When experiment undertakes to deal with such amazing rates of transmission as those of light or electricity, one of two things is indispensable; it must possess the means, either of operating over enormous distances of space, or of measuring excessively small intervals of time. When the propagation of light is under consideration, there is a free choice between the two methods. If we choose the first, which may be called the direct method, astronomy will supply ample spaces, and no extraordinary nicety of measurement in the other element is demanded. But the practicability of the second method, even when the spaces traversed by the light do not exceed the limits of the physical laboratory, has been demonstrated by Fizeau, Foucault and Cornu.

If we turn now from the propagation of light to that of electricity, it is obvious that nothing less than the largest lines of telegraph wire furnish the conditions required by the first method. On the 28th of February, and again on the 7th of March, 1869, the late Professor Winlock, of the Harvard College Observatory, sent electrical signals from Cambridge to San Francisco, and thence by other lines to Canada, and back again to Cambridge, over a loop of wire measuring 7200 miles. This long journey was performed by electricity in about two-thirds of one second; and no small portion of this brief interval was lost in bringing into action the thirteen repeaters which were interpolated into the circuit. In the determination of longitude by telegraphic signals, the transmission time of the signals comes out as an incidental result. When the signals are sent eastward, the apparent difference of longitude exceeds the real difference of longitude by the transmission time. When the signals are sent westward, the apparent is less than the true longitude by the same quantity. The average of the two values is the true difference of longitude, and half the difference of the two values is the transmission time of electricity. For example, in the campaign conducted by officers of the United States Coast Survey, in 1869-70, for the determination of transatlantic longitudes, I obtained the following results. The total transmission time between Brest, France, and Duxbury, Mass., by the way of St. Pierre, was 816 of one second. The total distance by cable is 3329 nautical miles; the distance from Brest to St. Pierre being 2580 nautical miles, and that from St. Pierre to Duxbury 749 nautical miles. When the differences of length, caliber and materials as between the two branches of the cable are all taken into account, I find that the transmission time between Brest and St. Pierre was 639 of a second, and between St. Pierre and Duxbury 177 of a second, so that the two branches were traversed, one at the rate of about 4000 nautical miles a second, the other at the rate of 4230 nautical miles a second. Wheatstone's remarkable experiments on the velocity of friction electricity, first published in 1834, offer an example of the second method of measuring great velocities. In this case, the experiment was made upon a length of only one quarter of a mile; and the exceedingly small fraction of time required by electricity

to traverse this short distance (amounting to only 2000 of one second) became distinctly measurable by the relative displacement which it produced in the images of two sparks, formed in a rapidly revolving mirror. Hence the hasty conclusion was adopted that the velocity of electricity was 288,000 miles per second. The immense discrepancy between this result and those afterwards reached by experiments on land and ocean lines of telegraph could not be overlooked, and an explanation was sought in the different tensions of friction and voltaic electricity. This explanation was unsatisfactory because direct experiments on telegraph wires appeared to indicate that the velocity of electricity was independent of the strength of the battery. The discrepancy itself vanishes, or changes its character, when attention is given to the law that the transmission time of electricity is proportional to the square of the distance. Wheatstone's experiment simply proved that electricity will go through one-quarter of a mile of wire at the rate of 288,000 miles per second, and that it would pass over only 268 miles of similar wire in one second. Now this is a much smaller velocity than is found by experiments on either land or ocean lines of telegraph; the reason being, probably, that in the inferences from Wheatstone's experiment no account has been taken of the intervals of air which separated the different branches of the conducting wire.

The theoretical law, already stated, viz.: that the transmission time increases with the square of the velocity, has been verified experimentally by Gaugain. He used two threads of cotton, each of which was 1.65 meters in length. When tried separately, the transmission time on each was eleven seconds. When they were placed end to end, so as to double the length, the time was fortyfour seconds.

As Wheatstone's experiment on the velocity of electricity has never been repeated, and as direct experiments upon telegraph lines are not numerous and are not likely to be rapidly multipled, and have not been hitherto very harmonious in their results, some other indirect method of conducting the investigation may be found of scientific value. For this purpose, I have availed myself of Lissajous' method of compounding the rectangular vibrations of two tuning forks, and amplifying the resultant motion, by the twice reflected beam of light, which afterwards enters a telescope.

