What do the authors mean? Their erudite and careful statements of the position of many foreign writers emphasize their failure to represent the position of the author of the "Origin of Species." The authors think that the problems and questions reating to sex, problems and questions carefully and ingeniously analyzed by them, "are in final synthesis all answerable in a sentence." Morphological questions are at base, they say, physiological; and physiological questions are ultimately referable to the metabolism of protoplasm, as Prof. Burdon-Sanderson pointed out last autumn. This metabolism is double: it consists on the one hand of anabolic, constructive, elaborative processes-processes attended with the storage of energy; and on the other hand of katabolic, destructive, disintegrating processesprocesses attended with the liberation of energy. These processes are complementary; in living protoplasm they seem for the most part coincident. Losing sight of the coincidence the authors have seized on the antithesis; the dea has grown upon them till they see a rhythm of anabolism and katabolism swinging through organic nature and producing-well, producing nearly everything. Take, for instance, secondary sexual characters. Males are frequently litbe, active, aggressive, gorgeously coloured and decorated. Females are often sluggish, vegetative, passive, and soberly coloured. These characters, according to Geddes and Thomson, occur because males have a male or katabolic diathesis, because females have a female or anabolic diathesis. "Brilliancy of colour, exuberance of hair and feathers, activity of scent glands, and even the development of weapons, are not and cannot be (except teleologically) explained by sexual selection, but in origin and continued development are outcrops of a male as opposed to a female constitution" (p. 22). It is impossible to follow in detail and state the innumerable objections to this explanation. Do the authors suppose a male diathesis explains the ascending series of horn and antler development? Can it in any way account for "interference" colours, which play so large a part in the adorning of males? Are women less female when they have radiant complexions and abundant tresses? What physiological reason is there for believing that skeletal weapons and scent glands, or the crystals in anthers, are due to the katabolism of "exuberant maleness," while menstruation and lactation are means of getting rid of "anabolic surplus?". Parthenogenesis occurs in groups of animals where the anabolic rhythm is dominant. Sex itself appears when katabolic conditions preponderate. And this is why flowers so often are situated at the end of the vegetative axis; this is furthest from the source of nutrition; the flower occupies a katabolic position, and is often the plant's dying effort (p. 226). Alternation of generations is a special example of the rhythm. Thus, but the authors do not cite this example in this connection, the tiny sexless and spore-bearing stalk parasitic on the moss-plant is the anabolic vegetative generation, while the conspicuous moss-plant is the sexual or katabolic generation-the generation peculiarly connected with starvation! It is obvious that the authors are nothing if not original. But the real value of the book must not be lost sight of in quotations from it. The chapters on the "Determination of Sex," on Sex Elements," and on "Growth and Reproduction," are very suggestive. But indeed, to biologists the greater part of the book and its theories must be useful and suggestive. It is only the general public that must be warned off. It is very much to be regretted that the authors have included a discussion of certain social and ethical problems absolutely unconnected with the title of their book. If such matters are to be discussed coram populo, it is only fair that explicit information should appear on the title-page. P. C. M. THE QUICKSILVER DEPOSITS OF THE Geology of the Quicksilver Deposits of the Pacific Slope. AMONG the numerous mineral treasures of California none are of more interest than the deposits of mercury ore which occur at intervals along the greater part of the Coast Range from the Mexican boundary to Clear Lake, in lat. 39° N., a distance of more than 200 miles. This region, together with the district of Steamboat Springs in Nevada, has been carefully examined by the division of the United States Geological Survey under the charge of Mr. G. F. Becker, and the results are now presented in another of the handsome quarto series of monographs published by Major Powell, the head of the Survey. The discovery of mercury in California preceded that of gold; the most productive locality, New Almaden, near San José, at the south end of the Bay of San Francisco, having been known for about 65 years, while the actual mining was commenced under a grant from the Mexican Government shortly before the cession of the country to the United States. In its earlier years the mine was extremely profitable, and the long judicial controversy that ensued before the title was satisfactorily established occupies a prominent place among the records of American mining litigation. The maximum production of 47,194 flasks of 764 pounds each was realized in 1865, but in 1886 it was reduced to 18,000 flasks, the total for the period 1850-86 being 853,259 flasks, or about twothirds of the produce of the Spanish Almaden. total