THE CHEMICAL NEWS

VOL. CXIX. No. 395

ON THE

PHOTOGRAPHIC SPECTRA OF METEORITES.*

By Sir WILLIAM CROOKES, OM, FRS, LL.D., &c.
(Concluded from p 55)

GNARENBURG (from Hanover, Prussia). The photographed spectrum of this aerolite extends from A 2714 to 5896, and includes in addition to the lines of iron 2 aluminium, 3 calcium, 31 chromium, magnesium, 3 manganese, 50 nickel, 2 potassium, 6 sodium.

Hvittis (from Abo, Finland). The photographed spectrum of this aerolite commences at A 2852 and extends to 5896; it includes in addition to the lines of iron I calcium, 30 chromium, 3 magnesium, 3 manganese, 31 nickel, 4 sodium.

Jelica (from Servia). The photographed spectrum of this aerolite commences at A 2852 and extends to 5895, and includes in addition to the lines of iron I calcium, 29 chromium, 3 magnesium, 3 manganese, 34 nickel, 2 potassium, 4 sodium.

Jhung (from Punjab, India). The lines shown in the photographed spectra extend from A 2883 to 5896, and include in addition to the lines of iron 2 aluminium, 2 calcium, 22 chromium, 3 magnesium, 3 manganese, 43 nickel, 2 potassium, 6 sodium.

Kansada (from Ness County, Kansas, U.S.A.). The photographed spectrum of this aerolite commences at A 2852 and extends to 4652. and includes in addition to the lines of iron i calcium, 26 chromium, I magnesium, 3 manganese, 45 nickel, 4 sodium.

Khairpur (from India). The photographed spectrum of this aerolite commences at A 2852 and ends at 5896, and includes in addition to the lines of iron 2 aluminium, 3 calcium, 29 chromium, 4 magnesium, 3 manganese, 41 nickel, 2 potassium, 6 sodium.

Launton (from Bicester, Oxfordshire). The whole of the 5 grm. of material available quickly burnt out and only short exposures were possible. The photographed spectrum extended from A 3302 to 5896, and includes in addition to the lines of iron 2 aluminiuin, 2 calcium, 21 chromium, 2 magnesium, 3 manganese, 20 nickel, 2 potassium, 6 sodium.

Mount Brown (from Evelyn County, New South Wales). The photographed spectrum of this aerolite extends from A 2599 to 6643, and includes in addition to the lines of iron 3 aluminium, 3 calcium, 28 chromium, II magnesium, 3 manganese, 55 nickel, 6 sodium.

Mount Dyrring (from Durham County, New South Wales). The lines found in the photographed spectrum of this aerolite extends from A 2482 to 4552, and includes in addition to the lines of iron 25 chromum, 17 magnesium, 3 manganese, nickel, 2 potassium, 6 silicon, 2 sodium.

Nammianthal (from South Accot, Madras, India). The photographed spectrum of this aerolite extends from A 2716 to 4652, and includes in addition to the lines of iron 2 calcium, 23 chromium, 3 magnesium, 3 manganese, 50 nickel, 2 sodium.

Parnellee (from Madras, India). The photographed spectrum extends from A 2788 to 5896, and includes in addition to the lines of iron I alumunium, 24 chromium, 3 magnesium, 3 manganese, 50 nickel, 2 potassium, 6

sodium.

Before discussing in detail the composition of these aerolites as revealed by their arc spectra it is well to record one or two points of general importance. Keeping the current and voltage constant, the number of lines recorded in the arc spectrum of a compound substance depends in large measure upon the time-length of the "exposure" and the sensitiveness of the photographic plate. For instance, if in an exposure of say five minutes there are a mass of lines, having some that are very much over-exposed and some that are only just visible, by making another exposure of the same material for double the time the very faint lines would appear stronger, and others that in the shorter exposure were not visible would come into view, and so on. But in practice a limit is fixed, governed partly by the demands of the strong lines and partly by the amount of the material at one's disposal. This latter limitation, the amount of material available, is very potent when dealing with such substances as rare meteoric

stones.

