met with in this country in attempts to rear even presentable "specimens" of such favourites in our smokebeladen and crowded cities and suburbs.

Then, again, the English reader gathers some information as to the Western popular names of trees, well known to him by very different ones; how many English people know what are the "cucumber-tree," the "yulan," the "buckeye," the "butter-nut," and the "button wood"? Bits of history also occur, and incidental notes on the rates of growth of various trees, their ages, &c. So that, after all, there are some dry facts in this singularly quaint and simply written talk about trees. We must not claim much for the work in this respect, however; and perhaps the chief reason we like the writing is because of its contrast to the empty and inflated style of too many of our native newspaper articles on similar subjects.

OUR BOOK SHELF.

Synopsis of Non-Metallic Chemistry. By William Briggs, B.A., F.C.S. Pp. 90. (London: W. B. Clive and Co.)

THIS book is intended for students preparing for the Matriculation Examination of London University. After the contents of an ordinary text-book have been studied, the reader is here supposed to find the more important points which have to be remembered, and which serve to recall the less important. Interleaved note-paper is provided, whereon facts readily forgotten may be recorded.

As is usually the case with such cram-books, little can be said in favour of the quality of the information supplied.

Formulæ are stated to be "arrangements of letters representing a molecule of a compound," and this definition is illustrated by regarding Fe,O, as denoting a molecule. The vapour-density of hydrofluoric acid is given as 10, and the solubility of hydrogen in water as practically the same at all temperatures from 1° to 20": neither statement is up to date. Such antiquated terms as basylous and chlorous, which are freely employed, might well be replaced; and to speak of distilling potassium perchlorate with strong sulphuric acid is inaccurate. The account given of fractional distillation is worthy of reproduction. The mixture of liquids is heated "up to the lowest of the boiling-points of the liquids present. The whole of that liquid (?) will be converted into vapour, and can be condensed in the usual way. On heating the remaining liquid up to the next boiling-point, we can separate another of the constituents, and so on until they are all separated out. The different liquids thus obtained must be redistilled to get them quite free from the others, small quantities of which may have been distilled over in the first process."

This last extract is typical of the bulk of the knowledge contained in the book, which, to say the least, savours more of the class-room than the laboratory.

A table of contents, a glossary, and three appendixes are provided. The last are concerned with the preparation and purification of substances and with the simpler chemical calculations. A list, with answers, of numerical examples set at the matriculation examinations is included.

Chemical Calculations. By R. Lloyd Whiteley, F.I.C. With a Preface by Prof. F. Clowes, D.Sc., F.I.C. (London: Longmans, Green, and Co., 1892.) THIS is still another addition to the numerous manuals on chemical arithmetic, and the points wherein it differs from its predecessors are somewhat difficult to discover.

Once more we have specific gravity observations in which corrections for temperature and air displacement are ignored; and although the author attempts to set right the prevalent misconception as to the meaning of density and specific gravity, it is questionable if he succeeds. According to him, specific gravity is always relative; it is the same magnitude as relative density. The use of absolute specific gravity-or, shortly, specific gravity with no temperatures of comparison attached-as denoting the weight of unit volume, is here overlooked. From a physical point of view, the definition of the absolute density of a gas as the mass of 11:16 litres is a needless complication. In ascertaining the percentage composition of a compound it is insisted that, first of all, the molecular weight must be calculated. The examples given to illustrate the rule include apatite, apophyllite, basic lead chromate, &c. The student is thus led to infer, here as elsewhere in the book, that the molecular weights of such bodies can be fixed.

The freezing-point and boiling-point methods of obtaining molecular weights are disposed of in two pages. No hint is given that the solutions must be dilute and non-electrolytic, if consistent results are to be obtained; or that, in general, the interpretation of the results of these methods is still subject to difference of opinion.

It is erroneous to state that "the alteration in the volume of a gas is proportional to the so-called absolute temperature," or to speak of "Dalton and Henry's law." Henry's law is distinct from Dalton's, and is the older by two years.

