426 M. H. Becquerel's Experimental Investigations 



the measurements can be carried very far, since it enables 

 the variations in coloration of the successive images to be 

 followed with gases as well as with a mass of glass. The 

 average wave-length of the luminous pencil can then be ap- 

 proximately determined in each case and compared with the 

 rotations obtained. The most simple means consists in making 

 a graphical representation. 



If the numbers inversely proportional to the squares of the 

 wave-lengths distinguishing each image be taken as the 

 abscissas, and the magnetic rotations observed for the ordi- 

 nates, we see that the typical points of the determinations 

 relative to the same substance approach very closely to a right 

 angle passing by the origin of the coordinates, which proves 

 that the rotations are very nearly in the inverse ratio of the 

 square of the wave-lengths, as may be verified by reference 

 to the preceding table. 



The numbers produced by the different images with different 

 colours may each serve for the determination of an average 

 right angle which, within the limits of the observations, repre- 

 sents the phenomenon with sufficient precision. 



Briefly summarizing, we see that, for five of the gases 

 studied, the magnetic rotation of the planes of polarization of 

 luminous rays having different wave-lengths is very nearly in 

 the inverse ratio of the squares of the w x ave-lengths of these 

 rays. The slight value of the rotations observed did not 

 allow me to ascertain whether these gases, like solids and 

 liquids, diverge a little from this simple law in proportion as 

 the rays become more refrangible. 



The nitrous oxide gives a rather less deviation than the 

 other gases. We may remark that, in order to make the 

 numbers agree with those given by the other bodies, it would 

 suffice to multiply the results by 1*05 ; and as this ratio 1*05 

 is exactly the ratio of the rays of white light to the rays of 

 yellow light, we may conclude that the inequality is due to 

 an increase of brilliancy in the luminous source at the moment 

 of the experiment with the white light. This explanation 

 appeared to me to be admissible ; but if such were not the 

 cause of the divergence observed, we may associate this fact 

 with that presented by oxygen, and which I will now discuss. 



Oxygen showed a remarkable anomaly. Previously, while 

 studying the magnetic rotatory powers of various bodies with 

 the yellow light, I had discovered that the presence of oxygen 

 in a combination tended to diminish the rotatory power of the 

 compound. I now find that gaseous oxygen has a feeble posi- 

 tive magnetic rotatory power, and that, in addition, it does not 

 disperse the planes of polarization of rays of different colours 

 as the other gases do. The numbers which relate to this body 



