356 Mr. Yngve Bjornstahl on Magnetic 



Statements of Experimental Results. 



Variations of double refraction with intensity. — At the 

 beginning of every series of experiments the well-cleaned 

 tube was tilled with water, and examined to see if there was 

 any double refraction, due to stress in the cover-glasses. It 

 is very difficult to avoid these stresses, which necessitate 

 re-cementing of the glasses. 



The tube filled with the colloid in question is placed 

 between the poles of the magnet. 



The phase-retardation increases with the intensity of the 

 field and is positive, i. e. it has the same sign as the double 

 refraction which one gets when glass is extended in the 

 direction of the lines of force. In Table I. A is the phase- 

 difference, expressed in minutes of arc per 2' 60 cm. length 

 of the liquid traversed, and H the intensity in g;iuss. In the 

 last column the ratio A/H 2 x 10 7 is tabulated (see PL XIII. 

 fig. 2). Disregarding the first values it is evident that this 

 ratio is constant. Thus the effect is proportional to the 

 square of the intensity and is rather great ; by way of 

 comparison I may mention that nitrobenzene *, exhibiting an 

 exceptionally great magnetic double refraction, gives a 

 phase-difference of 54' in a field H = 20,000 gauss. In the 

 case of sol No. 6 the corresponding quantity is 220'*3, 

 although the concentration of the sol is c = 2'5x 10~ 4 normal 

 or about 0*05 nig. gold per c.cm. 



Variation with concentration. — The relation between the 

 effect and the concentration is of fundamental importance 

 for the investigation. Many sols can be prepared only 

 between certain limits of concentration, while it is necessary 

 to make the measurements between other limits. Thus one 

 must proceed with the greatest care when diluting a solution.. 

 I began by observing that there was no proportionality with 

 regard to the first of a series of successive dilutions. The 

 delay of phase appeared to be less than could be expected 

 (see Table II. and PI. XIII. fig. 3 ; c is the concentration). 

 On further dilution the effect was proportional to tl e concen- 

 tration. This fact may possibly be explained by the sensi- 

 bility of the solutions to electrolytes. A sol was allowed to 

 settle and the water over the colloid was used for dilution. 

 A complete proportionality was obtained. 



The dispersion of the double refraction. — The change of the 

 effect with the wave-length was determined by inserting 

 monochromatic filters from Wratten and Wainwright, only 



* Cp. Cotton, Ann. tie Chimieet de Phys. t. xix. p. 153. 



