E [Voltage, or Electromotive force], identical with Potential on Electricity.

Absolute Units of Measurement. By W. MOON.

The disadvantages of the C.G.S. system of units are that the units are so small that one can form no conception of their value, and that owing to this smallness it is necessary to introduce a separate set of practical units for ordinary purposes of measurement. These disadvantages may be overcome by taking as fundamental units L and M larger or T smaller.

Of all the systems of metrical units that can be formed by varying L and T by multiples or sub-multiples of ten, that system is the best that is founded upon the 'Decimetre, Kilogramme, and Decisecond.' If the name 'Instant' were given to the Decisecond, the system could be spoken of as the D.K.I. system.

In the D.K.I. system'g'=981, or nearly unity, so that the weight of a kilogramme could be taken as the unit of force for rough calculations. This would be a great improvement, since the simplest way to conceive a force is as the weight of unit mass.

A prepared table shows that all the D.K.I. units are sufficiently large to be used for practical purposes, and that all the multiples and sub-multiples of them that would be required could be expressed by the usual Greek prefixes to their

names.

To express the large numbers required for insulation, resistances, and the small capacities of condensers, the Greek prefixes' omega' and 'omicron' might be used for powers of 10 and 10-1 respectively.

WEDNESDAY, AUGUST 26.

The following Papers and Report were read :

1. On the Measurement of Lenses.

By Professor SILVANUS P. THOMPSON, F.L.S.

The author described his focometer and some results obtained upon microscopic objectives and camera lenses. Wide-angled lenses were found in all cases to have the positions of principal planes inverted.

2. On a New Polariser. By Professor SILVANUS P. THOMPSON, F.R.S.

3. Some Experiments on a new Method for the Determination of 'v.' By A. G. WEBster.

The method is similar to that proposed by Maxwell with the title, Measurement of a resistance in Electrostatic Measure.' A condenser is connected in parallel with the two sets of quadrants of an electrometer, and both are connected in series with a battery and a high inductionless resistance. Contact being made and broken after a short time t by means of a Helmholtz pendulum-interrupter, the potential of the charge of the condenser and electrometer, measured by the first swing of the latter, is

P being the E.M.F. of the battery used, w the large resistance, c and y the respective capacities of the condenser and electrometer. A second set of experiments, c being disconnected, gave the value of y, which included the capacity of the leading wires and of an auxiliary condenser inserted for the purpose of making c+y and y more nearly equal.

By turning the micrometer screw of the pendulum-interruptor, and thus changing the distance between the contacts, t can be varied and a large number of points on the logarithmic curve found. By a process of calibration with the pendulum, it was found that the electrometer-throws were strictly proportional to the poten tials Po This calibration was made by taking a w so small that the exponential term vanished and measuring p for various po's.

The resistances used were made by ruling pencil lines upon finely-ground glass, upon the ends of which a thin film of platinum had been firmly deposited by burning in and soldering the connecting wires to this, giving a firm and reliable connection to the resistance.

The condenser was a large plate-condenser, 50 cm. in diameter, whose capacity was found by Kirchhoff's formula. Three capacities, 350-204, 770-513, and 998-459 cm. were used, with resistances of from one-half to five megohms.

The time-constant of the pendulum was found by the method of Pouillet for short intervals, by means of a ballistic galvanometer. One division of the micrometer was found to correspond to

1.1346 × 10-6 sec.

In the experiments, readings were taken at intervals between 100 and 2,000 micrometer-divisions.

A large number of observations was taken, in which all the measured quantities c, w, and t were varied.

The value of v arrived at was

2.987 x 1010 cm. sec-1

4. On the Magnetic Field in the neighbourhood of the South London Electrical Railway. By Professor W. E. AYRTON, F.R.S., and Professor RÜCKER, F.R.S.

Observations were made by means of a mirror galvanometer the period of which was 10 sec., and which was used as a magnetometer. The instrument was placed in two rooms about 70 and 180 feet respectively from the centre of the road, under which the railway runs at a depth of about 70 feet. It is believed that the earth was acting as the return portion of the circuit. In accord with this the instrument was found to be in continual vibration. The amplitude of the swing at the station nearer to the railway was often 50 mm., and the law of decrease with distance appeared to be inversely as the first power. It is, therefore, evident that experiments of the most ordinary accuracy could not be made within a very great distance of such a railway.

5. On the Periodic Time of Tuning-Forks maintained in Vibration Electrically. By Professor J. VIRIAMU JONES and T. HARRISON.

6. Magnetic Experiments made in Connection with the Determination of the Rate of Propagation of Magnetisation in Iron. By F. T. TROUTON.

7. On the Connection between the Crystal Form and the Chemical Composition of Bodies. The Symmetry of Crystals accounted for by the Application of Boscovich's Theory of Atoms to the Atoms of the Chemist. By WILLIAM BARLOW, F.G.S.

After mentioning that he read papers on the same subject at the meetings of the British Association at Aberdeen in 1885 and Leeds in 1890, the author states that he is now prepared to deal with the matter in a more general way, and to submit proof that the mutual interaction of different kinds of atoms present in

simple proportions is competent to produce the various kinds of symmetry exhibited by crystals if the fundamental doctrine of Boscovich is admitted-that the ultimate atoms are points endowed each with inertia, and with mutual attractions or repulsions dependent on mutual distances repulsion manifesting itself at the smallest distances and becoming infinite at infinitely small distances.

