to a certain height to afterwards decrease in temperature by radiation as the incandescent mass assumes first a vaporous and finally a solid condition. In the ascending branch are (1) nebula (bright lines); (2) stars with bright lines; (3) stars with little absorption, bright hydrogen lines; (4) stars with bright carbon, manganese, and zinc flutings; (5) stars with line absorption; at the top of the curve are stars with hydrogen absorption. In the descending branch are (1) stars with line absorption; (2) stars with carbon absorption. The sun is situated in the descending branch, as shown by the great thickening of the potassium line due to a concentration of the sun's atmosphere by condensation. Н. К. Т.

Galvanic Elements. By ROBERTS (Dingl. polyt. J., 267, 141). -The "permanganate element" is said to be free from chemical action when the circuit is broken. The solution consists of potassium permanganate, potassium dichromate, common salt, and sal-ammoniac. The negative electrode is a rod of zinc, the positive a carbon prism. The mean E.M.F. is 1-8 volts, the internal resistance 0.5 ohm. One element is sufficient to work an electric bell.

The "lead dioxide element" is prepared by mixing red lead with powdered potassium permanganate and sulphuric or hydrochloric acid. The mixture is poured into a mould containing carbon, and on solidifying forms a good porous conductor, which adheres firmly to the carbon, and has the same hardness. The electrode thus obtained is employed as a substitute for carbon with zinc and sulphuric acid, and yields a very constant current. By the application of a solution of salt and sodium dichromate still better results are obtained.

The electrolyte in the so-called "dry element" forms a thick, pasty substance prepared by mixing two saline compounds, each dissolved separately.

D. B.

The

Experiments with Lippmann's Capillary Electrometer. By J. H. PRATT (Amer. J. Sci., 35, 143–151).-A description of the instrument used, which is of very simple construction, is given. results show that it may be employed for measuring low potentials, up to 0.6 or 0.7 Daniell, and therefore for comparing the E.M.F. of different batteries, if only a known fraction of the current be used. Oxygen polarisation must be avoided, and only hydrogen polarisation employed. Under these conditions, the deflection of the meniscus may be taken as proportional to the E.M.F. for very low potentials, and for potentials up to 0.9 Daniell an empirical curve will show the relation between the two. The polarisation is complete, and no appreciable current passes through the electrometer until it is charged to a potential near that at which electrolytic action begins. The capacity of the instrument is very considerable as compared with that of the quadrant electrometer.

H. C.

Aëration Currents. By C. R. A. WRIGHT and C. THOMPSON (Proc. Roy. Soc., 43, 268-273; compare Trans., 1887, 672; Abstr., 1887, 1008).-If different aëration (spongy and smooth platinum and gold) plates are placed above the same oxidisable metal in dilute sulphuric acid or aqueous potash, and left until a con

stant aëration is obtained, and if they then are connected together, a current is set up, which lasts with decreasing intensity for several days, and which at starting has an electromotive force equal to the difference between the electromotive forces obtained by exposing each plate to the oxidisable metal. The passage of the current electrolyses the sulphuric acid, with the result that the aëration of the positive plate is reduced, whilst that of the negative plate is increased. If the current is interrupted for some time, the plates recover their original aëration. With silver plates in sulphuric acid, acetic acid, and aqueous ammonia, a stronger current is obtained, the silver passing into solution; moreover, the amount of silver deposited in a voltameter by the current is less than that dissolved in the cell. The difference is greater with feeble than with strong currents. If the silver is opposed to a powerful oxidising agent such as platinum in nitric acid, the current produced is stronger, but the discrepancy between the quantity of silver dissolved and that deposited remains. Other metals dissolve in the same way, thus mercury opposed to an aëration plate gives mercurous sulphate. Gold in potassic cyanide gives potassium aurocyanide. Palladium behaves in a similar manner. If platinum in chromic acid is substituted for the aëration plate, gold can be dissolved in hydrochloric acid, forming aurous chloride which subsequently decomposes. H. K. T.

Electrical Conductivity of Sulphur. By E. DUTER (Compt. rend., 106, 836-837).-Sulphur is a non-conductor at all temperatures below its boiling point, but at this temperature it is possible to pass an appreciable current through the liquid. The gold electrodes which were used became covered with a deposit which is under examination. C. H. B.

