Chapter IV: Appendix: To the List of the Meteorites (2)

[Sidenote: Pane 4l.]

34. The want of homogeneity and the structure of meteoric iron are beautifully shown by the figures generally called into existence when a polished surface is exposed to the action of acids or bromine; they are due to the inequality of the action on thick or thin plates of various constituents, [Sidenote: Etched figures.] the plates being composed chiefly of two nickel-iron materials termed kamacite and tænite. A third nickel-iron material, filling up the spaces formed by the intersection of these plates of kamacite and tænite, is termed plessite; it is probably not an independent substance but an intergrowth of the first two kinds.

In the Agram iron, investigated by Widmanstätten in 1808, the plates are parallel to the faces of the regular octahedron; such figures are well shown by the exhibited slice of the Toluca iron; different degrees of distinctness of such "Widmanstätten" figures are illustrated by specimens of Seneca River, Zacatecas, Charcas, Burlington, Jewell Hill, [Sidenote: Pane 4l.] Lagrange, Victoria West, Nelson County, and Seeläsgen. The large Otumpa specimen, mounted on a separate pedestal, furnishes a good example of the less distinct, and more or less damascene, appearance presented by the etched surface of some meteoric irons of octahedral structure.

The Braunau iron gives no "Widmanstätten" figures, but has cleavages parallel to the faces of a cube; on etching it yields linear furrows which were found (1848) by Neumann to have directions such as would result from twinning of the cube about an octahedral face; as illustrations of the "Neumann lines," etched specimens of Braunau and Salt River are exhibited. [Sidenote: Pane 4l.]

For meteoric irons of cubic structure the percentage of nickel is lower than 6 or 7; for those of octahedral structure it is higher than 6 or 7, and the plates of kamacite are thinner, and the structure therefore finer, the higher the percentage of that metal. A considerable number of meteoric irons, however, show no crystalline structure at all, and have percentages of nickel both below and above 7; it has been suggested that these masses have been metamorphosed, and that crystalline structure was once present, but has disappeared as a result of the meteorites having been heated, not merely superficially during their passage through the earth's atmosphere, but throughout their mass while travelling in outer space.

[Sidenote: Cooling of fused mixtures and of solutions.]

35. Though meteoric iron has been at some time, presumably, in a state of fusion, and its present structure is a result of the particular circumstances of the cooling of the liquid and afterwards solid material, attempts to produce such structures by the cooling of fused meteoric iron or artificial mixtures of nickel and iron have not yet been successful. It will be useful, therefore, to consider briefly some of the manifold changes which are found to take place during the passage of fused mixtures and of solutions to the solid state, and during the cooling of such solids to ordinary temperatures.

If a fused mixture of antimony and bismuth is allowed to cool, the solid which first separates is neither pure antimony nor pure bismuth, but a material which has a percentage composition depending on, though not identical with, that of the original mixture. The temperature for the beginning of the solidification is different for different proportions of the two metals, and is intermediate between 622° and 268°, the solidifying temperatures of antimony and bismuth, respectively; it approaches the latter more and more closely as the percentage of the bismuth is increased. The solid first separated is somewhat richer in antimony than the original mixture; the still fused part, therefore, is somewhat richer in bismuth than before, and does not begin to solidify till a lower temperature is reached; the temperature thus gradually falls, instead of remaining constant, during the solidification. In the cooling of such fused mixtures the changing composition of the part still fused has for effect a changing composition of the solid already separated; whence the slower the cooling of the fused material, the greater is the homogeneity of the final solid.

[Sidenote: Eutectic mixtures.]

A fused mixture of silver and copper behaves in a different way. When the percentage weight of the silver is 72, and that of the copper, therefore, is 28, solidification begins, not at a temperature between 960° and 1083°, the solidifying temperatures of silver and copper, respectively, but at a temperature below both, namely, 770°. The solid which first separates has the same percentage composition as the original mixture; the part still fused has thus itself the same percentage composition as before, and continues to solidify at the same temperature, and in the same way, until the solidification is complete. Such a mixture, having a definite composition and a definite temperature of solidification, was for a time regarded as a definite chemical compound with a complex chemical formula, but on microscopic examination the resultant solid is found to be heterogeneous; minute particles of the silver and copper are seen to lie side by side, the particles being granular or lamellar in form according to the circumstances of the cooling. If the percentage of silver is different from 72, whether it be higher or lower, the solidification begins at a higher temperature than 770°; whence the mixture containing 72 per cent. of silver has been conveniently termed _eutectic_ (i.e. very fusible); the term was suggested by Prof. F. Guthrie,[16] to whom our knowledge of the existence of such mixtures is due.

[Sidenote: Cooling of fused mixtures and of solutions.]

36. When the silver is in excess of 72 per cent., the excess of silver gradually collects together and solidifies at various parts of the cooling fused mass; the still fused portion thus gradually becomes poorer in that metal, and the temperature, instead of remaining constant, gradually falls during the separation of the solid. At length the percentage of silver in the fused portion falls to 72 per cent. and the temperature to 770°; the solid which now begins to form is no longer pure silver, but a material containing 72 per cent. of that metal; and it continues to have the same percentage composition as the surrounding liquid, and the temperature of solid and liquid to be 770°, until the solidification is complete. The final solid thus consists of blebs of silver scattered through a fine groundmass of eutectic mixture of silver and copper. Similarly, if the copper is in excess of 28 per cent., the final solid consists of blebs of copper scattered through a fine groundmass of eutectic mixture of silver and copper.

