48 



The mean of the 14 determinations, taken as one series, is 1.3402, ± 



.000056. With Morley's value for oxygen, 1.43896, ±.000038, the density 



ratio becomes — 



32:30.0124, ±.00168 



Combined with Gray's determination, the weighted mean is 

 32:30.0106, ± .00065 



Tlie density of gaseous ammonia, according to Leduc,' is 0.5971, re- 

 ferred to air as unity. But this figure represents only a single determina- 

 tion, with material of doubtful purity, and need not be considered further. 

 Guye and Pintza,^ with carefully purified ammonia, made five determina- 

 tions of density, which gave the subjoined results for the weight of the 

 normal litre : 



.77080 



.77069 



.77073 



.77099 



.77076 



Mean, .77079, ± .000035 



Perman and Davies ' made three series of determinations, by two meth- 

 ods; but the first series, with commercial ammonia, is to be rejected. 

 The other series gave the following figures for the weight of one litre of 

 the gas: 



7. 77. 



.7709 .77094 



.7711 .77094 



.7712 .77090 



.7713 .77088 



.7711 .77091 



.7713 



Mean, .77094 



Mean, .77115 



1 Ann. Chim. Phys. (7), 15, 39. 1898. 

 = Conipt. Rend., 141, 51. 1905. 



'Proc. Roy. Soc, 78A, 28. 1906. Perman and Davies give still othor density determinations 

 for different temperatures and pressures. 



