( '<>1 ) 



Much atleiilion sliould niereforc bc paid to tlic fVu-l that Bates ^) 

 has made iiieasiireiiienls witli Mohdiuiis of cyaiiijie, fiKdiisiiie, litmus 

 and aniline blue, from which it wouUl follow lliat these large 

 rotations did not exist, whereas Schmauss ') wilh these verv sub- 

 stances has found ^ery large rotations. According to Hates these 

 large differejices are caused by a source of error which arises from 

 the circumstance that for these measurements we make use of light 

 of which the intensity varies with the wave length '). He shows 

 that both with the iialf-shadow method, and wilh Ihat where we 

 adjust on a dark or a bright band in the spectrum, great ei-rors may 

 be made as soon as we arrive at a region where the intensity-curve 

 of the light used shows a considerable deci'ease, and that this may 

 produce apparently large rotations. 



As this source of errors might also occur in my measurements 

 with potassium fei-ricyanide, it seemed important to me to investigate 

 in how far this may have had a disturbing intluence, and thus in 

 how far the lai'ge rotations then found would ha\e to be ascribed 

 to it. 



With the method involving the use of a dark band in the spectrum, 

 the source of above errors comes to this, that as soon as the intensity 

 of the light on the two sides of the band is not the same, we are 

 inclined to wrongly adjust the middle of the band, and to displace 

 it too much towards the dark side. For we may suppose that for 

 an adjustment we, as a rule, search for two points on the borders 

 of the band which are of eqnal intensity and then adjust between 

 them. It must be noted that attention has been re})eatedly drawn to 

 this source of error ^) although, as far as 1 know, an experimental 

 investigation of the errors which may so arise was first made by 

 Bates ^). A theoretical solution would be possil>le in the manner 

 indicated by Bates, but this requires a knowledge of the intensity- 

 curve of the spectrum which is seen l\v the observer in the absence 

 of the magnetic rotation. Moreover Ave ought to know which of 

 the intensities on the edges is nsed by the observer to determine 

 the middle of the band, and this especially will partly depend on 

 the observer. An experimental determination may easily be made. 

 We need only produce a spectrum witli a movable dark band and 



1) Bates. Ann. d. Phys. (4) 12 p. 1091. 



S) ScHMAUss. Ann. d. Phys. (4) 2 p. 280 ; 8 p. 842 ; 10 p. 8.j3. 



3) Bates. Ann. de Phys. (4) 12 p. 1080. 



*) Gernez. Ann. éc. norm. 1 p. 12 (1864). 



Van ScHAiK. Thesis for the doctorate. Uti-echt 1882 p. 30. 

 «) B.\TE3 1. c. p. 1086, 