If the two metals are copper and antimony, instead of copper and silver, the results are more complicated; for the first two metals are capable of combining together to form a definite chemical compound represented by the formula Cu{2}Sb, and each of the metals forms a eutectic mixture with the latter. According to the percentage composition of the original mixture, the solid which first separates during cooling from fusion may be either copper or antimony or the compound Cu{2}Sb; the separation continuing, and the temperature falling, until the first eutectic proportion and its corresponding temperature are reached.

[Sidenote: Cooling of solutions.]

37. Analogous results are obtained during the cooling of solutions; for instance, during the cooling of a solution of sodium chloride (common salt) in water. A solution containing 23·5 per cent. of sodium chloride begins to solidify at -22° C.; the separating solid is not simple sodium chloride or simple ice, but has the same percentage composition as the original solution, and thus the temperature remains -22° until the whole material has become solid. On microscopic examination the solid is seen to be heterogeneous, and to consist of small particles of sodium chloride and ice lying side by side. If the percentage of sodium chloride is different from 23·5, whether higher or lower, solidification begins before the temperature has fallen to -22°. The characters of this particular solution are thus closely analogous to those of the eutectic mixtures described above. If the sodium chloride exceeds 23·5 per cent., the excess of sodium chloride begins to separate, and solidify, at various parts of the liquid, at a temperature higher than -22°; it continues to separate, and the temperature to fall, until the proportion of sodium chloride in the residual liquid is reduced to 23·5 per cent. and the temperature to -22°. Afterwards the separating solid has the same composition as the residual liquid (23·5 per cent. of sodium chloride), and the temperature remains constant, until the residual liquid has been wholly transformed into a solid fine-grained mixture of sodium chloride and ice. The final solid thus consists of large particles of sodium chloride dispersed through a fine groundmass consisting of eutectic mixture of sodium chloride and ice. Similarly, if the water is in excess of 76·5 per cent., the final solid consists of large particles of ice dispersed through a fine groundmass consisting of eutectic mixture of sodium chloride and ice.

The results of the cooling of a solution of ferric chloride are still more complicated; for this substance enters into chemical combination with water, and in no fewer than four different proportions. The solid which first separates from the cooling solution may thus, according to the percentage of ferric chloride, be either ferric chloride or water, or any one of the various compounds of the two; and to each pair of compounds nearest to each other in composition corresponds a different eutectic mixture and a different temperature for its formation.

[Sidenote: Cooling of solids.]

38. Some solid bodies, during cooling, show changes analogous to those observed in solutions, and are therefore termed "solid solutions." For instance, if a hot physically homogeneous solid obtained from the fusion of iron with carbon is cooled, there may result a separation in the solid of particles of either iron or cementite, the latter being a chemical compound of iron and carbon represented by the formula Fe{3}C; the particular substance separated depending on the percentage composition of the original solid. This separation continues, and the temperature falls, until the residual physically homogeneous material contains 0·9 per cent. of carbon and the temperature is 690°; the temperature then remains constant, although the body is surrounded by a cooling medium, until this residual physically homogeneous material has been wholly transformed into a fine-grained mixture of iron and cementite, containing 0·9 per cent. of carbon. This particular kind of mixture has been termed eutectic, though the transformation has taken place, not by solidification from fusion, but in a body which was already solid. Prof. Rinne has proposed for such cases the substitution of the term _eutropic_, thus avoiding the suggestion of fusion. The eutectic mixture of iron (or ferrite) and cementite is known as pearlite.

[Sidenote: Overcooling.]

39. Just as water may be cooled so quietly that it is still liquid at a temperature much below the normal freezing point, a mixture may be cooled in such a way as to pass much below the eutectic (or eutropic) point without the normal transformation taking place; it is then said to be overcooled. The equilibrium, however, is very unstable, and the transformation, once begun, takes place almost instantaneously throughout the whole mass.

[Sidenote: Crystalline structure of artificial iron.]

40. A structure analogous to that shown by the Widmanstätten figures, though on a finer scale, has been observed by Prof. J. O. Arnold and Mr. A. McWilliam[17] in cast steel containing 0·4 per cent. of carbon; the plates of iron (or ferrite) in the cast steel correspond to the plates of kamacite in meteorites. Further, it has been found that the plates in the cast steel disappear during the process of annealing; similarly, there are no Widmanstätten figures, and the structure of the material is granular, near the outer surface of an unweathered meteoric iron; presumably as a result of the high temperature to which the outer part of the mass has been raised during the passage of the meteorite through the earth's atmosphere.

[Sidenote: Structure of meteoric irons.]

41. At present it is generally imagined that kamacite and tænite are definite alloys, or perhaps solid solutions, of iron and nickel, the former being poor in nickel (6 or 7 per cent.) and the latter rich in that constituent (25 to 38 per cent.), that kamacite and tænite separate in succession from the molten mass or solid solution until the residual part is so rich in nickel that a eutectic (or eutropic) proportion is reached; the residual material then forms plessite, which, according to this view, is a eutectic (or eutropic) mixture of kamacite and tænite. But it is difficult to understand how the thin plates of tænite are deposited on the plates of kamacite, seeing that they contain more nickel than kamacite and plessite, and yet have an intermediate epoch of formation, prior to the epoch of formation of that tænite which is a constituent of the plessite; one suggestion is that the thin plates of tænite have been deposited on the plates of kamacite owing to the temperature having fallen well below the eutectic (or eutropic) point after the separation of the kamacite and before the eutectic transformation of the residual material has taken place. And Prof. Rinne[18] himself is of opinion that the Widmanstätten structure has been wholly developed in meteoric iron after the solidification of the mass; further, as the relations of the kamacite, tænite and plessite to the enclosed troilite indicate that the troilite was solid before the octahedral structure was developed, and as that mineral, under normal circumstances, solidifies at about 950°, he infers that the structure was developed below that temperature. In the case of the Jewell (Duel) Hill meteorite it was discovered by Dr. Brezina that, notwithstanding the pronounced octahedral structure, plates of troilite are embedded, not in accidental positions nor between successive octahedral layers, but parallel to the faces of the corresponding cube; whence Prof. Rinne suggests that this iron, now of octahedral structure, and possibly all others of a similar character, had a cubic structure at the epoch when they entered upon the solid condition. But, as both Prof. Rinne and Dr. Brezina[19] have pointed out, a fused mixture of nickel and iron, cooling undisturbedly in outer space, may have solidified at a temperature even below 950° and thus have been much overcooled.

