40 



prove that it enters undecompounded into composition with other bo- 

 dies. Having proved this, it would have been better if he had dropped 

 the term phlogisticated air : but though the language is hypothetical, 

 the ideas are precise : and with respect to hypothesis, a distinction must 

 be made between the art of communicating, and that of discovering 

 truth : I have noticed in my Address how important it is, for the sake 

 of clearness, and for the avoidance of prejudice, to discard from our 

 reasonings all hypothetical expressions, resting, like this of the sup- 

 posed combinations of phlogiston, on loose analogies. Nevertheless in 

 the mind of every discoverer a private reserve is made for the admis- 

 sion even of loose analogies, and for the idea that every body deemed 

 simple may prove to be compound; previous, for instance, to proof of 

 the composition of the alkalies, to have spoken in hypothetical lan- 

 guage of the oxide of potassium, would have been logically objection- 

 able ; but yet if Davy had not entertained the hypothesis, he would 

 never have made the discovery. 



The remaining experiments in this manuscript, bearing dates of 

 1784—5, are almost entirely devoted to the combustion of charcoal and 

 of the gas distilled from it, in various ways, with common air, oxygen, 

 and nitre, the chief object of which seems to have been to ascertain 

 whether that body is a compound. The conclusions at which Caven- 

 dish arrived were, that it contains no nitrogen, but that there was rea- 

 son to suspect that it might contain some hydrogen. 



His computation of the experiments which he had made in 1783 on 

 the gas distilled from charcoal is worthy of remark. It was not known, 

 till Cruikshank made the discovery in 1801, that carbon and oxygen 

 unite in other proportions than those which form carbonic acid ; nor 

 was it known, till Henry stated it in 1 805, that carbonic oxide, or oxy- 

 gen in any form, existed in the gas from moist charcoal. Cavendish 

 however deduced the last of these facts from his experiments ; he com- 

 puted this gas to consist of carbon, hydrogen, and water, stating the 

 water at such an amount, as fully to represent the amount of oxy- 

 gen in the gas*, and to account for the presence of free hydrogen 

 without the corresponding oxygen which would have been due to the 

 decomposition of water, he proposes two suppositions — " From this ex- 



* It was probably from these experiments that Cavendish derived in part 

 his opinion before alluded to, that water is a constituent part of inflammable 

 air ; since it appeared to be so in that kind which is obtained from charcoal. 

 It is remarkable, that Cruikshank, after his discovery of carbonic oxide, should 

 have overlooked its existence in this gas, and like Cavendish, should have com- 

 puted the gas to contain water, in the proportion of 9 grains in 14§. 