After referring to the principal views which have been put forward as to the nature of the molecules or units of crystals he goes on to argue that stable equilibrium of a group of atoms endowed with Boscovich's properties is evidently found in that disposition of the atoms which gives the repulsions greatest play; that it is, in fact, the arrangement in which the packing is closest, or, in the language of modern conceptions, the arrangement in which the potential energy of the system is a minimum.

He then proceeds to answer the question, What grouping of a concourse of atoms will give closest packing? first pointing out that the answer depends on whether the atoms are of different kinds, and, if they are, on the numerical proportion of each kind present, and also on the relative magnitude of the spaces they occupy, or, in other words, on their relative capacities for repelling or being repelled.

For simplicity sake, he takes first the imaginary case of atoms confined to the same plane, and points out that if there are two kinds of atoms present in equal numbers, one of which exercises a feebler repulsion than the other, their repulsions may be so proportioned that closest packing will be attained when one kind of atom lies at the angles of a system of equal squares fitted close together, the other at the centres of the same squares.

He then applies similar reasoning to cases of atoms not in the same plane, and, after remarking that atoms which are all of one kind will pack closest when their centres have the relative situation of the centres of a close-packed assemblage of equal globes-a familiar example of which is found in the stacking of cannonshot-he states that the more general case of the closest packing of two or more kinds of atoms is approximately depicted by the closest packing of globes, if the globes are of different sizes, to represent the effects of the difference in the repulsions exercised by the different atoms.

After saying that the nature of the grouping in which stable equilibrium is found will depend on the ratio between the lengths of the radii of the globes employed, the author traces the nature of the grouping for several particular values of this ratio.

He points out that not only holohedral groupings corresponding to the simpler forms of the crystallographic systems can be obtained in this way, but that the more complicated partial symmetry of hemihedral and tetartohedral forms are

also to be obtained.

As examples of the latter he gives a grouping in closest-packing that has the precise symmetry of zinc-blende ZnS, which, according to Groth, crystallises in the tetraedrische hemiedrie of the cubic system, and another grouping that has the precise symmetry of cuprite Cu2O, which, according to Groth, crystallises in the plagiedrische hemiedrie of the cubic system. The numerical proportion of the spheres of different radius employed is, in each case, that of the atoms present in the molecule of the compound represented.

Polar-pyroelectric phenomena and circular polarisation are, the author points out, associated with peculiarities of the internal symmetry of the groupings, which correspond in outward symmetry with the bodies displaying these phenomena.

The grouping is portrayed by beads of different colours suspended in space in the symmetrical manner requisite in each case.

The author concludes his paper by referring to some geometrical properties of the symmetrical systems of the crystallographer which he has discovered by an extension of the methods adopted by Bravais and by Sohncke, and which have greatly facilitated his work in finding symmetrical groupings to fit the forms and composition of a variety of different substances.

8. Report of the Committee on the Volcanic and Seismological Phenomena of Japan.-See Reports, p. 123.

9. On Phenomena which might be Observable if the Hypothesis that Earthquakes are connected with Electrical Phenomena be entertained. By Professor JOHN MILNE, F.R.S.

It seems reasonable to assume that superheated high pressure steam escaping at a volcanic focus A through fissures to a region B-A and B being more or less insulated by partially non-conducting material-might also result in the development of large quantities of electricity, followed ultimately by violent discharges.

If a conductor C electrically connected with the surface, say the ocean, is separated from B by partially non-conducting matter I, then BIC may be regarded as a condenser, and the charges at B and C are intensified. Discharges might also take place between B and C, and the charges at A, B and C would act inductively at the points a, b, c, of the surface respectively nearest to them.

The phenomena related to the above hypothesis are as follows:

1. Earthquakes and Earth-currents.-From a comparison of observations at 700 stations in Japan, there seems to be no connection between earthquakes and abnormal disturbances on land lines. It would, however, seem that any subterranean discharge-as, for instance, between A and B-must produce simultaneous change of potential at a and b, and that therefore no change of current should be expected. Thus the hypothesis is not opposed to the facts.

2. Connection between Earthquakes and Volcanoes.-Most earthquakes do not originate at volcanoes, but below the sea or on the coast-line where flat ground suddenly slopes down below a deep ocean. From the hypothesis we should expect that the greatest electric stress, and therefore the greatest disruptive stress, would be between B and C.

3. Potential at Hot Springs.-Measurements made at seven springs in the same valley extending from Yumoto, 100 to 200 feet above sea-level, to Ashinoyu 3,000 feet above sea-level. Between a hot spring and the earth, 10 to 100 yards distant, the difference at the foot of the valley is about 05 of a volt, but at high elevations, where the water is most sulphurous, the difference rose to 0.6 volt.

The difference of potential between 6 and a neighbouring point ought to increase when b is a volcanic vent. The existence of sulphurous water, however, is not to be overlooked.

4. Variations in Potential between Water-Bearing Strata and the Superincumbent Surface. For more than 100 days a continuous photographic record was taken of the difference of the potentials of the water in a well 30 feet deep, and of a point on the surface of the earth 25 yards off. At the time of three small earthquakes deflections equivalent to 2 or 3 volts were observed, but they may have been due to mechanical disturbance.

10. Experimental Study of a Curious Movement of Ovoids and Ellipsoids. By Professor LECONTE.

11. On Vowel Sounds. By Dr. R. J. LLOYD.-See p. 796.

12. A Latent Characteristic of Aluminium. By Dr. A. SPRINger.

According to the author's investigations aluminium is remarkably adapted for use in the construction of sound-boards by possessing an elasticity capable of sympathetic vibration uniformly through a wide range of tone-pitch, and by the absence of higher partial tones during vibration.