jet is returned back again and although the efficiency is small, when compared with a pump, still the Injector has the advantage of working while the engine is at rest.

The same principle is applied occasionally in the Water Injector and the Ejector, where, for instance, a large body of water, in the form of leakage or water ballast, is to be rapidly cleared out. A somewhat similar instrument, although quite different in principle, is that called the Pulsometer, which is really a revival of the Marquis of Worcester's and Savory's Fire Engine, where the pressure of steam acts directly on the surface of the water. To check the great condensation a piston was introduced, and hence our modern steam engine.

Hot Air Engines are described in chapter xi., and here the mathematical theorems for Perfect Gases receive their most beautiful applications, so that formal treatises on Thermodynamics usually treat this part of the subject at length. Our author dismisses it in about eight pages, with a short description of the principal systems, as, unfortunately, all the practical objections against the use of Superheated Steam are intensified tenfold in the Hot Air Engine. Ericsson once fitted a steamer to cross the Atlantic with engines on this principle: they were very cumbrous although the boilers were dispensed with; and the experiment did not lead to further imitation. An exception must be made in favour of the Gas Engine, as the only practical application of the Hot Air Engine; the author works out the theory, and comes to the remarkable conclusion that the efficiency of the Gas Engine Cycle does not depend, as in ordinary Thermodynamics, on the difference of temperatures so much as on the degree of expansion and compression.

The author reaches the real part of his subject in chapter xii., where he discusses the theory of the Actual Steam Engine, as we really find it working, in the mill, mine, and on the railway or steamer.

Here Hirn appears as the great authority on the careful records of what takes place in the actual engine (chapter xvii.).

"The measurement of quantities of heat, especially when it has to be done in an engine at work, is an operation of great difficulty; and it was not till 1862 that it was shown experimentally by Hirn that h, the heat emitted, is really less than H, the heat received by the engine" (Maxwell, “Theory of Heat ").

The example of Hirn has been followed up of recent years by careful and long-continued experiments on steamers and pumping engines in regular work, and the results of the most important of these tests receive careful description and analysis, in chapters xv.-xviii.; a preliminary chapter, xiv., giving a detailed account of the best procedure and instruments required in Testing Steam Engines.

The book will be found indispensable, not only by designers of Steam Engines, but also by writers of abstract treatises on Thermodynamics, as restraining their mathematical development within reasonable limits of actuality, and as directing their analytical powers in a useful direction.

A. G. GREENHILL.

BRITISH flies.

An Account of British Flies (Diptera). By the Hon. M. Cordelia E. Leigh, F.E.S., and F. V. Theobald, B.A., F.E.S. Vol. I., Part I. (London: Elliot Stock, 1891.)

THE

`HE reader involuntarily glances back at the title of this work when the first words that meet his eye on the front page are: "One of the branches of science that has advanced with rapid strides during recent years is geology...." To commence with Fossil Diptera, and to enumerate the families (and some of the genera) members of which are found preserved in the earth's strata, before either families or genera have been in the least degree defined, is a somewhat novel way of beginning. When the work is completed, students will find it useful to transfer chapter i. to the end. The second chapter, entitled "Classification of Diptera, with an introductory account of the ancient and modern classification of Insecta," contains much matter of interest to entomologists in general, although it is questionable whether the authors have arranged their material in either the most attractive or the most methodical form. The classification of the Diptera it is intended to follow is that of Verrall, published in 1888, in which the order is divided into two great sections-the Orthorrhapha and the Cyclorrhapha; the Nematocera and Brachycera being included in the former, and the Proboscidea with the Eproboscidea in the latter. The Aphaniptera (now included in Nematocera) form the subject of the third chapter, in the course of which this first part terminates. The structure and metamorphosis of Pulex are discussed at some length, and certain species are described in detail. Some uncertainty seems to exist in the authors' minds as to how many of them are engaged upon the work, for they use both "we" and "I." This calls to the recollection Cruikshank's picture, "In which there is Antagonism of interest yet Mutuality of object."

It is not possible from a perusal of the first thirty-two pages to form a fair idea as to the general character of the work. It may be stated, however, that it appears to be written for those who are already entomologists, a. familiarity with entomological science on the part of the reader being assumed by the authors. Considerable trouble has evidently been taken in consulting authorities whose works are accessible only to the few. That there is plenty of room for a good treatise on the British Diptera will readily be admitted, and if the authors should have something new to tell about such genera as Chlorops, Oscinis, Cecidomyia, and Hylemyia, so much the better. Part I. is illustrated by five woodcuts.

