On inspecting the above Table, it will be seen that the thermal effect produced at slow velocities was five times as great as with the bare bulb. This increase is evidently due to friction. In fact, as one layer of wire was employed, and the coils were not so close as to prevent the access of air between them, the surface must have been about four times as great as that of the uncovered bulb. At high velocities, it is probable that a cushion of air which has not time to escape past resisting obstacles makes the actual friction almost independent of variations of surface, which leave the magnitude of the body unaltered. In conformity with this observation, it will be seen that at high velocities the thermal effect was nearly reduced to the quantity observed with the uncovered bulb. Similar remarks apply to the following results obtained after wrapping round the bulb a fine spiral of thin brass wire. Table VII.—Bulb wrapped with a spiral of fine brass wire. The thermal effects on different sides of a sphere moving through air, have been investigated by us experimentally by whirling a thin glass globe of 3.58 inches diameter along with the smaller thermometer, the bulb of which was placed successively in three positions, viz. in front, at one side, and in the rear. In each situation it was placed as near the glass globe as possible without actually touching it. thermal effect must necessarily suffer diminution; unless indeed it gives rise to increased resistance. TABLE VIII.-Smaller Thermometer whirled along with glass globe. The effects of fluid friction are strikingly evident in the above results, particularly at the slow velocities of 3 and 7 feet per second. It is clear from these, that the air, after coming in contact with the front of the globe, traverses with friction the equatorial parts, giving out an accumulating thermal effect, a part of which is carried round to the after pole. At higher velocities the effects of friction seem rapidly to diminish, so that at velocities between 23 and 38 feet per second, the mean indication of thermometers placed all round the globe would be nearly constant. Our anticipation (written before these latter experiments were made), that a complete verification of the theory propounded at the commencement was impossible with our present means, is thus completely justified. It may be proper to observe, that in the form of experiment hitherto adopted by us, the results are probably, to a trifling extent, influenced by the vortex of air occasioned by the circular motion. We have on several occasions noticed the effect of sudden changes in the force of wind on the temperature of a thermometer held in it. Sometimes the thermometer was observed to rise, at other times to fall, when a gust came suddenly on. When a rise occurred, it was seldom equivalent to the effect, as ascertained by the foregoing experiments, due to the increased velocity of the air. Hence we draw the conclusion, that the actual temperature of a gust of wind is lower than that of the subsequent lull. This is probably owing to the air in the latter case having had its vis viva converted into heat by collision with material objects. In fact we find that in sheltered situations, such for instance as one or two inches above a wall opposite to the wind, the thermometer indicates a higher temperature than it does when exposed to the blast. The question, which is one of great interest for meteorological science, has hitherto been only partially discussed by us, and for its complete solution will require a careful estimate of the temperature of the earth's surface, of the effects of radiation, &c., and also a knowledge of the causes of gusts in different winds. XXII. "On the Thermal Effects of Longitudinal Compression of Solids." By J. P. JOULE, Esq., F.R.S.; and "On the " In the further prosecution of the experiments of which an outline was given in the Proceedings' for January 29, 1857, the author has verified the theory of Professor Thomson, as applied to the thermal effects of laying weights on and taking them off metallic pillars and cylinders of vulcanized india-rubber. Heat is evolved by compression, and absorbed on removing the compressing force in every substance yet experimented on. In the case of metals, the results agree very closely with the formula in which e, the longitudinal expansion by heat under pressure, is considered the same as the expansion without pressure. It was observed, however, that all the experimental results were a little in excess of the theoretical, and it became therefore important to inquire whether the force of elasticity in metals is impaired by heat. In the first arrangements for this purpose, the actual expansion of the bars employed in the experiments was ascertained by a micrometric apparatus,-1st, when there was no tensile force, and 2nd, when a weight of 700 lbs. was hung to the extremity of the quarter-inch rods. The results, reliable to less than one-hundredth of their whole value, did not exhibit any notable effect of tensile force on the coefficient of expansion by heat. An experiment susceptible of greater delicacy was now tried. Steel wire of th of an inch in diameter was wound upon a rod of iron of an inch in diameter. This was heated to redness. Then, after plunging in cold water, the spiral was slipped off. The number of convolutions of the spiral was 420, and its weight 58 grains. Its length, when suspended from one end, was 6.35 inches, but on adding to the extremity a weight of 129 grains, it stretched without sensible set to 14.55 inches. The temperature of the spiral thus stretched was raised or lowered at pleasure by putting it in, or removing it out of an oven. After several experiments it was found that between the limits of temperature 84° and 280° Fahr., each degree Centigrade of rising temperature caused the spiral to lengthen as much as 00337 of an inch, and that a contraction of equal amount took place with each degree Centigrade of descending temperature. Hence, as Mr. James Thomson has shown that the pulling out of a spiral is equivalent to twisting a wire, it follows that the force of torsion in steel wire is decreased 00041 by each degree of temperature. An equally decisive result was obtained with copper wire, of which an elastic spiral was formed by stretching out a piece of soft wire, and then rolling it on a rod of an inch in diameter. The spiral thus formed consisted of 235 turns of wire, of an inch in diameter, weighing altogether 230 grains. Unstretched, it measured 6.7 inches, but with a weight of 1251 grains attached to it, it stretched, without set, to 10:05 inches. Experiments made with it showed an elongation of 00157 of an inch for each degree Centigrade of elevation of temperature, and an equal shortening on lowering the temperature. The diminution of the force of torsion was in this case 00047 per degree Centigrade*. * Since writing the above, I have become acquainted with M. Kupffer's researches on the influence of temperature on the elasticity of metals (CompteRendu Annuel, St. Petersburgh, 1856). He finds by his method of twisting and transverse oscillations, that the decrease of elasticity for steel and copper is 000471 and 000478. Very careful experiments recently made by Prof. Thomson, indicate a slight increase of expansibility by heat in wires placed under tension.-August 1. J. P. J. Professor Thomson has obligingly furnished me with the following investigation : On the Alterations of Temperature accompanying Changes of Pressure in Fluids. Let a mass of fluid, given at a temperature t and under a pressure p, be subjected to the following cycle of four operations in order. (1) The fluid being protected against gain or loss of heat, let the pressure on it be increased from p to p+w. (2) Let heat be added, and the pressure of the fluid maintained constant at p+w, till its temperature rises by dt. (3) The fluid being again protected against gain or loss of heat, let its pressure be reduced from p+w to p. (4) Let heat be abstracted, and the pressure maintained at p, till the temperature sinks to t again. At the end of this cycle of operations, the fluid is again in the same physical condition as it was at the beginning, but, as is shown by the following considerations, à certain transformation of heat into work or the reverse has been effected by means of it. In two of these four operations the fluid increases in bulk, and in the other two it contracts to an equal extent. If the pressure were uniform during them all, there would be neither gain nor loss of work; but inasmuch as the pressure is greater by w during operation (2) than during operation (4), and rises during (1) by the same amount as it falls during (3), there will, on the whole, be an amount of work equal to a de, done by the fluid in expanding, over and above that which is spent on it by pressure from without while it is contracting, if de denote a certain augmentation of volume which, when a and dt are infinitely small, is infinitely nearly equal to the expansion of the fluid during operation (2), or its contraction during operation (4). Hence, considering the bulk of the fluid primitively operated on as unity, if we take dv =e, dt to denote an average coefficient of expansion of the fluid under constant pressure of from p to p+w, or simply its coefficient of |