organs besides foliage are found, it being by no means absolutely certain that because the foliage is identical the species are so. The discussion raised by Prof. Haughton, and continued by Prof. Duncan and Mr. Wallace, seems therefore hardly worth prolonging, since it is based upon an assumption that is only probably correct. But even if the identity were proved, a single species is not satisfactory evidence of former temperature.

I am indebted to Mr. Winslow Jones for the only information that I have yet obtained about the growth of either species in England. He recollects a small tree of A. excelsa, growing near the water's edge in a garden on the upper portion of Falmouth Harbour, which he believes died three years ago. He has seen flourishing trees at Naples, Cintra, Malta, and Algiers, but even Northern Italy seems beyond the range of successful cultivation. Of the two A. Cunninghami seems the more tender, though possibly its le s symmetric growth may have excluded it from many gardens. In Madeira it grows generally best close to the sea and in sheltered places.

Lindley was mistaken in regarding the two species as one. All the needle-leaved (Eutacta) section of Araucaria are certainly closely allied, for the species, however distinct in other respects, possess two kinds of foliage, that of the young plants being identical in all : yet otherwise the species are clearly and distinctly marked off from each other.

It

With further regard to the identification of the Bournemouth foliage with Araucaria, I find that Massalongo1 gives an excellent photograph of the same foliage from Chiavon, in North Italy, and of an immature cone consisting of 250 scales. Although existing Sequoias have cones with from 16 to 20 scales, Schimper says: "Il est sans aucun doute un Sequoia et peut-être identique au S. Sternbergii. Les cônes ont la plus grande ressemblance avec ceux du S. gigantea" (Pal. Végétale," vol. iii. p. 573). I am beginning to lose all faith in the so-called science of paleo-botany as worked out by our Teutonic brethren. Not only is the above quɔtation an absurdity, for which Heer is responsible, but I fail to see any good evidence to support the change made by Heer from Araucaria Sternbergii to Sequoia Sternbergii. The foliage is more Araucaria like than Sequoia-like, and has been found associated with an Araucaria cone, but never with any Sequoia cones. has nothing to do with the Icelandic foliage, neither with the Upper Miocene foliage from Turin, nor that from Bilin nor Oeningen. The true Araucaria Sternbergii characterises a wellmarked horizon, that of the Newer Eocene or Oligocene in Central Europe, and has been found at Barton in Hampshire; it differs from the Middle Eocene form (A. venetus, Mass.) of England and Italy in the needle-like leaves hugging more closely to the branchlet, as the latter differs in its turn from the Araucaria of the Grès du Soissonnais, which has needles very widely opened out. This progressive change may have taken place pari passu with the changing climate. At Sheppey, where foliage is plentiful, I have met with a beautifully-preserved axis of an Araucaria cone with the basal scales attached, exactly as we find them in the existing species.

Now with regard to Mr. Wallace's letter, I pointed out in NATURE, vol. xix. p. 126, that the Tertiary fossil plants, even of the Eocene, require at most an increase in temperature of 20°, and that the land connection between Europe, Greenland, and America, which there is reason to suppose existed then, would, by shutting out Arctic currents, have produced more than the required increment. If this theory appeared for the first time in my article, however clumsily I may have worded it, and if it has been of use to Mr. Wallace, it is only fair that the fact should be acknowledged, while if it has escaped his notice he will perhaps pardon my now drawing his attention to it. At the same time the publication of the Tertiary flora of North-East Siberia, which I had not then seen, and of Saghalien, has modified the views I put forward in a manner which I trust I may shortly find time to explain. J. STARKIE GARDNER

Chalk

MR. WALLACE's theory that chalk was deposited in comparatively shallow water requires careful examination before it is accepted by geologists. I do not think he has given sufficient evidence to bear out his views which are necessary to his theory of continents.

Mr. Wallace cites the resemblance between chalk and Globigerina-ooze, namely

The similarity of the minute organisms found to compose a "Speci nen photographicun." Verona, 185). Plate xxi.

