March 15, 1900J 



NATURE 



48] 



In Part i. of this paper, read to the Society on February 2, 

 1899, the object of the experiments was stated to be the deter- 

 mination of the magfiilude of the thermo-electric effects ob- 

 tained from any one metal strained and unstrained. The re- 

 sults then given were obtained from two wires of the same 

 material, one wire being previously drawn through a draw- 

 plate, so as to reduce it in size from No. 18 standard gauge 

 (0"I22 cm. diameter) to about No. 24 standard gauge (0*0559 

 cm. diameter). 



The metals for which results were given in Part i. were 

 copper (six specimens), lead (two specimens), platinoid, German 

 silver, reostene and manganin.^ 



The present paper gives results of similar experiments made 

 on specimens of commercial ^ and pure lead, obtained from 

 Messrs. Johnson and Matthey ; and specimens of annealed steel, 

 of aluminium, and of nickel. 



B. — "Thermo-electric difference between free wires and wires 

 previously permanently elongated by longitudinal stresses." 



Attempts were made to determine the thermo-electric dif- 

 ference between free wires and wires previously permanently 

 elongated by a longitudinal stress. It was found difficult to 

 elongate the hard wires permanently to any appreciable extent 

 before they broke. 



The greatest percentage permanent elongation that could be 

 got in hard drawn copper, manganin, nickel and German silver 

 was 07, 0*5, 07 and 0*5 respectively. The thermo-electric 

 difference between the stretched and the unstretched wires was 

 then determined, and the results are given. 



C. — "Thermo-electric difference between free wires and wires 

 under stress, producing (i) temporary elongation, (2) permanent 

 elongation." 



The hot junction was kept permanently at steam temperature 

 during each set of experiments by an arrangement described. 

 Increasing weights were added on to the wire to produce (i) 

 temporary elongation, (2) permanent elongation. Three read- 

 ings of the galvanometer were taken : (i) with a weight on the 

 wire, (2) with a weight off, and (3) with the circuit broken. A 

 heavier weight was hung on, and other three readings taken, and 

 so on to the heaviest weight used in the experiments. 



The readings of the galvanometer were in the same direction 

 for all the wires tried with weights on and off, except for 

 soft copper and iron. The greatest permanent elongation pro- 

 duced in any of the hard copper wires experimented on was 

 0"I7 per cent., and for this permanent elongation the reading on 

 the galvanometer was in the same direction for weights off and 

 on, though always greater for the latter. 



For the soft copper wire the readings were in the same direc- 

 tion for weights on and off up to a permanent elongation of 

 I per cent. After a f)ermanent elongation of 472 per cent, the 

 current with weight on was 0'00i03 mikroampere per degree 

 from stretched to unstretched through the hot junction, while 

 with the weight off the current was 0-00075 mikroampere per 

 degree from unstretched to stretched through the hot junction. 



For iron wire the current was in the same direction for 

 weights on and off up to a permanent elongation of 0*35 per 

 cent. ; but after a permanent elongation of 341 percent, the 

 current with weight on was 000461 mikroampere per degree 

 from unstretched to stretched through the hot jnnction, and 

 with weight off 00069 mikroampere per degree from stretched 

 to unstretched through the hot junction. 



In " Mathematical and Physical Papers," vol. 2, p. 270, § 109, 

 Kelvin says : — " I have thus arrived at the remarkable conclusion 

 that when a permanent elongation is left after the withdrawal of a 

 longitudinal force which has been applied to an iron or copper 

 wire, the residual thermo-electric effect is the reverse of the 

 thermo-electric effect which is induced by the force, and which 

 subsists as long as the force acts." 



It seems (i) that for small longitudinal strain in copper or in 

 iron the direction of the current through the hot junction is the 



1 Dr. Anderson, Chemical Laboratory, the University, Glasgow, gave me 

 the following analyses for reostene and for manganin :— 

 Reostene. 

 Si ... 

 Fe ... 

 Ni ... 

 Mn ... 



98-30 



98-585 



- Dr. Anderson analysed the commercial lead, and found it contained 

 99-12 per cent, of lead. 



NO. 1585 VOL. 61] 



same, whether the force which produced the permaneut strain is- 

 on or off, (2) that as the permanent elongation is increased by 

 increased longitudinal forces, a stage is reached which gives 

 zero current when the forces are removed, and (3) that for 

 greater longitudinal forces and permanent elongations the direc- 

 tion of the current is opposite with the pulling forces off and on. 

