Double Refraction of Gold Colloids. 363 



provisional approximation.) Comparing the flowing experi- 

 ment described above with the effect in a magnetic field, I 

 will make an attempt to determine the orientation o£ the 

 particles in the latter case. In the first experiment the 

 double refraction was negative when we used as standard a 

 crystal whose axis was parallel to the direction of the flow. 

 In the magnetic field, the double refraction is positive. 

 Hence it follows, that the ellipsoid tends to turn its major 

 axis (or one of the majores) in a direction perpendicular to 

 the lines of force. The symmetry of the orientation is not 

 identical in both experiments but only analogous, being 

 represented in the magnetic field by an axial vector, in the 

 field of flow by a polar one. However, it is obvious that we 

 are justified in drawing the above conclusion. 



Considering, then, the cause of the orientation, we first 

 disregard the direct influence of the field. It is obvious 

 from the account given on page 362 that the Brown ian 

 movement per se has no directive influence. 



For the Brownian rotatory movement * the mean square 

 A of the angle of rotation in a time t for a sphere of radius a 



is given by A 2 =^- x i— — », where B, is the general gas 



& J N 47r^a 3 ' & to 



constant, T the absolute temperature, N the constant of 

 Avogadro, r f the viscosity of the fluid. As a consequence 

 of this movement Foucault's currents are induced in the 

 more or less massive particles, but it is difficult to imagine 

 that these currents would produce any orientation. But 

 they will always damp the motion to a certain extent. 



Before drawing any further conclusions, I shall mention 

 some propositions concerning forces on magnetic bodies. 

 In the field between two poles of a magnet, the former being 

 symmetrical about the axis and the equatorial plane, an 

 elongated ellipsoid of revolution of isotropic material sets 

 itself axially if it is paramagnetic, equatorially if it is 

 diamagnetic. The actual terminology is really due to this 

 fact. 



In a uniform field the relations are quite different f . Let 

 us consider an ellipsoid of isotropic material, the perme- 

 ability of which is /u,, to be surrounded by a medium of 

 permeability /jl . Let the axes of the ellipsoid be a>b>c 

 respectively. Under the influence of the field H (com- 

 ponents H 0a , H 06 , H 0c ) the field H in the ellipsoid also 



* Einstein, Ann. dor Pln/s. (4) xix. p. 379 (1906). 

 t Cp. Lord Kelvin, Phil. Mag., March 1855, 



