A. E. Bostwick — Absorption Spectra of Mixed Liquids. 471 



carbonate with the alkali present ; and that, for quantities of 

 from 1 to 1*5 grams, this excess may be considered as one- 

 tenth of one per cent of the salt taken. 



For an analysis of sodium bicarbonate or baking soda, it is 

 therefore sufficient to determine the loss by ignition, the H 2 0, 

 and the alkali in the ignited residue to obtain the total amount 

 of C0 2 in the sample, adding to this amount an amount of C0 3 

 equal to O10 per cent of the weight of the samples taken. 

 The method is much more rapid, far easier and quite as accu- 

 rate as the distillation process, and may prove of value for 

 technical purposes. 



Akt. XL VIII. — Preliminary Note on the Absorption Spectra 

 of Mixed Liquids ; by Arthur E. Bostwick. 



In 1865 Prof. Melde, in experiments on the absorption spec- 

 tra of mixed liquids,* proved, as he thought, that they were not 

 formed by simple addition of the component spectra, but that 

 the bands shifted, a large band of one liquid seeming, in gen- 

 eral, to attract a small band of the other, and more strongly as 

 the proportion of the former in the mixture was increased. 

 Dr. Shuster, in discussing these observations, f shows that 

 where a small band falls on the slope of a large one the effect 

 of optical addition is to shift the apparent maximum of ab- 

 sorption, and concludes that " Prof. Melde .... would have 

 observed exactly the same phenomena if he had put his two 

 liquids in front of each other instead of mixing them to- 

 gether." The experiments herein described were undertaken 

 to lind out whether or not the effect observed by Melde was 

 due to the cause alleged by Dr. Shuster ; they appear to 

 prove that while a small part was so due, the bands were 

 shifted principally by a true action of one liquid on the other. 



Melde's method consisted simply in observing the spectra 

 of the liquids separately, and then that of their mixture. He 

 made no attempt to observe that of one liquid in front of the 

 other. The method of the writer was to observe the spectrum 

 obtained by passing the light through both vessels,- and then 

 that obtained by mixing the liquids and pouring them again 

 into the same vessels, thus securing exactly the same conditions 

 except that of mixture. Common test tubes were first used, 

 but an irregularity in results led to the discovery that the 

 spectrum differed according as one liquid or the other was 

 next the slit, each tube acting as a cylindrical lens. Cells 



* Pogg. Ann., cxxiv and cxxvi. 



f British Association Report on Spectroscopy for 1882. 

 Am. Jour. Sci.— Third Series, Vol. XXXVII, No. 222.— June, 1889. 

 30 



