[ 208 ] 



XVIII. Intelligence and Miscellaneous Articles. 



To the Editors of the Philosophical Magazine. 



_ University College, Galway, 



Gentlemen,— j une is, 1912. 



I" AM anxious to correct a statement made in a paper in the May 

 number, p. 751, that the effect of exposing a metal to ultra- 

 violet light is to make it more electro-positive to other metals. As 

 a matter of fact, the contrary is the case : the metal becomes more 

 electro-negative. It is easy to show, for example, that the contact 

 difference of potential between aluminium and copper can be 

 reduced almost to one half of its original amount, by exposing 

 the aluminium to ultra-violet rays. On the other hand, Kontgen 

 rays, at any rate according to some experiments made in the case 

 of zinc, have the effect of making a metal more electro-positive. 



Yours faithfully, 



Alexander Anderson. 



THE QUATERNIONIC FORM OF RELATIVITY. 



To the Editors of the Philosophical Magazine. 

 Gentlemen, — 

 The appearance of Prof. Silberstein's paper entitled " The Qua- 

 ternionic Form of Eelativity " in the May issue of the Phil. Mag. 

 is a welcome sign that continental mathematicians, who have 

 already largely availed themselves of various systems of vector 

 notation, are perhaps awakening to the suitability of quaternions 

 in such a connexion. Minkowsky (see footnote p. 79, Nach. Gott. 

 1908) must have had such an idea, but decided in favour of the 

 matrix notation. The quaternion form of the Hertz -Heaviside 

 equations was given by the present writer at the British Association 

 Meeting at Dublin (1908), and had been given by him in lectures 

 for some years previously. An application of quaternions to the 

 Eelativity Principle will be found in a paper, vol. xxix. Section A, 

 No. 1, Proc. Irish Academy (read Feb, 1911). Other writers, such 

 as Somerfeld, subsequent to Minkowsky, have used the four- 

 dimensional vector. The obvious defect of this latter method is 

 that it involves a second kind of vector having six components. 

 Beyond this, however, the quaternion has the advantage of being 

 asymmetrical, the time-scalar occupying a different position from 

 the space-vector. It is thus more in touch with real phenomena. 

 For no matter what view we take of relativity, the physical 

 methods of measuring time are quite different from those of 

 measuring spaces, and the flexibility of the quaternion product 

 allows us to put our results at once in a form suitable, if need be, 

 for numerical computation. 



Yours truly, 



Arthfb W. Conway. 



