February 5, 1920] 



NATURE 



617 



staff and equipment, and the encouragement of re- 

 search. 



A SCHEME for the establishment of a University 

 Bureau in the City of London in connection with the 

 University of London commerce degree is described in 

 the Times of January 30. It is proposed that the 

 bureau shall assist in the suitable and wide employ- 

 ment of commerce degree students and graduates in 

 all branches of trade and commerce throughout the 

 country and assist employers in all matters affecting 

 the tniining and employment of all such students and 

 g^raduates. An initial sum of 50,000!. has been set 

 aside for the purpose of establishing the bureau on a 

 proper footing. 



SOCIETIES AND ACADEMIES. 

 London. 

 Royal Society, January 22. — Sir J. J. Thomson, 

 IK-esident, in the chair.— 'Prof. E. G. Coker and K. C. 

 Chakko : The stress-strain properties of nitro-cellulose 

 and the law of its optical behaviour. The physical 

 properties of nitro-cellulose are studied from its 

 behaviour in tension, whereby values of Young's 

 modulus and Poisson's ratio are obtained and the 

 form of the load-extension curve is determined. The 

 optical properties of the transparent material are 

 observed, with special reference to its behaviour under 

 load ; and it is shown, by observations with a com- 

 parison beam not stressed beyond elastic limit, that 

 the relative retardation produced between the two 

 components of a polarised beam is consistent with a 

 linear optical stress law, which holds up to stresses 

 of about twice those at the elastic limit. These 

 results are confirmed by observations of the 

 retardation bands produced in a polarised spectrum 

 by a beam under uniform bending moment. The 

 stresses and strains are deduced on the assumption of 

 a linear stress-optical law, and stress-strain curve so 

 obtained is found to agree with purely mechanical 

 measurements of a tension member. — .S. Marsh : 

 Alternating-current electrolysis. The behaviour of 

 Ijlatinum, gold, and nickel electrodes during the pas- 

 sage of an alternating current of 25 to 80 cycles per 

 second has been examined. The electrolytes employed 

 weredilute sulphuric acid and barium hydrate solution. 

 Curves representing the relation between volume of 

 gas evolved and time of passage of current are of two 

 distinct types : (a) One type resembling " saturation- 

 current " curves in radio-activity. This effect is shown 

 clearly in the cases of platinum and gold in an acid 

 electrolyte. (b) :\ second type in which the rate of evolu- 

 tion of gas falls off with time until ultimately a steady 

 rate of evolution sets in, decreasing in value with 

 increasing frequency of the alternating current. Two 

 possible explanations of the phenomena are discussed : 

 fa) .Adsorption of hydrogen at an electrode during one 

 half-period, followed by recombination with oxygen in 

 the succeeding half-period, (b) Oxidation of the elec- 

 trode by the oxygen of one half-period, followed by 

 reduction of the oxide by the hydrogen of the suc- 

 ceeding half-period. Evidence is given that the oxida- 

 tion theory successfully explains the effects with gold 

 and nickel. In the case of platinum it is believed 

 that oxidation plays a prominent rdle, though adsorp- 

 tion also may be effective in this case. It is shown 

 that the electrodes have an initial surface activity in 

 promoting recombination, which activity increases 

 (a) with frequency of alternation, and (6) up to a 

 maximum value with the time of passage of the 

 current. If the current density is less than that corre- 

 sponding to this maximum activity, then sooner or 

 later evolution of gas ceases. If the current density 



NO. 2623, VOL. IO4I 



is greater, then after a time gas is evolved at a steady 

 rate.— VV. H. Eccles .and J. H. Vincent : The variations 

 of wave-length of the oscillations generated by three- 

 electrode thermionic tubes due to changes in filament 

 current, plate voltage, grid voltage, or coupling. When 

 electrical oscillations are sustained in a circuit com- 

 prising inductance and electrical capacity by aid of a 

 three-electrode thermionic vacuum tube of the kind 

 used in wireless telegraphy, the frequency of the 

 oscillations and the wave-length of the radiation 

 depend principally upon the values of the inductance 

 and the electrical capacity, but also partly upon the 

 resistance in the oscillatory circuit; upon the voltages 

 of the various batteries in use; upon the temperature of 

 the filament supplying the electrons; upon other pro- 

 perties of the vacuum tube ; and upon the coupling 

 between portions of the circuit associated with the 

 arid and the anode. The object of the present investi- 

 t'alion was to study experimentally the effects of 

 altering each of the chief variables, with the view of 

 finding the conditions most favourable for the pro- 

 duction of continuous oscillations of constant fre- 

 quency. For this purpose two circuits were sustained 

 in oscillation at nearly the same high frequency, 

 namely, about 120,000 vibrations per second, and the 

 niidible beat between these frequencies was observed. 

 Then changes made in one circuit alone caused varia- 

 tions of freauency that were measurable by acoustic 

 observation of the beat-note. The preliminary inves- 

 ti nations showed that variation of the heating current 

 of a filament was the most fertile source of erratic 

 dianfes of frequency, and resulted in the discovery 

 that increase in the filament current of one tube pro- 

 duced at low values of current a decrease of frequency, 

 and at higher values an increase of frequency, while 

 at a certain value of filament current the frequency 

 had a stationary value. This phenomenon provides a 

 method of setting an oscillation generator so as to 

 produce a vibration of frequency constant to less than 

 one part in 100,000. Provided with this knowledge, 

 the other problems enumerated above were attacked. 

 In an apparatus in which the inductance was eight 

 millihenries, the electrical capacity 250 electrostatic 

 units, and the wave-length 2750 metres, it was found 

 that raisint; the voltage of the anode battery from 

 iT,Q to 140 increased the wave-length bv 6 metres, and 

 raising the grid voltage bv i increased the wave- 

 lent'th about 10 metres. The coupling in the circuit 

 produced larffe effects bv its variation. — S. D. 

 Carother.s ; Plane strain : the direct determination of 

 stress. It is pointed out in the first part of the paper 

 that in plane strain the stresses, if determined directly, 

 are usually obtained bv the aid of the well-known 

 stress function method. The problem is usually that 

 of findin.rt a function y satisfying v,*v=o throughout 

 the body, with suitable values of x over the various 

 boundaries. The most general value of x '" Cartesiati 

 co-ordinates appears to be 



X = .\fi + Bxfi + Cy.$ + D(x'+yY9), 



where .\, B, C, and D are any constants and 

 ft is any plane harmonic function. It is shown 

 that for any orthogonal co-ordinates the stresses 

 derived by the stress-function method when 

 applied to x = (x' + y')9 can always be resolved into two 

 distinct sets, while in the case of Cartesian co- 

 ordinates the stresses can be split up into four distinct 

 sets. In view of the foregoing, the present paper has 

 for its object the determination of the various sets of 

 stresses which might legitimately occur in a state of 

 plane strain, expressed in the simplest possible terms, 

 with the view of rendering the buildine: up of a given 

 state of stress a manageable operation. The paper 

 sets forth the usual stress equations of equilibrium. 