2 Actually described as Araucarites, a useless modification in this instance.

considerable portion of both deposits; several species of Globigerina appearing to be identical in the chalk and the modern Atlantic mud; the presence of Coccoliths and Discoliths in both formations; the abundance of Sponges in both; the presence of Porifera vitrea, the nearest representative of the Ventriculites of the white chalk; the resemblance of the forms of Echinoderms; and without attempting to reconcile these with a shallow sea-deposit, he proceeds to state the case on the other side. This consists of the difference in analysis between chalk and Globigerina-ooze, the former containing more carbonate of lime and less alumina, the presence of silica in the Globi gerina-ooze being perhaps counterbalanced by the flints in the chalk. greater proportion of alumina certainly points to different conditions, which Mr. Wallace considers to be that chalk is the very fine mud produced by the disintegration of coral-reefs, and mentions a deposit resembling chalk at Oahu in the Sandwich Islands and the deposit in several growing reefs, without however attempting to show that there is any probability that the remains found in these would bear any resemblance to the Sponges and Echinoderms of the chalk, or why we find no remains of these Cretaceous coral reefs.

The

Mr. Wallace does not state in what the greater resemblance between chalk and Globigerina-ooze of shallow over deep water consists, but he looks on it as 66 weighty evidence."

Mr. Gwyn Jeffries, he says, finds all the Mollusca of the chalk to be shallow-water forms, many living at forty to fifty fathoms, some confined to still shallower waters, while deep-sea forms are absent. The late Dr. S. P. Woodward considered that Ammonites probably lived in water not over thirty fathoms ; and these facts are as difficult to reconcile with Mr. Wallace's views that chalk was deposited in a sea of not over a few thousand feet as in a deeper sea.

The rareness of corals and absence of coralline beds of the age of the Lower or Upper Chalk is an important objection to the theory that chalk was deposited similarly to the Oahu chalk, the beds of Maestricht and Faxoe being above the chalk, and the former are not even conformable with it.

The point I think is still an open one, whether we shall accept Mr. Wallace's views that chalk was deposited in a comparatively shallow sea and not very far from land, or in a deep sea, the immense break between the chalk and Eocene beds giving ample time for very considerable alteration to have taken place in the disposition of land in the interval. I send this letter in the hope that a discussion on the point may elicit new facts bearing on the subject. S. N. CARVALHO, JUN.

8, Inverness Terrace, Kensington Gardens, W.

On Estimating the Height of Clouds by Photography and the Stereoscope

THE great practical value of meteorological science and the desirability of extending its usefulness by the collection of data relating to atmospheric current will perhaps be sufficient excuse for asking attention to anything likely to promote this end.

In studying the currents and other peculiarities of the atmosphere a method of estimating the height, motion, and character, as also the position with respect to each other, of each stratum of cloud, is a requirement of almost paramount importance, the value of the means employed being proportional to the number of particulars provided in its record, and the facility with which any set of observations can be compared to another at any future period. With such ever-changing subjects as clouds in constant motion, and having no strongly-defined marks, the use of theodolites is almost out of the question, and the sextant and mirror process for similar reasons would be a very tedious operation.

These considerations have induced me to endeavour to make use of photography and the stereoscope, the former to secure a couple of simultaneously-exposed photographs at the extremities of a base line, and the latter to observe them reproduced apparently solid for the respective distances of the points composing the picture to be measured when superimposed on a scale of distances and placed in it. The base line is thus practically reduced to the width of the eyes, and the difficulties arising from motion eliminated.

The recording apparatus consists of a base 50 or 100 feet long, constructed of wood and turning on a pivot at the centre of its length, its extremities being suitably supported by a framework of wood or other material upon which they could easily roll. The small cameras for the ends of this are each to be hinged at the back of its base to a second board having a graduated quadrant and rackwork erected from one of its sides for adjusting

the camera to any degree of altitude. These supplementary boards are then ivoted at the centre of part of a divided circle, previously inlaid in the wood at the extremities of the base line, in such a manner that a line passing through the axis of the lenses would cut the pivots. The cameras thus furnished can be adjusted with ease to any vertical or horizontal angle. These angular adjustments of the two instruments must always coincide, with the slight exception that the horizontal ones must make internal angles with the base included between them, or, in other words, the lenses of both require to be directed to a point opposite to the centre of the base line.