OUR BOOK SHELF.

Principles of Agriculture. Edited by R. P. Wright, F.H.A.S. (London: Blackie and Son, 1891.) THE raison d'être of this little volume is to be found in its "tail," where are reproduced the questions set in the Science and Art Department Examinations in the Principles of Agriculture during the last eleven years. The title-page ought to state, but it does not, that this is a revised edition of a book that was published some years ago. This fact is only discoverable from the preface. The original edition was arranged in three

parts, whilst the current edition is in four parts. The added part is somewhat of a jumble, inasmuch as it is supplementary of each of the first three parts. The scheme of the book is not apparent from the list of contents, and this omission results in confusion. Whilst, however, the arrangement of the book is bad, the matter is good. In skilful hands, indeed, the material which is here accumulated might have been very attractively presented. At p. 132, a dozen pages are commenced on the pests of the farm, whilst another dozen pages devoted to the same subject begin on p. 180. At p. 71, the reader enters upon 30 pages about manures, and at p. 167 he gets a further dozen pages also upon manures. And so on.

With reference to the fixation of nitrogen by leguminous plants, mention is made of the presence on the roots of these plants of "little bag-like enlargements, or tubercles as they are called." It is unfortunate that this effort should be made to associate the pathological term "tubercle" with these structures. The word "nodule" is much preferable, and is not less explanatory.

Despite the fact that the book has been written to enable candidates to "pass an examination," it is as useful and trustworthy a little treatise of the kind as we have seen.

Elementary Trigonometry. By J. M. Dyer and Rev. R. H. Whitcombe. (London: George Bell, 1891.) THE title of this book is on all fours with the contents. The work is well adapted for school use. The explanations of book-work are clearly expressed, and the text is amply illustrated by a store of exercises. Sufficient ground is covered to meet the wants of average Army pupils.

We have detected errata in the text on pp. 21, 30, 36, 59, 61, 62, 65, 67, 74, 80, 101, 136, 153. The major part of the proof-sheets has been carefully gone over, but occasionally, as we have indicated, the authors have nodded. The printing in places, in our copy, is defective. But these faults only slightly mar a work which treats a hackneyed subject with all the freshness one can look for in an elementary text-book.

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.]

Opportunity for a Naturalist.

SINCE the completion of " Argentine Ornithology," in which was given an account of the 434 species of birds then known to occur in the Argentine Republic, Mr. Arthur Holland, of the Estancia Espartilla, and Mr. J. Graham Kerr, of the Pilcomayo Expedition, have made excellent contributions to the same sub. ject, and have added some 30 species to the Argentine avifauna. But much more remains to be done, and, in continuation of the work, I am now anxious to get a good series of birds from Uruguay, the fauna of which, so far as we know it, does not appear to differ materially from that of its neighbouring Republic. For this purpose I have made arrangements with a friend to take in a naturalist at his Estancia, near Minas, about sixty miles from Monte Video, and am looking for a qualified collector to occupy the post. His necessary expenses will be met, but his further remuneration must de end, more or less, on the results obtained. May I ask the aid of NATURE to make known this eligible opportunity for a young naturalist who can make good birds'-skins, and is anxious to pass a few months in a foreign clime? P. L. SCLATER.

3 Hanover Square, London, W.

Warning Colours.

MR. BEDDARD, in his letter published in NATURE of November 26 (p. 78), calls attention to Dr. Eisig's suggestion that

those bright colours of animals which have hitherto been regarded as of warning significance are merely the substances which confer the unpleasant taste, and that therefore the older The writer interpretation is unnecessary and in fact erroneous. furthermore implies that Dr. Eisig's views are not alluded to by those who have written upon animal colours, because they have escaped their attention. There is, however, another possible explanation of such neglect, and one which in my own case is certainly the correct one-viz. that the views in question appear to be so inherently improbable that a large body of confirmatory evidence is required before they demand attention. I do not by this mean to suggest that the unpalatable attribute may not possess a bright colour: this is certainly often the case, especially with the secretions expelled by many insects when they are irritated. But it is highly improbable that these facts afford any refutation of the theory of "warning colours "—that is, of the view which regards the bright and conspicuous colouring as an indication (in mimetic forms a false indication) of some unpleasant attribute, whether associated or unassociated with the colour itself. And as regards the bright colours of Lepidopterous imagines, such association is, to say the least of it, entirely unproved. It by no means follows that the yellow colouring of the brimstone and other butterflies is disagreeable in flavour because it "is due to a substance formed as a urinary pigment." And the relation of many animal colours to these pigments by no means necessarily implies unpalatability. Again, it would be impossible to regard merely as a coincidence the fact that the substances in question almost invariably produce a conspicuous appearance, and, furthermore, produce it in a variety of ways. Such an appearance is, as is well known, not merely due to the individual colours, but to their mutual arrangement and relationship. It is due, moreover, to a variety of physical principles, for the production of white is very different from the production of the colours which are so often contrasted with it. Con

