M. Turretini stated in studying a certain hydraulic question he had arrived at a solution interesting from an electrical point of view. The question was in arranging the utilisation of the motive power of the Rhône at Geneva to maintain a constant hydraulic pressure whilst a reservoir was being filled or emptied. This problem was solved by placing on the channel a centrifugal pump moved by a turbine driven directly by the water under the pressure of the channel. The apparatus has been working for two years, keeping up a constant pressure day and night. He thinks that in central stations worked with accumulators, an electric arrangement for collecting the loss of the charge of the accumulators whilst discharging would constitute an advance on the present system.

M. Ch. Jacquin exhibited an arrangement for measuring easily, day by day, the insulation of the nets of a central station.

He also gave an account of unpublished experiments carried out in simple cables for alternating currents, in order to determine the loss by induction, Foucault's currents and hysteresis in the armature of such a cable. A wattmeter with an alternating dynamo gave the total power, P t d, dissipated in the cables, and with a continuous current dynamo the power expended R 12 in the conductor.

Third Section.-Telegraphy, Telephony, Signals.

M. Palaz read a paper on the relations of vicinity of telephonic and industrial nets. He drew the conclusions that industrial electric nets ought to have a complete metallic circuit, as well insulated as possible. The telephonic nets should be constructed of double wire, in order to permit the ulterior development of inter-urban telephony, and the simultaneous and undisturbed working of telephonic and industrial nets.

M. Vaschy gave some practical instructions on the effect of industrial wires upon telephonic nets. The effects are of two kinds : leakages and phenomena of induction. Leakages occur rarely, and are easily suppressed when the insulation of the wires is normal, and the distance of the wires is not below 1 metre. Induction may be produced either by continuous current dynamos or by alternating currents.

A vote was given in favour of the adoption of a double wire for urban and inter-urban telephone lines. The President said that another question seemed to arise from the discussion, that of the regulation of industrial circuits. M. Vaschy thought that this question belonged rather to Section II., and M. Raymond was of opinion that it would be unjust to enforce a metallic closing in circuits for lighting or for the conveyance of motive power. Such an arrangement would interfere with the development of electric industry, which is of greater interest than that of the telephone. The section voted for adjournment.

M. d'Infreville called the attention of the section to the following method of utilising lines: He takes two circuits to form of them a new closed circuit, each of the original circuits constituting a wire of the new circuit, and so on. The system has not been applied in the United States.

M. Belugon said that the system in question had been laid before the Consultative Commission of Posts and Telegraphs, but had not yet been tried.

M. Piérard had tried the system with bad results. The current was bifurcated into several wires, having different currents of self-induction, and which consequently affected speech differently.

M. Mercadier described a monotelephone or electromagnetic resounder. Only the first harmonic is allowed to be produced. The diameter of the neutral line is 0.68 x D, D being the diameter of the plate. Three holes are pierced into it, which may be introduced into the rods and fixed at will, and the bobbin may be more or less approximated to the vibrating plates.

SESSION OF THURSDAY, AUGUST 29TH, 1889.

M. Best brought for distribution a treatise on the applications of electricity in Mexico,

[SEPTEMBER 13, 1889.

M. Chaye spoke on the utility of discovering an indi cating relay for submarine applications. The best instrument is the telephone associated with the microphone. The microphone is placed in a watertight metal case or kept suspended in the water. In this manner the sounds of submarine explosions or the sounds of submerged bells might be recognised. ship on her way might thus make known her presence in foggy weather, either by the mere fact of her movement or by means of signals. It would, however, t very useful to have an automatic indicator, which would supersede the necessity of constantly listening t the telephone.

In his experiments the author finds that three or for Leclanché telegraphic elements furnish the most co venient current for the apparatus and the circuits em ployed.

The microphones consist merely of one or tw vibrating pencils of carbon, the variations of intensit of the current being relatively considerable.