[Sidenote: Tænite possibly a eutectic mixture.]

42. In the course of a recent elaborate investigation of the changes of the magnetic permeability of the Sacramento meteoric iron with changing temperature, Mr. S. W. J. Smith[20] has been led to infer that the magnetic behaviour can only be explained by imagining the meteorite to consist largely of plates of nickel-iron, containing about 7 per cent. of nickel (kamacite), separated from each other by thin plates of a nickel-iron constituent (tænite), containing about 27 per cent. of nickel and having different thermo-magnetic characters from those of kamacite; he suggests, however, that tænite is not a definite chemical compound, but is itself a eutectic (or eutropic) mixture, and consists of kamacite and a nickel-iron compound containing not less than 37 per cent. of nickel. And he points out that, while the tænite mechanically isolated from meteorites for analysis has approximately the lower percentage (27 per cent.), the tænite chemically isolated through the prolonged action of dilute acid (which would remove much of the admixed kamacite) has a higher percentage, which in several cases approximates to 40 per cent.

[Sidenote: Few siderites have been seen to fall.]

43. The Siderites _actually observed to fall_, or found soon after a luminous meteor had been seen, or a detonation heard, by people in the neighbourhood, reach only the small number of nine; they are, Agram, Charlotte, Braunau, Victoria West, Nedagolla, Rowton, Mazapil, Cabin Creek, and N'Goureyma. The remaining specimens in collections of Siderites are presumed to be of meteoric origin by reason of the peculiarity of their appearance and chemical composition, and of the characters of the material in which they have been found (Art. 7).

[Sidenote: Siderites of large size.]

The large Cranbourne meteorite, mounted in a special case in the Pavilion, before rusting weighed 3-1/2 tons. The two largest known were found in Western Greenland and Mexico, respectively, and are both of very irregular shape. The Greenland mass is 11 feet long, 7-1/2 feet wide, and 6 feet thick, and its weight, which had been variously estimated at from 50 to 100 tons, has been determined to be 36-1/2 tons; the mass had long been known to the Eskimos, and was inquired after by Captain John Ross in 1818; it was shown by a native to Lieutenant Peary in 1894, who afterwards transported it from Melville Bay to New York; it is now preserved in the American Museum of Natural History in that city. The Mexican mass is 13 feet long, 6 feet wide, and 5 feet thick, and has an estimated weight of 50 tons; it is the property of the Mexican Government, and is still lying at El Ranchito, near Bacubirito, Province of Sinaloa.

[Sidenote: The iron found at Ovifak is probably of terrestrial origin.]

44. The difficulty of distinguishing an iron of terrestrial from one of meteoric origin was rendered very evident by the prolonged controversy as to the origin of the large masses of iron, containing one or two per cent. of nickel, and weighing 9,000, 20,000, and 50,000 lbs., respectively, found in 1870 by Baron N. A. E. Nordenskiöld on the beach at Ovifak, Disko Island, Western Greenland.

A careful examination of the rocks of the neighbourhood shows that the basalt contains nickeliferous iron disseminated through it, and that the large masses of iron, [Sidenote: Pane 4m.] at first thought to be meteorites, are very probably of terrestrial origin, and have been left exposed upon the seashore through the weathering of the rock which originally enclosed them. Some of the malleable metallic nodules extracted from the basalt were found to contain as much as 6·5 per cent. of nickel. In 1880 Professor K. J. V. Steenstrup[21] found ferriferous basalt _in situ_ in three different parts of the island. At Assuk (Asuk) the enclosed balls of iron reach a diameter of nearly three-quarters of an inch. Some assert that the basalt and the nickel-iron have been expelled together from great depths below the earth's surface, while others consider that the nickel-iron is due to the reduction of the iron-compounds in the basalt by the passage of the lava through the beds of lignite and other vegetable matter found in the vicinity.

[Sidenote: Other terrestrial irons.]

[Sidenote: Pane 4m.]

45. With the Ovifak iron in the case are shown other specimens of iron which have been brought by various explorers from West Greenland, and were formerly thought to have had a meteoric origin. The discovery of ferriferous basalt, not only _in situ_ in several places, but also deposited in a Greenlander's grave (1879) along with knives (similar to those given to Captain John Ross in 1818) and the usual stone tools, renders it clear that the Eskimos were not dependent solely on meteorites for their metallic iron, as had long been supposed.