spicuous effects are furthermore often gained without the use of pigment, as in the brilliantly metallic pupa of Euplaa core and of Mechanitis lysimnia. Hence the contention that the bright colour of distasteful insects is a mere incident of chemical composition which has been selected on other accounts is so inherently improbable that it would require a large body of evidence to support it.

But perhaps the strongest argument against the view is that it creates such an artificial distinction between inedibility due to mere unpalatability, and that due to other unpleasant attributes. Mr. Beddard would probably admit that the conspicuous colouring of the skunk, the coral snake, and the wasp possesses a true warning significance; and yet he would interpret the black and yellow colouring of the larva of the cinnabar moth or the pupa of the magpie moth (both known to be unpatable) in an entirely different way, and would deny that it possesses a warning meaning.

In addition to these considerations, the undoubted existence of an unpalatable quality not residing in the superficial pigments is quite clear in many brightly coloured insects. The irritating hairs and odoriferous secretions of many Lepidopterous and Hymenopterous larvæ, and the evil-smelling yellow fluids which exude from Coccinellida and from many conspicuous butterflies are examples.

The recent investigations of the distinguished Russian naturalist Portchinsky (II. "Coloration marquante et taches ocellées, leur origine et leur développement," St. Petersburg, 1890) have, among other things, shown us the distinct manner in which the colours which attend unpalatability are displayed by the insect when it is disturbed. He thus explains some of the cases of "shamming death" which are so often alluded to in works on insects (the other cases being explained by the necessity for concealment). Two examples which he adduces are so interesting, and have so important a bearing on this discussion, that I cannot resist the temptation of reproducing them here, especially as Portchinsky's paper, being written in Russian, is almost unknown in this country. I have, however, been

most kindly helped by my friend Mr. Morfill, and now possess a complete translation, which I hope soon to publish. The female of Spilosoma mendica possesses black and yellow legs, and, when disturbed, it folds its limbs and drops to the ground, generally falling on its back, so that the contrasted colours are displayed (see Fig. 1). In the closely allied Spilosoma urtice the dorsal surface of the abdomen is black and yellow, and this insect, when irritated, raises its wings and curves the abdomen downwards so that the colour is conspicuous. Furthermore,

make trial of them upon a material which will ensure their ultimate rejection. But if the colour has not a meaning as such, there is no reason why this spot should be attacked in preference to any other part of the exposed surface; and the existence of the colour as a covering to the most vital parts seems to indicate that it acts as a warning away rather than in the reverse manner.

The fact that brightly coloured animals are frequently attacked does not seem to me to be a great difficulty. The really important point is whether the enemy remembers the attack, and is assisted in identifying the unpalatable species by its bright colours. Many experiments seem to show that this is so. Certainly Mr. Beddard will not assert that the majority of insect-eating animals fail to know and recognize a wasp without tasting it. Again, the question is really, as Mr. Titchener implies in his interesting communication, one of "comparative palatability"; and there is no doubt that insect eating animals when sufficiently hungry will attack and sometimes devour insects which they would ordinarily reject. Furthermore, an animal which naturally prefers a varied insect food, and which is fed in confinement largely on other substances and partially on a monotonous insect diet, may be expected to be less scrupulous than it would be in the wild state. I may state, however, that the most intelligent insect-eating animals, such as the marmoset, hardly ever make mistakes; their suspicion being at once aroused by any trace of a warning colour.

It is well known that we chiefly owe the theory of warning colours to Mr. A. R. Wallace. My own conviction of its entire validity rests upon the results of a prolonged series of experiments, of which only a part has been published. I believe that I conducted these experiments fairly, that my mind was open, and that I had no personal bias in the matter at all, either in favour of or against the theory. And I can confidently make the same claim on behalf of others who have experimented in the same manner-such as Mr. Jenner Weir, Prof. Weismann, and M. Portchinsky. I may allude especially to the writings of the last-named authority, as they are the most important as well as the most recent contribution to the theory which we owe to Mr. Wallace.

