of the Plane of Polarization of Light. 315 



o / 



In the first case there was a rotation of +11 25 



,, second ,, ,, +12 27 



Difference .... +12 



If the platinized glass was placed between the poles without 

 using the sulphide of carbon — 



o / 



The rotation obtained was +111 



And for the glass with iron +2 13 



Difference +12 



Exactly the same as before. 



We see that the rotation of the sulphide of carbon which is 

 in as close contact with the iron as the glass is, is not at all 

 influenced by it. We cannot therefore avoid the conclusion 

 that the difference 1° 2' in both cases is the rotation produced 

 by the iron. 



III. Dispersion produced by Rotation, Specific Rotatory Power, 

 and 'Circular Double Refraction of Iron, Cobalt, and Nickel. 



Since lately there has frequently not been sufficient sun for 

 the purpose, I have not been able to make extended experi- 

 ments on the rotation of the plane of polarization for the dif- 

 ferent colours produced by iron, cobalt, and nickel. Never- 

 theless I have been able to recognize with certainty that iron 

 has an anomalous rotatory dispersion. Light which had 

 passed through a tolerably dark red glass was rotated per- 

 ceptibly more powerfully than blue light obtained by passing 

 the sun's rays through ammoniacal copper-sulphate solution. 



Corresponding experiments with cobalt and nickel gave 

 only a feeble rotatory dispersion. Experiments in which the 

 red- and blue rays of a somewhat impure spectrum were 

 employed showed also an anomalous rotatory dispersion in 

 the iron. 



The specific rotatory power of iron was determined with as 

 good a mirror as possible. 



The thickness of the iron film obtained by weighing the 

 mirror before and after coating it with iron with the battery 

 was found to be 0'000055 millim.; the rotation for the mean 

 rays of the spectrum produced by glass and platinum was 

 1° 37', and that produced by glass, platinum and iron 3° 25': 

 consequently that produced by iron 1° 48'; the thickness of 

 the platinized glass was 1*61 millim. Consequently the iron 

 rotates some 32,000 times more than the glass used. 



Y2 



