duced by a Bourbouze lamp with a refracting system of flint-glass, (2) to study the variations of this spectrum with the temperature of the source, and (3) to observe also the absorption spectra of various bodies and their variations with the temperature of the source. Among other important facts observed, these experiments appear to prove finally a variation in the distribution of heat in the spectrum with the temperature, and also to show that flint-glass becomes less diathermanous as the temperature falls.

Aymonnet has also examined the specific absorbing power of bodies for radiant heat, using a thermo-pile and prism, the solution to be examined being placed between them. From his results he concludes that the atomic absorbing power appears to be constant, first, for all elementary bodies dissolved in the same menstruum, and, second, for all these bodies when existing in compounds of analogous chemical constitution.

The same author has studied calorific spectra by a modification of the ordinary method. He concludes that heat spectra contain easily recognized minima, that these minima are periodic, that they change their position when the source of heat is varied, that these variations are also produced by absorption, and that by these absorption changes much light is thrown upon the mechanism of solution.

The radiometer continues to be the subject of extensive experimentation. Among the papers which have appeared upon it, one of the most noteworthy, perhaps, is that of Mr. Crookes, in Nature, in which he says: "The results I have obtained seem to show conclusively that the true explanation of the action of the radiometer is given by Mr. Johnstone Stoney, according to which the repulsion is due to the internal movements of the molecules of the residual gas." He gives a number of highly interesting experiments with this instrument. Alvergniat seems to have made an experimentum crucis with the radiometer. By making the vanes of aluminum and silver, and by maintaining the globe during exhaustion at 400° C. in the vapor of sulphur, he obtained a vacuum so perfect that there was no rotation. On admitting a trace of air, however, rotation recommenced. Salet has modified the instrument in a very simple way in order to show the correctness of the molecular bombardment theory.

The vanes are fastened immovably to the glass support, and near them moves a disk of mica suspended from its centre. Exposed to light, the disk is caused to rotate rapidly by the molecules projected from the black surface.

Garbe has discussed the radiometer from the standpoint of theory, assuming that in this case the sum of the moments of different points relatively to a given axis is constant. From this three conclusions follow: 1st, the containing envelope being free to move, and the vanes starting from rest, when equilibrium is attained the rotations of the two will be in opposite directions with a velocity inversely as their moments of inertia; 2d, the vanes having a certain initial velocity, if the apparatus be left free, either (a) the globe will revolve in the opposite direction (in case the vanes revolve more rapidly), (b) it will remain stationary (in case the velocity of the vanes remains constant), or (c) it will revolve in the same direction (in case this velocity lessens); 3d, the apparatus being inverted and its parts thus becoming fixed, the action and reaction are equal, and no motion takes place under any conditions.

In a paper read to the Royal Society, Stoney has discussed the method by which heat is transferred across the vacuous spaces in Crookes's radiometers. He considered the laws under which this transfer of heat takes place, and showed that they are different from the already known laws of radiation, convection, conduction, and contact. Hence he suggests that this newly discovered mode of conveying heat should be called penetration. Numerous observations made more than thirty years ago by De la Provostaye and Desains, but not then understood, as well as more recent ones of Dulong and Petit, and of Grove, are readily interpreted by means of these newly discovered laws of heat.

Stoney and Moss have experimented to determine the relation of the force which moves the radiometer-and which they call "Crookes's force"-to the tension of the residual gas, and the influence of variations in the distance between the reacting surfaces. They find that with a residual tension of five millimeters there is a reaction through a space of at least ten millimeters; that at distances of twenty to eighty millimeters the force seemed to vary inversely as the tension; and that it appeared to be nearly independent of the dis

tance when the tension exceeded twenty millimeters. They observed, moreover, sensible deviation from the law of inverse squares at most of the tensions.

Rood has described some very ingenious experiments on the radiometer, which show most conclusively that the theory which supposes the motion to be due to a reaction between the blackened surface of the vanes and the containing envelope is the true one. A two-vane mill with blackened surfaces of aluminum, and carrying a small magnet, was prepared, and before one of these surfaces was placed a screen of mica, also attached to the suspending wire. The whole was placed in a flask, which was exhausted to 0.25 millimeter. Light falling upon the unprotected vane alone, caused a deflection of 3.23°; upon the protected, 0.10°. When it fell on both there was a deflection of 2.38° in favor of the unprotected disk—thus proving that when reaction is prevented between the walls and the vanes no revolution takes place. The author also devised an experiment for measuring this repulsion. Experiments were also made showing that motion. under atmospheric pressure is due to currents.

