are effective as well as metal plates. It varies in some inverse ratio with the distance, and is distinctly produced when the coils are 1 m. apart. In the above experiment, the larger spark may be either short and dense, or long and zig-zag, and every part of it is effective. The smaller spark, however, should be short (between knobs); the seat of the action upon it appears to be in the neighbourhood of the cathode or negative pole. The influence is reciprocal; that is, the smaller spark also favours the larger. The action is propagated in straight lines, like light, and may be reflected from polished surfaces. It may also undergo refraction; but its refrangibility (roughly measured by means of a quartz prism) is much greater than that of the violet rays. Most solid substances are opaque to it; amongst these glass, paper, agate, and mica, even in the thinnest sheets, are noticeable. Amongst crystalline substances, copper sulphate, topaz, and amethyst are opaque to it; but it is transmitted by sugar, alum, calc-spar, and rock salt; transparent gypsum and rock crystal transmit it perfectly. Amongst liquids, water transmits it freely; sulphuric and hydrochloric acids, alcohol, and ether, less so; whilst melted paraffin and petroleum, benzene, bisulphide of carbon, ammonium hydrosulphide, and coloured liquids generally, stop it completely. Solutions of potassium, sodium, and magnesium sulphates, are fairly transparent to it; those of mercuric nitrate, sodium thiosulphate, potassium bromide and iodide, are very opaque. Amongst gases, air, hydrogen, and carbonic anhydride are very transparent; chlorine, and bromine and iodine vapours, partially so; and coal-gas and nitric peroxide

very opaque.

Even an ordinary candle-flame may produce effects similar to those described, and may cause the reappearance of sparks between the terminals of an induction-coil after they have been drawn so far apart that the discharge has ceased. Similar effects are produced by the luminous flames of gas, wood, and benzene, and the non-luminous flames of alcohol, carbon bisulphide, and the Bunsen burner. Incandescent platinum, and the flames of sodium, potassium, sulphur, and phosphorus, and of pure hydrogen, are without effect. The effective rays are more refrangible even than the so-called photographic rays; for the latter are not sensibly absorbed by coal-gas.

CH. B.

Specific Heat of Liquid Carbon Compounds. By R. SCHIFF (Gazzetta, 17, 286—303).—In a former memoir (Abstr., 1887, 7), the author has shown that the variation of the specific heats in a homologous series of compounds is expressible either by a straight line, or a small number of parallel lines. In the equation for the mean specific heat Ct-v = a + b1+v), the coefficient b remains constant for all the terms of a homologous series, whilst the coefficient a either remains constant for all or several terms of the series, or varies per saltum with the molecular weight. These conclusions are confirmed in the present memoir, in which determinations are given of the specific heats of the ethereal salts of the chloro-substituted acetic acids and of the succinic acids, as also of certain aromatic compounds. The specific heats of ethereal salts of the propyl and allyl series are also compared. Thus it will be seen in the table below that the value

for b in all the ethereal salts of mono-, di-, and tri-chloracetic acid is a constant = 0.00038, in those of the succinic acids = 0·00066, in aniline and toluidine = 0·0007. Further, the value for a in the allyl salts of propionic, butyric, and valeric acids is a constant = 0.433, but in other cases it varies somewhat indefinitely. From these results, the author concludes that neither isomerism nor change of molecular weight produces definite variations in the specific heats.

Heat of Formation of Zinc Ethyl. By GUNTZ (Compt. rend., 105, 673-674).-The zinc ethyl was decomposed by dilute sulphuric

acid

ZnEta liq. + nH2SO, diss. ZnSO, diss. +

2C2H, gas

develops +79.8 Cal.

With hydrochloric acid, the number obtained is +780 Cal., but when these numbers are corrected for the heats of dissolution of the zinc salt in the excess of acid they become +808 and 77-8 respectively. Taking the mean of these values, it follows that

ZnEt, liq. + 2H2O liq. ZnH2O, pptd. +

2C.H. gas

develops +57.8 Cal.

the great development of heat explaining the rapidity with which this decomposition takes place. Further

Zn solid + 2C2H, gas = ZnEt, liq. + H2 gas absorbs -43·2 Cal. and

Zn solid+C, solid + H10 gas = ZnEt, liquid absorbs -318 Cal. The formation of zinc ethyl from its elements or from zinc and the hydrocarbon is accompanied by a very considerable absorption of C. H. B.

heat.

Relation of Gases to the Laws of Marriotte and Gay-Lussac. 1 dv By C. PUSCHL (Monatsh., 8, 327-337).-If a = be the coefficient v dt

of expansion, and c = — the compressibility, we have―

1 dv
v dp

When h = 0, Marriotte's law holds, and the gas will be more or less compressible than the law requires, as h is greater or less than 0.