To find the direction of the sound the microphoni plate is inclosed in a lead box with thick sides in which there is a window. On turning this box we per ceive a maximum intensity as soon as the box is turned in the right direction.

M. Samuel read his paper on Munier's Multiple Printing Telegraph. M. Munier has come upon a practical arrangement for grouping several Hughes' apparatus in multiplex on the same wire. The number of contacts per apparatus is only seven (two of which go to earth) in place of 28, and the number of emissions per letter never exceeds two. Six letters are printed by a single emission of the current. The speed of the apparatus can be carried to a mean maximum of 15 rotations per minute. Hence with a quadruplex, 6 letters are produced per minute, in place of 180 as in the simple Hughes. M. Munier obtains these result by means of a new principle: the principle of keys fractionation.

M. Dumont described his electric disc. A motor drives a vertical shaft which causes the disc to revolve by means of an endless chain. The mechanism is too complicated to be described here. The disc has been acting successfully at the Rainy station since 1886.

The President introduced the question whether five or three minutes should be taken as the unit of a telephonic conversation?

M. Banneux thought it necessary to distinguish between local conversations and those at great and very great distances. It is impossible to foresee a pruri the value of the words great distance.

Major-General Webber asked if M. Banneux had any statistics on the influence of telephonic lines on the number of telegraphic messages?

M. Banneux, in reply, stated that he could not abso lutely say whether the telephone had or had not injured the telegraph.

Fourth Section.—Electro-Physiology.
The section did not sit.

SEPTEMBER 13, 1889.]

ormulæ of M. Leduc and Mr. Tomlinson gave an ccount of his researches on nickel. The results btained agree, as in the case of bismuth, with this law: he alteration of the resistance is proportional to the quare of the intensity of magnetisation.

M. Huron dwelt on the relation between the variation f the resistance and the mechanical properties.

M. Stoletow summed up the results of his experients on electric discharges caused by radiations actino-electric phenomena), and added some critical emarks on the different hypothesis made to explain hese phenomena.

M. Pilleux announced the following law:

If a thermo-electric couple is working between the emperatures, T and T', if we call k the mean coefficient f variation of electric conductivity of one of two etals between the temperatures, T and T', between which the couple' works, and k' that of the other metal, we have:

being the electromotive force of the couple.

He adduced numerous examples of the verification of his law.

Sir W. Thomson said that if we consider two or three ure metals, as, e.g., iron, copper, silver, and an alloy ike German silver, we do not find this law exact.

M. Marin made a communication on the movement of fluids and the applications to the theory of elecricity.

M. Trouvelot presented photographs of electric sparks. M. Gouré de Villemontée has examined the influence of gas in the equalisers of potential based on the fall of water. By the method of reduction to zero he has letected a very distinct influence of carbonic acid.

Sir W. Thomson said that this effect might be due to à modification of the surface of the metal by the action of the gas, and it is probable that the effect will be much greater with the active gases, oxygen or sulphuretted hydrogen, or by taking nitrogen on the one hand and oxygen or atmospheric air on the other. The polish of the surface must have an action in itself.

M. Drzewiecki showed the advantage of considering lines of force in place of the flow of force, for purposes of instruction.

M. Raverot said that when the absolute systems of the British Association were established they avoided taking force as a fundamental unit. As regards dimensions, there are, on the contrary, the usual dimensions of mechanical force (M, L, T), which serve as the point of departure. This contradiction arises from an investigation published in 1882 by Clausius, who, wishing to rectify the British Association system of electrostatic dimensions in accordance with the views of Ampère and the experiments of Rowland, has given a new system of dimensions. There has been a fourth system pointed out by Helmholtz to complete the series of possible alternatives.

M. Courtois presented photographs relating to the effect of the passage of the current in the wires.

M. Pellat said that to know the absolute precision of the electro-dynamometer it would be necessary to construct a second standard instrument of the kind which he had already produced.

Sir W. Thomson much wished that M. Pellat would make a second instrument to verify the former, and requested that the Congress should give a vote to this effect. This proposal was adopted unanimously.

