tallic acid-feebler in its relations as an acid than arsenic acid, and volatilizing at a lower temperature than the latter. Arsenious acid, moreover, volatilizes at a temperature below that required by metallic arsenic.

In the antimony column, the oxyd SbO, is usually viewed as a weak base, but seems also to be capable of uniting as a feeble acid to the alkalies, and even of expelling carbonic acid from their carbonates (Liebig). The iso-dimorphism of SbO, and AsO, is well established. SbO, is volatile at quite a moderate temperature, while metallic antimony requires at least a white heat to vaporize it. SbO, is a body of distinctly acid properties. Both SbO, and SbO, are converted by heating in the air into SbO, -the so-called antimonious acid-which seems therefore to be the most stable oxyd when strong bases and acids are not present. It is most probable that, as Frémy maintains, Sbo, is not itself an acid, but that a so-called alkaline antimonite is in fact a mere mixture of an antimoniate with the compound of antimonic oxyd and alkali (2SbO,=SbO2+SbO ̧).

In the bismuth column, the teroxyd is homologous as a base with teroxyd of antimony, but shows little tendency to play the part of an acid with even the strongest bases. This oxyd and the metal itself are volatile at high temperatures. BiŎ, also seems to be devoid of acid properties, but the compound Bio, probably exists, and is homologous with SbO,, forming alkaline salts of little stability.

Comparing now ruthenium and osmium with the above recognized members of the arsenic group, we find first that both metals form protoxyds, which are feeble bases, as are probably the corresponding compounds of the other members of the group. We next meet with the sesquioxyds, whose formula is exceptional in the series, but for neither metal has this grade of oxydation been obtained in the free state and pure, and in the case of osmium its existence may be gravely doubted. Anhydrous Ru,O, is supposed by Claus to be formed during the roasting of metallic ruthenium in the air at a high temperature, but only on the ground that the absorption of oxygen slackens when about enough has been taken up to form this compound, and that the proportion necessary for the binoxyd is never fully attained. Claus, however, describes a sesquichlorid with which double salts are formed by the chlorids of potassium and ammonium, and we must therefore assume a sesquioxyd also. Sesquioxyd of osmium is quite unknown in the separate state, and the belief in its existence is founded solely upon the preparation by Berzelius of a dark brown substance, supposed to consist of the sesquioxyd united to ammonia, which, dissolved in hydrochloric acid, yields a brown compound, supposed to be the sesquichlorid of osmium and ammonium. Neither of these, however,

can be crystallized, nor has the constitution assigned to either been supported by an analysis. The so-called ainmonio-sesquioxyd detonates when heated, (sometimes with much violence, as I have noticed in removing by heat the deposit of this substance which forms on the end of a retort-neck during the distillation of osmic acid into a receiver containing ammonia,) and hence is probably analogous to fulminating platinum, containing perhaps the binoxyd of osmium. The binoxyd itself is a feeble base, the characteristic color of whose salts in solution is yellow as is the case with the corresponding compounds of iridium. Similar remarks apply to the binoxyd of ruthenium-probably the body obtained, as we have shown, by Frémy in crystals. The teroxyd of osmium is the body supposed to have been isolated in the experiment described at the beginning of this paper. Its position as a feeble acid, capable, however, under some circumstances, of playing the part of a base, its fusibility and volatility (greater, apparently, than those of osmic acid, as nitrous acid is more fusible and volatile than hypo-nitric), its probable crystallization in octahedrons of the regular system, in which arsenious acid and teroxyd of antimony are also found, all tend to indicate homology with the other teroxyds of the arsenic group. The general relations of ruthenic acid, so far as these are known, place it in a similar position. Just as we find hyponitric acid (NO) and antimonious acid (SbO,) to be the most stable of the higher oxyds of nitrogen and antimony, so the well known osmic acid (OsO) seems to be the grade of oxydation which osmium. most readily assumes and retains when not in contact with bases. OSO, and ŎSO, (the latter as described by Frémy) seem scarcely capable of existing in the separate state; when set free from their salts they soon pass into OsO,; while it may as well be doubted that the latter ever exists as a distinct acid in combination with bases as that NO, or SbO does so. No so-called osmiate has ever been analyzed; the saturating capacity of the acid, if it be such, is unknown; when free and in solution in water it has no acid reaction, it does not displace carbonic acid from the carbonates, and it is itself expelled by heat from most of its supposed compounds, and is separated in part by water even from potash and soda. No compound of OsO, with a base has been obtained in crystals, while Frémy states that he has crystallized the alkaline salts of both OsO, and OsO,. RuO, and RuO, are as yet unknown.

3

The tendency throughout the whole arsenic group is manifestly to the production of the acid compounds MO, and MO,, the former the more fusible and volatile body, the latter the stronger acid. In addition we have some cases of the protoxyd (MO), a feeble base, and the binoxyd (MO,), a body of still more feebly basic properties, verging upon the acids. All other grades of

oxydation, so far as they exist at all, may perhaps be correctly viewed as compounds of the preceding inter se. The stability of the oxyd (MO) in the separate state is remarkable-its formula

is one of rare occurrence.