The tuning forks and telescope are permanently fixed to a baseboard, so as to preserve their adjustment. Each tuning fork is provided with an electro-magnet, in order to maintain its vibration for a long time. The tuning forks, when vibrating independently, are nearly in unison, each making about 128 vibrations in one second. When the electro-magnets are brought into action, by a voltaic current circulating continuously through them and a standard tuning fork, furnished with an electro-magnet and a break-circuit attachment, the first two forks are forced into exact unison with the standard, and, therefore, with each other. Under these circumstances, the resultant orbit seen in the telescope is in

variable. If the instrumental corrections for the two electro-magnets are equal, this orbit will be the first of the series for the unison; that is, an oblique straight line. If this is not the case, it will be convenient to make it so, by introducing resistances at the proper place in the circuit. Then, the apparatus is ready to be put to the work of measuring the velocity of electricity. An additional length of resistance coil is introduced, sufficient to change the orbit to some other in the series. The best one to select is the straight line which inclines in the opposite direction. The new orbit proves that one of the forks begins a vibration by half a period behind the other fork; which, in this particular case, is of one second. This fraction of a second is the transmission time for the passage of the current through the additional resistance coil. Unison forks of higher pitch would register smaller fractions of time. So would also forks, in which the ratios of vibration were less simple; but the orbits would be inore complex and could not be observed with the same precision as the straight lines.

I have perfected the apparatus, just described, to such an extent as to feel assured of its adaptation to the purpose which has been specified. But I wish to make a larger number of observations, upon different lengths of resistance and under various combinations, before I give numerical results. I propose, hereafter, to subject in this way to experimental trial, the theoretical law that the transmission time increases with the square of the distance, and that the velocity is inversely as the distance. If this law holds good, the unit time and the unit velocity may be found for a unit distance, or a unit resistance, and then the time and the velocity can be computed for any other distance or resistance. This unit time and unit resistance must be accurately calculated from a combination of all the results of the various experiments. It is also desirable to ascertain the time and velocity for coiled and uncoiled, for naked and covered conductors; as also for air lines and ocean lines. It is to be observed that, in all cases, the time and velocity ascribed to the passages of the electricity apply to that amount of electricity which is required to work the receiving instrument.

SCIENTIFIC INTELLIGENCE.

I. CHEMISTRY AND PHYSICS.

1. Problems in Chemical Dynamics.-In continuing his valuable researches in thermo-chemistry, BERTHELOT has developed some important facts in chemical dynamics. He finds that sodium butyrate when crystallized contains three molecules of water, all of which it loses in a dry vacuum or when heated to 110° C. The last half molecule of water is very persistent; so that by careful management, a definite hydrate of this composition

can be isolated. If now, these salts be dissolved in 120 parts of water at 6°, the anhydrous salt dried at 110° sets free 427 calories, the same salt dried in a vacuum 4.21 calories, the lower hydrate 366 calories, and the ter-hydrate 3:44 calories. Hence (1) the anhydrous salt is identical, however dried; and (2) heat is set free when a salt already abundantly hydrated, is dissolved in water. From the above numbers also, it appears that the union of the half molecule of water (liquid) with one molecule of the anhydrous salt sets free 0.58 calory; while the subsequent union of this with the two and a half other molecules of liquid water, sets free only 0.22 calory. If these values for liquid water be converted into those for water in the solid state by subtracting from them the heat of fusion of water, 0-715 for each half molecule, then the curious fact appears that the union of the first half molecule absorbs 0·135 calory, that of the subsequent two and a half 3:55 calories, while that of the three together is 3:49 calories; or in other words the union of solid water to solid sodium butyrate to form a crystallized hydrate, causes a considerable absorption of heat, contrary to the general fact. Consequently it is clear that the formation of hydrated sodium butyrate at a temperature at which water is liquid, i. e., above zero, must be attended with the evolution of heat, while the same hydrate produced with solid water, below zero, would cause an absorption of heat in its production. Berthelot calls attention to the change of sign in the heatrelations produced by combination at different temperatures as being a fact of the same order as that observed in allotropic elemental changes, such as for example, those of sulphur. The thermic relations then of allotropic changes of state are thus closely approximated to those of a chemical reaction properly so called; the stability of the bodies formed being intimately related to the changes of sign in the heat-relations attending their transformation.--Ann. Chem. Phys., V, vi, 433, Dec., 1875. G. F. B.