From an examination of the whole set of aerolite spectra it appears that the proportion of nickel to iron is generally constant. There is a nickel line at λ 3619'391 and an iron at 3018919; these two lines form a close pair, and in twenty-seven out of the thirty spectra they remain in the the general character of the spectrum. But in three, same relative intensity, being faint or strong according to Bustee, El Nakhla, nnd Aubres, while the iron line shows faint. (In connection with this proportionality between as usual, the nickel line is either absent altogether or very the lines of iron and nickel, I have examined the spectra of a number of siderites, and in the majority of instances the proportionality remains constant; but there are a few cases in which the nickel line is fainter than that of iron and others in which it is stronger).

In several of the aerolites in my list it will be seen that the total number of lines recorded is below the average. The aerolite Bustee is an instance. (Through the kindness of the Trustees of the British Museum and of Dr. Prior, three successive portions of this aerolite, making rapidity with which it burnt away in the arc prevented me 12:171 grm. in all, were placed at my disposal). The from giving the length of time for each exposure necessary to bring out many of the fainter lines. A rather interesting point came out in connection with this aerolite.

In the early work my stock of material amounted only to about 2 grm.-and the length of time given for an exposure was necessarily short-it was noticed that although the stronger chromium lines came out in good intensity, the nickel lines were absent. Having obtained more material I was able to get photographs with longer exposures; the nickel lines could then be seen, although as a whole much more faintly than those due to chromium. This gave rise to some experiments upon the relative photographic intensity of the two metals under the conditions in which I was working.

The chromium spectrum is fairly rich in lines. Under the conditions prevailing in these experiments, in addition to a large number of faint lines there are recorded about one hundred strong ones, and of these there are three groups, which might be called the dominant lines of the elements; the strongest of these groups is composed of the three lines at A 4254'50, A 4274 97, and λ 4289'90; the next strong group is also three lines at a 3578.840, 3593 633, and A 3605 478, and the faintest consists of two lines, A 4862 02 and A 4870'9.

Some pure chromium was taken and reduced to powder, some pure electrolytic nickel was prepared in the same way, and a mixture was made of silver, kaolin, and yttria, with 10 per cent each of Ni and Cr. This was pressed into a button and the spectrum taken as in the case of the aerolite. It was at once seen that the

chromium groups came out in greater intensity than the adjacent nickel lines.

And to this was added an equal weight of pure silver powder. The arc spectrum was then photographed under conditions similar to those of the aerolites.

It was seen that the chromium groups were still the most prominent lines in the spectrum, those of nickel being comparatively faint.

Keeping the other ingredients constant and gradually reducing the proportion of chromium several more experiments were made, and it was found that when the chromium had been so reduced that it was only 0.16 per cent, the lines A 4862.02 and A 4870 9 were no longer visible; a further experiment with electrodes containing only o'i per cent chromium distinctly showed the other two groups.

These experiments show that it is easy to detect the presence of chromium in an arc between electrodes that contain only 5 parts Cr in 10,000.

It is not easy to see why in subjecting the two elements nickel and chromium to the heat of the electric arc it should produce so much more intense atomic disturbance in one case than in the other-the melting points of the two elements are not very different-it may be, however, that the volatilisation points differ. I have pursed this matter in connection with the examination of a large number of meteoric irons (siderites) that will form the subject of a further communication.

(Without anticipating this communication on siderites, I desire to point out the remarkable fact that although the chromium lines arpear, sometimes very strongly, in all my aerolite spectra they do not appear in any of the siderites I have so far examined-except in one instance, that of "Zacatecas " Although of course the nickel lines are always visible-from the spectroscopic examination of a number of specimens of iron, to which I had added decreasing amounts of chromium, I found that the dominant lines of chromium were quite visible even when the amount present was no more than o'0175 per cent. This is due to the fact that chromium is absent from the nickelirons but almost universally present in the aerolites in the form of chromite. In Aubres, &c., there is probably more chromite than nickel-iron. In the magma which produced meteorites all the Cr was converted into oxide or sulphide, giving rise to the mineral chromite or daubreelite and leaving no Cr for the nickel-iron).

Although these experiments only make it possible to form an approximate estimate of the amount of chromium present in these aerolites it would seem to lie between o6 per cent and o'r per cent.

In the previous experiments the proportion of nickel which was kept the same in all mixtures was evidently in considerable excess of that contained in any of the aerolites; the experiments were therefore continued, blocks made in the same way as in the experiments with chromium were used as electrodes for the arc. The proportion of nickel was from 2 per cent down to o'04 per cent-photographs of the spectrum were made under the usual working conditions. The portion of the spectrum taken was that containing the chromium group, A 3578.840,

A 3593 633, A 3605 478, and the closely adjacent nickel lines; it was found that when the amount of nickel was reduced to o'04 per cent the line 3619'391 was only just visible.

