40 STUDY OF ABSORPTION SPECTRA 



towards the violet. Some other assumption of a more fundamental 

 character is necessary to explain the phenomenon under consideration. 



If the non-hydrated salts decreased the association of the solvent, 

 and thereby produced an absorbing system freer to vibrate, it seems 

 that most of the points brought out by the transmission curves could 

 be satisfactorily explained. We should then expect the solution to be 

 more opaque than the solvent at the centers of the bands. In case 

 there was no shift in the bands, the same relations should hold for the 

 other wave-lengths of light investigated. The above suggestion is at 

 present to be accepted only tentatively, as offering a possible explana- 

 tion of the results obtained. 



We are now applying a method which we hope will decide whether 

 the above assumption is or is not correct, whether the presence of a 

 non-hydrated salt does or does not diminish the association of water. 



Frequent reference has been made to the peculiar characteristics of 

 the transmission curves for magnesium nitrate. It was desired to see 

 whether the salt itself had any absorption. Some magnesium nitrate 

 was prepared nearly anhydrous by repeated crystallization from abso- 

 lute alcohol, and the salt was then dried in a vacuum over phosphorus 

 pentoxide for several days. It was dissolved in acetone and a 1.07 

 normal solution prepared. This was studied spectroscopically in the 

 same manner as has already been outlined for aqueous solutions. 



Fifteen millimeters of acetone showed a faint band at 0.9/z, having 

 less than 10 per cent absorption. Another quite similar band appeared 

 at 1.03/-1, part of which was probably due to water. At 1.22^ there is 

 a fairly intense band absorbing about 60 per cent of the light. The 

 absorption again reaches a maximum close to the point in the spectrum 

 where our measurements for the 1.25/i water-band ended. Using such 

 depths of the solutions that they contained as much of the solvent as 

 had been employed to obtain the transmission of 15 mm. of the pure 

 solvent, it was hoped, if magnesium nitrate did have any absorption 

 of its own, that it could be detected. From 0.7/4 to 1.3/z, which covers 

 most of the spectrum studied in this work, the percentage transmissions 

 for both the solution and the solvent were very nearly the same. There 

 was, therefore, no indication of any absorption due to the dissolved 

 substance. 



Beyond 1.3ju, as we approached another intense band, some very un- 

 usual differences were noted. Until more work can be done with pure 

 acetone in this region, and with anhydrous magnesium nitrate, we do 

 not feel justified in drawing any conclusion from these results. 



Anhydrous magnesium nitrate was also dissolved in some very care- 

 fully prepared absolute ethyl alcohol, the concentration of the solution 

 being 0.53 normal. It was desired to learn whether the same kind of 

 results would be obtained as had manifested themselves with water 

 as the solvent. Table 14 is for 20 mm. of alcohol. It will be noticed 