By

Electrical Conductivity of Concentrated Nitric Acid. E. BOUTY (Compt. rend., 106, 654-657).—The addition of water to fuming nitric acid at first causes an increase in conductivity almost proportional to the quantity of water added, and this proportionality persists through a much wider range than the increase in conductivity resulting from the addition of nitrates. From 0.076 mol. H2O to 1478 mol. H2O the mean increase in conductivity is 0-456 for H2O, the unit being the conductivity of normal nitric acid, the specific resistance of which at 0° is 4.59 ohms.

When the increase in conductivity produced by salts (6.955 per equivalent) is compared with that produced by water (0-456 per equivalent), the water is less active than the salts in the ratio of 1: 15.25, and this ratio would be still lower if the pure acid were the basis. It would seem that the electrolyte produced by the first additions of water contains a somewhat large number of molecules of water in each molecule of electrolyte.

The addition of small quantities of dehydrating substances, such as phosphoric anhydride or sulphuric acid, to the nitric acid increases its conductivity, probably owing to the formation of complex electrolytes containing both acids.

The determinations of the conductivity of concentrated nitric acid

itself were not very satisfactory, owing to the action of the acid upon the vessels containing it, and consequent variations in resistance. Probably the electrolytic molecules are not the same with different degrees of dilution. A study of the phenomena of polarisation seems to indicate three distinct phases in the electrolysis, namely, the phase of the formation of the electrolyte corresponding with a composition represented by the formula N2O,,4H.O, the E.M.F. of polarisation being 005 to 0.2 volt, and the increase in conductivity per equivalent of water 0-456, whilst the principal product at the cathode is nitrogen peroxide; (2) the phase of dissociation extending somewhat beyond the point of maximum conductivity, with an E.M.F. of polarisation of 06 to 09 volt, the product at the cathode being a complex mixture containing nitrous acid; (3) the phase extending from the point of maximum conductivity to high degrees of dilution, the E.M.F. of polarisation being from 16 to 18 volt, whilst hydrogen is the only product at the cathode.

C. H. B.

Influence of the Composition of Glass on the Depressionphenomena of Thermometers. By R. WEBER (Ber., 21, 1086 — 1096). Tables are given showing the composition of various kinds. of glass, and the depression which took place after various lengths of time. The results show that the quantity of silica present does not affect the depression, nor has the quantity of lime, which varies from 10 to 15 per cent., any appreciable effect. Neither a large amount of lime nor of silica can counteract the injurious effect caused by the presence of both potash and soda. The quantity of alumina has relatively little effect on the depression, but its presence is of use in promoting the liquefaction and thorough melting of the glass.

F. S. K.

Conduction of Heat in Liquids. By C. CHREE (Proc. Roy. Soc., 43, 30-48).—A flat-bottomed dish is supported just in contact with the surface of the liquid, and into this, water at about 75° is poured, the temperature at a known depth below the surface being then ascertained at intervals by measuring the resistance of a platinum wire suspended horizontally 26 cm. beneath the bottom of the dish. From these measurements, the conductivity is deduced. Two series of experiments are distinguished; in the first, the water in the dish was syphoned off after a certain interval of time had elapsed; in the second, it was allowed to remain in the dish until the end of the experiment. In the first series, the conductivities measured are, the units being centimetre and minute

For water at 18° 0·0747.

For sulphuric acid, sp. gr. 1·054, at 20·5° = 0·0759.

For sp. gr. 110, at 2010 = 0·0767.

For sp. gr. 1.14, at 193°

For sp. gr. 1.18, at 21°

0.0765.

0·0778.

From which it appears that the presence of a very considerable quantity of sulphuric acid produces an extremely small change in the conductivity for heat of water.

For carbon bisulphide at 15° = 0·0322.
For methylated spirit at 19.5° = 0·0354.
For paraffin oil at 19° 0.0264.

The method and theory are not sufficiently accurate to allow of any value being attached to the third significant figure in the above. The results in the second series were

For methylated spirit at 18°

= 0.0346.

For paraffin oil at 20° = 0·0273.

For turpentine oil at 18° = 0·0189.

The results on the whole agree with Weber's (Ann. Phys. Chem. [2], 10, 103, 304, and 472).