The information gained from these mixtures of chromium and of nickel made it possible to obtain an approximate estimate of the quantities of these elements present in the three aerolites, El Nakhla, Aubres, and Bustee. A mixture containing chromium o'25 per cent and nickel 0:04 per cent gave a photograph which for Cr-Ni was practically identical with that of Aubres.

The estimation made in this way cannot be more than a good approximation on account of unavoidable variations, due to intensity of light and time of exposure; also to conditions of photographic development, which cannot be exactly controlled, but the indications given by the relative intensities of closely adjacent Ni and Cr lines when they occur on the same film are very much more exact, and the result proves that in these aerolites the element chromium exists in greater abundance than nickel. To this fact must be added that of the almost total absence of chromium in the familiar nickeliferous irons of the siderites.

In a general survey of the spectrum analyses of the 30 aerolites which forms the subject of this paper, the most striking fact is their similarity in composition, and the small number of elements represented. Making full allowance for wide differences in the photographic activity of the arc spectra of the elements, it is remarkable that we only see with certainty the lines due to some ten bodies, and of these ten, four only-iron, chromium, magnesium, and nickel - appear to be present in quantity, With three exceptions, Bustee, El Nakhla, and Aubres, the proportion between these elements appears to be practically the same in all.

This suggests that these earthy meteorites must have a common origin, and that origin might be due to the disruption of a body in which the process of cosmial evolution has been completed; in short, may we not conclude that the aerolites are fragments of a finished and cooled planet. It is possible that we have in our museums fragments of a world unrealised- world that at one time had its place between Jupiter and Mars in our planetary system.

From the results of my unfinished notes on the spectrum analyses of meteoric irons, to which I have already referred, it would appear that either the siderites are of a different origin, or that they have constituted the solid nucleus or core, from which by some process at present unknown the chromium and other elements had been separated, leaving the magnetic elements iron and nickel as the familiar ferro-nickel meteorites.

IS THE ELECTRICAL CONDUCTIVITY OF THE ELEMENTS CONDITIONED BY THE PRESENCE OF ISOTOPES?

A SUPPLEMENTARY NOTE.*

By F. H. LORING.

If the percentages of isotopes, which were given in my previous paper, be plotted additively with the resistivities, also given, a straight-line curve is obtained, showing a relation which seems to imply that the theory of wholenumber isotopes as applied to the ordinary elements is not without some foundation of fact. The accompanying diagram shows the relation.

Commenting on the diagram, cadmium is an exception in not fitting in with the percentage rule suggested, but as there is a peculiarity in Northrup's temperature-resistance

To be read in conjunction with my paper in the CHEMICAL NEWS, July 11, 191, cxix., 14.

5 to 5 as given in the CHEMICAL NEWS (1914, CX., 25)on the foregoing value gives one that falls exactly on the straight-line curve B. The resistivity of thailium at fusion is obtained from Vincentini and Omodei's paper (Atti. Accd. Torino, 1889-90, xxiii., 30), but these investigators give no series of values which can be plotted, so that some uncertainty obtains with regard to the use of the single value calculated from the data given.

It will be seen that in those cases in which the -0values are taken (see inset diagram) the percentages for the lower-mass isotopes are selected in order to give

carves for zinc and cadmium, and the curve of the former is not confirmed by Tsutsumi, the value for cadmium may be queried. Should, however, the abnormal characteristic of cadmium eventually be established, then a closer study of the exception might prove of interest. Gallium contracts on melting, and the resistance of this element in the state of maximum density close to the melting-point is 27.2 x 10-6 (A. Guntz and W. Broniewski, Comptes Kendus, cxlvii., 1474). Plotting 50 per cent-which is the most regular value involving an isotopic proportion of

continuity to the curves B anic. The choice of resistivity is, however, in accord with the principle of nearest state, since there is likely to be nearer comparableness if the values be taken when the elements are in their most dense state at fusion. The example of gallium given above makes this point clear. The plotted values for the resistivities of cadmium, tin, and lead are those of Northrup (loc. cit., previous paper). Tsutsumi's values for tin and lead are respectively 23.8 and 48.1. There is therefore little to choose between the values of these investigators. The