2nd. From the moment the meteor entered the atmosphere, it would lose velocity. The resistance which the air offers to so rapid a motion, is enormous. If meteorites be admitted to come in general from meteors, it may be added that they rarely enter the ground more than two or three feet. They do not strike the earth with a velocity at all comparable to that which meteors are known to have, in the higher regions. They lose almost all their velocity in passing through the atmosphere.

A careful examination of all these observations leads me to believe that the actual velocity was as great as 36 miles a second. If we consider the resistance of the air, and then make as large an allowance for errors of observation as can reasonably be made, it seems almost impossible that it could have entered the atmosphere with a velocity less than twenty-one miles. The parts of the earth directly under the meteor, were by the earth's motion in its orbit, and on its axis, moving in a line inclined 89° 31' to the path of the meteor, with the velocity of 19.023 miles. If the velocity of the meteor in this path was 21, its velocity relative to the sun would then be a little more than 281 miles. If the meteor had been moving in a parabolic orbit around the sun, it would have had from the combined action of the earth and sun, a velocity of 27.9 miles a second. If, therefore, as I think, can hardly be doubted, the meteor entered the atmosphere with a velocity not less than 21 miles, it must have been moving in a hyperbolic orbit.

We have been accustomed to consider the solar system as filled with small planetoids, millions of which, each day, come into the atmosphere, and are burnt up, causing the shooting stars. Now we find that we must, in all probability, add one, and no doubt innumerable other similar bodies to the stellar spaces. It opens a new view of creation.

It must not hence be imagined, that the meteors and shooting stars all come from the stellar spaces. The periodicity of the August and November meteors, shows plainly that they are from permanent members of the solar system.

This meteorite did not come from the moon. If we could suppose a lunar volcano to throw out a body with such an enormous velocity, that body must come to the earth, nearly from the direction of the moon. But the moon was at that time about 120° from the direction of the meteor's path.

The recent researches, respecting the transformation of motion into heat, throw some light on the subject of shooting stars. When these bodies come into the atmosphere, the motion they lose is transformed into motion of the air, heat, light, sound, and probably other forms of energy. If it was all transformed into heat, it would be easy to compute the amount due to the loss of a given velocity. If they have a motion of their own, and their directions are subject to no law, it is easily seen that the average

velocity is much greater than 19 miles a second. A body weighing one pound, and moving 25 miles a second, has momentum sufficient to raise (25×5280)÷2g=271,500,000 pounds one foot. By Joule's equivalent the raising of 772 pounds one foot, corresponds to the heat necessary to raise one pound of water one degree Fahrenheit. If the capacity of the meteoric substance for heat is 0.2, (that of iron is 0.12,) the loss of a velocity of 25 miles would be equivalent to heating (271,500,000÷0·2)÷772= 1,760,000 pounds of the substance one degree Fahrenheit, if the whole of the motion was transformed into heat. A very small fraction of this heat would doubtless suffice to burn up, or dissipate, any substance whatever.

It is often urged, that the shooting stars cannot be solid bodies, since of the millions that daily enter the atmosphere, so few come to the ground. The above calculation shows that the heat generated may be ample to vaporize or dissipate them.

The shooting stars need not in general be large bodies. The apparent size is due to irradiation, and indicates, not amount of matter, but rather amount and intensity of light. Thus the stars though often spoken of as mere points have disks. The diameter of stars of the first magnitude was estimated at 2′ by Tycho Brahe. The telescope has shown that this disk is spurious. If these stars are equal in size to the sun, Tycho's estimate makes their diameters 50,000 times too great.

It has been estimated that the light of the sun's surface is four or five times as great as that of the same surface of the lime in the calcium light. It is also estimated that the light of the sun is 20,000,000,000 times that of Sirius. A simple calculation shows that an inch globe as brilliant as the calcium light, would give at over 100 miles distance as brilliant a light as a star of the first magnitude. The estimates which are used as the basis of calculation are confessedly very vague, yet they show that a very small body may furnish as much light as a shooting star. Such a body would naturally burn up without passing through the atmosphere.

I can therefore see no reason, as some persons do, to make a marked distinction between the different classes of meteors. Those which furnish meteorites, those which explode with a loud report, and those of all degrees of brilliancy which are not heard to explode, all seem to belong to one class, and to differ from each other no more than substances on the earth. That some are solid and others aëriform is not impossible. Differences of chemical constitution, size, velocity, and orbit exist, and these may account for the variety of appearance.

Note. Since the above was in type, the meteor of July 20 seems to furnish better data for proving that meteors sometimes come from the stellar spaces.

AM. JOUR. SCI.—SECOND SERIES, VOL. XXX, No. 89.–SEPT., 1860.

ART. XVIII-Crystalline form not necessarily an indication of definite Chemical Composition: or, on the possible variation of constitution in a mineral species independent of the Phenomena of Isomorphism. By JOSIAH P. COOKE, Jr., A.A.S., Professor of Chemistry and Mineralogy in Harvard College.*

2

IN a memoir presented to the American Academy of Arts and Sciences in September, 1855,† I described two new compounds of zinc and antimony which I named stibiobizincyle and stibiotrizincyle, on account of their analogy in composition to the metallic radicals of organic chemistry. The symbols of these compounds are Sb Zn and Sb Zn3; and they are distinguished by the high perfection of their crystalline forms, the last being still further characterized by a most remarkable property of decomposing water quite rapidly at 100° C, I stated in the same memoir that crystals of these two compounds could be obtained containing proportions of zinc and antimony differing very widely from those required by the law of definite proportions; and I also traced out the relation between the composition of the crystals, and that of the menstruum in which they are formed. It is my object in the present paper to consider the bearing of these facts, already fully described, on the idea of mineral species, and to offer a few suggestions which I hope may be of service in determining the true chemical formulæ of many minerals, and thus in simplifying the science of mineralogy. But in order to render myself intelligible, it will be necessary to recapitulate very briefly the facts in question, referring to the original memoir for the full details.