Mr. Skey announced in 1885 the discovery of terrestrial nickel-iron in New Zealand. Grains of the alloy (Awaruite), containing as much as 67·6 per cent. of nickel, are found in the sand of the rivers flowing from a range of mountains composed of olivine-enstatite rocks, in places altered to serpentine: similar particles have been found in the serpentine itself. Similarly, in the sand of the stream Elvo, near Biella, in Piedmont, and of the river Fraser, British Columbia, grains of nickel-iron containing 75 or 76 per cent. of nickel have been found: and in the placer gravel of a stream in Josephine and Jackson Counties, Oregon, U.S.A., large quantities of waterworn pebbles, which enclose an alloy (Josephinite) of nickel and iron containing 72 per cent. of the former metal, have been met with. Professor Andrews many years ago established the presence of minute particles of metallic iron in some basalts; Dr. Sauer has lately found a single nodule of malleable iron of the size of a walnut in the basalt of Ascherhübel, in Saxony; Dr. Hornstein has described large nodules of (nickel-free) iron found in basalt in a quarry at Weimar, near Cassel; Dr. Beckenkamp has described nodules of metallic iron found in clay at Dettelbach, near Würzburg; and Dr. Johnston-Lavis has announced the find of an enclosure of metallic iron in a leucitic lava of Monte Somma; Dr. Hoffmann has noted the occurrence of minute spherules of brittle iron both in perthite and quartzite in Ontario; Dr. Hussak has recorded the discovery of metallic iron in an alluvium of Brazil, and Dr. Högbom has found it associated with topaz, quartz, felspar, and other minerals, in limonite from an unspecified place in South America; two minute grains of iron were found by Mr. Osaka in the débris of an agglomerate at Nishinotake, Japan.

[Sidenote: The stony matter of meteorites.]

46. The stony part of the siderolites and aerolites is almost entirely crystalline, and in most cases presents a peculiar "chondritic" or granular structure, the loosely coherent grains being composed of minerals similar to those which enclose them, and containing in most cases minute particles of iron and troilite disseminated through them: glass-inclusions are found to be present. The minerals mentioned above as occurring in meteorites are such as are very characteristic of the more basic terrestrial rocks, such as dunite, lherzolite and basalt, which have been expelled from considerable depths below the earth's surface.

47. Several attempts to classify aerolites according to their mineralogical constitution have been made, but it cannot be said that any of them is very satisfactory; seeing that even in the same stone there may be much difference in its parts, a perfect classification on such a basis is scarcely to be hoped for.

[Sidenote: Chondritic aerolites.]

About eleven out of every twelve of the stony meteorites belong to a division to which Rose[22] gave the name of _chondritic_ (_chondros_, a grain): they present a very fine-grained but crystalline matrix or paste, consisting of olivine and enstatite or bronzite, with more or less nickel-iron, troilite, chromite, augite and anorthic felspar; through this paste are disseminated round chondrules of various sizes (up to that of a walnut) and with the same mineral composition as the matrix; in some cases the chondrules consist wholly or in great part of glass.[23] In mineral composition chondritic aerolites approximate more or less to terrestrial lherzolites. Some meteorites consist almost solely of chondrules, others contain only few; in some cases the chondrules are easy separable from the surrounding material. Of the chondritic division Knyahinya, Pegu, Muddoor, [Sidenote: Pane 4n.] Seres, Judesegeri, Khiragurh, Utrecht and Nellore (pane 4p) afford good illustrations.

[Sidenote: A carbonaceous group.]

A few meteorites belonging to this division are remarkable as containing carbon in combination with hydrogen and oxygen. Of these the Alais and Cold Bokkeveld meteorites [Sidenote: Pane 4n.] are good examples: the former has a bituminous smell; it yields sulphates of magnesium, calcium, sodium and potassium, if steeped in water.

[Sidenote: Aerolites without chondrules.]

[Sidenote: Pane 4o.]

48. The remaining aerolites are not chondritic, and they contain little or no nickel-iron; of these we may specially mention for their mineral composition the following:--

_Juvinas_ and _Stannern_, consisting essentially of anorthite and augite.

_Petersburg_, consisting of anorthite, augite and olivine, with a little chromite and nickel-iron: both Juvinas and Petersburg may be compared to terrestrial basalt.

_Sherghotty_, consisting chiefly of augite and maskelynite.

_Angra dos Reis_, consisting almost wholly of augite; olivine is present in small proportion.

_Bustee_, of diopside, enstatite and a little anorthic felspar, with some nickel-iron, oldhamite and osbornite.

_Bishopville_, of enstatite and anorthic felspar, with occasional augite, nickel-iron, troilite and chromite.

_Roda_, of olivine and bronzite.

_Chassigny_, consisting of olivine with enclosed chromite, and thus mineralogically similar to a terrestrial dunite.

[Sidenote: Is there a periodic recurrence?]

49. The importance of the examination and classification of meteorites, with a view to a possible recognition of _periodicity_ of fall of specimens presenting the same characters, need only be mentioned to be appreciated: such a determination is, however, rendered very difficult by the close similarity of structure and composition presented by the great majority of the aerolites of the large chondritic division.

[Sidenote: Few aerolites are known which have not been seen to fall.]

