gin of the appearances in all three of the cases is ascribable to the same cause: viz., to a wave-like, vertical motion produced by the air on the surface of nearly still water, combined with feeble horizontal currents (acting in directions nearly at right angles to each other) upon the bottom.

Amherst College, Oct. 15, 1866.

ART. XV.-On a Theory proposed by Fresnel, and on a mode of measuring the average size of very fine particles; by OGDEN N. ROOD, Prof. of Physics in Columbia College.

IF the light from a candle-flame be received on a ground glass surface, so obliquely that the incident ray makes only a very small angle with the glass surface, the light will be copiously reflected, and a bright uncolored image of the flame will be seen by reflexion. As the angle made by the incident ray is increased, the reflected image becomes first yellow, then red, and finally disappears altogether.

Fresnel has attempted to account for this fact, on the ground that the more refrangible rays, having shorter wave-lengths, are caused to interfere by a difference of path, which is still too small to effect complete interference in the case of the longer waves of red light; the difference in path, depending on the depth of the minute scratches on the surface of the glass, and on the angle which the ray makes with this surface.

As it is not difficult to measure approximately the angle at which the red ray ceases to be reflected, it would be easy to put this theory to the test of experiment, if the average depth of the scratches on the ground surface were known.

The impossibility of obtaining such measurements, has hitherto prevented this theory of the action of finely roughened surfaces on light, from being either confirmed or overthrown.

Some time ago, while experimenting on a plane polished surface of glass which had been smoked with lamp-black to complete opacity, I was surprised to find that the lamp-black surface, at a great obliquity, reflected all the rays of light with much brilliancy, so that it resembled in appearance a polished surface of metal or glass. With less degrees of obliquity the reflected light was yellow, red, and finally disappeared altogether.

The lamp-black surface in this experiment was obtained from burning paraffine, and it was found that the red ray ceased to be reflected at an angle of 18°, reckoning from the glass surface. The source of light was a small gas-flame and the experiments

* Pogg. Annalen, Bd. xii, p. 210.

were made in a darkened room at night, the glass plate with its lamp-black surface, being attached to the axis of a graduated circle, the lamp-black having been removed from the upper half of the plate, so as to allow the proper adjustments to be made with the aid of the naked glass surface. I then attempted to measure with the microscope the average size of the smaller and more numerous particles of lamp-black; the result obtained was that they varied in size from 000018 to 000012 of an inch. Several months afterwards, I made a calculation to ascertain what the difference in the path of the interfering rays would be, using these data, and what relation this difference bore to the length of a wave of red light.

Assuming the dimensions of the particles of lamp-black to be the same in all directions, we have the annexed construction. BD will represent the diameter of a particle of lamp-black, the ray AB is reflected from its sur

B

below; X is the angle made by the light with the plate, and the difference in path of the two rays will evidently be equal to CB -AB, a quantity readily found by calculation.

Taking the size of the lamp-black particles to be equal to 000018 of an inch, the difference in path of the two rays for an angle of incidence of 18° is 000011, while the wave-length of the line C in the red space is nearly 000026 of an inch. This shows that the difference in the path is not far from half a wavelength of red light, if the larger of the two estimates of the size of the particles of lamp-black is employed.

I then made a new set of experiments relative to the angle at which the red ray disappears, using as before lamp-black from paraffine. This was found to vary somewhat in different portions of the same plate, as is seen in the table below.

20°.75, 18°.75, 20°, 20° 20°.1

=

21°, New microscopic measurements on the size of the lamp-black particles were made with a different microscope, the value of the micrometer not being known; it was estimated that the size of the smaller and more numerous particles varied from

to

of an inch, but that there were more particles approaching the first number than the second, a circumstance of which I have not taken any advantage in the following calculation.

Taking the mean of these determinations, and combining it with the mean of the first determination, we obtain for the mean size of the particles =0000146 of an inch.

3

AM. JOUR. SCI.-SECOND SERIES, VOL. XLIII, No. 127,-JAN., 1867.

The average angle of disappearance of the red ray being 20°, there results a difference of path =0000998; that is, the difference of path is to the wave-length of red light nearly as 10 to 26.

When the difficulty of obtaining an approximate measurement of the size of the particles of lamp-black is considered, it is surprising to see how nearly the calculated difference in path approximates to half a wave-length of the light in question.

I found that a surface of magnesia produced by smoking a glass plate to opacity with burning magnesium wire, also reflected light in the same way at very oblique incidences. It was ascertained that the final tint was red, and that the red rays themselves disappeared at 11°. The size of the particles of magnesia was estimated at 000036. Long after the measurements had been obtained I calculated the difference of path for the interfering rays; this was found to be,

wave-length of C

⚫000014 ⚫000026

giving a still nearer approximation to a difference of a wavelength.

These experiments then seem to point out the correctness of Fresnel's theory, and we should I think be justified in reversing the process, and using the angle of disappearance of the red ray, in connexion with the known wave-length of this ray, for the purpose of calculating the average size of small particles or the average depth of fine scratches or furrows.

I give below the calculated values of the average size of the particles of lamp-black and magnesia.

Some experiments were made on the angle of disappearance of the red ray with lamp-black produced by the burning of dif ferent substances; where the figures are connected by a bracket it is intended to indicate that the two angles were obtained from the same portion of the plate.

It would appear from these last experiments that the average size of the particles of lamp-black from burning camphor is somewhat greater than from paraffine, while in the case of "burning-fluid" the particles are smaller.

New York, Dec. 4th, 1866.

SCIENTIFIC INTELLIGENCE.

