236 



NATURE 



[July 5. 1894 



noticed. The " notes " say that ina'discussion on the paper, Mr. 

 C. S. Peirce proposed that the term " Galileo " be applied to 

 the unit of acceleration in the C.G.S. system. We also find 

 in them an account of the proceedings at the centenary cele- 

 bration of the birth of Lobachevyky hy the Physico- Mathe- 

 matical Society of the University of Ka/in. Further we learn 

 that Lambert's essay (./. our notice of the Biilhlin for Decem- 

 ber 1893) '* '° ^* incorporated in a volume entitled "Die 

 Theoric du Parallellinien " (Teiibner, of Leipzig), to be edited by 

 Drs. P. Stoeckel and F. Kngel. The prime factor will he the 

 " 5rst book of the marvellous work by Saccheri, ' Euclid 

 vindicated from every fled:,'" in which (in 1733) the two 

 hypothese; which, besides Euclid's, are possible are developed, 

 and all the results obtained which have been ascribed to 

 Legendre. There is a list of new publications in higher and 

 applied mathematics. I 



WiedtmanH s Annalen tier Physik tind Chemie, No. 7. — [ 

 Further electro-optical experiments, by J. Elster and H. Geitel. [ 

 The capacity of thin layers of sodium, potassium, and rubi- 

 dium applied to the walls of vacuum tubes of promoting the 

 passage of a current when illuminated differs for different 

 colours. For long waves, rubidium is the most, and potas- , 

 slum the least sensitive. If the layers are illuminated by 

 polarised light the current intensity is greatest when the plane 

 of polarisation is perpendicular to the plane of incidence. 

 Electric oscillations of small period can be transferred to rare- 

 fied gas by illumination in presence of an alkali metal.— .\ new 

 phenomenon attending the passage of electricity through badly 

 conducting liquids, by O. Lehmann. This is a description of 

 the formation of halos round the electrodes in a solution of pig- 

 ments in water thickened with gelatine, sugar, or glycerine. 

 Considerable disturbance is produced where the different 

 coloured halos meet, while the rest of the solution remains 

 undisturbed. — Experiments with Tesla currents, by K. Himsiedt. 

 The author gives an account of methods by which Tesla's 

 experiments can be repeated with ordinary laboratory apparatus. 

 High potential and rapid oscillations were produced by a 

 Lecher wire combination used for producing Hertz oscillations. 

 —On the demonstration of Hertz's experiments, by P. Drude. 

 The author avoids the necessity of a high tension accumulator, 

 as used by Zehnder, by allowing the sparks of the resonator to 

 discharge an electroscope charged by a dry pile. The )ioint 

 behind the concave mirror is put to earth ; also one pole of the 

 dry pile, the other pole being connected with the electroscope 

 and the sphere behind the mirror. When sparks pass, the 

 leaves of the electroscope collapse partly or totally. "This may 

 be shown to a large audience by projecting an image of the 

 electroscope on to a screen. — The change of phase of light by 

 reflexion at thin films, by W. Wernicke. Under the name of 

 "optical phase analysis " the author describes a method of detect- 

 ing exceedingly minute impurities on the surface of polished 

 glass or glass covered with a thin layer of gelatine. The 

 influence of the pl.iy of cohesive force upon free molecules as 

 regards their optical properties is investigated for pigments and 

 the me;als, with especial reference to silver. 



SOCfET/ES AND ACADEMIES. 

 London. 



Royal Society, Februarys. — "Thermoelectric Properties 

 of Salt Solutions." liy George Frederick Emery, late Scholar 

 of Trinity College, Cambridge. 



In a circuit formed by a metallic wire and a solution an electro- 

 motive force is developed pro(iortionalcly to the dlfl'^erencc of 

 temperature between the junctions. The solution to be examined 

 is put in a U-tube wiih an electrode and thermometer In each 

 limb, round one of which is a hot water-jacket. 4 Is = E. M.F. 

 perl C, unit 8 being IC* volt. Experiments were made with 

 acetate chloride and sulphate of zinc, and sulphate nitrate and 

 acetate of copper. 8 varied considerably with concentration. 