The cameras also require their rapid exposing shutters to be electrically connected, to ensure the pair of sensitive plates being impressed at the same instant, and each dark lide employed to have a fine wire strained at its centre from top to bottom imme diately in front of the prepared plate, and as close as possible to it without touching. The transparent lines produced in the developed negatives by these wires will constitute the zero of distance of any pair, and during the operation of reading off must be made to agree with similar ones on the scale of measurements obtained as follows:

Upon a large cardboard rule a number of squares in fine black lines, one inside the other, and each one slightly out of the centre of its predecessor to the right hand, the outside square being then divided with a line at a tenth part of its diameter to the left of its centre. This line will indicate the zero of the scale. After placing a distinguishing mark or number in the corner of every square for purposes of identification, the cardboard will be ready to be photographed and reduced at the same time to the intended size of the cloud negatives. Two transparent positives copied from this and observed when placed side by side in a suitable stereoscope with the edges representing the left-hand one of the cardboard together, will appear in that instrument with the lines composing the zero only a few inches away, and the squares as a succession of vertical planes commencing some distance from that and receding from the eye in the order of greater to less, each one representing its own distance in space.

To find the value of these distances it will be necessary to focus the two cameras upon some terrestrial objects whose distances can be measured by any of the known methods, and negatives taken. The two resulting landscapes, when placed in the stereoscope, each superimposed face to face upon its respective scale, and the fine vertical lines of the whole made to Occupy one apparent distance, an operation offering but little difficulty, every object or point of the landscape will be found to stand out in the vertical plane suited to its own distance, the relation between them being noted for the values found by measurement of the one to be marked upon the other. As a scale prepared thus would be of no value for any other angle at which the cameras might be placed, it would be most convenient to make use of two or three angles only, more being quite unnecessary, and prepare a scale for each, or one with a reference table of values for the respective angles would suffice. Again, in respect of altitudes. As the terrestrial measurements would only be absolutely accurate for those of clouds in the zenith, or of them, if it were possible, from the earth's centre in any direction, the tables of reference would have to include calcu lated corrections for altitude, or the graduations could be valued for the most useful degrees by experimental means.

It will be gathered from the above that the constancy of length of the base line can be ascertained, and corrected if necessary, by taking a couple of views of the same landscape for comparison with the preceding pair; slight fluctuations of length would not how ever be of much consequence in dealing with the com. paratively coarse measurements of thick masses of cloud floating in so short a distance as the few miles of atmosphere capable of forming them consists.

To ascertain the height of clouds photograph a pair of negatives, and place these in the stereoscope with a pair of scale plates agreeing with the angle at which they were taken, and adjust as for the landscapes described above. The data required may then be read off by noting the vertical plane each stratum occupies.

Prints of these negatives should afterwards be made for the particulars of height, direction of motion of the respective layers, point of compass, wind rate, state of barometer, thermometer, and general remarks upon the weather, to be recorded upon them for comparison or circulation.

Meteorological observatories fitted with such an addition to their present splendid collection of instruments would have their

Correction of an Error in "Island Life"

My friend Dr. Günther has kindly called my attention to an extraordinary error at p. 322-323 of my "Island Life," where I state that the Loch Killin Charr (Salmo Killinensis) inhabits a lake in Mayo County, Ireland; instead of a small lake in Inverness-shire, 2000 feet above the level of the sea, as given in Dr. Günther's original description in the Proceedings of the Zoological Society, 1865, p. 698. On referring to my MSS. notes for this part of my work, I find that the babitat was first correctly given, but subsequently scored out and altered to the erroneous Irish locality! Why this was done I cannot now discover; and I can only regret that I should have fallen into so palpable an error, and request such of the readers of NATURE as possess my book to make the necessary alterations. ALFRED R. WALLACE