The crystals both Sb Zn and Sb Zn3 can be obtained with great readiness. It is only necessary to melt together the two metals in the atomic proportions, and when the metals are fully alloyed, to proceed exactly as in crystallizing sulphur. The melted mass is allowed to cool until a crust forms on the surface, which then is broken, and the liquid metal remaining in the interior poured out. On subsequently breaking the crucible, the interior is found lined with magnificent metallic crystals, which, when not tarnished by oxydation have a silver-white lustre. In the course of my investigations on these compounds, crystallizations were made, or attempted, of alloys, differing in composition by one half to five per cent, according to circumstances, from the alloy containing 95 per cent of zinc, to that containing 95 per cent of antimony; but only two crystalline forms were observed, that of Sb Zn2 and that of Sb Zn3. The crystals of the

* Communicated by the Author.

+ Transactions of the American Academy of Arts and Sciences, New Series, vol. v, p. 337. This Jour. [2], xx, 222.

*

two compounds both belong to the trimetric system; but they differ from each other, not only in their crystallographic elements, but also in their whole "habitus." Stibiotrizincyle crystallizes in long acicular prisms, which group themselves together into larger prismatic aggregates; while stibiobizincyle crystallizes in broad plates, which twin together on an octahedral face, and form a very characteristic cellular structure. This very striking difference in the character of the crystals proved to be an important circumstance in the investigation, as it enabled me to distinguish with certainty between the two compounds, even when the faces of the crystals were so imperfect that a measurement of angles was impossible.

3

The most remarkable result of the investigation, and the one to which I wish to direct especial attention, is the fact that each of the two crystalline forms was found to be constant under very wide variations in the per-centage composition of the crystals. As this is a point of great importance, it will be necessary to enter more into detail, considering in the first place the crystals of Sb Zn3. The crystals of this compound are obtained in the greatest perfection from an alloy containing the two metals in just the proportions represented by the formula, namely, 42.8 parts of zinc, and 57-2 parts of antimony. They are then comparatively large, generally aggregated, and, as the three analyses cited in the accompanying Table indicate, they have the same composition as the alloy.

On increasing gradually the amount of zinc in the alloy up to 48.7, the crystals continued to have the composition of the alloy; and the only difference which could be observed in their character was that they were smaller, and more frequently isolated. Between these limits the whole mass of the alloy exhibited a strong tendency to crystallization; and by pouring it, as it cooled, from one vessel to another, it could be crystallized to the last drop. On increasing the amount of zinc in the alloy to 50·7 per cent, the amount of zinc found in the crystals was uniformly less than it was in the alloy; but no closer relation between the two could be detected, owing, undoubtedly, to the unavoidable irregularity in the crystallization of the alloys which contained more than 50 per cent of zinc. This arose from a peculiar pasty condition which the liquid mass assumed at the point of crystallization. Definite crystals, however, were obtained from an alloy of 60 per cent zinc containing 55 per cent; above this the crys

:

tals became less and less abundant, and gradually faded out, although the alloy of 86 per cent of zinc exhibited a radiated crystalline texture; and a trace of this structure could still be discovered even in the alloy containing only 4 per cent of antimony. It was very interesting to trace the gradual fading out of the crystalline structure, as the character of the phenomenon was entirely analogous to that which may be noticed in many crystalline rocks.

Finding that the crystalline form of Sb Zn3 was constant under so great an increase of the proportion of zinc in the crystals, it might be supposed that, on returning to the alloy of 42.8 per cent of zinc and increasing the amount of antimony, we should obtain crystals containing an excess of antimony; but so far is this from being true, that the slightest excess of antimony entirely changes the character of the crystallization. On crystallizing an alloy containing 41.8 per cent of zinc, not a trace of any prismatic crystals could be seen; but in their place there was found a confused mass of thin metallic scales, which, as will soon be shown, are imperfect crystals of Sb Zn3. Thus it appears that, although perfectly formed crystals of Sb Zn3 can be obtained containing 55 per cent of zinc (that is, 12 per cent above the typical proportions), they cannot be made to take up the slightest excess of antimony.

Let us pass now to the crystals of Sb Zn2. In order to obtain crystals having the exact typical constitution, it was found necessary to crystallize an alloy at least as low as 31.5 per cent of zinc. At this point large compound crystals are obtained corresponding to the large crystals of Sb Zn3; and the same was true of alloys down to 27 per cent of zinc. Between these two limits (namely, alloys of 31.5 and 27 per cent of zinc) the crystals formed were found to have the theoretical composition of Sb Zn2, indicating of course a tendency towards this point; but on increasing or diminishing the amount of zinc in the alloy beyond these limits, the composition of the crystals immediately began to vary in the same direction as that of the alloy. The crystals of Sb Zn2 containing an excess of zinc are smaller and more frequently isolated than those having the exact theoretical composition. A similar fact, it will be remembered, is true of the crystals of Sb Zn3.

At the alloy of 33 per cent of zinc, the definite crystals of Sb Zn2 begin to disappear, and are succeeded by thin metallic scales, which are obviously imperfect crystals of the same form. This was established, not only by the obvious law of continuity noticed in the different specimens (the perfect crystals gradually passing into the scales), but also by the peculiar mode of twining, which was the same with the scales as with the large crystals, forming the peculiar cellular structure already referred to. Moreover, the angle between two scales thus united was found