I may also take this opportunity of replying to a very similar objection raised by some reviewers against my book on the "Colours of Animals, their Meaning and Use, &c." They point out that I have not alluded to Eimer's work on the comparison of the wing markings of Papilionida, and they assume that his paper has, therefore, escaped my attention. But Eimer's paper has no bearing whatever on the value of colour in the struggle for existence, and this is the subject of my book, as anyone can infer from the preface, or even from the title. For this reason I was also compelled to omit reference to what I venture to regard as the far more important work of Weismann on the development of the colours and marking of caterpillars, and of Dixey on the wing-markings of Vanessida and Argynnida, as well as a very large proportion of my own work, which is a continuation of that begun by Weismann, and was, in fact, inspired by it.

Oxford, December 15.

EDWARD B. POULTON.

My friend Prof. Meldola has drawn my attention to a communication by Mr. F. E. Beddard in NATURE of November 26 (p. 78), in which the view is expressed that the brimstone butterfly (Gonepteryx rhamni) is rendered protected or unpalatable by the yellow pigment of its wings being due to a substance formed as "a urinary pigment," and that the coloration is "a consequence of the deposition in the integument of bitter pigments.' The following objections may be urged against the view that this coloration, said to be of the nature of a "urinary pigment," affords any protection whatever.

Gonepteryx rhamni itself has its female much paler than the male and of a greenish-white hue, whilst the wings in both sexes are of a leaf-like appearance, which can only be due to the process of natural selection, and can scarcely have been exercised in the direction of "protective resemblance" if the insect was already unpalatable by the "urinative" nature of the yellow pigment of its wings.

Yellow Lepidoptera have certainly no immunity from the attacks of birds; on the contrary, the scanty records we possess of these onslaughts go to prove that the contrary is the case. The late Mr. P. H. Gosse observed one of the greenlets (Vierosylvia calidris) to pursue a species of Terias in Jamaica ("Birds of Jamaica," p. 194). In Southern India, Mr. E. L. Arnold found the principal victims of the green bee-eaters to be specimens of Terias hecate (" On the Indian Hills," vol. i. p. 247-48). Quite recently in the Transvaal I have observed the wagtail, Motacilla capensis, to pursue and devour the yellow Lithosiid moth, Binna madagascariensis.

But the facts of "mimicry" seem to effectually dispose of the supposition. In South Africa, the yellow black-margined Papilio cenea affords by its females the most striking examples to prove the non-protective value of this coloration; for the females respectively mimic those two well-known "protected butterflies," the blackish Amauris echeria and the reddish Danais chrysippus, whilst, to add to the negative evidence, the yellow male has been seen by Mr. Weale to become the prey of the flycatcher, Tchitrea cristata.

On the Amazons, Mr. Bates has long since shown that the yellow and black Leptalis orise mimics the markings-even to the colour of the antennæ and the spotting of the abdomen-of the protected or unpalatable Methona psidii. Russell Hill, Purley, Surrey.

W. L. DISTANT.

A Difficulty in Weismannism.

IN his communication of November 28 (NATURE, December 3, p. 102), Prof. Hartog asks us to believe that Weismann, in a letter from which he quotes, insists (1) that the Ahnenplasmas "not completely unchangeable," and (2) that "each Ahnenplasma unit corresponds to an individual of the species itself; and if put under suitable trophic conditions would, singly, reproduce such an individual."

are

Assuming that thesis II. adequately represents the Freiburg Professor's latest views, and that a few sentences detached from their context are to be depended upon, we must, it seems to me, conclude, with Prof. Hartog, that he has unearthed an inconsistency, and, what is of more importance, shown that the shuffling process is not only unnecessary, but that a new significance must be found for it.

I am, however, still inclined to believe that hypothesis B is the one upon which Weismann has founded his theories of heredity and sexual reproduction. The hypothesis, however, should take account of the variability, slight though it may be, of the Ahnenplasmas. We agree to call the Ahnenplasmas Protozoan, simply because we have no conception of the kind and amount of the variation they have undergone since they parted company with the unicellular organisms in which they originated. We have no reason, however, to believe that the external causes which led to their variation in unicellular organisms are powerless to affect them now that they are localized in the reproductive cells of multicellular ones.

Prof. Hartog, moreover, while relinquishing the idea of the variability of the offspring of the lioness, endeavours from another point of view to attack Weismannism on the plane of hypothesis B. Is he, too, sceptical as to Weismann's adherence to hypothesis A, or does he simply wish to overwhelm the so-called disciples?

In either case, several objections may be made to his argument. In the first place, we object most emphatically to any