We have dwelt on some of the points which seem to call for criticism. On the other hand, the book has its good features. The problems are numerous, carefully selected, and well arranged. Contents, answers, and index are supplied. It seems to us, however, that instead of being as good as several of its kind already in existence, it, as a new book, should have been better. The Year-book of Science. Edited for 1891 by Prof. T. G. Bonney, D.Sc., LL.D., F.R.S. (London: Cassell and Co., 1892.)

ALL who have any sympathy with scientific pursuits will heartily welcome the appearance of this epitome of the more important results of the investigations which were published during the past year. Scientific inquiry now covers so much ground that all men of science must be more or less specialists, and it is difficult for them to keep in touch with the developments in other branches through the usual channels, although it frequently happens that an advance in one subject may throw light upon and induce investigations in another. There are also many engaged in practical pursuits who require a convenient means of determining how far contemporary researches may be technically applied.

With a well selected staff of contributors, the editor has attempted to meet the wants of all by the present volume, which is divested as far as possible of technicalities. The scope of the work is sufficiently defined by the following paragraph from the editorial :

"It is almost needless to remark that this volume is not intended to be a record or catalogue of papers. The endeavour of its projectors and compilers has been to select those memoirs, in each several department, which appeared to be of somewhat exceptional interest, either by throwing light on special difficulties or by being suggestive of further advances."

In a work of this kind strict impartiality is essential, and we see no reason to suppose that the various contributors have abused the power vested in them. On the whole, the production is very satisfactory, and the improvements which the editor contemplates for the next volume will make it more so. One can only wonder that science has had to wait so long for a year-book of its

own.

Handy Atlas of Modern Geography. (London: Edward Stanford, 1892.)

IT would be difficult to obtain a small atlas more complete than this. Every place of any importance appears to be represented on one or more of the thirty coloured maps. The degrees of latitude and longitude are subdivided into parts of five minutes each, so that the positions of places, the names of which are not engraved, can be easily and accurately located by reference to the alphabetical list at the end. This list is a comprehensive one. It gives the latitude and longitude of the principal mountains, rivers, capes, bays, islands, towns, and villages, and forms an excellent supplement to a very good atlas.

LETTERS TO THE EDITOR.

The Editor does not hold himself responsible for opinions expressed by his correspondents. Neither can he undertake

to return, or to correspond with the writers of, rejected manuscripts intended for this or any other part of NATURE. No notice is taken of anonymous communications.]

Aurora.

A VERY brilliant display of aurora was seen here last night, the 25th inst. At about 9.25 p.m. a number of red streamers proceeded from a length of some 110° in azimuth along the northern horizon, and extended upwards for (on an average) 30°. The length of the streamers varied quickly, sometimes shooting upwards for 70° from the horizon. In the course of five minutes the red streamers gave way to white or yellowish white ones, narrower and more sharply defined than the red ones. At 9.40 p.m. there was a decrease in the brilliancy of the phenomenon, but at 9.45 p.m. long red streamers again appeared for a few minutes, which again shortly gave way to a brightness of the horizon only. Close to the horizon the colour was white, or nearly so, the whole time. The apparent point of convergence of the streamers was far south of the zenith, say 30°. GEO. M. SEABROKE.

Temple Observatory, Rugby, April 26.

PROBABLY many of your readers witnessed the brilliant display of the northern lights between nine and twelve o'clock last night, the 25th, but it may be as well to call attention to it, as being the finest display seen here for many years. Appearing soon after nine o'clock, the luminous arc and the radiating beams, sometimes rose and orange coloured, presented a varied and beautiful spectacle until close upon midnight, when they faded away.

The most noteworthy features of this display were the vividness and height of the arc, which reached an angle of about 13° above the horizon, whilst the beams were visible up to 51°. The whole expanse of the arc from east to west was about 93°, and the duration of the phenomenon a little under three hours. ARTHUR MARSHALL.

Cauldon Place, Long Row, Nottingham.

A FAIRLY distinct aurora was visible here on the northern horizon last night. I first observed it at 9.15, when the streamers appeared somewhat less bright than the Milky Way. Ten minutes later one streamer, about 15° west of north, brightened considerably, and appeared of a pale reddish-yellow tint. It fluctuated in intensity, and soon became less bright. The streamers, which inclined slightly to the west of the vertical, extended to about 30° to 40° above the horizon.

