monstrated. It does not appear, however, to equal for elegance of construction or for universality of application the similar set of apparatus devised by the late Professor Willis, of Cambridge, and known by his name. This latter apparatus, by the way, deserves to be much more generally known in this country. As constructed by the Worcester Free Institute, it is invaluable in demonstrating mechanical principles.

Kimball has published an important paper on the variation of friction with velocity, in which, curiously enough, he harmonizes the statements of Morin and Coulomb, that the coefficient of friction does not vary with velocity, with that of Bochet, that it decreases as the velocity increases, and of Hirn, that it increases as the velocity increases, simply by showing that each is true at some given velocity. For very low velocities the coefficient is small; it increases at first rapidly, then slowly, until at a certain rate of speed it reaches a maximum; beyond this point increase of velocity decreases friction. The results of the above-named experimenters are explained by showing that Morin and Coulomb operated at velocities where the coefficient is near a maximum, and so obtained constant results; while Bochet operated at high and Hirn at low velocities. Kimball used in his measurements sliding friction down an inclined plane, sliding friction at uniform velocities on a horizontal plane, friction of belts on the surface of cast-iron pulleys, and friction of wrought-iron journals in boxes or bearings of dif ferent materials. The practical bearings of his results are highly important.

Sir William Thomson has described in Nature the results which he has obtained with his new astronomical clock, devised in 1869 with a view to improve both the compensation for changes of temperature and the form of escapement. The latter is a modified Graham's dead-beat escapement, the escapement-wheel consisting of only one tooth, being simply a piece of fine steel wire attached to a collar fitting loosely upon the shaft, and driven by friction from it, the shaft being connected with a suitable train of wheel-work with uniform motion, moving a trifle faster than the keeping of accurate time requires. To the lower portion of the pendulum-bob two pallets are attached, near the end of

the escapement-wire, so that at each semi-revolution of the shaft the wire, if too fast, will strike the pallet, and be retarded till the pendulum swings clear of it, the motion of the collar being thus governed by the pendulum. In the clock in the author's house, the arc of vibration does not exceed half a centimeter on each side of the vertical. As to the compensation, the zinc and platinum compensation at first adopted have been discarded, and mercury and glass substituted with the most satisfactory results.

Higgs has described a simple motor for preserving a pendulum in vibration during the course of an experiment. The pendulum is suspended through the coil of a Siemens galvanoscope, and automatically so breaks and closes the circuit that the deflection of the needle attached to the suspending rod upward or downward keeps up the motion.

2. Of Liquids.

Amagat has published in full his memoir on the compressibility of liquids, in which he especially considers the effect of temperature and of pressure upon the coefficient of compressibility. The apparatus consisted of a hollow iron rectangular base containing mercury, on the top of which was a pump to give the pressure (the piston being worked by a screw), a manometer closed at top and surrounded by a cylinder containing water, and a piezometer, the latter consisting of a bulb tube to contain the liquid to be examined, the open end being inverted and cemented into an opening in the iron base, and the bulb extending up into a chamber with glass sides, containing water at any desired temperature. The results, so far as the question of temperature is concerned, are in complete accord with theory, the coefficients increasing with the temperature. With regard to the influence of pressure, the author finds that within wide limits of pressure, and quite independently of the variation due to temperature, the coefficient always diminishes as the pressure increases. Thus for ether at 100°, the coefficient from 8 to 14 atmospheres was 0.000560, and at 30 to 36 atmospheres it was 0.000474, while at 13.7° it was 0.000168 and 0.000152 respectively.

Millar has made some experiments on the relative density of liquid and solid iron. He finds that pieces of pig-iron

placed in melted metal at first sink, but in a few seconds rise again and float on the surface. Flat bars of cast iron carefully laid on the surface continue to float. A solid ball 2 inches in diameter, lowered into the metal by a fine wire, disappeared completely at first, but rose in a few seconds and floated, with about half an inch diameter of surface exposed. Since in foundry practice is allowed for linear contraction of cast iron, the author believes that the finally cooled solid is denser than the molten metal; but as the sharpness of iron castings points to an expansion on solidification, he also believes that the contraction in cooling more than counterbalances the expansion during solidification. This view of the case is fully supported by the experiments on floating above described.

Sire has devised a new form of apparatus for demonstrating the hydrostatic paradox of Pascal. It consists as usual of three containing vessels, one cylindrical, the other two conical, the first with its base upward, the second with the base downward; but in the new apparatus the three are cemented at bottom into rings, giving their bases absolutely the same area. Below these rings are three glass cylinders communicating with each other, and filled with mercury. On filling the vessels with water, and opening communication between them to equalize the level in them all, the mercury in all the cylinders below is observed to stand at exactly the same height.