H. C.

Bibasic Glyceroxides. By DE FORCRAND (Compt. rend., 106, 665 -667, and 746-749).-When sodium ethoxide and monosodium glyceroxide in molecular proportions are heated together for some time in presence of an excess of alcohol, and the liquid evaporated to dryness in a current of hydrogen at a temperature not exceeding 180°, the product contains 29-23-29.86 per cent. of sodium, and hence only about half the sodium in the ethoxide combines with the glycerol. If, however, evaporation takes place at 180-190°, decomposition is practically complete. Below 100°, the compound

CH,NaO3, EtOH

is formed; at 100-105°, or even 120° in dry hydrogen, the product is C3H,NaO3, EtONa; at 180-190° the bibasic glyceroxide, C,H,Na2O1, is obtained. Potassium methoxide and monopotassium glyceroxide do not react at 180° even after several hours.

The heat of solution of disodium glyceroxide, C,H,Na2O, at 10° is +14.48 Cal.

C3H-NaO, solid + Na solid = H gas +
C3H6Na2O3 solid..

solid + C3H,Na2O, solid. ..

CH,NaO, solid + NaOH solid = H2O solid

+ C3H,Na2O3 solid ....

and inversely

C3H Na2O3 solid + H2O liquid = NaOH solid+CH&Na2O3 solid

develops +29.91 Cal.

The last result explains the instability of the compound in moist air. In the case of monosodium glyceroxide, this reaction is endothermic, and hence the compound is much more stable, and can even exist in solution in a partially dissociated condition. The reaction C,H,NaO, solid + EtONa solid = C,H,Na2O, solid + EtOH liquid absorbs 2:08 Cal., is endothermic for the solid compounds, but becomes possible under the complex conditions of dissociation resulting from the action of heat. The following result, C,H,NaO,,EtOH solid EtONa dissolved in nEtOH liquid = CH,Na2O3 solid +

(n + 2)EtOH liquid, absorbs -16.64 Cal., explains the fact that disodium glyceroxide is only formed at a high temperature. The corresponding reaction for the monobasic glyceroxide is exothermic, and this compound is formed at the ordinary temperature.

C, (diamond) + H。 gas + Na, solid + Os gas = C,H,Na,O, solid, develops +239-62 Cal.

The author was unable to obtain any alcoholate of disodium glyceroxide. The residue at 120° is shown by its stability in a vacuum at 105°, and its heat of solution (+13·52 Cal. at 10°) to be the compound C,H,NaO, ETONA.

The heat of formation of this compound from its proximate constituents is +1:34 Cal., and in the production of disodium glyceroxide it is first formed with development of heat, but at a higher temperature decomposes with absorption of -35 Cal.

The heat developed by the action of a second molecule of sodium hydroxide in dilute solution on glycerol in dilute solution is only +0.22 Cal., compared with 0.515 developed by the action of the first molecule, but even this value is higher than the corresponding thermal disturbance, +0.04-0·11 Cal. in the case of alcohols. The action of sodium, sodium oxide, or sodium hydroxide on glycerol develops +14 Cal. more than the corresponding reaction with monosodium glyceroxide. The thermal disturbances resulting from the action of sodium, sodium oxide, and sodium hydroxide on monosodium glyceroxide are in each case about 2:0 Cal. lower than for the corresponding reaction with ethyl alcohol, a result which explains why, when two atoms of sodium act on a mixture of ethyl alcohol and glycerol, one atom is substituted in each compound, the bibasic glyceroxide only being formed at a high temperature.

C. H. B.

Freezing Mixture. By I. A. BACHMAN (Amer. Chem. J., 10, 45 -47). The spent acids (mixture of equal volumes of strong nitric and sulphuric acids) from Grove's battery, mixed with snow or broken ice till it forms a thin paste, gives a temperature of -30°.

H. B.

Relative Densities of Hydrogen and Oxygen. By LORD RAYLEIGH (Proc. Roy. Soc., 43, 356-363).-Regnault's method was adopted, but the globe was not cooled by ice, being simply inclosed during weighing in a wooden case, and the average temperature taken. The hydrogen was evolved from amalgamated zinc and platinum in sulphuric acid in a vacuum, the hydrogen being liberated by connecting the zinc and platinum plates, the gas was purified by potash and mercuric chloride, and dried with phosphoric anhydride. The oxygen was obtained from potassium and sodium chlorates. An important correction, overlooked by Regnault, is made in observing the compression which the globe suffers when exhausted; this the author determines by experiment, and from the theory of elastic shells. The ratio found is 15.884 (Regnault, 15.964), the difference being mainly due to the above correction. With Scott's ratio of atomic volumes 1.9965 (Abstr., 1888, 411), the value 15 912 is obtained for the ratio of atomic weights. H. K. T.