Natural Science for Women

WILL you allow me to supplement your kindly reference to the instruction in physical science given to women in Bedford College, London, by the statements that for the last two sessions a class in biology has been conducted there by Mr. Charles Stewart of St. Thomas's Hospital Medical School. The course of study is in every sense a practical one, with special reference to the Preliminary Scientific and First B.Sc. examinations at the University of London, and the best testimonial to the excellence of the instruction in these various subjects is furnished by the remarkable success during the present year of the Bedford College pupils at the University examinations, a success not less marked in the Science than in the Arts examinations. ALFRED W. BENNETT

Movements of Leaves

:

A YEAR ago we had in our conservatory a healthy young plant of Acacia mollissima. It bore no flowers, but consisted of a simple axis adorned with the soft feathery leaves of its genus, which closed up at night. Our gardener however thought it would improve in appearance if it could be made to bear a few branches; and with that view he cut it back. His end was achieved a new stem shot up from the section, and graceful limbs were thrown out in turn by it. But along with this a strange result followed the fresh leaves borne by the new stem and by the branches now closed at night, while the old leaves below the section ceased to do so. These lower leaves have long since fallen off, but the upper ones kept to their habit, and at the present time all fold up at dusk save a few of the very oldest, which only partially shut, or, in one case, do not shut at all. When our plant was cut back it stood three feet high; now it stands seven: which shows that the vigour of the plant as a whole in no wise diminished by the operation. Chislehurst, December 23

DUS

M. L. ROUSE

ON DUST, FOGS, AND CLOUDS1 UST, fogs, and clouds seem to have but little connection with each other, and we might think they could be better treated of under two separate and distinct heads. Yet I think we shall presently see that they are more closely related than might at first sight appear, and that dust is the germ of which fogs and clouds are the developed phenomena.

This was illustrated by an experiment in which steam. was mixed with air in two large glass receivers; the one receiver was filled with common air, the other with air which had been carefully passed through a cotton-wool filter and all dust removed from it. In the unfiltered air the steam gave the usual and well-known cloudy form of condensation, while in the filtered air no cloudiness whatever appeared. The air remained supersaturated and perfectly transparent.

The difference in the behaviour of the steam in these two cases was explained by corresponding phenomena, 1 Abstract of a paper read to the Royal Society of Edinburgh, December 20, by Mr. John Aitken. Furnished to NATUKE by the Council of the Society.

in freezing, melting, and boiling. It was shown that particles of water vapour do not combine with each other to form a cloud-particle, but the vapour must have some solid or liquid body on which to condense. Vapour in pure air therefore remains uncondensed or super-saturated, while dust-particles in ordinary air form the nuclei on which the vapour condenses and forms fog or cloudparticles.

This represents an extremely dusty condition of the air, as every fog and cloud-particle was formerly represented by a dust-particle, which vapour by condensing upon it has made visible. When there is much dust in the air but little vapour condenses on each particle, and they become but little heavier, and easily float in the air. If there are few dust specks each gets more vapour, is heavier, and falls more quickly.

These experiments were repeated with an air-pump, a little water being placed in the receiver to saturate the air. The air was then cooled by slightly reducing the pressure. When this is done with unfiltered air a dense cloudiness fills the receiver, but when with pure air no fogging whatever takes place, there being no nuclei on which the condensation can take place. In this experiment, and in the one with steam, the number of cloudparticles is always in proportion to the dust present. When the air is nearly pure and only a few dust-particles present, then only a few cloud-particles form, and they are heavy and fall like fine rain,

The conclusions drawn from these experiments are: (1) that whenever water vapour condenses in the atmosphere it always does so on some solid nucleus; (2) that dust-particles in the air form the nuclei on which the vapour condenses; (3) that if there was no dust there would be no fogs, no clouds, no mists, and probably no rain, and that the supersaturated air would convert every object on the surface of the earth into a condenser on which it would deposit; (4) our breath when it becomes visible on a cold morning, and every puff of steam as it escapes into the air, show the impure and dusty condition of our atmosphere.

