448 



Experlmenls on Jmmoma. 



be expected, the same tendency. The sixth experiment lit 

 the table is the one whicli has been just described. 



From an attentive examination of the forcfroing table, it 

 will appear that the resuhsare not perfectly uniform, though 

 perhaps as nuich as can he expected from thcMiature of the 

 experiments. Thus the. proportion of permanent gases ta 

 the an)monia decomposed (the nitrogen being aclu?.!lv 

 mca-ured, and the hydrogeri estimated by doubling the 

 Dxyaen expended) maybe observed to difi'er considerabiv ; 

 the highest product beine 1931, and the lowest 180-2, from 

 100 of ammonia. There can scarcely be a doubt, however, 

 that this u.uii of coincidence is owinc to the same cause 

 .'is that which I have aireadv assiancil for the varialjlc pro- 

 portions of permanent gas, which-are obtained from equal 

 (juantities oi ammonia h-.' elee;rization. And, accordingly, 

 1 have found that the evolved gases, as ascertained by com- 

 bustion, bear the smaUest proportion to the ammonia 

 when most pains have been taken to obviate the presence 

 of nioisturc. The Kiwc.-l number, thereiore, is to be as- 

 sumed as most correct ; l)ul other circumsianees being con- 

 sidered, 1 believe the second cxperintent furni:-hes the n)ost 

 accurate data for deierrnlning the composition of anmionia. 

 The same explanation will p.ppdv to the diflVrent propor- 

 tions of oxv<ien i':us required i(n- liie salutation of 100 mea- 

 sures of au)mo!nj, the variation no doubt arising frou) the 

 uncertAinlv of il'.e ipuiuuty of alkaline gas which is actually 

 burned.. The prop.-.itioa ef oxygen to amiiionia^ \A.bich I 



believe 



