364 Prof. A. Griinwald on the Spectrum of Water-vapour 



ammonia, NH 3 , and of methane, CH 4 , at the low temperature 

 of electric discharge (to avoid dissociation) or the absorption - 

 spectra of these gases. These spectra must extend in the 

 ultra-violet portion as far as wave-lengths 1728 x 10~ 7 millim. 

 and 1383 x 10~ 7 millim., and, if possible, still further. 



Hitherto, so far as nitrogen is concerned, I have only suc- 

 ceeded in showing that it is in its simplest chemical condition 

 (like oxygen) a compound of the modified hydrogen H/ with 

 a substance W in equal volumes without condensation ; and 

 that the latter substance W must have an entire large group 

 of atoms in common with oxygen, inasmuch as a group of 

 more than forty N-rays (*. e. their wave-lengths) may be 

 transformed by multiplication by the factor ^ into a corre- 

 sponding group of O-rays (i. e. their wave-lengths) . 



The volume-formula of nitrogen has most probably the same 

 form as that of oxygen, only written with smaller indices 

 because the atomic weight of nitrogen (14) is less by 2 than 

 the atomic weight of oxygen. 



In carbon also I was able, by simple comparison of the 

 spectrum with those of H 2 0-vapour, of H, and of 0, to 

 establish the existence of the primary substance b in at 

 least two different chemical conditions, as well as the occur- 

 rence of the primary substance c, which is also contained in 

 oxygen. 



(5) u. In the experiments which I made to determine the 

 factor of reduction (0*6336, very nearly Jg) for the substance 

 a in H to the H 2 0-spectrum, I found that this factor is the 

 mean or chief reduction-factor of the four atoms which in H 

 are combined with the substance b ; but that further special 

 reduction-factors to the H 2 0-speCtrum belong to the separate 

 atoms of these four, of which I succeeded in determining two, 

 viz. j and ^§, and in verifying them, at least so far as they 

 could be compared with the known H 2 0-spectrum. 



If then X be the wave-length of an H-ray of the elementary 

 line-spectrum which is due to the substance a in H, then 

 0*6336 X (empirically, theoretically more nearly Jg X), f X, 

 and ^- J X are with considerable accuracy three different wave- 

 lengths of the HjjO-spectrum, if the latter is known with 

 sufficient exactness. 



/3. If, on the other hand, X be the wave-length of an H-ray 

 which is produced by the substance b combined in H with the 

 element a, then not only must |X be a wave-length of the 

 H 2 0-spectrum according to (2), but also according to (4) /3 (or 

 (4) 7, as the case may be) f f X and J^ X must be two wave- 

 lengths of the O-spectrum; further, §|xffX and §2 X 5 9 ^ 

 must be two wave-lengths of the H 2 0-spectrum ; and recog- 