50. Attention has been already directed to the fact that although many masses of meteoric iron, some of them like that of El Ranchito, near Bacubirito, in Mexico, weighing very many tons, have been found at various parts of the earth's surface, very few of them have been actually observed to fall: in the case of the stony meteorites just the opposite holds good, for they are never very large, and few are known which have not an authenticated date of fall. This may be due to the fact that a meteoric stone is less easily distinguished than is a meteoric iron from ordinary terrestrial bodies, and will thus in most cases remain unnoticed unless its fall has been actually observed; while, further, a quick decomposition and disintegration must set in on exposure to atmospheric influences. The smaller size of the meteoric stones may be due to the greater ease with which they break up on the sudden increase of temperature of their outer surface, consequent on their entry into the earth's atmosphere. The largest meteoric stone preserved in a Museum is one which fell as part of a shower at Knyahinya, Hungary, in 1866: it weighs 647 lbs. and is at Vienna. A larger stone (723 lbs.) fell at Tabory, Russia, in 1887, but was broken to pieces by the impact on the earth; fragments of a still larger single stone, weighing at least 1244 lbs., were found near together at Long Island, Kansas, U.S.A., but the fall was not observed.

[Sidenote: The chondrules and their matrix.]

51. If we now examine minutely the structure of the meteoric stones, it will be seen that almost all of them appear to be made up chiefly of irregular angular fragments, and that some of them bear a close resemblance to volcanic tuffs. In the large group of chondritic aerolites, chondrules or spherules, some of which can only be seen under the microscope while others reach the size of a walnut, are embedded in a matrix, apparently made up of minute splinters such as might result from the fracture of the chondrules themselves. In fact, until recently, it was thought by some[24] that the chondrules owe their form, not to crystallisation, but to friction, and that the matrix was actually produced by the wearing down of the chondrules through collision with each other either as oscillating components of a comet or during repeated ejection from a volcanic vent of some small celestial body. Chondrules have been observed, however, presenting forms and crystalline surfaces incompatible with such a mode of formation, and others have been described which exhibit features resulting from mutual interference during their growth.

The crystallisation of the chondrules is independent of their form, and must have started, not at the centre, but at various places on their surfaces; Dr. Sorby[25] argued that some at least of the chondrules must once have fallen as drops of fiery rain, and have assumed their shape in an atmosphere heated to nearly their own temperature. The chondritic structure is different from anything which has been observed in terrestrial rocks, and the chondrules are distinct in character from those observed in perlite and obsidian. After much study, Dr. Brezina[26] lends his weighty support to the hypothesis that the structural features of meteorites are the result of a hurried crystallisation: and Prof. Wadsworth[27] accepts the same interpretation.

[Sidenote: Some meteoric materials appear to have been altered since their consolidation.]

[Sidenote: Pane 4o.]

52. Since the time of their consolidation some meteoric stones, as Tadjera, appear to have been heated throughout their mass to a high temperature: and in the case of Orvinio, Chantonnay, Juvinas, and Weston, fragments are cemented together with a material having the same composition as the fragments themselves, thus giving rise to a structure resembling that of a volcanic breccia. Others seem to have experienced a chemical change, for some of the chondrules in Knyahinya and in Mezö-Madaras, when examined with the microscope, are found to be surrounded by spherical and concentric aggregations of minute particles of nickel-iron, perhaps due to the reducing action of hydrogen at a high temperature. Others, as Château-Renard, Pultusk and Alessandria, present what in terrestrial rocks would probably be called faults: in some cases the fissures are seen to have been filled with a fused material after the chondrules have been broken and one side of the fissure has glided along the other. These peculiarities of structure suggest that the small body which reaches the earth is only a minute fragment of a much larger mass. It has been suggested that the chondritic structure is of metamorphic origin, and a mere result of enormous pressure on the stony material during the passage through the earth's atmosphere; according to still another view, the structure, though metamorphic, is of extra-terrestrial origin, and due to the quick cooling of a tuff-like stone which has been partially melted, for instance, by the heat from a neighbouring new star or by traversing the hot vapours on the limits of an old one.

[Sidenote: Do meteorites reach our atmosphere as clouds of gas or dust?]

53. The idea that meteorites arrive at our own atmosphere, not as fragments of rock, but as mere clouds of gas or dust, has been recently revived and again discarded. According to this hypothesis, the air, instead of dispersing the entering cloud, acts in the contrary way, and in a few seconds of time presses the particles together to form solid bodies. This idea is open to various objections, and in any case one can scarcely understand how large masses of iron, presenting a wonderful regularity of crystalline structure, can have been the result of so hurried a process: and if we once grant that the irons enter the atmosphere as solid bodies, it is difficult to believe that the same is not the case with the stones.

* * * * *

[Sidenote: Where do meteorites come from?]

54. From the above it will be evident that the old hypotheses that meteorites are terrestrial stones which have been struck by lightning, or carried to the sky by a whirlwind, or are concretions in the atmosphere, or are due to the condensation of a dust-cloud coming from some volcano, or have been shot recently from terrestrial volcanoes, are inconsistent with later observation; it may be granted that the bodies reach our atmosphere from outer space. From what part or parts of space do they come? Their general similarity of structure and chemical composition, and more especially the presence of nickeliferous iron in almost every one, suggest that most, if not all of them, have had a common source, and that they are chips of a single celestial body.

[Sidenote: Probably not from the sun, nor from the moon, earth, or other planet.]