Hasler has proposed a new water meter, the peculiarity of which consists in the mode of counting. Upon the axis of the revolving drum is a steel bar magnet, which revolves close to the partition separating it from the counting wheels. Upon the axis of the lowest of these wheels is a second magnetized steel bar, smaller than the former. This is carried round by the larger bar solely through its magnetic attraction, and so effects the registration.

Trowbridge has made a series of ingenious experiments on vortex-rings in liquids, analogous to the smoke-rings of Thomson and Tait. Applying to this case the general equations of vortex motion, he draws the conclusion not only that all liquids falling upon the free surface of liquids from such a height that the surface of the liquid is not too much disturbed to enable the drop to be acted upon symmetric

ally by the forces at the free surface will form rings, but also that a vortex movement can arise in the process of diffusion by a variation in density and pressure without the aid of initial angular velocities. The apparatus employed to produce the rings consists merely of a small glass tube, slightly smaller at one end, having a bit of cotton wedged in nearer the larger end, over which a piece of rubber tube is slipped. The apparatus being filled by means of the mouth with liquid, it can be ejected in such a way as to form the rings either at or beneath the surface of the liquid.

De Romilly has made some curious experiments on capil lary action. He finds that if a bell-jar be covered at bottom with a cotton netting whose meshes are from one-eighth to one-twelfth inch in diameter, water drawn up into it will remain suspended, a well-pronounced meniscus being observable at each mesh. Moreover, although the strength of capillary attraction diminishes with the temperature, the water in the jar may be boiled by placing a Bunsen burner beneath the netting without falling through it. Special apparatus is needed to maintain the level, which the author figures and describes in his original paper.

De la Grye has studied the changes of form which are produced when two liquids of different densities are superposed and rotated with different velocities. If the more viscous of the two be uppermost, as in the case of oil and water, the oil becomes thinner in the centre, and if a more viscous liquid still, as a solution of gutta-percha in benzene, be used in place of oil, the appearances presented recall remarkably those of sun spots. If, however, the more viscous liquid be below, as, for example, oil and alcohol, the upper layer becomes thicker in the middle. It would hence appear that if the solar spots are formed by centrifugal force, the photospheric layer must have more cohesion than the gaseous substratum beneath it, and than the overlying chromosphere.

3. Of Gases.

Mendelejeff has made an extended investigation into the accuracy of Boyle's law of gaseous compression, special apparatus being used for the purpose, in which all possible causes of error were eliminated, and which allowed the most perfect accuracy of measurement. The experiments were

made at pressures varying from 700 to 2200 millimeters. The results obtained confirmed the conclusions of Regnault, although showing numerical differences in the values obtained, and proving, for instance, that the deviations of air from Boyle's law are even less than appeared before. But the most important result of the researches is that the divergences from Boyle's law, shown by the air being negative at pressures above the mean atmosphere, as was observed by Regnault, proved to be positive (volume decreases slower than pressure increases) at pressures below it. We must, then, conclude that the air experiences a change of compressibility at a certain pressure about the mean of that of the atmosphere; and this conclusion is supported by the circumstance that such a change has been also noticed in carbon dioxide and sulphurous oxide gases, but at pressures far lower than is the case for air. Only for hydrogen does the divergence continue positive for all pressures. Altogether we must conclude that the deviations from Boyle's law are far more complicated than has been suspected.

Romilly has communicated to the French Physical Society the results of his experiments on the use of a jet for aspirating and condensing gases. He finds, 1st, that the jet should be placed at a distance from the receiving tube equal to about four times the diameter of this latter tube; and 2d, that the conical opening of 6° given by Venturi for water is the best angle for air also. The first point is proved by finding that a gasometer is filled in two minutes forty seconds when the jet is placed in the interior of the receiving cone, in eight seconds when it is withdrawn a little from this cone, and in three seconds when it is removed four times the diameter of the cone. If a cone of 8° be substituted for the Venturi cone of 6° in the above experiment, sixteen seconds. is required; and twenty-four seconds is necessary if only an opening in the thin walls of the vessel is used. Moreover, Romilly finds that the maximum effect is not obtained when the jet is central. The point of maximum effect varies with the distance between the jet and cone, the locus of these points constituting an ellipse.

Frankland has presented to the Royal Society a paper on the transport of solid and liquid particles in sewer gasesa subject of great hygienic importance in reference to the

E