 The value for pure water appears to be about 8 6, Iml cannot 

 be mcmired directly ; for some salts 8 increases with concen- 

 tration 1 ,r ,,rlirrs It diminishes. In all cases examined the 

 carp ' from hot to cold through the solutions. With 



'inc 'nated zincelectrodcs ucreused; with thccopper 



sails the clgtir'jde used was a fine wire projecting from the 

 end of a diawn-out glasi tube. Values of 3 for mixed sails 



NO. I 288, VOL. 50] 



seemed to show that differences from the water value are 

 qualitatively but not quantitatively additive. Thus 5, starting 

 Irom about S'6, tends for moderate concentrations to a nearly 

 constant value for each salt. M. liouty, with very strong 

 solutions of zinc chloride, found that 5 rapidly diminished, 

 whence the entire curve for all concentrations between zinc 

 chloride and pure water would have a point of inflexion. If, 

 keeping the salt a fixed quantity, we use mixtures of two 

 solvents, we get a complete curve for 5. E.xperiments on i per 

 cent, of cadmium bromide give good results with all solvents 

 used ; with mixtures of methyl alcohol and water with alcohol, 

 it gave the following values : — 



Methyl .ilcohol I 



per cent. 5 



lOO'O I1'3 



9o'o iro 



81-3 1076 j 



700 10-4 - io'5 



SCO IO'27 



30-0 9-86 



'S-7 964 



100 S'9 



o 815 



In the first we have an inflected curve never far from the mean, 

 in the second a small admixture causes a large drop in S. These 

 two pairs of solvents represent two classes. The alcohols mix 

 quietly without chemical action, while alcohol and water mix 

 with evolution of heat, and change in bulk. 



A few experiments, believed to be entirely novel, were made 

 on the E. M.F. in a circuit composed of two kinds of liquid with 

 junctions at different temperatures. Zinc sul|)hate 4 per cent, 

 and weak zinc chloride gave E/(/' - /) = 136 \ 10"' V. 



Zinc acetate and zinc >ul hate gave E'/' - / = 0'8 x 10'*. V, 

 I 05 >. 10"* V, I 13 >, 10"' V, mean value = I x to"' V. 



Lastly, measurements of the Peltier effect at a metal-liquid 

 junction were made with various apparatus with fairly consistent 

 results. 



For 15 per cent, copper sulph.ite and copper, different 

 measurements gave the heat evolved per unit =H=0'I992, 

 0-I927, 01956, o 207S, 02091, 01952. 



The last and best gives 11/ 1" = 6'83 x 10" ■• = 5 for the solu- 

 tion. Cupric nitrate with 5 = 6'I4 gave H = 01764, 

 H/T = 6'i. Thus these thermoelectric eflects are of a 

 reversible nature. 



May 10. — "The total eclipse of April 16-17, 'S93. Report 

 of results obtained with the slit spectroscopes." By Captain 

 E. II. Mills, R.E. 



This paper deals with the results obtained from the photo- 

 graphs of the spectrum of the eclipsed sun taken in lirazll and 

 Africa at the total eclipse of .Xpril 1S93. 'he instruments 

 employed, of which there were four, were slit spectroscopes of 

 the ordinary type, and were each arranged to take one photo- 

 graph during totality. Of the four resultini; photographs two 

 were |).irlially unsuccessful and were not measured. The two 

 others each show a strong prominence spectrum, and on both 

 sides of this a continuous coronal spectrum, in which latter 

 are seen a number of very faint lines. The wavc-lenglhs of 

 these lines were determined by using the known lines in the 

 prominence spectrum .as reference points, and from these con- 

 strucling an interpolation curve. The coronal lines, whose wave 

 lengths were thus fixed, were, in almost all cases, apparently 

 identical with lines which had been observed at previous 

 eclipses, instruments of a similar type having been employcil 

 at the eclipses of 1882, 1883, and 1SS6. 



The prominence spectrum, as shown on the photographs, 

 extends from w.l. 3667 to w.l. 5316. It is chiefly remarkalilc 

 lor the extended hydrogen scries, there being eight lines beyond 

 the lineal w.l. 3699, the wave-lengths of which are given as 

 3692 5, 3687, 3682, 367S, 3675, 3O72, 3669 5, 3667. 



" Kesearchcs on Modern Explosives" (preliminary com- 

 munication). Hy William Macnab and E. RIstori. 



A series of experiments with explosive compounds has been 

 undertaken by the authors for the purpose of stiirlying cliemicil 

 reactions at high temperatures and pressure', and of clucidatii' 

 certain thermal constants relating chiefly to the specific heat *>\ 

 gases under such conilitions. Nitroglycerin, nitrocellulose, anil 

 several combinations of these two bodies, which are used as 

 smokeless powders, have been chiefly employed in these ex|)eii- 

 mcnis. The results given in this communication relate princi- 