 It seems, in fact, that the permanent elongation must exceed a 

 definite limit to produce the reverse thermo-electric effects which 

 Kelvin observed with the longitudinal force on and removed. 

 I hope to further investigate this point and to report the results- 

 to the Society. 



Physical Society, March 9.— Prof. Everett, F.R.S., Vice 

 President, in the chair. —A paper on the damping of galvano- 

 meter needles was read by Mr. M. Solomon. The solution of 

 the equation of motion for a magnetic needle, swingiiig in a< 

 uniform magnetic field, points to the conclusion that the ratio of 

 the period to the logarithmic decrement is independent of either- 

 the moment of the needle or the strength of the controlling 

 field, and is simply a function of the damping coefficient and the 

 moment of inertia of the moving system. This ratio should 

 therefore be constant if these latter quantities are constant. Ejt- 

 periments to test the constancy of period to logarithmic decre- 

 ment have been conducted at the Central Technical College at 

 various times since 1891, and they have invariably pointed to ai 

 variation in the value of the ratio. The object of the present 

 paper is to discover the cause of this variation. It may be due 

 to an alteration in the moment of inertia or to an alteration in 

 the damping coefficient. If the control magnets are either 

 directly above or directly below the needle, there is no chance 

 of any change in moment of inertia. The damping coefficient 

 depends on three things : (i) Viscosity of the air ; (2) viscosity 

 of the suspension ; and (3) eddy currents. The author has- 

 carried out experiments with a galvanometer on open circuit, and\ 

 finds a constant value for the ratio. The viscosity of the air and 

 suspension therefore cause no variation. Upon closing the cir- 

 cuit and repeating the experiments, the value of period over 

 logarithmic decrements alters. The variation is therefore due to 

 eddy currents. The damping factor due to eddy currents may 

 vary owing to three causes: (i) Change in moment of needle 

 due to change in field strength ; (2) effects of self-induction ;; 

 (3) effects of rise of temperature on the resistance of the coils. 

 The author points out that the two latter causes would tend to- 

 alter the ratio in the wrong direction, and he therefore concludes 

 that the variation is due to an alteration in the strength of the 

 swinging needle produced by altering the strength of the cpn- 

 trolling field. Mr. Blakesley said it was interesting to note the 

 fact that the ratio of period to decrement was independent off 

 the controlling field. In the case of a condenser discharging 

 this ratio is independent of the capacity ; in the case of a 

 tuning fork, of the rigidity ; and in the case of water oscillating 

 up and down in a U-tube, of the acceleration due to gravity. 

 Mr. Rosenbaum said that the ratio considered was constant irv 

 the case of a Nalder D'Arsonval galvaiKMneter. Mr. Solomort 

 said that his arguments did not apply to a galvanometer of this^ 

 description, because the swinging system was not a magnetic 

 needle but simply a coil. — A paper on the distribution of a gas 

 in an electric field was read by Mr. G. W. Walker. The 

 author has considered a gas as consisting of a number of mole- 

 cules each containing two atoms of equal mass, one positively 

 and the other negatively charged with electricity. When under 

 the action of electrical forces some of the molecules split up, 

 and we arrive eventually at a steady state in which there is a 

 definite number of undissociated molecules and of free positive 

 and free negative atoms. Treating the problem as one-dimen- 

 sional the potential at any point is expressed in general by 

 elliptic functions, and is therefore periodic. Applying the re- 

 sults to the case of a vacuum tube, it is found that there is super- 

 imposed upon the gradual fall of potential along the tube 

 minor periodic variations which it is suggested are connected 

 with the striae of discharge. Both the matter density and the 

 electric density are periodic along the tube. If the places of 

 maximum matter density coincide with the places of minimum 

 electrical action, then whether luminosity is due to collisions or 

 recombinations there will be maximum luminosity at these 

 points. In general these points do not coincide, and thus the 

 positions of maximum luminosity are not clearly defined. The 

 analysis leads to the conclusion that the distance between the 

 striae is inversely proportional to the density of the gas and to 

 the current strength, and these facts have been experimentally 

 verified. — Mr. C. E. S. Phillips exhibited a surface tension^ 