The source of the fine atmospheric dust was then referred to, and it was shown that anything that broke up matter into minute parts would contribute a share. The spray from the ocean, when dried and converted into fine dust, was shown to be an important source. matter also probably contributed a proportion. Attention was then directed to the power of heat and combustion as a source of this fine dust.

Meteoric

It was shown that if there is much dust then each particle only gets a little vapour condensed upon it, that when the particles are numerous they become but little heavier, and easily float in the air, and give rise to that close-packed but light form of condensation which constitutes a fog, and therefore whatever increases the amount of dust in the air tends to increase fogs, and that when the dust-particles are not so numerous the cloud-particles are larger and settle down more quickly.

It was shown that by simply heating any substance, such as a piece of glass, iron, brass, &c., a cloud of dust was driven off, which, when carried along with pure air into the experimental receiver, gave rise to a dense fog when mixed with steam. So delicate is this test for dust that if we heat the one-hundredth of a grain of iron wire the dust driven off from it will give a distinct cloudiness in the experimental receiver, and if we take the wire out of the apparatus and so much as touch it with our fingers and again replace it, it will again be active as a cloudproducer. Many different substances were tried, and all were found to be active fog-producers. Common salt is perhaps one of the most active.

Heat, it is well known, destroys the motes in the air, and it might be thought that flame and other forms of combustion ought to give rise to a purer air. Such however is not the case. Gas was burned in a glass receiver,

and supplied with filtered air for combustion, and it was found that the products of combustion of pure air and dustless gas gave rise to an intensely fog-producing atmosphere. It may be mentioned here that the fogproducing air from the heated glass, metals, and burning gas were each passed through the cotton-wool filter, and the air was in all cases made pure, and did not give rise to cloudiness when mixed with steam

It will be seen that it is not the dust motes which are revealed to us by a beam of sunlight when shining into a darkened room, that form the nuclei of fog and cloudparticles, as these may be entirely removed by heat, and yet the air remain active as a cloud-producer. The heat would seem to break up the larger motes which reflect the light into smaller and invisible ones. When speaking of dust, it is to these infinitesimally small and invisible particles we refer. The larger motes which reflect the light will no doubt be active nuclei, but their number is too small to have any important effect.

It is suggested, and certain reasons are given for sup posing, that the blue colour of the sky is due to this fine dust.

The

Other experiments were made to test the fog-producing power of the air and gases from different sources. air to be tested was introduced into the experimental receiver and mixed with steam, and the relative densities of the fog produced were noted. It was always found that the air of the laboratory where gas was burning gave a denser fog than the air outside, and that the air outside varied, giving less fog during wet than during dry weather. The products of combustion of gas burned in a Bunsen flame, a bright flame, and a smoky flame, were all tested and found to be about equally bad, and all much worse than the air in which they were burned. Products of combustion from a clear fire and from a smoky one gave about equal fogging, and both much worse than the air of the room.

Experiments were made by burning different substances. Common salt when burned in a fire or in alcohol flame gave an intensely fog-producing atmosphere, but burned sulphur was the most active substance experimented on. It gave rise to a fog so dense it was impossible to see through a thickness of 5 cm. of it.

The vapours of other substances than water were tested to see if they would condense in the cloud form without nuclei on which to deposit. All the substances experimented on, which included sulphuric acid, alcohol, benzole, and paraffin, only gave a cloudy condensation when mixed with ordinary unfiltered air, and remained perfectly clear when mixed with filtered air, all these acting like water vapour.

Before referring to fogs, which have now become so frequent and aggravated in our large towns, it was pointed out that caution was necessary in applying the results of the experiments.

The conditions of a laboratory experiment are SO different, and on so small a scale, that it is not safe to carry their teaching to the utmost limits and apply them to the processes which go on in nature. We may, however, look to the experiments for facts from which to reason, and for processes which will enable us to understand the grander workings of nature.

