274 ]\/[ r> R # Hosking o?z the 



When ellipse and circle intersect, 



0= +0-58601, 0= +0-31945. 

 When ellipse is outside circle, 



x= +0-60200, y^-- +0-30250. 

 The real intersections on the axes are 

 v^ + v 2 x 2 -\-v- d = 0, 

 \,z i + 2^ + v 3 = 0. 



When ellipse and circle intersect, 



x= +0-58619, *=0-44278. 

 When ellipse is outside circle, 



a?= +0-60257, t/ = 0-43628. 



As the difference in lengths of axes of the crystals has 

 been much exaggerated in the diagrams, only the relative 

 positions of these points and lines are shown. 



A word should be said regarding the physical interpretation 

 of the parts of the locus hitherto omitted. A supposititious 

 case of refraction, which cannot be realized, may be given 

 to explain the existence of the dotted branches of the curve. 

 If a source of light were situated at the origin of coordinates, 

 emitting an ordinary and extraordinary ray in opposite 

 directions, then the locus which would give coincidence of 

 direction to these rays is the dotted curves indicated. 



I have not been able to interpret the presence of the 

 singular points on the axes. When the intersections on 

 the axes are singular points they always lie inside both the 

 ellipse and the circle. The two tangents to the ellipse and 

 Circle, which are the wave-fronts of the rays of light, are 

 consequently imaginary. It seems, then, as if the rays also 

 must be imaginary. 



XXII. Viscosity of Solutions. By R. Hosking, Kernot 

 Research Scholar in the University of Melbourne*. 



IT is evident that, in the immediate future, the study of the 

 viscosity of liquids is destined to prove as helpful in the 

 investigation of the molecular kinetics of liquids as the 

 corresponding study of gaseous viscosity has been found to 

 be in the kinetic theory of gases. 



Already a large amount of work on the viscosity of liquids 

 has been published, but much of the pioneer work has been 

 * Communicated by Mr. W. Sutherland on behalf of Prof. 