Second Section.-Industrial Applications. M. A. Potier in the chair.

M. Faure addressed a tribute of gratitude to the memory of Planté, and indicated his conception of a perfect accumulator. It is composed of two nonattackable plates uniformly covered with active matter. The active matter plays here also the part of conductor, and the portion which acts thus is five or six times as large as that which takes or furnishes oxygen. The heat resulting from the oxidation and the sulphatation of the reduced lead gives the measure of the maximum disposable energy, and all that takes place on the peroxidised plate merely diminishes the energy.

303

M. Hillaret insisted on the conicity of the cables of the so-called loop circuits for obtaining a constant difference of potential at the ends of the recipients.

M. Roux, on behalf of M. Raverot, presented a paper on compound machines and variable speed.

The installation of compound dynamos requires a special care by reason of the necessity of making them turn exactly at the speed for which the compoundage has been obtained by the constructor.

A means of utilising a compound dynamo presents a certain practical interest, and consists in a very simple electrical arrangement.

We may conceive naturally that by varying the derived resistance either by hand, or rather by the regulator of the motor, we should realise the regulation of the compound working with the desired precision. We will even add that this is the real method of regulation to be employed, the machine being compensated by construction for the smallest admissible speed.

M. Desroziers asked how the regulation is obtained in the case of a single lamp. He had studied this question, and is in possession of a more complete arrangement which gives the full solution of the problem for considerable variations of speed.

M. Raverot replied that experience had confirmed his theory.

M. Ch. Jacquin, after having rapidly reviewed the representative curves of the working elements of alternating current machines without self-induction and with self-induction, and of transformers without iron, showed the curves relative to a transformer containing iron or the real transformer. These curves give the values of the working elements of a transformer of 7,500 watts-i.e., the intensities, the powers furnished and expended, the yield, &c.—as a function of the time, the resistance and the yield of the secondary circuit. This latter curve is the most interesting, as it permits us to have characteristic curves of a transformer of any system and power whatever.

Prof. Forbes presented his current meter for continuous or alternating currents. The apparatus is composed of three parts-a conductor, a small windmill, and a system of toothed wheels. The conductor consists of two concentric rings connected by fine wires. The mill is entirely of mica. The heat liberated by the current in the conductor gives rise to currents of air by convection, which move the mill. There is no clockwork nor any variable contact in this meter, which is equally fit for continuous and variable currents of any speed. Another type is made with a small weight as motor, which communicates the movement to the apparatus for feeble currents. The type presented works with current of 3 to 36 ampères, and its resistance is 0.01 ohm. The maximum loss in volts is 0.4 volt.

M. E. Hospitalier, referring to alternating current machines, the use of which becomes more and more extended, without insisting on the creation of new names demanded as exact definition of the quantities

and their corresponding units which are often met with in the study of alternating currents.

M. Decaux explained the results obtained by a series of experiments on the action of the electric light upon dyed tissues or upon pictures. The use of luminous electric sources will permit us to determine more easily the resistance of different colours to light.

Mr. Richard Frères exhibited several apparatus, an indicator of speed and various meters.

M. Kapp remarked that all the communications were of great interest, but that the object of the congress is chiefly to discuss on conventions and notations, and urged that the proposals made by M. Hospitalier should be discussed at once.

M. Naze, who had requested to give a communication on the Gaulard transformers, withdrew his demand, and

304

distributed a pamphlet summarising the points he had intended to bring forward.

At the request of M. E. Boistel, the President expressed his opinion on the manner of explaining the reaction of the armature in machines. This reaction is determined by the default of symmetry of the magnetic field with reference to the line of brushes. The magnetic flux across a section of the armature is variable even in a machine at rest with the intensity of the current, I, which traverses this armature for every section other than those which are parallel or perpendicular to the axis of symmetry of the field.

