TRANSACTIONS OF SECTION A. 515 



cient accuracy to justify its adoption as final He thinks it wrong by nearly one 

 per cent. The principal measurements are : — 



Joule (1878) 4ir.2t 



Hirn 4 2485 



VioUe 4-2701 



Regnault 4-2877 



Rowlands 



This ought to be a reason to re-e.-camine the question, and with our existing 

 knowledge of the ohm and the ampere we ought to be able to derive the Joule 

 with an exactitude that will satisfy the scientitic world. 



15. ExJiihition of Ley den Jars with Multiple Fracture. 

 By J. T. BoTTOMLEY, F.E.S., and Sir A.uchibald Campdeli,, Bart. 



16. 071 SparJcless Electro-magnets. By Professor S. P. Thompson, Ph.D. 



17. On the Correspondence between the Molecular Refraction, Dispersion, 

 and Maqnetic Rotation of Carbon Compounds. By Dr. J. H Glad- 

 stone, F.R.S., and Dr. W. H. Pekein, F.R.S. 



The molecular refraction, or refraction equivalent, of chemical compounds has 

 been the subject of many recent investigations. It is fully recognised that ' the 

 molecular refraction of a compound is the sum of the molecular refractions of its 

 constituents.' It is also well known that very fundamental differences of chemical 

 structure, such as change of atomicity, or of the mode of combination of such ele- 

 ments as carbon and oxygen, cause a difference in the refraction value. 



The molecular dispersion is the difference between the molecular refraction of a 

 body for two different parts of the spectrum, for instance the solar lines A and H. 

 A series of dispersion values may thus be determined for the elements ; and these 

 are analogous to the refraction values, but not strictly proportional to them. It is 

 also found that wherever a difference of chemical structure causes a differ- 

 I'nce in value in regard to refraction, it does so in regard to dispersion. This differ- 

 ence is generally much larger in proportion to the actual amount. 



The molecular magnetic rotation of organic liquids has lately been examined 

 minutely, and this has revealed the Axct that a corresponding series of vahies may 

 be determined for their elements or constituents, and also that the same differences 

 of chemical structure which produce clianges of refraction or dispersion produce 

 analogous changes in the power of rotating the polarised ray. This shows another 

 point of connection between an electro-magnetic effect and the velocity of light. 



18. On Kerr's Magneto-optic Phenomenon : its Latcs and Application for 

 Measuring purposes. By H. E. J. G. DU Bois, Ph.D. 



The rotation of the plane of polarisation on normal reflection from magnets was 

 studied with a special view to its close relation with their magnetisation. The 

 metals were used in the shape of prolate ellipsoids of revolution, on to which small 

 reflecting planes had been ground in different positions. They were magnetised in 

 a coil and their moments measured by the ordinary magnetometric method. The 

 following are the principal results: — 



1°. The rotation is proportional to the normal component of temporary or re- 

 sidual magnetisation, 



2°. For the ratio of these quantities the name 'Kerr's Constant' is proposed, 

 m honour of the eminent discoverer of these important phenomena. 



3°. The temperature variation of Kerr's Constant, if any, is so small that its 

 sign And value could not bo determined; it cannot be more than a few per cent. 



I, L 2 