Volpicelli has also given the results of some radiometric experiments. He finds, for example, that a freezing mixture applied to the upper half of the globe causes a rotation with the non-blackened face foremost, as when radiant heat is used; but when applied to the lower half, the rotation takes place in the inverse direction, the blackened faces being in advance. In the latter case, radiant heat brings the mill to rest. The whole globe being plunged in a hot liquid or in a freezing mixture, there is no motion.

Crookes has given the name Otheoscope to a form of the radiometer in which, regarding the blackened surfaces of the vanes as the heater and the glass the cooler, and deducing from theoretical considerations that the latter rather than the former should be the moving body, a blackened fixed surface is so arranged that the stream of molecules driven from it shall impinge upon the transparent vanes and drive them round. In this instrument, unlike the radiometer, the glass envelope plays no part other than a preserver of the rarefaction. At the Royal Society's May soirée six otheoscopes of different forms and thirteen new radiometers were exhib

LIGHT.

1. Reflection.

Govi has proposed the use of thin layers of gold-leaf for obtaining good transmitting and reflecting surfaces in optical experiments. If upon the oblique face of a right-angled isosceles prism a very thin layer of gold be deposited by means of an alkaline solution of gold chloride and aldehyde, and then the prism be cemented by Canada balsam to a second and similar prism, a cube is obtained containing in its interior a surface of gold inclined at 45° to two opposite surfaces. By means of such a cube two images are seen-one by light transmitted through the film, which is of a pale green color; the other by light reflected from the film, and which is yellow. The cube thus becomes of excellent service as a camera lucida, etc. Govi proposes to place such a cube on the front of the telescope of a cathetometer, and then to compare directly the object to be measured with the equally distant scale by means of the direct image of the one and the reflected image of the other.

Wright has continued his researches on the volatilization of metals by the electric spark in vacuo, and has successfully applied the method to the production of mirrors. The glass to be metalized is placed in an exhausted globe, and a shower of sparks rained upon it from the negative electrode made of the metal to be used until the deposit was sufficiently thick. Platinum appears to be the best metal for specula, a perfect coat being deposited on a plate two centimeters in diameter in twenty to thirty minutes, the vacuum for the purpose being from 1.5 to 1.75 millimeters, and made on hydrogen. This layer was found to be 0.000174 millimeter thick, or one fourth of the length of a wave of red light. The author thinks this process may be brought into general use in the arts, the polish of the metallic surface being exquisite, far surpassing that obtained artificially.

Miller has described a new form of Wollaston's reflecting goniometer, which appears to him to have numerous advantages. There is nothing new in the leading principle involved in the instrument, the changes being mainly in the mechanical construction.

2. Refraction.

Mascart has made a research upon the refractive power of gases. A beam of light was sent through a collimator to two plates of plate-glass connected together at right angles; the halves of the beam were bent right and left by refraction through the glass, then passed parallel through two copper tubes containing the gases, and after refraction by a second system of glass plates placed in reverse directions, the halves were united again, and the beam passed through a slit to a system of prisms, then to a telescope. If the pressure in one of the copper tubes varied, the phases of the two parts of the beam became different, and from the number of fringes displaced the refraction of the gas could be determined. The influence of pressure was examined, then the refractive power for different wave-lengths, then the influence of temperature, and from these data the absolute refractive power was deduced. The figures obtained range from 0.1387 for hydrogen and 0.2706 for oxygen to 0.7036 for sulphurous oxide and 0.8216 for cyanogen. The refraction of a mixture is the sum of the refractions of its components; but that of a compound gas is in general greater than that of a mixture.

De Waha has proposed a new and simple mode of measuring the index of refraction of liquids. In a rectangular glass tank a piece of silvered glass is supported at any convenient angle to one side, this side being also silvered to one half its height. The tank being placed in the centre of a divided circle, a beam of light from a narrow slit is allowed to fall horizontally upon the side of the tank and normal to it, and then upon the piece of silvered glass, the circle being turned until this ray is also normal. The angle read off on the circle is the angle of the prism. The liquid to be examined is then poured into the tank, and the beam of light is so adjusted that its incidence in the liquid upon the silvered glass surface is normal. In this condition of things, the angle of refraction is the angle of the prism. Measuring then directly the angle of incidence, and dividing its sine by that of the angle of the prism, the refractive index is obtained.

Govi has suggested a mode of varying the focus of a microscope without touching the instrument or the object, and