The currents of Foucault and the retardations of magnetisation, or hysteresis, augment the reaction of the armature in the generating machine but diminish it in the recipients.

has not been able to receive a special name.

The following names have been proposed but have not been accepted by a sufficient majority intensity, virtual, equivalent, mean, quadratic, practical, useful, effective, thermic, indicated.

The apparent resistance or fictitious resistance of a circuit is the factor by which we must multiply ✓ (12) mean to have (E) mean.

The President opened the discussion on the names of the poles of an accumulator, and the section resolved that in every accumulator the positive plate is that which is connected to the charging machine and which becomes the positive pole during the discharge.

Third Section.

M. Fribourg in the chair.

The subject of the length of a telephonic conversation was resumed. According to M. de la Touanne, in the circuit Paris-Brussels, there were from May 5th to May 15th, 285 conversations of less than three minutes, and 91 of more than three minutes. Hence he judged that a limit of three minutes is sufficient. It was suggested that conversations between persons at more than 15 kilometres from each other should be considered as interurban, and that the time for a conversation should be five minutes. M. Raymond proposed to reduce the time to three minutes, which was put to the vote and carried. It was also resolved that all telephonic communications should be considered as interurban if

[SEPTEMBER 13, 1889

between subscribers or public offices forming part of different groups.

The President raised the question, should the nowber of successive conversations allowed to one perso be limited? The majority appeared to take the affirms tive view.

Dr. J. Maier proposed the word phone to designate a telephonic message. M. Wenmann proposed the word telepheme, but no resolution was arrived at.

M. Wenmann handed in a paper on the influence of telephony on telegraphic activity. He contended tha it was not sufficient to compare the total number of telegrams during the years before and subsequent t the introduction of the telephone. In a case of the system of annual subscription, the number of telegrams had decreased by 60 to 90 per cent.

M. Baudot gave an account of some recent improve ments in his multiple printing telegraph. His new apparatus is to be seen at the Exhibition.

M. d'Infreville described a system which is patente neither in France nor England, and which gives very good results for annulling the disturbing action of neighbouring lines.

M. Bourdin applies the Gaulard and Gibbs transformers to telephony. He substitutes a small trans former for the bobbin of the microphone. He hope thus to reduce the diameter of the inter-urban telephonic wires.

M. Dumont left at the office an account of his method of sending the time by telegraphic wires.

Fourth Section.-Electro-Physiology.

M. Trouvé presented an interrupter giving the sam intensity to the induced current whatever the number of interruptions per second.

M. Wetzler made a communication on capital punish ment by electricity. The best results are obtained by means of alternating currents.

M. Chavez read a memoir in which he attributed ar important part in electrogenesis to capillary action, the part of which in osmose intervened independently of capillarity.

MM. Courtoy and Lagrange described in a communi cation the phenomena produced in conductors by the passage of the discharges of condensers.

At the general closing meeting held on August 31st. under the presidence of M. Mascart,

All the resolutions voted by the different section: were unanimously adopted, save the one of the thin section which fixed the limit of time for a telephot conversation at 3 minutes, which the Congress con sidered as having too administrative a character to b dealt with in full session.

M. Mascart finally delivered an eloquent address, in which he paid very prominent honour to Sir Willia

Thomson.

ELECTRICAL STANDARDS.

Report of M. PELLAT

(Master of the Conferences at the Faculté des Sciences, Paris) The most convenient method generally for obtaining the absolute value of a certain quantity consists in com paring this quantity to a standard, i.e., to a quantity of the same kind equivalent to a unit or a known multip of the unit. The choice of standards, and the determi nation of their value in absolute units, are consequently of the highest importance in exact measurements. Thus we can easily conceive how a certain conducte at a constant temperature may furnish a standard of resistance, how a certain condenser may furnish a standard of capacity; but there are quantities, such as the intensity of a current or a quantity of electricity, of which it is difficult to have true standards. We get out of the difficulty in this case by means of apparates which always give the same indication when they are traversed by a current of the same intensity, or by the discharge of the same quantity of electricity, and which, as they render the same service as a true standard merit the name of standard apparatus. In electro magnetic measurements (the only ones that it is ex

SEPTEMBER 13, 1889.]

pedient to enter into here) there are only three quanities, the value of which can be obtained directly in absolute units. These are :

1. The coefficients of mutual induction or of selfinduction, the value of which can be obtained by simple measures of length (dimensions [L]).