I watched them till 9. 50, when they seemed fading in intensity, and when I looked again at 10.30 they had disappeared entirely.

Thought Cot, Brentwood, April 26.

ARTHUR E. BROWN.

Pigments of Lepidoptera. THE appearance of Mr. F. Gowland Hopkins's letter on this subject in the last issue of NATURE (p. 581) demands a brief explanation from me-although it is not easy to reply satisfactorily within narrow limits-and the more so since Mr. Hopkins appears to have somewhat misunderstood my standpoint.

Let me first acknowledge the courteous tone of Mr. Hopkins's letter, and express my sense of the value to myself of criticism from his pen, and the more so since I have been labouring under the disadvantage of being practically entirely uncriticized so far-a disadvantage that I have not failed to appreciate.

Now, Mr. Hopkins remarks: "Mr. Coste's experiments are very useful as forming a method of classifying these pigments; but... they are of far too empirical a nature for any considerations as to the constitution of the bodies to be based upon them."

Here it is that Mr. Hopkins appears to have missed the point If he will do me the favour to refer to the of my work. detailed account of my experiments in the Entomologist, passim, I think that he will find it tolerably clearly emphasized that my interest in this work, so far, has been almost entirely biological. I stated expressly in my opening article that my object had been to discover, if possible, the genealogies of the colours, and to obtain evidence (so far as coloric characters could afford it) of the phylogenetic relations of allied species: and I may perhaps varieties of whose occurrence in the natural state I have since add that the results obtained have enabled me to predict several been informed. So that Mr. Hopkins is mildly reproaching me because my work does not tend in a direction at which it was not originally aimed, while he is at the same time good enough to admit that it is of some use for the end at which it was aimed. However, it was only to be anticipated that one could not go very far without becoming involved in the further question as to the constitution of the pigments; but here I was met by three considerations. In the first place, I was anxious to obtain first of all as much as possible of what Mr. Hopkins designates 'empirical" evidence as to the reactions and classification of the pigments before making any researches at all into their constitution; secondly, the amount of material at my command was far too scanty for any even approximate analysis; and in the third place, shortly after my experiments had been commenced, my attention was drawn to an abstract of a paper by Mr. Hopkins on the constitution of the yellow pigments. Finding that he was already in possession of the field here, I felt almost bound to leave this part of the subject alone, at least for the present; and I think that I may say that I have on the whole taken exaggerated care not to extend my experiments into that quarter where Mr. Hopkins was engaged, or to avail myself of the discoveries that he had already made, in order to trespass on his investigations. Putting aside my provisional suggestions as to the nature of the "reversion effect," it has only been at a comparatively recent stage of my work, and in consequence of experiments that have not yet been published, that I have at all turned my attention to the constitution of the pigments; these results being such as would have compelled me to consider the question even had I heard nothing of Mr. Hopkins's work. I hope that this explanation will put me right in Mr. Hopkins's eyes, and will satisfy him that he has considerably misunderstood the spirit of "some remarks [perhaps clumsily expressed by me] made at the close of the last article"; and that it will also satisfy him as to the question of priority. I had no thought of questioning Mr. Hopkins's priority in his own work, and the less so since I have throughout been under the impression that we were working mainly on different-though sometimes adjacent-lines.

must not so far trespass upon your space as to criticize Mr. Hopkins's criticisms upon the "reversion effect"; but I will ask him kindly to examine the detailed accounts of the "reversion" experiments which I gave in the Entomologist, since his remarks appear to me somewhat to ignore the evidence there brought forward: and at the same time I may remark that his statements as to the constitution of the yellow pigments appear to me hardly to invalidate, but rather indirectly to confirm, the suggestions made by me as to the reversion reaction with red pigments. The new information that Mr. Hopkins promises in his closing paragraph I shall look forward to with great interest. April 22. F. H. PERRY COSTE.

