direct correspondence with him. 
The date of the reading of Cavendish’s Memoir was the 15th 
of January, 1784. The volume of the Philosophical Trans- 
actions, of which this Memoir forms a part, did not appear till 
about six months after. 
Lavoisier’s Memoir (volume of the Academy of Sciences for 
1781)* had been read in November and December, 1783. 
Various additions were made to it afterwards. The publica- 
tion took place in 1784. 
This Memoir contained a description of the experiments of 
June, 1783, at which Lavoisier announces that Blagden was 
present. Lavoisier adds, that this English physicist informed 
him “that Cavendish having already burnt inflammable air in 
+ closed vessels, had obtained a very sensible quantity of water ;” 
but he nowhere says that Blagden informed him of the con- 
clusions that Cavendish had inferred from those same experi- 
ments. 
Lavoisier declares, in the most express manner, that the 
weight of the water is equal to that of the two gases that were 
ignited, unless, contrary to his own opinion, a sensible weight 
be assigned to the heat and to the light that were disengaged 
during the experiment. 
This account does not agree with that of Blagden, which, 
according to all probability, was written as a refutation to 
Lavoisier’s relation, after the reading of Cavendish’s Memoir, 
and before the volume of the Academy of Sciences had reached 
England. ‘This volume appeared in 1784, and assuredly it 
could not have reached London, either when Cavendish read 
his paper to the Royal Society, or still less when he wrote it. 
We must remark, besides, that in the passage of Cavendish’s 
* The date 1781 appears to be a clerical error for 1784.— Translator. 