2. The resistance of a conductor, the value of which may be given by measures of length and measures of time (dimensions [L T-1]).

3. The intensity of a current, the value of which may be given by the measurement of a force and by measurements of ratios of lengths (dimensions [F] = [L* M* T-1]). It is better to call the standards of these three quantities principal standards, since these are the only ones the value of which can conveniently be determined without having recourse to other electrical standards. In contradistinction, we will call derived standards the standards of the other electrical values, which, moreover, practically resolve themselves into standards of electromotive force and capacity.

PRINCIPAL STANDARDS.

1. Standards of Coefficients of Mutual Induction or of Self-Induction.

We will confine ourselves to remarking that it would be advantageous to have standards of coefficients of mutual induction, such standards not being as yet in existence, or, at any rate, not in use in laboratories.

2. Standards of Resistance.

At the International Conference of Electrical Units, in the first session in 1882, it was decided that a standard of resistance should be constructed equivalent to the ohm. Appeal was made to the savants of the various countries represented at the Conference to determine what length must be given to a column of mercury at 0°, having a section of 1 square millimetre, in order to represent the theoretical ohm (10-9 C.G.S. units).

During the second session, in 1884, the Conference, after having taken into consideration the reports made upon this subject,* fixed this length at 106 centimetres (?); the standard ohm is the resistance of a column of mercury having a section of 1 square millimetre, and a length of 106 centimetres, at the temperature of melting ice.

M. Benoit constructed, at the International Office of Weights and Measures, for the Administration of Posts and Telegraphs, four prototype standards in mercury representing the legal ohm.† These standards each consist of a rectilinear tube, having a section of about 1 square millimetre, debouching at its extremities into large vessels, containing, like the tube, very pure mercury; the whole rests upon a plate of brass, which enables the apparatus to be plunged into melting ice, or into a bath of a known temperature. The glass tube was calibrated and gauged with all the exactness that it is possible to attain, and cut to the required length (taking into account the correction due to the expansion of the lines of force in the mercury of the vessels) in order that the standard might represent the legal ohm.

M. Benoit also constructed for the Administration of Posts and Telegraphs a large number of secondary standards formed of a glass tube bent five times, with vertical branches, and terminated by two large tubes forming a funnel, filled like the small tube with pure mercury. The resistance of these secondary standards was obtained by electrical comparisons with the prototype standards. It was these secondary standards that served to measure the metallic resistance of French manufacturers.

In England another method has been adopted. The specific resistance of mercury had been determined as a function of the ohm of the British Association

Proceedings of the International Conference for the Determination of Electrical Units, p. 43.

+Construction of prototype standards of electrical resistance of the Adminstration of Posts and Telegraphs, by Réné Benoit (Gauthier Villars, 1885.)

305

(B.A.U.) by several physicists, amongst others, Lord Rayleigh and Mrs. Sidgwick. From this datum, the committee of the British Association have taken :

1 legal ohm = 1.0112 B.A.U.

The English legal ohm has rather a higher resistance than the French legal ohm; the difference is 0005 according to the comparison made by Mr. R. T. Glazebrook. In a recent study of the specific resistance of mercury, Messrs. Giazebrook and Fitzpatrick+ have compared the prototype standards of the ohm of the British Associations to the resistance of a column of mercury at 0° 1 millimetre in length, and having a section of 1 square millimetre. They found that the resistance of this mercurial column is equal to 95352 B.A.U.; consequently:

We see by this table that the Siemens and Halske standards agree well with other mercurial standards, taking into account the slight error which seems to exist, according to Messrs. Glazebrook and Fitzpatrick's experiments, in the experiments of Lord Rayleigh and Mrs. Sidgwick, which would bring the first figures on the table to a value almost equal to unity.

