ABSORPTION SPECTRA AND CHEMICAL CONSTITUTION. 159 



crystallised. The oxalate was distilled with pure potassium hydroxide and 

 the ammonia absorbed in pure distilled water, the spectrum of which was 

 photographed on the same plate as that of the ammonia solution. Much 

 greater thicknesses of liquid were examined than in previous experiments. 



A layer 200 mm. thick of a solution containing 10'6 per cent, of ammonia 

 prepared in this way from oxalate transmitted all rays to ^/A. 3638 (/\2749), 

 but the spectrum was feeble from \IK 2738 (\ 3652) to i/X 3638 (\ 2749). 

 No band was visible. A layer 100 mm. thick transmitted the rays to 

 V\ 4323 (\ 231 3), but the spectrum was very feeble beyond VX 3904 (X 2561). 



From another portion of the purified oxalate the liberated ammonia 

 was passed into optically pure hydrochloric acid ; the ammonium chloride 

 recrystallised several times was then examined, the solution of the salt 

 employed having the same thickness of layer and containing the same 

 amount of ammonia as that previously used in determining the position 

 of the absorption band in ordinary ammonia. It now showed no trace of 

 selective absorption, the spectrum being continuous to '/X 4666 (X 2143) 

 with a scarcely perceptible weakness at the extreme ultra-violet end. Pure 

 ammonia may therefore be obtained without difficulty by the decomposition 

 of a crystallised ammonium salt such as the oxalate. 



Ammonia obtained from Hijdroxylamine. 



Ammonia obtained by the reduction of hydroxylamine was next 

 examined. Hydroxylamine hydrochloride was reduced with a zinc- 

 copper couple and the ammonia distilled into pure hydrochloric acid ; 

 the ammonium chloride thus obtained was subsequently purified by 

 recrystallisation. 



A layer of 150 mm. of a solution containing 2*5 grams ammonia in 

 100 c.c. distilled water showed a continuous spectrum to ^/X 4411 (A 2267) ; 

 the spectrum is weak from '/X 3886 (\ 2573), but there was no indication 

 of selective absorption. 



As therefore neither ordinary ammonia, which has been carefully 

 purified by the above method, nor ammonia obtained by the reduction 

 of hydroxylamine, shows selective absorption, we conclude that the 

 absorption band of ordinary ammonia is due to the presence of 

 traces of foreign substances which distil over with it from the gas 

 liquor. 



We next endeavoured to ascertain the nature, and estimate the 

 amount, of the impurity to which the band of ordinary aqueous am- 

 monia is due. The position of the band seemed to indicate the pyridine 

 bases as the most likely cause of the absorption, and, in fact, we 

 found that a layer of 150 mm. thick of a solution containing 7"68 grams 

 of pure ammonium chloride (equivalent to 2*5 grams of ammonia) and 

 '00001 gram of pyridine in 100 c.c. water, showed almost exactly the 

 same amount and character of absorption as a layer of ordinary aqueous 

 ammonia of the same thickness and strength. 



In a further experiment we found that the addition of the same 

 amount of pyridine (in the form of hydrochloride) to 100 c.c. of distilled 

 water produced an identical result, the spectrum being hardly distin- 

 guishable from that of ordinary aqueous ammonia.' It follows, there- 



' In this connection it is interesting to note that, although a solution contain- 

 ing 0000001 gram pyridine in 100 c.c. distilled water no longer showed an actual gap 

 in the spectrnm, there was a perceptible weakening of the lines of that portion of the 

 spectrum in which the band of ordinary aqueous ammonia occurs. 



