March 20, 1913] 



NATURE 



69 



and E is Young's modulus. Hence by determining 

 in and E and also by measuring" the changes in 

 thii kness of a stressed plate, the sum of the principal 

 stresses may be evaluated as an average throughout 

 the thickness of the plate. Having obtained the sum 

 and difference, it is a simple matter to state the values 

 of p, and p, separately. A new form of instrument 

 is described in the paper, specially devised for measur- 

 ing small changes in thickness of a stressed plate. 

 This instrument is partly optical, readings being 

 obtained by means of a ray of light reflected from a 

 mirror which is rotated partially by the strain to be 

 measured. One millimetre on the scale is equivalent 

 to two millionths of an inch change in the lateral 

 dimensions of the specimen. A number of experi- 

 mental determinations are given in the paper and 

 show verv concordant results. 



COLLOIDS AND THEIR VISCOSITY. 



SPECIAL interest attached to the meeting of the 

 Faraday Society, held on Wednesday, March 12, 

 in view of the distinguished foreign guests who took 

 an active part in the proceedings. These included 

 Prof. Pauli (Vienna), Dr. Wolfgang Ostwald (Leip- 

 zig), Prof. Victor Henri (Paris), Prof. Freundlich 

 (Brunswick), and Prof. Nernst (Berlin). 



The meeting took the form of a symposium upon 

 colloids and their viscosity, and the afternoon session 

 was opened by Dr. Wo. Ostwald, who, in an intro- 

 ductory address of a general character, showed the 

 importance of viscosity measurements as a means of 

 study of the colloidal state. In the course of his 

 remarks, which were fully illustrated with examples, 

 he laid special stress upon the need for kinetic, as 

 opposed to static, methods for the investigation of 

 heterogenous systems, and in this connection also 

 emphasised the value of viscosity measurements. An 

 illustration of this principle w T as immediately afforded 

 by the communication of Profs. Freundlich and 

 Ishizake on the rate of coagulation of Al(OH) 3 -sols 

 as measured by the viscosity change, the results 

 of which were in complete accord with those of Paine 

 upon copper oxide-solutions, using a totally different 

 method. The following empirical formula proved to 

 express the experimental results of coagulation by 

 potassium salicylate with great exactness : — 



dx/dz = 2Kz{i + bx)(i-x) 2 , 

 where K is a constant depending on the concentration 

 of the electrolyte, z represents time, and x the amount 

 of precipitated particles, the latter taken as propor- 

 tional to the increase in viscosity. From the equation 

 in its more general form, 



dx dt = KJ(U)(i-x) 2 , 

 Freundlich and Ishizake drew the following conclu- 

 sions. The term (1 — x) 2 suggests the coagulation 

 process to be primarily a reaction of the "second 

 order " in which the colloidal particles may be supposed 

 to unite in pairs, the cause for which union is to be 

 found in an asymmetry of their electric charges (ex- 

 pressed in the term /(U)) due to unequal degrees of 

 electrolyte-adsorption. The degree of asymmetry was 

 found to be proportional to the time z, to the number 

 of precipitated particles, and to an exponent of c, the 

 concentration of electrolyte thus: — j(\3) = \c1z(i + bx), 

 where *, q, and b are constants. 



Prof. Pauli directed attention to the importance of 

 viscosity measurements in the study of " emulsoid " 

 colloids in a survey of the chief results obtained in his 

 own school, showing what important generalisations 

 as to the ionisation and degree of hydration of pro- 

 teins in solution had been arrived at by this means. 

 Ills experiments proved, for example, that at the iso- 

 NO. 2264, VOL. 91] 



electric point, where, by definition, the ionisation of 

 the protein is a minimum, a close correlation existed 

 between that property and (1) osmotic pressure, optical 

 rotation, viscosity, arid imbibition of water, all of 

 which reached their lowest value, and (2) precipit- 

 ability by alcohol which was at its maximum. With 

 increase in concentration of protein ions, caused by- 

 addition of either acid or alkali, a corresponding rise 

 was found to occur in the value of the first set of 

 properties and a fall in the precipitability. 



The evening session was chiefly devoted to a dis- 

 cussion of the factors concerned in the viscosity of 

 colloidal solutions and the interpretation to be placed 

 upon the viscosity value. Mr. Emil Hatschek de- 

 veloped a mathematical theory of the viscosity of two- 

 phase systems, showing that for "suspensoid" equally 

 with "emulsoid" colloids, viscosity depended upon the 

 volume-ratio of the two phases, and was independent 

 of the size of the colloidal particles. In the case of 

 the former, as shown also by Einstein and Bancelin, 

 the viscosity increased in linear ratio with the volume 

 of disperse phase, while in the case of "emulsoid" 

 coljoids the viscosity of the system was equal to 



s'A volume of system, 



-■/,- — ' where A = ratio : — ; -. — r . — \ , 



vA-i volume of disperse phase 



the viscosity of the continuous phase being taken as 

 unity. Experimental support was adduced in both 

 instances, and interesting confirmation also obtained 

 lor the above formula in the case of paraffin soap- 

 solution emulsions, where viscosity had been deter- 

 mined by means of Couette's apparatus, and direct 

 measurement could be made of the volumes of both 

 phases. Prof. Henri gave a critical survey of the 

 various direct and indirect methods available for 

 volume-measurement of colloidal particles. He showed 

 that, among the indirect methods, that of Perrin 

 (based on the distribution with depth of colloidal 

 particles after settling), and that of Ravleigh (bv 

 measurement of the intensity of light after lateral 

 diffusion through colloidal solutions) were among the 

 most trustworthy, since in the formulae used for cal- 

 culation of r, the radius of the colloidal particles, the 

 term r was raised to the third and sixth power re- 

 spectively. As a result of work with Fe(OH),-sols 

 Prof. Henri, expressed the view that apart from 

 the question of phase-ratio, or size of colloidal par- 

 ticles, the arrangement of the latter might have a 

 very important influence upon the viscositv of the 

 system. 



An interesting discussion followed, in which, among 

 others, Dr. Ramsden, Dr. S. B. Schryner, Dr. 

 McBain, and Dr. C. J. Martin took part. In the 

 absence of the chairman. Dr. R. T. Glazebrook, the 

 chair was taken bv Mr. Emil Hatschek. 



ATMOSPHERIC HUMIDITY AND 

 TEMPERATURE. 

 TWO papers on the psychrometer formula, re- 

 •l printed from recent Proc. Roy. Soc, Victoria 

 (vols. xxiv. and xxv.), discuss a modification, proposed 

 bv Dr. Ekholm, of the Stockholm Meteorological 

 Office, to be made in Regnault's formula for the 

 wet- and dry-bulb hygrometer, which would have im- 

 portant consequences if confirmed. The formula so 

 modified would be i- = ij/-AB(/-l'), where x and / 

 are respectively the actual vapour-pressure and 

 the saturation vapour-pressure at the tempera- 

 ture t' of the wet bulb. A is the ordinary 

 psychrometric constant, and 17 the coefficient, less than 

 unity, inserted by Ekholm to allow for a supposed 

 diminution of vapour-pressure at the surface of the 

 wet bulb due to hvgroscopic action of the material 

 covering- it. The first paper, by Dr. E. F. J. Love 