The average of the figures submitted to the Conference of 1844 for the length of the mercurial column representing the true ohm is 106-04. The figures given by the null method gave an average considerably lower than those furnished by other methods. M. Mascart has shown since that this method admits of several corrections which had not been made, and which tended to raise the value. If the figures obtained by the null method are suppressed, the average is raised to 106-15. A ifting of the methods employed would raise the figures still higher and bring it to 106.25.

Since 1884, a large number of other determinations

"On the Specific Resistance of Mercury," Chemical News, No. 1,172 (May, 1882); Phil. Trans., Part I. (1883). +"A Comparison of the Standard Resistance Coil, &c.," Phil. Mag., 5th series, Vol. XX., p. 343 (1885).

+

"On the Specific Resistance of Mercury," Phil. Trans. Vol. 179 (1888).

§ These standards, to the number of eight, are deposited at the Cavendish Laboratory at Cambridge. The mercurial standards of M. Siemens were first constructed in 1868. In 1881-2, the old standards having nearly all disappeared through accidents, a very careful reconstruction of the mercurial standards was made with rectilinear tubes of glass gauged and calibrated by the Imperial Committee of Weights and Measures, at Berlin. It is these new standards that are here designated by the name of the Siemens and Halske standards.

[

Journal de Physique," 2nd Series, Vol. IV., p. 101 (1885).

of the ohm have almost all concurred in furnishing figures approximating to 106-3 (the number submitted to the Conference of 1884 by Messrs. Mascart, de Nerville and Benoit, by Mr. Glazebrook and by Lord Rayleigh was 106-28). Amongst these experiments we may mention those of Mr. Rowland, made under eminently favourable conditions, and which gave the figures 106 32; those of M. F. Kohlrausch,† 106-32, and the quite recent experiments of M. Dom,‡ 106-24, made by the null method, but with the necessary corrections made; the experiments of M. Wailleumier, § 106-27, made by the excellent modification of the Lorenz method indicated by M. Lippmann.

To sum up, these different measurements concur in showing that by taking 106-3 as the length of the column of mercury representing the theoretical ohm, we shall probably only commit an error of less than oth.

3. Standards of Intensity of Current.

Most frequently the intensity of an electric current is measured into an absolute value by one of the three following methods: 1st, by the tangent galvanomoter ; 2nd, by M. F. Kohlrausch's method; 3rd, by electrodynamometers, and especially balance electro-dynamometers. These last instruments give the intensity of the current by a weight, by the help of the formula i = A √ pg, in which p represents the number of grammes required for the weight (p g) to make equilibrium with the electro-dynamic force, and A a constant depending on the form of the instrument and determined by measures of length; these latter could be made with an arbitrary unit, since only ratios of length enter into the expression of A. Thus the balance electro-dynamometers come under the category of standard apparatus. Moreover, by the simplicity of the measurement, we can, with these instruments, obtain a degree of accuracy higher than that given by the other methods.

Messrs. Joule, Cazin, Mascart, Helmholtz, &c., have constructed balance electro-dynamometers with a view to particular researches ; each of these instruments might have been considered as a prototype standard, but no copy has been made of them.

There are, however, in existence standard apparatus giving in absolute value the intensity of a current. Sir W. Thomson is the inventor of a series of balance electro-dynamometers intended, some for measurements of currents of low intensity (milli-ampère, centiampère), and others for the measurement of currents of mean intensity (deci-ampère, ampère) or currents of high intensity (hecto-ampère, kilo-ampère). These instruments are widely known in England and in a few other countries.