For a gas at the critical temperature and compressed from an ordinary density, pu diminishes with the pressure, and h is therefore negative. As the pressure approaches the critical pressure, h diminishes and ultimately = -co. Further compression causes h to increase, and for a certain pressure h = 0. Here Marriotte's law holds. For greater pressures h is greater than zero, and the departure from Marriotte's law is in the reverse direction.

When the temperature of the gas is higher than the critical, h reaches a finite minimum during compression. This minimum increases with the temperature, and for a certain temperature the minimum value of h will be zero, and then both h = 0, = 0. At still higher tem

dh dp

peratures the minimum value of h will be positive, and pv will continually increase, as happens for example in the case of hydrogen. It appears therefore that the temperature for which h = 0,

dh

= 0

dp simultaneously is the highest possible at which Marriotte's law can be exactly obeyed. The present method of consideration also shows that at a sufficiently low temperature hydrogen behaves like ordinary gases.

If we trace a curve with ordinate h and abscissa p, there will be a point on it for which h is a minimum. This point will lie above or below

dh dp

the axis of abscissæ, according as the temperature exceeds or falls short of that for which h = 0, =0. For the lower temperatures, the curve meets the axis in two points. At the point with the smaller abscissa, d(pv) dh is a maximum for = h = 0, and also is negative. At dp the other point pv is similarly a minimum. Consequently for all gases a maximum value of po will precede a minimum, and at a certain temperature above the critical temperature the two will coincide.

ро

dp

for minimum values of pv, the pressure increases as the temperature diminishes. Since Amagat found in the case of oxygen the pressure at the minimum of pv decreases with rise of temperature, we infer this gas was nearly at that temperature where the maximum and minimum of pu coincide,

Similar reasoning is applied to the consideration of Gay-Lussac's law, and results of a very similar nature are deduced.

C. S.

Highest Boiling Point of Fluids. By C. PUSCHL (Monatsh., 8, 328-341).—A gas or fluid can be reduced to such a condition that v being its specific volume, p its pressure,

Thus the zero value of e is a minimum for the critical state. If the temperature is lowered, the minimum value of e reached during compression is negative, and there must therefore be for every temperature two values of v, for which e = 0. At the greater value of v, de is positive, and p is a maximum, whilst the contrary holds for the

dv

less value. When the volume is greater than that for the maximum

VOL. LIV.

с

of p, the substance exists in the form of vapour, and when the volume is less than that for the minimum value, the substance is fluid. Between the two, the substance is in an unstable state.

A saturated vapour may be compressed, so that e changes from a positive value to zero, and then the tension is a maximum = p'. And in a similar manner, by sufficiently reducing the pressure on a fluid, the tension may be made a minimum = p". For tensions <p", the fluid will be in an unstable condition.

dp dt

Suppose that a substance at a temperature above the critical fills a rigid vessel of such volume that the substance is of the critical density. If the temperature be lowered, will be positive. Thus whilst at the critical temperature the vapour-tension pp", at a lower temperature p is <p". But in accordance with what precedes, the fluid and its vapour are now in unstable equilibrium. Also p" at a sufficiently low temperature is negative. Thus there must exist a temperature somewhere below the critical when p again = p". At this point, when the temperature is lowered, the substance suddenly appears liquid, and on the other hand with rise of temperature, it completely changes into vapour. This temperature is therefore defined to be the highest boiling point of the liquid. If the mass contained in the vessel be greater or less, the temperature of the highest boiling point will not be affected. C. S.

The Relation of Hydrogen to Marriotte's Law. By C. PUSCHL (Monatsh., 8, 374-377).—As in his previous papers, the author traces the increase or decrease of the quantities, a, the coefficient of expansion, apv and h, by considering the signs of their differential coefficients. Since, according to Regnault's experiments, not only pv but also h increases as the pressure increases up to 20 m., he concludes that in order that h may have a minimum at greater compression, it must have first a maximum at a pressure above 20 mm. If hydrogen be allowed to expand, h will diminish, and we may conjecture that when the pressure is sufficiently small, h will be equal to zero. In this case, pv is a minimum. Since is positive, this minimum value will occur at greater pressures for diminishing temperature. In gases for which pv reaches a maximum through rarefaction, a minimum will occur on further rarefaction, and with diminishing temperature the maximum and minimum will approach towards coincidence. C. S.

dh

dt

Evaporation and Dissociation; Continuous Changes from the Gaseous to the Liquid State at all Temperatures. By W. RAMSAY and S. YOUNG (Phil. Mag. [5], 24, 196-212).-A further proof of the correctness of the formula p = bT-a (where p is the pressure, b and a are constants for each separate volume of 1 gram, and T is the absolute temperature), is furnished by the behaviour of methyl and ethyl alcohols, under a very wide range of temperature, pressure, and volume. Tables are given showing the correspondence of numbers calculated by the above formula with those experimentally determined.