362 Mr. Yngve Bjornstahl on Magnetic 



dubious and partly abandoned in favour of the theory of 

 double strata*. In addition The Svedbergf has shown that 

 a magnetic field does not perceptibly influence the Brownian 

 movement. Hence, if we suppose the particles to be spherical, 

 there is no possibility of explaining the magnetic double 

 refraction. 



If a thin sheet of fluid is caused to flow between two 

 parallel planes a short distance apart, it settles down under 

 certain circumstances into a steady state of flow, so-called 

 laminar motion. Let us consider the forces on a bar-shaped 

 body in the liquid. The difference of velocity between two 

 adjacent strata, i. e. the gradient of velocity, exerts an 

 orienting couple on the rod. This has a tendenc3 T to settle 

 its axis in the direction of the stream-lines. Diesselborst, 

 Freundlich, and Leonardt ± placed a sheet of colloidal V 2 5 -sol 

 between two crossed nicols and found double refraction 

 in the streaming fluid. They also showed that the colloid 

 consisted of bar-shaped particles, and inferred that the double 

 refraction was due to orientation. 



Combining this method with the Brace-apparatus, I have 

 investigated gold sols prepared in accordance with the same 

 method as was used by Pihlblad § and "Westgren ||. I have 

 found that the sols show double refraction, acquiring 

 properties analogous to an uniaxial crystal, whose axis is 

 parallel to the direction of the flow. The double refraction 

 is negative, and has the same sign for a high disperse sol 

 2r=36 /jbfM as for a low disperse one 2?' = 200 jifi. Hence we 

 must conclude that the shape of the particles in these 

 colloidal gold solutions cannot be represented by a spherical 

 symmetry. 



There is another phenomenon which points to the same 

 conclusion : the gold sols show a rather great double 

 refraction in an electric field If. 



Thus, establishing the aspherical form of the particles, we 

 may refer the magnetic double refraction to their orienta- 

 tion ; the existence of a time of relaxation can be explained 

 in this way. I will return later to the discrepancy from 

 Mie's theory. 



Let us suppose the shape of the particles to be represented 

 by ellipsoids of revolution. (This assumption is only a 



* Cp. Smolucliowski, Krak. Anz, 1903, p. 182. 

 t Die Existenz der Molehule, Leipzig, 1912, p. 99. 

 t Loc. cit. § Loc. cit. \\ Loc. cit. 



fi Cp. Phys. Zeit. xxi. p. 137 (1920). 