2. Action of Light on Silver Bromide, colored and uncolored. -H. VOGEL has given a resumé of the results of his recent experiments upon the chemical action of light upon silver bromide both pure and when mixed with some coloring matter. He finds: (1) that pure silver bromide shows by sufficiently long exposure to a strong light, a sensitiveness even to the ultra-red rays-having obtained plates showing not only the line A but a line beyond this, at a distance equal to that between A and B. Silver chloride is also sensitive as far as A and silver brom-iodide even beyond. (2) To the substances already mentioned, which increase the sensitiveness of silver bromide for the special rays which they absorb, may be added methyl-violet and cyanin, the latter increasing remarkably the sensitiveness for the orange. (3) In place of putting the coloring matter into the collodion as formerly, Vogel now prefers to flow the previously prepared plate with an alcoholic solution of the coloring matter which is then allowed to dry. (4) Experiments are necessary to determine the strength of these alcoholic solutions, since when they are too strong, the light is seri

ously weakened before reaching the collodion. If, however, the prepared plate be exposed to the spectrum from the back side, this difficulty will be avoided. Moreover, in this way the action of imperfectly transparent coloring matters may be tested.-Ber. Berl. Chem. Ges., viii, 1635, Jan., 1876.

G. F. B.

3. Corrosion of Platinum Stills by Sulphuric Acid.-In 1862, SCHEURER-KESTNER communicated to Dr. Hofmann the results obtained by him in concentrating sulphuric acid in stills of platinum, which were published by the latter in his Report. The figures there given having been criticised as exaggerated, the author now publishes further facts upon this question. From 1851 to 1861, 4309 tons of sulphuric acid were concentrated to 66° B. in an alembic, the body of which weighed 40 kilograms. The entire loss of this part of the still was 12295 grams or 2.859 grams for each ton of acid. To destroy the nitrous products which were the cause of this large loss, ammonium sulphate was added in amount just sufficient for the purpose. In 1862, 1843 tons of acid were concentrated in the still, with a loss of 2490 grams; being 1.22 grams of platinum for each ton of acid, a marked decrease. From 1864 to 1875, 17516 tons of acid (of 1000 kilograms each) were concentrated to 66° in a still the body of which weighed 50 kilograms. The acid contained no nitrous compounds, and only sulphurous acid. The loss of the still was 16178 grams, or 0.925 grams to the ton of acid. To produce acid therefore of 66° B. containing 94 per cent H2SO4, there is a loss to the still per ton of acid of one gram when nitrous compounds are absent, and of 2 to 3 grams when they are present. These numbers are much increased however, by carrying the concentration above 66°. In a still weighing 30 kilograms, 180 tons of acid were produced, containing 97-98 per cent real acid. The still lost 1092 grams platinum, or 607 grams per ton of acid. In producing 47 tons of acid of 99 per cent, there was a loss of 8.80 grams platinum per ton of acid. An analysis of the acid itself showed 8.38 grams of platinum to the ton, in solution in it. To the figures here given for the loss of the body of the retort, about 13 per cent should be added for the other parts. It appears then that this loss of platinum in concentrating sulphuric acid is actual, and that it is a chemical not a mechanical one. The use of a platinum-iridium alloy for the stills prevents to a large extent this action, but the brittleness and consequent fragility of the alloy is a serious objection to it.-Bull. Soc. Ch., II, xxiv, 501, Dec.,

1875.

G. F. B.

4. Production of a Secondary Hexyl Alcohol.-OECHSNER DE CONINCK has studied the products obtained by the hydrogenation of ethyl-butyryl, a mixed acetone discovered by M. Friedel among the products of the dry distillation of calcium butyrate. To obtain the ketone, two kilograms of this salt were distilled in portions of 150 grams, and gave 660 grams of distillate, which when fractionated yielded a little butyral, considerable methylbutyral, 80 grams of a limpid highly refractive liquid of a strong