55. Dr. Sorby suggested that they are probably ejected from the sun itself, though this is difficult to reconcile with the fact that some of them are easily combustible. Others, among whom we may mention Laplace, have suggested that they come from volcanoes of the moon which are now active; but the suggestion, although mathematically sound, has no physical basis, for, so far as one can discover, active volcanoes do not there exist: and Sir Robert Ball[28] has virtually excluded the lunar volcanoes, which were active in times now long past, by pointing out that if a projectile from the moon once misses the earth, its chance of ever reaching the earth is too small to be worthy of mention. It has further been shown that, although the explosive force necessary to carry a projectile so far from one of the smaller planets that it will not return, is not very large, yet the initial velocity requisite to carry the body as far as the earth's orbit is so considerable, and the chance of hitting the earth so slight, that a more probable hypothesis is, to say the least, desirable. If these bodies have been shot from volcanoes of any planet, Sir Robert Ball is himself inclined, upon mechanical grounds alone, to believe that the projection was from our own in bygone ages; for as such projectiles, having once got away from the earth, would take up paths round the sun which would intersect the earth's orbit, every one of them would have a chance of some time or other meeting with the earth again at the point of intersection, and of appearing as a meteorite. The size and initial velocity requisite for the escape of a projectile through a lofty atmosphere would be enormous: even then the difficulty would still remain that meteorites generally differ, both in structure and material, from anything known to have been ejected from existing terrestrial volcanoes. To meet these difficulties, Sir Robert has speculatively suggested that the matter was expelled before the surface of the earth became solid, and at a time when there was as much activity in the terrestrial planet as there is now in the material of the sun itself.

Nor is it probable that they are portions of a lost satellite of the earth, or are due to a collision of two planets; for in each of these cases we should expect to have received some of the larger fragments which must at the same time have been produced.

Much light is thrown on the history of meteorites by the discovery of a relationship with shooting stars and comets.

[Sidenote: Shooting or falling stars.]

56. The meteorite-yielding fireball, referred to in Art. 17, is not the only luminous meteor, apart from lightning, with which we are acquainted. On a clear dark night any one can see a star shoot now and then across the firmament: it is estimated that on the average as many as fourteen are visible to a single observer every hour. Are the _shooting_, or, as they are often called, _falling stars_ products of our own atmosphere, or do they, like the meteorites, come from outer space? In 1794 Chladni, in the memoir already referred to, gave reasons for believing that a meteoritic fireball and a shooting star are only varieties of one phenomenon.

[Sidenote: The November star-showers.]

57. But long after the cosmic origin of meteorites had been generally acknowledged, the atmospheric origin of the shooting stars was still asserted, and it was not till the wondrous star-shower of November 12-13, 1833,[29] that the cosmic origin of any of the shooting stars was finally established. During that night upwards of 200,000 shooting stars, according to a rough estimate, were seen from a single place; and the remarkable observation was made at various localities, widely distributed over North America, that the apparent paths of the shooting stars in the sky, when prolonged backwards, all passed through a point in the constellation Leo: this point of radiation appeared to rotate with the heavens during the eight hours for which the shower was visible.

Hence it was manifest that the star-shower was independent of the earth's rotation and must therefore have come from outer space; that the radiation of the paths was only apparent and due to perspective; and that, relatively to an observer, the flights of all the shooting stars were really parallel to the direction of the apparent radiant point.

On the same day of November in each of the three following years the shower was repeated though on a less grand scale, and the constancy of the radiant point was confirmed: similar small showers had been seen also in 1831 and 1832 before the radiation had been noticed. Though in the years immediately before and after 1831-6 no remarkable display of November meteors took place, it was remembered that a similar shower had been chronicled by Humboldt and by Ellicott, as observed by them on November 12, 1799; and a study of ancient documents revealed the fact that a grand star-shower had been recorded several times in October and November since A.D. 902, the date having gradually advanced, during that long space of time, from the middle of October to the middle of November.[30] The only sufficient explanation of the observed facts is that a swarm of isolated small bodies, solid and non-luminous--meteorites in fact--is moving in an orbit round the sun, completing the circuit in 33-1/4 years; the orbit intersects that of the earth, and the earth meets the swarm at the place of intersection. The isolated bodies or meteorites become luminous, as already explained in Art. 17, after their entry into the earth's atmosphere. The swarm can be only a few hundred thousand miles thick, for the earth, travelling through space at the rate of 66,000 miles an hour, passes through the densest part in 2 or 3 hours, and through the whole in 10 to 15 hours: its length, however, must be enormous, amounting to hundreds of millions of miles; for, although the meteorites move with a velocity of twenty miles a second, the swarm takes 5 or 6 years to pass the place of intersection with the earth's orbit, thus causing star-showers, more or less dense, during that number of years.

Contrary to expectation, no large November star-shower occurred either in the year 1899 or in the years which have since elapsed.

Schiaparelli has shown that the unequal attraction of the sun for the individuals of a swarm of meteorites moving round it would scatter them along the orbit, and in the course of time produce a more or less complete ring; if this intersects the earth's orbit an annual star-shower must ensue.

[Sidenote: The August star-shower and its comet.]

58. A small annual star-shower occurs, in fact, on August 10-11,[31] and has been observed since A.D. 830: it radiates from a point in the constellation Perseus. Schiaparelli calculated in 1866 the orbit and motion of the meteorites producing it, and was surprised to find that the numbers corresponded exactly with those calculated for one of the recently observed comets; in other words, a comet was moving in the path of the meteorites, and at exactly the same speed. At the same time Schiaparelli gave numbers defining the motions of the meteorites which would cause the periodic November star-showers.

[Sidenote: Star-showers related to comets.]

59. Immediately afterwards, when the numbers calculated by Oppolzer for the orbit of the comet discovered by Tempel were published, it was seen that they were really identical with those already calculated by Schiaparelli for the orbit of the meteorites of the November star-shower, and that here again a comet and a swarm of meteorites were moving in exactly the same path at exactly the same rate.

Almost immediately afterwards it was shown that the radiant points of the small star-showers of April 20-21 and November 27-28 both correspond to the orbits of known comets.

It was evident that these could not be accidental coincidences, and that the comets and the attendant swarms of meteorites are closely related to each other.