I. CHEMISTRY AND PHYSICS.

1. On a new form of magneto-electric machine.-When the armatures of an ordinary magneto-electric machine with permanent steel magnets are wound with coarse wire, currents of electricity are obtained which are capable of developing magnetism in an electro-magnet. It is easy to see that the magnetism thus developed may in its turn be made to generate a current of electricity, and that this again may induce magnetism in a second and larger electro-magnet, and so on alternately. Mr. H. Wilde has availed himself of this principle to construct magnetoelectric machines of extraordinary power. As the author's descriptions. are not very clear, even with the aid of figures, we shall content ourselves with giving the general construction of the apparatus and the results obtained with a particular machine. In this machine the genera tor or primary source of the electric current was a magneto-electric machine consisting of six small permanent magnets weighing only one pound each and capable of lifting collectively a weight of, at most, 60 ĺbs. The current from this excites an electro-magnet weighing three tons, the total weight being about four and a half tons. The armatures are driven at a uniform velocity of 1500 revolutions per minute by means of a steam-engine and a very strong leather belt. With this machine pieces of iron rod fifteen inches in length and one-fourth of an inch in diameter were melted. With an intensity armature a light was obtained between points of gas carbon sufficiently intense to cast shadows from the flames of street lamps at a distance of a quarter of a mile. It is easy to see that by passing the current derived from the electromagnet through another and larger electro-magnet a vast increase of electric force could be obtained, of course at the expense of a greatly increased motive power. With an unlimited increase of motive power an unlimited increase of electric force could be obtained, as in fact the whole machine is to be regarded as a means for transforming heat into mechanical power, and this last into electricity. It is to be regretted that the author has given no precise data from which the amount of electricity set free can be determined with precision. The quantity of water decomposed per minute, with the expenditure of a measured amount of mechanical work, is what we require in order to form a correct estimate of the value of the apparatus, as compared with that of other electro-motors. In any case, however, it is safe to predict a brilliant and useful future for the new apparatus.-Proc. of Royal Society, XV, 107.

W. G.

2. On the synthesis of chlorid of thioxyl.-WURTZ has made the very interesting observation that chlorid of thioxyl, S2O2Cl2 or SOCI,, may be produced by the direct union of anhydrous hypochlorous acid with sulphur, Cl2O+S=SOCI2.

2

The vapor of hypochlorous acid is passed into chlorid of sulphur holding sulphur in suspension, and the operation is discontinued as soon as the sulphur has entirely disappeared. The chlorid of thioxyl may then be separated from the chlorid of sulphur by distillation. Chlorid of thioxyl

as thus prepared is a colorless liquid which has a penetrating odor reminding of sulphurous acid and chlorid of sulphur. Its density at 0° is 1675, and its boiling point 78° at 746 mm. Water resolves it into chlorhydric and sulphurous acids, SOCI2+H2O=2HCI+S02.

Liquid hypochlorous acid explodes on contact with sulphur, and it is for this reason that the action must be moderated by suspending the sulphur in chlorid of sulphur and keeping this at a temperature of -12° C. From the above it is clear that Cl2 may unite directly with a radical, a fact which stands related to the observation of Carius that HCIO unites directly with certain hydrocarbons.- Comptes Rendus, lxii, 460.

2

W. G.

3. On a new series of hydrocarbons.-SCHORLEMMER has discovered among the products of the distillation of cannel coal, besides the homologues of marsh gas and benzol, other hydrocarbons attacked by concentrated sulphuric acid. When the oil, after treatment with the acid, is distilled, the oils of the benzol and marsh gas series pass over first and there remains a black tarry mass. If this mass be distilled, a thick brown liquid with an offensive smell passes over between 300° and 400°. By repeated distillations with caustic alkali and with sodium a series of carburets may be obtained with the general formula (CH2-2)2; of these the author describes €12H20, €14H24, €16H28 These are all colorless oily highly refractive liquids, having a faint peculiar smell resembling that of the carrot or parsnip root. These oils unite with bromine to form colorless heavy liquids easily decomposed by heating. A molecule of oil takes up two atoms of bromine. Strong nitric acid dissolves the oils, forming nitro-compounds which with tin and chlorhydrie acid give organic bases. With sulphuric acid and bichromate of potash the oils yield carbonic, formic, acetic, and perhaps other acids. The author considers it certain that the oils of this series are polymers of the acetylene series.-Annalen der Chemie und Pharm., cxxxix, 244. W. G.

4. On the compounds of tantalum.-MARIGNAC has published the conclusion of his researches on niobium and tantalum, the first part of which has already been noticed in this Journal. To determine the atomic weight of tantalum, pure crystallized fluotantalate of potassium, KF. TaF was treated with concentrated sulphuric acid and carried finally to a temperature of 400° C. On boiling with water, bisulphate of potash is dissolved out and sulphate of tantalum left in small granular crystals, which by strong ignition yield tantalic acid. The bisulphate of potash is brought by evaporation and ignition to the state of sulphate and weighed as such. Four analyses closely agreeing with each other gave the number 182.3 as the atomic weight of tautalum; a molecule of tantalic acid has therefore the formula Ta25, and the molecular weight 444-6. The analysis of the fluotantalate of ammouium leads to the number 182 as the atomic weight of tantalum, and this number is adopted by Marignac as most probable. The author remarks that the difference between the atomic weights of niobium and tantalum, which belong to the same natural family, is the same as that between the atomic weights of the closely allied metals, tungsten and molybdenum, namely, 88. Tantalic acid forms two classes of salts, in one of which it is monobasic and in the other quadribasic. The first class have the formula