14 



which he made in 1781, of the atmosphere under all circumstances, at 

 diiFerent times of the day, in town and country, in summer and winter, 

 by which he determined its composition more accurately than any of 

 his contemporaries, and with a precision which has scarcely since been 

 exceeded : thus, with a knowledge of the specific gravities of the gases, 

 and of the weight of common air, he was in a condition to have com- 

 pared the correspondence of the weight of the gases consumed in the 

 combustion, with that of the fluid produced. But this experiment 

 had a weak side, in the practical difficulty of collecting the fluid : he 

 therefore took a more certain method of examining the question by 

 volume instead of weight, by ascertaining whether the production of 

 the fluid was accompanied by the total disappearance of the com- 

 bining gases : to a given bulk of atmospheric air, he added, in success- 

 ive experiments, a gradually decreasing volume of hydrogen gas, and 

 found a point at which the computed volume of oxygen entirely 

 disappeared. But there was yet a possibility of error : the fluid 

 produced might contain something besides water: he analysed it, and 

 found that the water was pure. Not yet satisfied, he repeated the ex- 

 periment in a simpler form, by burning the hydrogen with oxygen, in 

 place of common air ; and here a difference little to have been expected 

 appeared, for, on analysing the fluid he found it to contain not water only, 

 but nitric acid ; he traced the acid to its source in the small portion of 

 atmospheric air with which the gases chanced to be contaminated, and 

 inferred that the oxygen and nitrogen which it contains, unite under 

 certain circumstances to form nitric acid. Thus he was led to the dis- 

 covery of the cause of its diminution when traversed by the electric 

 spark, and from the residuary defect of the experiment he completed 

 the solution of the problem. 



The experiment by which Cavendish had in 1781 ascertained the con- 

 version of oxygen and hydrogen into water, Priestley repeated in an 

 imperfect manner in 1783 ; and since it is this repetition which M. 

 Arago has mistaken for the first proof of the composition of water, 

 listen. Gentlemen, to Priestley's own preface to the account he 

 gives of it : " Still hearing," he says, " of many objections to the con- 

 version of water into air, I now gave particular attention to an expe- 

 riment of Mr. Cavendish's concerning the re-conversion of air into wa- 

 ter, hy decomposing it in conjunction with inflammable air" He then 

 relates the precautions he took in repeating this experiment, expresses 

 his wish that he had a nicer balance, and tells how he collected the 

 fluid by wiping the inside of the glass with filtering-paper; but it does 

 not appear, either from his own statement or the still more particulai- 