[Sidenote: Comets.]

60. An intimate connection between, if not complete identity of, meteorites, shooting stars and comets, had indeed long been suspected. Astronomers were convinced that comets, though occasionally of enormous size, are always of extremely small mass, since they pass by the earth and other planets without sensibly disturbing their motions; the comet of 1770 passed through the system of Jupiter's satellites without any perceptible action upon them: it has been calculated that the mass of a small comet may be about eight pounds. Again, the light of a comet, like that of a cloud or planet, was seen to be partially polarised: hence part, at least, must be reflected sunlight, for the plane of polarisation passes through the sun's place. Further, stars of very small magnitude have been seen not only through the tail, but even through the nucleus, of a comet without any apparent alteration of position by refraction: hence it was inferred that a comet is not a continuous mass, but consists of particles so far distant from each other that a ray of light may pass through the comet without meeting a single one of them. Such a constitution likewise accounts for the absence of phases of the reflected light: for although only half of each particle will be directly illuminated by the sun, the remaining half will receive light irregularly reflected from the particles more distant from the sun.

Among others, Chladni in 1817 had referred to the great similarity in the motions of comets and meteorites: Olmsted, in 1834, had calculated the orbit of a comet which would cause the November star-shower; his results were wrong owing to the assumption that the shower was annual: Cappocci, in 1842, gave reasons for believing that a meteorite is a small comet: Reichenbach, in 1858, in a most elaborate paper,[32] sought to prove that a comet is a swarm of meteorites; that each chondrule of a meteorite had once been an individual of a cometary swarm, and owes its rounded shape to frequent collision with its fellows; that the rest of the stone consists of the broken splinters thus produced; and that the brecciated aspect of many meteorites is due to collisions in the denser part or nucleus of a comet. As already pointed out in Art. 51, later modes of investigation have led petrologists to reject this method of accounting for the rotundity of the chondrules.

[Sidenote: Other star-showers.]

61. In addition to the few radiant points which correspond to swarms moving in orbits identical with those of known comets, there are numerous radiant points which have not yet been recognised as related to existing comets, and may possibly be due to swarms produced by the dispersal of comets along their orbits; indeed, it has been inferred from observation of shooting stars that on the average there are no fewer than fifty distinct radiant points, and therefore showers, for any night of the year. But there are still others of which there is yet no satisfactory explanation. A cometary swarm is thin, and is passed through in a few hours; the stars are seen to radiate from the corresponding point of the sky for only that length of time: but there are other radiant points which have a duration of several months, and this is the case notwithstanding the constantly changing direction of the earth's motion in space.[33] Since the position of the radiant point in the sky as seen by a terrestrial observer depends not only on the direction in which the swarm is moving, but also on the velocity and direction of motion of the observer through space, it is easily seen that a radiant point having a fixed position during some months corresponds to something quite distinct from a cometary swarm. It has been suggested by Mr. W. F. Denning (1899) that in some cases a long-continued radiant point may really be due, not to a single swarm, but to successive swarms not physically associated with each other. On the other hand, Professor H. H. Turner has shown that the average effect of the earth's attraction on a meteorite passing near it is to change only the _position_ in our orbit at which we meet the meteorite (i.e. the time of year), not the relative-direction of motion or the relative speed; hence, a swarm of such meteorites must be spread out, in the course of ages, into a succession of rings, all of them equally inclined to the earth's orbit, but intersecting it at different places; the radiant point will then be of long duration. Professor A. S. Herschel[34] made the suggestion that the radiant points of long duration may have resulted from the passage, in bygone epochs, of quickly moving streams of cosmical matter through a ring of small bodies circulating, as satellites, round the earth.

[Sidenote: Daily and yearly maxima of shooting stars.]

62. The rotation of the earth round its axis is such that the part furthest from the sun, for which it is therefore midnight, is moving in the same direction as the earth in its orbit; whence, at the part of the earth most forward in the orbit it is sunrise, and at the part most backward it is sunset. Thus, as Schiaparelli pointed out, the meteorites which enter the atmosphere in the first half of the night are more or less following the earth in its orbit, and have their velocity relative to the earth diminished by the earth's own motion of translation; they are thus less likely to produce shooting stars than those which enter the atmosphere in the second half of the night and are travelling more or less oppositely to the earth as it moves in its orbit, and have their relative velocity increased. Hence, if the directions of flight of meteorites were uniformly distributed in space, the number of shooting stars hourly visible at one place, a number which would be constant if the earth were at rest, would gradually vary during the night, reaching a maximum about 3 A.M.

Also, as the point in space towards which the earth is moving in its orbit varies in height above the horizon during the year, being highest in autumn and lowest in spring, the number of shooting stars hourly visible at one place will gradually vary from night to night, reaching a maximum in the former season and a minimum in the latter, if the directions of flight of the meteorites be themselves uniformly distributed in space.

[Sidenote: The breaking up of comets.]

63. The history of Biela's comet[35] is of great interest as throwing light on the relationship of comets and swarms of meteorites. Though already observed in 1772 and in 1806, this comet was not recognised as periodic till it was seen by Biela in 1826, when its orbit was determined. On its returns in 1832 and 1845 it was found in its calculated positions, but in the latter year was seen to be double, a small comet being visible beside a larger one. Vast changes took place during the time the companions were visible. The smaller one grew both in size and brightness, each threw out a tail, the smaller threw out a second tail, afterwards the larger showed two nuclei and two tails, then the smaller became the brighter of the two companions; next three tails were shown by the primary, and three cometary fragments were visible round its nucleus. On the next return, in 1852, the two comets were farther apart, one being more than a million miles ahead of the other. The next favourable return was to be in 1866, and the orbit was by this time so well known that the positions of the two companions could be calculated beforehand with great precision; owing to the changes which had been visibly taking place, the arrival of the comets was looked forward to with great interest by astronomers. But neither in 1866, nor on the next occasion in 1872, were they to be seen in their calculated positions, and a careful examination of the whole sky failed to lead to their discovery.