I WAS about to pen some remarks on Mr. Perry Coste's recent articles on this subject, when a letter from Mr. Gowland Hopkins in the last number of NATURE (p. 581) expressed substantially the same views as those which I had arrived at. I write now rather to support Mr. Hopkins in his strictures than to offer any fresh criticisms of my own. The articles on "Insect Colours" published in these columns are, as the author

Double Orange.

THE abnormality in a Maltese orange described in NATURE of April 7 (p. 534) would appear of common occurrence in the Queensland or South Australian fruit. A friend assures me that in a case recently received from Australia, 80 per cent. of the contents showed small oranges, more or less perfect, either embedded in the pulp or in the rind. The quality of the fruit I observed was in no way affected. It would, however, be interesting to obtain further testimony. Although the small oranges may not affect the commercial value of the fruit, their presence must be undesirable in the groves where perfection is sought.

Katrine, Surbiton.

GERALD B. FRANCIS.

ON THE LINE SPECTRA OF THE ELEMENTS, THE HE distribution of the lines in the spectra of the elements is by no means so irregular as it might seem at first sight. Since Lecoq de Boisbaudran, in 1869, discovered the general plan in the spectra of the alkali metals, a number of interesting facts have been brought to light, which will probably one of these days find their mechanical explanation, and will then greatly advance our knowledge of the molecules.

Mechanical explanations of some of the facts have been attempted already. Lecoq de Boisbaudran explains the fact that the rays of the alkali metals are, on the whole, less refrangible the greater the atomic weight, by observing that the oscillations of a body suspended in a given elastic medium will become less frequent when the mass of the body is increased. This explanation, however, seems to me to remain rather vague and unsatisfactory as long as it does not lead to any numerical results that agree with the observations. Taken literally, it makes the oscillation-frequency inversely proportional to the square root of the atomic weight, which is far from being

the case.

A second well-established fact has received different explanations by Julius 1 and by Johnstone Stoney. It has long been observed by Hartley that in the spectrum of several elements a number of doublets or triplets of lines appear, the oscillation-frequencies in each doublet or triplet differing by the same amount. Recent measurements by Prof. Kayser and myself have confirmed this observation. Julius believes that this phenomenon is due to a cause analogous to the combination tones in the theory of sound.

If two rays, with oscillation-frequencies a, B, combine with other rays, p, q, r, s, to oscillation-frequencies

r + a + a r+ B s + B,

the same difference a B will occur several times. That the doublets under consideration are in many cases remarkably strong is accounted for by the fact that the intensity of the combination tone is proportional to the product of the intensities of the primary tones, so that it must become very strong when the amplitude of the primary tones is sufficiently increased.

Johnstone Stoney gives a different explanation of the doublets. He supposes that the path of the molecule from which light emanates is an ellipse, which by disturbing forces is gradually changed, and he shows that on this supposition, instead of one ray, two rays or more would originate, and the oscillation-frequencies of these rays would differ by an amount depending on the rate of change of the ellipse. If now, instead of the ellipse, the path of the molecule is any other curve, it can be considered as consisting of a number of superposed ellipses, all of which change in the same way on account of the disturbing forces. To each of the ellipses a doublet of lines corresponds, and the oscillation-frequencies of each

doublet differ by the same amount. In this explanation I do not understand the decomposition of the arbitrary curve in a series of superposed ellipses. For the movement is supposed not to be periodical, and Fourier's theorem then would not apply, at least the periods of the superposed ellipses would not be definite, as long as there are no data except the arbitrary curve itself.

Besides, both Johnstone Stoney and Julius only try to explain one of a number of regularities that have been observed in the spectra of the elements. A plausible suggestion about the movement of the molecules ought to explain more than one of the observed phenomena. I think it may be useful to point out the other regularities that have been observed in the distribution of lines, and for which as yet no mechanical explanation has been

attempted.

