SECTIONAL TRANSACTIONS. 



SECTION A.— MATHEMATICAL AND PHYSICAL 



SCIENCE. 



Thursday, September 7. 



(For references to tlie publication elsewhere of communications entered in the 

 following list of transactions, see p. 408.) 



1. Discussion on The Origin of Magnetism. Opener, Prof. P. 



Weiss (see p. 411). 



2. Prof. Sir J. Larmor, F.R.R. — On the Structural Significance 



of Optical Rotatory Quality. 



It is known that there are crystals, such as quartz, which exhibit high optical 

 rotation, whereas the same substance in amorphous (fused) condition is devoid of it. 

 It must then be due to chirality (right and left handedness) in the crystalline forma 

 which is not present optically in the constituent molecule. A molecule may thus be 

 chiral in geometrical form without being chiral in optical structure. The manner in 

 which crystalline form alone can thus be effective for optical rotation has been analysed 

 {Roy. Soc. Proc. 99A, Jan. 1921, pp. 7-11) on the hypothesis that the crystalline unit 

 is the charged ion, not the bipolar molecule. This view of crystal structure was origin- 

 ated and confirmed over a wide range by X-ray ultra-optical analysis in the hands of 

 W. H. and W. L. Bragg and others. The point now made is that it suffices to account 

 for the optical rotation of substances such as quartz, which otherwise it would seem 

 hardly possible to understand. 



It is explained [loc. cit. p. 7) that owing to the screw structure as regards the ionic 

 crystal elements, an impressed electric field produces a slight magnetisation which is 

 a linear function of its time-gradient : and that is the efficient cause of the optical 

 rotation. For isotropic screw quality the two vectors would be proportional, say 

 A = A' . d¥/dt : and then the optical rotation per unit depth for radiation of period 

 2tz/2} would be Tcp'Vc. The optical rotation is, however, dispersive : therefore the 

 ionic twist constituting the' magnetisation A involves inertia as well as structural 

 elastic reaction, and thus it has a period of free oscillation. The relation 

 A = k . dVjdt is thus to be replaced, to include variety of jieriods of the light, by 

 one of type 



The previous form therefore still holds if 



instead of I: being constant. The rotatory power thus assumes the form -- ,j r-.' 



which is equivalent to /(X- — 17), the free period of crystalline ionic twist being 

 c~' V I/. For aeolotropy there may be more than one such period, up to three : the 

 discussion may follow concisely the lines of ' either and Matter,' p. 350. In the 

 Bakerian Lecture for 1921, Prof. T. M. Lowry has adduced evidence that for various 

 substances one such term adequately represents the course of the dispersion. 



This considers only optical rotation of crystalline origin. But in most substances 

 part of the rotation persists in the amorphous state and in .solution : therefore the 

 molecule is itself chiral in its ionic structure, and this essential property also contri- 

 butes its part. The free periods of chiral twist may or may not be included among 

 the prominent free radiative periods of the molecule. The dispersion would require 

 more than one term completely to represent it. 



In a recent important summary {Comptes Bevdn.?, July, 1922, p. 174) L. Lang- 

 chambon reports that, for various organic substances when strictly purified, while the 

 1922 B B 