It having been shown that vapour, by condensing on the dust-particles in the air, gives rise to a fogging, the density of which depends on the amount of fine dust in the air; the more dust the finer are the fog-particles, and the longer they remain suspended in the air. It having been also shown that all forms of combustion, perfect and imperfect, are producers of fog nuclei, it is concluded that it is hopeless to expect that, adopting more perfect forms of combustion than those at present in use, we shall thereby diminish the frequency, persistency, or density of our town fogs. More perfect combustion will, however, remove the pea-soup character from the fogs

and make them purer and whiter, by preventing the smoke which at present mixes with our town fogs and aggravates their character, and prevents them dissolving when they enter our rooms. Smoke descends during a fog, because the smoke particles are good radiators, and soon get cooled and form nuclei on which the water vapour condenses. The smoke thus becomes heavier and falls. This explains why falling smoke is often a sign of coming rain. It indicates a saturated condition of the atmosphere.

Sulphur when burned has been shown to be an intensely active fog-producer. Calculation shows that there are more than 200 tons of sulphur burned with the coal every winter day in London, a quantity so enormous as quite to account for the density of the London fogs. It is suggested that some restriction ought to be put on the amount of sulphur in the coal used in towns.

Before utterly condemning the smoke and the sulphur, it was pointed out that it would be necessary thoroughly to investigate and fully to consider the value of smoke as a deodoriser, and also the powerful antiseptic properties of the sulphurous acid formed by the burning sulphur. The air during fogs is still and stagnant. There is no current to clear away the foul smells and deadly germs that float in the air, which might be more deadly than they are, were it not for the suspended soot and burned sulphur. We must therefore be on our guard lest we substitute a great and hidden danger for an evident but less evil.

ON THE SPECTRUM OF CARBON

ALTHOUGH fifteen years have passed since the possibility of one substance possessing more than one spectrum was first suggested by Plucker and Hittorf, the question of the existence of double spectra cannot yet be considered as decided. One of the elements to which multiple spectra have been attributed is carbon, which was at one time supposed to possess four different spectra of these one has been shown to be due to oxide of manganese, a second to oxides of carbon, the origin of a third (obtained only from oxides of carbon) has hardly been discussed (though it may prove to be one of the true carbon spectra), and the other "carbon" spectrum-the best known of all-is the one first attributed to carbon by Attfield, but ascribed to acetylene by Angström.

In a paper read before the Royal Society, and of which an abstract is given in NATURE, vol. xxii. p. 620, Professors Liveing and Dewar describe experiments to prove that this spectrum is that of a hydrocarbon, and not of carbon itself; and also that certain blue bands, best seen in the flame-spectrum of cyanogen, are due to compounds of carbon and nitrogen, and not to carbon itself. They attribute to hydrocarbon (amongst others) the yellowishgreen group, which we will call y, of wave-lengths from about 5635 to 5478, and the emerald-green group, which we will call 8, of wave-lengths from about 5165 to 5082; and they attribute to nitro-carbon the two blue groups of wavelengths 4600 to 4502 and 4220 to 4158, which we will call and respectively.

As these result are directly opposed to my own experience, I have thought it necessary to repeat two of the experiments described in my paper on the carbon spectra in the Philosophical Magazine for October, 1869, under such conditions as to exclude (as far as lay in my power) all trace of hydrogen in the one case, and of nitrogen in the other.

The difficulty of supposing carbon to be present in the state of vapour at any temperature which we can command seems to be the chief reason why so many investigators think it necessary to attribute the spectrum in question (with experimental evidence or without it) to compounds of carbon. I am not aware that Ångström ever gave any experimental proof of his assertion that this spectrum was caused by acetylene.