(1) The doublets and triplets existing in the spectrum of an element can be arranged in series which show an appearance of great regularity. These series seem to be analogous to the over-tones of a vibrating body. But they possess a remarkable peculiarity, which, as far as I know, is without analogy in the theory of sound. The difference of two consecutive oscillation-frequencies decreases as these increase, and there seems to exist a finite limit to the oscillation-frequencies of a series. If n represents integer numbers, the oscillation-frequencies of a series may with great accuracy be represented by the formulaA Bn2 Cn→4,

where A, B, C are positive constants. B has nearly the same value for all the series of the different spectra. A is the limit towards which the oscillation-frequency tends, when n increases.

(2) For elements that are chemically related, the series are distinctly homologous, both in appearance of the lines and in the values of A, B, C, and with increasing atomic weight shift towards the less refrangible end of the spectrum. Homologous series have been observed in the following groups of elements :

Lithium, sodium, potassium, rubidium, cæsium;
Copper, silver;

Magnesium, calcium, strontium;
Zinc, cadmium, mercury;
Aluminium, indium, thallium.

In the first two and in the last group the series consist of doublets, while in the remaining two groups they consist of triplets. Thus we may say that the spectrum shows a relationship between the elements similar to that between their chemical properties. It is interesting to note that magnesium forms a group with calcium and strontium, and appears more nearly related to them than to zinc, cadmium, and mercury.

(3) The doublets and triplets in each group broaden as In the first group the the atomic weight increases. difference of oscillation-frequencies is nearly proportional to the square of the atomic weight. The constant difference of the oscillation-frequencies in the doublets and triplets may also be noted in the values of A, B, C. For a series of doublets or triplets we have two or three different values of A, but only one value of B and one value of C.

(4) In each of the spectra of sodium, potassium, rubidium, and cæsium, a series of doublets has been observed, in which the oscillation-frequencies do not differ by a constant amount, the difference diminishing inversely proportional to n1. For these series A and B have only one value each. The least refrangible doublet of the series has the same difference of oscillation-frequencies

as the doublets in the other series of the same element. In the spectrum of lithium there is a homologous series of single lines. All the lines of these series have the same 'Lithium has here to be excepted, whose lines are all single.

character; they are strong and easily reversed, and in all of them the first doublet is situated on the less refrangible side of the spectrum, and all the others in the violet and ultra-violet. The series shift towards the less refrangible side with increasing atomic weight.

For further details the reader is referred to the following memoirs-Kavser and Runge," Ueber die Spectren der Elemente," Abhandl. der Berl. Akademie, 1890-92; Rydberg," Recherches sur la constitution des spectres d'émission des élements chimiques," Kongl. Svenska Vetenskaps-Akademiens Handlingar, Bandet 23, No. 11, C. RUNGE.

1890.

ABERRANT FOSSIL UNGULATES OF
SOUTH AMERICA.

TILL
ILL within the last few years palæontologists and
zoologists were being continually startled by the
discovery of strange forms of extinct Ungulates which
rewarded the researches conducted in the Tertiary rocks
of the United States. The animals thus brought to the
notice of the scientific world have, to a very large extent,
modified our conceptions of the relationships of the
various groups of hoofed or Ungulate Mammals to one
another; and have led to the very general adoption of
the view of the ordinal unity of all these multifarious
types. Several of them, indeed, so far as we may judge
from their mere skeletons, indicate signs of a transition
between the Perissodactyle and Proboscidean modifica-
tions of Ungulate structure; but none of them tend in
the least degree to break down the hard and fast line of
demarcation between the Perissodactyle (odd-toed) and
Artiodactyle (even-toed) modifications, which is main-
tained throughout all the known Tertiary deposits of the
Old World. Moreover, after a little "shaking down,"
the whole of these North American Ungulates, with the
exception of the curious Rodent-like Tillotherium, fall
fairly well into their places in the Ungulate order;
although some of the earlier and smaller types present
indications of close affinity with the common stock from
which we may presume both Ungulates and Carnivores
to have taken origin.

