Magnetic Rotary Dispersion of Solutions g 



In addition to the above wave-lengths the solutions were ex- 

 amined for many other points in the spectrum. The method be- 

 ing a differential one and not dependent on successive readings 

 of the absolute rotation of solvent and solution, the phenomena 

 sought for, if present, could be observed directly. Hence a solu- 

 tion could 1)e examined in a very short time. The differences 

 between the solvent and solution, for wave-lengths not greatly 

 absorbed by the latter, could have been made much smaller, had 

 it been necessary, by increasing the sensibility of the polarizing 

 system. 



The differences noted are irregular and apparently are merely 

 experimental errors. With a polariscope giving consecutive set- 

 tings within 0.005°, considerable difHculty would naturally be ex- 

 pected in eliminating the various sources of error sufficiently to 

 make this sensibility available. Even after the elimination of 

 double refraction and mechanical errors there are the effects of 

 bleaching and perfect compensation for the absorbing solution 

 to be met. Unless the liquid in both the solution tube and the 

 compensating cell is perfectly fresh a dift'erence in rotation of 

 perhaps 0.01° will be observed. At the time, for instance, the 

 data for the anilin blue solution at wave-length 517 i^ij. were 

 taken there was actually present a difference of 0.01° between 

 solution and solvent. Upon looking for it at some subsequent 

 time it would perhaps have vanished. The above data ^eem to 

 show that there is not sufficient anomalous magnetic rotary dis- 

 persion in the solutions studied to be measured with the means_ 

 available. Mr. Williams, fellow in physics, in a forthcoming 

 paper has studied the dispersion of alcoholic fuchsin solutions by 

 means of channeled spectra. He finds no anomaly in the dis- 

 persion curve of a fuchsin solution thirty-six times the concen- 

 tration 0.000024. Hence, if we accept the various formulae, for 

 example Becquerel's where 



which have been presented as criteria for the magnitude of the 



273 