In France, M. Pellat has had a balance electrodynamometer constructed by M. Carpentier, which differs from those employed up to that time in that the small bobbin with vertical axis forms one body with the beam of the balance, and is placed in the midst of a long horizontal bobbin. The small bobbin having only a single layer of fine wire, and being far from the ends of the large bobbin, the irregularities of winding, which are always produced at the ends of a bobbin with several layers of wire, are here without influence. The measurements of length necessary in order to determine the constant, A, of the apparatus were made at the instance of the International Bureau of Weights and Measures. Supposing all the possible errors of the dfferent measurements to be added up numerically, we find that the error in the constant, A, does not exceed 6th. The error in the weights is, moreover, much less.

[SEPTEMBER 13, 1

M. Pellat* has also had copies of his absolute electrodynamometer constructed by M. Carpentier, to whic he has given the name of standard ampere: they only differ from this last instrument in that the mont bobbin bears several layers of wire, and that the fiel bobbin is much shorter. The constant, A, of the standar ampères is determined by comparison with the absolne electro-dynamometer; this determination is made with an exactitude of roth. The constant, A, is, moreover, absolutely independent of the temperature.

DERIVED STANDARDS.

1. Standards of Electromotive Force.

A standard of electromotive force is constituted by an element of a battery. The difference of potentia between the two poles of the battery in an open circus, which measures its electromotive force, is determine! in absolute value by comparison with the difference of potential, e, that exists at the two ends of a know resistance, r, traversed by a current of known intensity, i (e ir). The comparison can be made either by the charge of a condenser or, which is much more exact, by opposing the element of the battery to the electromotive force developed at the two ends of the resistance by the passage of the current, and making this latter vary until the two electromotive forces are equal. Unfortunately, no element of a battery studied by the exact method that we have just indicated pos sesses an electromotive force that is absolutely invariable with the time.

The standards usually employed are the Daniell sulphate of copper element or its various modifications (Callaud battery, Post Office standard, &c.), the Latimer-Clark element of mercurial sulphate, and the Gouy element of binoxide of mercury.

Unless excessive precautions are taken in its construction and preservation,† the Daniell element cannot serve as a standard in measurements of great precision. The electromotive force of various elements constructed in the same manner, but with material derived from different sources, may vary from 100 to 1.14 v at least. And, what is still more serious, the same element, when once mounted, does not maintain an invariable electromotive force owing to the inevitable alterations in the surface of the metals. The best means of getting a Daniell cell with an almost constant electromotive force consists in mounting the element with new solutions before using it, and closing the circuit for a little while, so as to coat the electrode with copper; the circuit must be opened a few minutes before taking the measurement. The Latimer-Clark and Gouy elements, in which pure amalgamated zinc and mercury serve as electrodes, are far preferable, the metals not being altered, or at least the alteration which takes place after a long time on the surface of the zinc not modifying the electromotive force. The LatimerClark element may be constructed in two different ways (1) by employing as electrolyte a paste formed of mercurial sulphate imbibed from a solution of sulphate of zinc; (2) by employing a weak solution (15 p., 100 for example) of sulphate of zinc and depositing mercurial sulphate in a powder on the surface of the mercury.‡

According to the unpublished researches of Messrs. Potier and Pellat, the liquid Latimer-Clark element (1.465 v. on an average) is preferable, for many reasons, to the Latimer-Clark element with the electrolyte in the form of a paste; the electromotive force is less

*"Bulletin de la Société Internationale des Electriciens." Vol. V. (May, 1888).

See on this subject the work of J. A. Fleming, "On the Use of Daniell's Cell as a Standard of Electromotive Force" (Phi Mag., August, 1885.) Notwithstanding the opposite conclusion by the author, Mr. J. A. Fleming's very careful article shows that it is difficult to get with the Daniell cell a good standard of electromotive force.

Instead of depositing mercurial sulphate, we can produce it by the electrolysis of the sulphate of zinc in the element itself, by causing it to be traversed by a current going from the mercury to the zinc. We are thus sure to obtain pure mercurial sulphate. (Unpublished researches of M. Potier.)