On the other hand, the evidence that the spectrum is due to carbon is that first stated by Attfield, that if these lines "are absent in flames in which carbon is absent, and present in flames in which carbon is present," if they are "observable equally in the flame of the oxide, sulphide, and nitride as well as in the hydride of carbon," and if "present whether the incandescence be produced by the chemical force, as in burning jets of the gases in the open air or by the electric force, as when hermeticallysealed tubes of the gases are exposed to the discharge of a powerful induction-coil," then they "must be due to incandescent carbon vapour"; and if this is borne out by experiment the conclusion that the lines are due to carbon (as gas, liquid or solid) cannot be resisted, whatever may be the apparent impossibility of volatilising or even liquifying carbon, even by the most powerful current of electricity directed through it.

We must bear in mind how very small a quantity of a gas is often sufficient to give us a spectrum, and when the carbon spectrum is obtained by the decomposition of olefiant gas or cyanogen by passing sparks through the gas, the carbon certainly exists as gas in the compound which is decomposed, and before the liberated atoms unite together to form the molecules of the solid, there is surely no impossibility in their existing for the moment as gas-as gaseous carbon.

Ön an examination of Professors Liveing and Dewar's paper to ascertain the experimental evidence upon which the bands γ and & are attributed to hydrocarbon and not to carbon itself, we find it stated that "the green and blue bands characteristic of the hydrocarbon flame seem to be always present in the arcs, whatever the atmosphere. This is what we should expect if they be due, as Angström and Thalèn suppose, to acetylene, for the carbon electrodes always contain, even when they have been long heated in chlorine, a notable quantity of hydrogen."

Since then it is impossible to completely expel hydrogen from the carbon-poles, we must reject all the experiments in which the electric arc was observed in atmospheres of different gases, although "the green and blue hydrocarbon bands were seen more or less in all of them."

Turning then to other methods of producing the spectrum, we find it stated that in the flame of carefullydried cyanogen "the hydrocarbon bands were almost entirely absent" (they should have been entirely absent); "only the brightest green band was seen, and that faintly." Hence we are to infer, I suppose, that the bands y and d, so brilliant in the flame of cyanogen in air or oxygen, are due to the accidental presence of hydrogen (see the extract from Morren's paper, NAture, vol. xxii. p. 7. Dibbits also speaks of this spectrum as "by far the most magnificent" he has seen).

Next we have the experiment of burning hydrocyanic acid, in which, as we have hydrogen present, we expect to find the hydrocarbon bands brilliantly developed. But we find the result stated as "very much the same as that of cyanogen." The flames of hydrogen and sulphide of carbon, and of hydrogen and carbonic oxide, do not give the hydrocarbon bands (their spectra being continuous); a mixture of hydrogen and carbon tetrachloride gives them faintly, and a mixture of hydrogen and chloroform gives them strongly.

In all this we have no proof of the point in question, nor even any special probability that the bands are due to hydrocarbon; and yet, in the face of experiments in which the spectrum is obtained from cyanogen, when care has been taken to exclude hydrogen, we are asked to attribute the bands to the hydrocarbon formed by combination with some trace of hydrogen (as water or otherwise), supposed to be present as impurity. In the same way the presence of the bands and obtained under circumstances when nitrogen has been intentionally excluded, is to be explained by “the extreme difficulty of

removing the last traces of air." So that in the case of cyanogen with a trace of hydrogen present, the spark persists in giving us the spectrum of hydrocarbon; and when we have naphthalin with a trace of nitrogen present, it gives us the spectrum of nitrocarbon ! Attfield states that the spectrum in question is obtained from pure dry cyanogen. "The ignition of the gases having been effected in air, it was conceivable that hydrogen, nitrogen, or oxygen had influenced the phenomena. To eliminate this possible source of error the experiments were repeated out of contact with air. A thin glass tube one inch in diameter and three inches long, with platinum wires fused into its sides and its ends prolonged by glass quills having a capillary bore, was filled with pure dry cyanogen, and the greater portion of this gas then removed by a good air-pump. Another tube was similarly prepared with olefiant gas. The platinum wires in these tubes were then so connected with each other that the electric discharge from a powerful induction-coil could pass through both at the same ime. On now observing the spectra of these two lights 'n the simultaneous manner previously described, the