At the present time the wave of discovery of new forms appears to be passing from the northern to the southern half of the New World; so that while the paleontologists of the United States are to a great extent engaged in the important task of revising and completing the preliminary work of the last twenty years, their confrères in Argentina are almost flooding scientific literature with descriptionssometimes, it is to be feared, rather crude and hasty ones -of a number of new or hitherto imperfectly known forms of extinct mammals. This descriptive work has been mainly undertaken by Messrs. Ameghino, Burmeister, and Moreno. Unfortunately, however, the greater part of it is still in the form of preliminary notices, unaccompanied by illustrations; while on several points the three describers above mentioned are by no means in accord, and it is quite clear that unnecessary names have frequently been published. There is, indeed, one large illustrated work published by Dr. Ameghino; but since, so far as we are aware, there is only a single copy (in the Natural History Museum) in England, palæontologists have not the opportunity of paying it that attention in private study which its importance demands.

In spite, however, of these drawbacks, the information at present before us-imperfect though it be-introduces us to several groups of extinct Ungulates totally unlike any found in all the rest of the world put together, and which are of especial interest as tending to a certain extent to break down the distinction between Perissodactyles and Artiodactyles. It should be observed, before proceeding further, that the explorations conducted in

Patagonia and various parts of Argentina have shown that the deposits containing mammalian remains, instead of being exclusively of Pleistocene age, comprise a large portion of the Tertiary period, probably extending down at least as far as the Oligocene; although the exact correlation of the different beds with European deposits is probably premature.

With these preliminary observations, and asking our readers at the same time to bear in mind that a considerable part of our knowledge is still in a very imperfect and crude condition, we propose to glance at some of the peculiarities presented by the more remarkable forms of Ungulates described from the deposits in question.

Since the date of the publication of the results of Darwin's voyage in the Beagle, we have been gradually acquiring a knowledge of the structure of that remarkable South American Ungulate known as Macrauchenia, of which the complete osteology has been described by Burmeister. This animal, which had the general proportions and size of a horse, conforms in several respectsmore especially in having three-toed feet, in which the middle (third) digit is symmetrical in itself-so markedly with the Perissodactyles, that by common consent it has been generally regarded as an extremely aberrant member of that group. The molar teeth are, indeed, more like those of the Rhinoceros and Palæotherium than of any other Old World Ungulates, while the infolding of the enamel of the crowns of the incisors is a character known elsewhere only in the horses. The absence of any gap in the dental series, and the nearly even height of the teeth, are characters in which Macrauchenia agrees with the Old World Anoplotherium. Perissodactyle affinities are indicated by the presence of a third trochanter on the femur; but in certain peculiarities in the anklejoint this animal differs from all typical Perissodactyles, and agrees with the Artiodactyles. Moreover, a certain peculiarity of structure in the vertebræ of the neck is repeated elsewhere only in the camels and llamas, which form an isolated group of Artiodactyles. In the complete closure of the orbit by bone, Macrauchenia resembles the horses and many Artiodactyles; but in the narial aperture being situated on the top of the skull between the orbits (whence the nostrils were probably produced in the form of a proboscis), it is absolutely peculiar.

There are thus many indications that, while Macrauchenia is a specialized form that can in no sense be regarded as the ancestral type from which Perissodactyles and Artiodactyles have originated, it retains certain generalized features which were probably directly derived from such ancestral stock.

Among the Ungulates discovered in Patagonia is one named Proterotherium, which was at one time referred to the Artiodactyles, but subsequently placed among the Perissodactyles. In the skull, so far as can be gathered from Ameghino's description, the orbit is closed, as in Macrauchenia, but the narial aperture appears to have had the normal position. The molar teeth are so like those of true Perissodactyles that they were originally described under the name of Anchitherium; but the rest of the dentition is very peculiar. Thus, in the upper jaw there appears to have been only a single pair of incisors in the premaxillæ, these being pyramidal and obliquely truncated like the canines of the pigs; and as there were no canines, it may be inferred that there was a long toothless interval in the jaw. In the lower jaw there were two pairs of incisors, and no canines. The lower molar teeth were inserted by four distinct roots-a feature unknown in any existing Perissodactyle, although occurring in the pig. In the limbs, both the front and hind feet were furnished with three complete toes, much resembling those of Hipparion; the ankle-joint is, however, said to resemble that of the Artiodactyles. have no information as to the third trochanter of the femur. On the whole, this genus appears to indicate a

We