The affinity of all the elements of the group for oxygen is considerable; it is so even in the case of osmium and ruthenium, usually placed among the noble metals. Dumas (Traité de Chim. app.) states that osmium does not oxydize at common temperatures, nor even at 100° C., but I have obtained conclusive evidence that oxydation may go on slowly even at the ordinary atmospheric temperature. The paper label and the cork of a tube containing pure metallic osmium have in the course of several years become blackened, precisely as organic matter is by the fumes of osmic acid, the black tint on the paper decreasing from the mouth of the tube along the outside. A piece of white paper in which some black platinum residue had been wrapped, was strongly stained in the immediate neighborhood of the powder in the course of a few weeks. The same effect is distinctly observable even upon the paper label placed inside a tube of native iridosmine (Siberian) in the usual coarse grains-a specimen which has lain among other minerals, and has never been placed near any artificial preparations of osmium. Osmium, like arsenic and antimony, is clearly capable of slowly taking up oxygen at common temperatures. At a red heat, roasting in a current of air affords, as is known, a good method of obtaining osmic acid from the iridosmine of platinum residues-just as by similar roasting arsenious acid is prepared from the native arseniurets.

It would be a matter of much interest to compare osmium with its supposed homologues under circumstances in which we should expect it to play an electro-negative part. Frémy has announced his belief in the existence of an osmiuretted hydrogen, but such a body has not yet been isolated and described. Compounds of the metal with ethyl, methyl, &c., would be well worth examination, and it is not unlikely that such might be prepared from a body which in some states of combination exhibits such a high degree of volatility.

The earlier experiments of Deville and Debray upon the platinum metals seemed to have shown that both osmium and ruthenium could be volatilized, at exceedingly high temperatures, without previous fusion; if this were confirmed, a strong point of resemblance with arsenic would be made out, but it appears from a more recent paper that osmium at least may be fused and obtained as a perfectly compact mass, the apparent volatility of the metal being due doubtless to previous oxydation, the crucibles used being permeable to air. We have seen, as regards arsenic and antimony, that their oxyds are more volatile than the metals themselves.

It is lately stated that osmium may be obtained in crystals by the same means as those used for boron and silicon, but I have as yet seen no account of the form which it assumes.

Deville has furnished another interesting fact with respect to osmium, by determining the density of the vapor of osmic acid, which he has found 8.88. This, if we take the generally received atomic weight for osmium, gives the atomic volume

=

131.6 =14-82, indicating a condensation to 2 vols. If we now

8.88

calculate back to the theoretical atomic weight we get (14.57x 8.88)-32-97-38, a number closely approaching 97, which, as we have seen, brings the equivalent of osmium into simple and harmonious relation with those of the other elements of the arsenic group.

The specific gravity of fused metallic osmium having been lately determined by Deville =214, there can be little doubt that all the metals of the platinum family possess the same atomic volume when in the free state, about 46 or 47; the specific gravity of ruthenium is not yet known with accuracy, but such experiments as have been made render it improbable that it will prove an exception. This number is about one-fourth the mean. of the at. vols. of the long recognized members of the arsenic group, but these latter differ so widely among themselves that the comparison is of little or no value. It would be desirable to get a good determination of the density of osmic acid in the solid state, so that its at. vol. might be calculated and compared with that of antimonious acid.

The specific heat of osmium, so far as its value as a physical character goes, opposes the introduction of this element into the arsenic group. It has been determined by Regnault='03063; multiplying now by the equivalent 97, we have the product, 2.9711, thus placing osmium in the list of the elements (including the majority) for which the product of sp. ht. by at. wt. is nearly 3, while for phosphorus, arsenic, antimony and bismuth the product thus obtained is twice as great, or about 6. In this respect, however, osmium probably resembles nitrogen-the latter examined, as it necessarily is, in the gaseous form.

It is to be hoped that the conducting power for heat and electricity of compact osmium will soon be examined; nothing is as yet known of these characters.

* Phosphorus,

Arsenic,

Antimony,

Bismuth,

31

Lastly, as regards the magnetic relations of the element-it is placed, with some doubt, by Faraday in the paramagnetic class; the metal and its protoxyd were found to act feebly in this sense, while pure osmic acid is said to have shown itself clearly diamagnetic. The strongly diamagnetic character of phosphorus, antimony and bismuth would render a re-examination of this point interesting. Arsenic, however, is said to be very feebly diamagnetic, and is placed by Faraday close to osmium in the list of metals examined, though on the opposite side of the line of magnetic neutrality or indifference.

Reviewing, now, the united physical and chemical characters of osmium, and comparing them with those of the generally recognized members of the "arsenic group," we are, I think, justified in concluding that here this curious metal should be placed in a natural arrangement of the elements-while important distinctions seem to separate it from some, at least, of the platinum metals, with which it is usually associated and described. Tuscaloosa, Ala., Nov. 1, 1859.

ART. VII.-The Comas and Tails of Comets; by Prof. W. H. C. BARTLETT, U. S. Military Academy at West Point.

COMETS have, at all times, been objects of curiosity and wonder; and the question in regard to the nature of their luminous appendages, has exercised the speculative ingenuity of philosophers from the earliest records of astronomy. Everything written about them is read with interest, and the most extravagant theories in respect to their constitution and the laws of their being find a ready favor with the public. They are still among the enigmas of the heavens. Among the recent and remarkable efforts at solution, is one by the ablest mathematician of the country, perhaps of the age: and granting the premises, there is no avoiding the conclusions of the comprehensive and searching analysis for which this eminent man is so remarkable. But the assumption, that the attractive energy which summons a comet from the depths of space to the presence of the sun, retains its nature unchanged and strengthens with the diminution of distance for a part of the approaching mass, and yet reverses its character and becomes repulsive for another part, in order to obtain material to build up the tail, appears so unsupported by the analogies of nature as to give to his results the taint of improbability. Indeed, a theory which demands such an exercise of faith in matters of science, and from such friends, can only inspire doubt, and should yield the place it has too long occupied