To air pun

of combustion-tubing (a) drawn out at both ends. In this it was repeatedly heated to the temperature of incipient decomposition whilst a current of dry air was drawn over it. One end of the tube was then closed by fusion at the point g, and the other bent round and fitted, as shown in the figure, to a U-tube (6) containing phosphoric anhydride the discharge-tube was interposed between this U-tube and a second U-tube d also containing phosphoric anhydride, the other branch of which was connected to one end of a vertical tube e of more than thirty inches in length, the lower end of which passed into mercury contained in the bottle f, the upper portion of which could be exhausted by means of the air-pump. The connections with the U-tube were made by means of perforated india-rubber stoppers, and the joints were surrounded during the experiment by melted paraffin.

characteristic lines of the hydrocarbon spectrum were found to be rigidly continued in that of the nitrocarbon. Moreover, by the same method of simultaneous observation the spectrum of each of these electric flames, as they may be termed, was compared with the corresponding chemical flames, that is with the oxyhydrocarbon and oxynitrocarbon jets of gas burning in air. The characteristic lines were present in every case."

"The spectrum under investigation having then been obtained in one case when only carbon and hydrogen were present, and in another when all elements but carbon and nitrogen were absent, furnishes to my mind sufficient evidence that the spectrum is that of carbon."

Morren also adopted this method of producing the spectrum by taking the spark of an induction coil in a sufficiently rapid current of pure cyanogen at atmospheric

pressure.

I have again repeated this experiment with cyanogen under conditions which would seem to ensure that the gas should be dry (see also Phil. Mag., 1875).

The cyanogen was prepared by heating pure cyanide of mercury, which was finely powdered and placed in a piece

The apparatus having been exhausted, the mercuric cyanide was heated till the apparatus was filled with cyanogen at atmospheric pressure; it was then again. exhausted and again filled with cyanogen. After having been thus exhausted and re-filled five or six times, the spectrum of the spark between the wires at c was examined at various pressures. The spectrum figured in my paper in the Philosophical Magazine for October, 1869, was obtained, the groups y and 8, with which alone we are at present concerned, being the brightest in the whole spectrum. Next careful search was made for the red hydrogen line. The cross-wires of a one-prism spectroscope were accurately adjusted to the red line, as seen in a hydrogen vacuum tube, and the spectroscope was then directed upon the spark in the cyanogen. No trace of the line could be observed.

A second experiment was devoted to the examination of the spark in an atmosphere of naphthalin vapour, from which nitrogen had been excluded as far as possible, in order to ascertain whether the bands and 0, which Professors Liveing and Dewar attribute to cyanogen, would be produced. Professors Liveing and Dewar are somewhat in error in saying that I laid much stress on the occurrence of these bands in carbonic oxide. They were never obtained very brilliantly from carbonic oxide (except under pressure), but they are obtained brilliantly from a naphthalin vacuum tube. I have obtained them also from a vacuum tube containing pure marsh-gas (my note-book remarks" very bright"), and as confirmation by an independent observer, I would remark that Plücker maps them in the spectruin of a vacuum tube containing methyl.

The vacuum tube in this second experiment contained pure solid naphthalin fused on the sides of the tube; this was placed in position so that the upper end passed through one hole in an india-rubber stopper into a flask filled with carbon dioxide; a vertical tube of thirty inches length passed through the second hole in the stopper of the flask, and its lower end dipped below mercury. The whole of the vacuum tube except the lowest portion was surrounded by a wider tube containing melted paraffin.

When the apparatus had been arranged, the experiment was commenced by passing a rapid current of carbonic acid through the vacuum tube, so as to fill the flask and escape through the mercury. After passing the gas for a considerable time, the lower end of the tube was closed by fusion, the naphthalin all melted down into this end, where it was made to boil violently, while the paraffin was maintained at a temperature of about 220° C. After the current of naphthalin vapour had lasted some time, the upper end of the tube was closed by fusion, the tube removed and cooled, and its spectrum examined. It gave a spectrum in which the groups and were plainly