The connexion between several comets and meteoritic swarms having in the meantime been established, it was now surmised that Biela's comet might have been scattered along part of its path, and that some evidence of the dispersal might perhaps be obtained on the next occasion, November 27, 1872, of the passage of the earth across the comet's orbit. In fact the star-shower of that date, with a radiant point corresponding to the orbit of Biela's comet, was observed to be much more dense than usual, the stars shooting across the sky at the rate of a thousand an hour for several hours.

[Sidenote: Passage of the earth through a comet.]

64. Klinkerfues, a German astronomer, was struck with the idea that if this star-shower were really due to the passage of the earth through a moving swarm of meteorites, the latter might possibly be visible as it departed from our neighbourhood. The swarm having come from a radiant point in the northern sky, after passing the earth would need to be sought near the opposite point in the southern sky; he telegraphed, therefore, to the Madras observatory, asking Pogson, the astronomer, to search for the swarm in the direction opposite to the radiant point. The search was successful; on two mornings a small comet was distinctly seen, and on the second morning it showed a tail with an apparent length equal to one-fourth the apparent diameter of the moon. Bad weather came on, and the comet got away without being again seen. The two Madras observations agree with a motion in the orbit of Biela's comet, and show that the earth had passed excentrically through the small comet seen by Pogson. This small comet was probably a third fragment of Biela's, for it was 200 million miles behind the calculated position of the first two. From these two observations it is inferred that a swarm of meteorites, though only manifesting itself by a star-shower when passing through the earth's atmosphere, at some distance from us may be visible as a comet by reflected sunlight.

[Sidenote: Fall of a meteorite during a star-shower.]

65. A dense star-shower[36] recurred on the same day of the month (November 27) in 1885, the principal part being over in six hours. The hourly number visible at one place at the time of greatest density was estimated at 75,000. In the densest part of the stream, the average distance of the individuals from each other was about twenty miles.

During this star-shower a piece of iron weighing about 8 lbs. was seen to fall at Mazapil in Mexico:[37] in external characters and chemical composition it is similar to the other meteoric irons: the simultaneity was probably accidental.

[Sidenote: The reason of its rarity.]

66. It may be asked why, if star-showers are caused by the entry of solid bodies into our atmosphere from without, there is only one authentic instance of material being actually seen to fall and being picked up during such a shower. As it is absolutely beyond question that star-showers do come from outer space, we can seek an explanation only in the size or speed of the entering individuals, or in the nature of their material. A sufficient reason is to be found in the small size of the individuals; for the meteorites which actually reach the ground rarely weigh more than a few pounds, and are often quite minute; a small diminution of the original individual would thus ensure its complete destruction before the planetary velocity was exhausted: that the individuals of a swarm are extremely minute follows from the fact that the total mass of the biggest swarm is small, while the number of the individuals seems almost infinite.

[Sidenote: Large and small luminous meteors essentially similar.]

67. Between the small silent shooting star visible only with the telescope and the large detonating meteorite-yielding fireball there is every gradation; during the star-showers themselves many fireballs of great size and brilliancy are seen, while the smaller individuals appear in no way different from the solitary shooting star. The luminous meteors, large and small, are in the upper atmosphere, few higher than 100 miles, few lower than 30 miles from the earth's surface; they all have velocities of the same order of magnitude, comparable with that of the earth in its orbit; in each there must be a solid body, as is proved by the long path in the sky, for attendant gas or vapour would be immediately scattered or burnt; large and small present similar varieties of colour, and leave similar luminous trails; examination with the spectroscope teaches us that the light of the meteors is such as would result from the ignition of such meteorites as have actually reached the ground. The frequent absence of detonation may likewise be due in many cases to the small size, or small relative velocity, of the entering meteorite.

[Sidenote: The light of a comet.]

68. That part of the light of a comet is reflected sunlight is confirmed by examination with the spectroscope, in which instrument is seen a feeble continuous spectrum crossed by dark lines, identical with those afforded by the direct light of the sun. But a comet is also more or less self-luminous; for, in addition to the continuous spectrum, there are bright flutings and bright lines to which much attention has been given. The three ordinary bright flutings were found by Sir William Huggins in 1868 to be identical with the spectrum obtained when an electric spark is passed through olefiant gas, and they are now recognised as due to carbon. The carbon is presumed to be combined with hydrogen, sometimes also with nitrogen; in the case of comets approaching very near the sun, the lines of sodium, and others which have been supposed to be iron-lines, are seen.[38]

[Sidenote: Tait's suggestion.]

69. The discovery made by Schiaparelli proves, as already pointed out, that there is a relationship between comets and meteoritic swarms; Schiaparelli himself held the view that a comet and its attendant swarms are merely of identical origin. In 1869[39] Tait discussed, from a purely dynamical point of view, the question as to whether the swarm of meteorites attending a comet may not really be part of the comet itself; he showed that many cometary characters can be mechanically explained on the assumption that comets are really swarms of small meteorites, and pointed out that the self-luminosity may be produced by the heating of the individuals through collision with each other.