produce of the Californian mines, which was about 80,000 flasks in 1877, declined to 30,000 in 1886. The The second mine in point of importance, known as New Idria, is about 70 miles in a south-easterly direction from New Almaden, the ore, cinnabar, occurring under conditions similar to those in the latter mine-namely, in very irregular groups of fissures in metamorphic strata, which pass into others containing Neocomian fossils of the genus Aucella. These were succeeded by other Cretaceous and Tertiary formations up to the Miocene, the close of the latter period being marked by an upheaval and the commencement of volcanic activity. The ore deposits are closely related to the latter, and are probably nearly all, if not entirely, of post-Pliocene origin. In the Clear Lake region, in lat. 39° N., which adjoins the group of volcanic cones known as Mount Konocte (or Uncle Sam) hot springs and solfataras are abundant in a small area of basalt of comparatively recent origin. The most important of these, known as the Sulphur Bank, was at first worked for sulphur, but, on getting below the surface, cinnabar was found in the decomposed basalt, and for some years it produced large quantities of mercury, up to 11,152 flasks in 1881; but latterly the yield has fallen off, being only 1449 flasks in 1886. The Redington Mine, adjoining Knoxville, about 25 miles south-east of Clear Lake, was discovered in making a cutting for a road, and has been worked since 1862, and has produced nearly 100,000 flasks of mercury, a quantity which has only been exceeded by the mines of New Almaden and New Idria. In 1886 the yield had fallen to 409 flasks, the immense irregular body of ore at the surface having changed in depth to some narrow veins following fissures in the metamorphic Neocomian strata. These are to a large extent converted into serpentine; and a black opal, known as quicksilver rock, accompanied the ore, which was remarkable as consisting largely, in the upper workings at least, of amorphous black sulphide of mercury, or meta-cinnabar, a mineral that was there recognized in quantity for the first time. This deposit is considered to be the result of the action of hot springs in connection with an adjacent mass of basalt-springs which are now dormant except in so far that sulphur gases are given off and sulphur crystals are deposited in the old . workings, where a comparatively high temperature, exceeding 100° F., prevails. The Steamboat Springs in Nevada, near the Comstock lode, have been also studied by the author. These, although presenting no deposits of commercial value, are interesting from the light they cast upon the phenomena of the formation of mineral veins, and have therefore been carefully investigated by several observers, including the late Mr. J. A. Phillips, F.R.S., and M. Laur, of the École des Mines. The author considers that the main source of the ore in the Comstock lode is the diabase forming the hanging wall, and that the mineral contents were extracted from this pre-Tertiary eruptive mass by intensely heated waters charged with alkaline carbonates and sulphides rising from great depths, and that a similar origin may properly be attributed to all the cinnabar, pyrites, and gold found in the mercury-mines of the Pacific slope, having been brought in as solutions as double sulphides of metal and alkalies. The original source must have been either the fundamental granite of the country, or some infra-granitic mass, it being extremely improbable that they were extracted from any In connection with this subject, the author has made a series of interesting experiments on the relations of the sulphide of mercury to that of sodium, which show that mercuric sulphide is freely soluble in aqueous solutions of sodium sulphide, although the contrary has repeatedly been asserted. Mercuric sulphide may be precipitated from sulpho-salt solutions in many ways, particularly by excess of sulphuretted hydrogen, by borax and other mineral salts; by cooling, especially in the presence of ammonia, and by dilution. In the latter case, a certain quantity of metallic mercury separates as well as the sulphide, indicating one of the methods by which the native metal has been produced in Nature. volcanic rock at or near the surface. In addition to the mines specially described, the author has extended his study of the subject to a consideration of the principal mercury-mines other than those of America, partly from personal investigation in Spain and Italy, and partly with the help of other observers and published accounts. He expresses a very decided opinion against the supposed substitution origin of the Almaden deposits, considering them to be essentially of a vein-like character, the cinnabar being deposited in fissures or interstitial cavities in sandstone previously existing. This latter conclusion is substantially similar to that arrived at by the late Mr. J. A. Phillips and the present writer, in a microscopic study of the Almaden ores made some years since. The details of the foreign deposits have been very carefully collected, the comparatively new discoveries of Avala in Servia, and Bakmuth in Southern European Russia, being included. The latter mine, which, at the time the book was completed, was not at work, has since become of considerable importance. The ore, cinnabar, occurs as an impregnation of a bed of carboniferous sandstone from 14 to 17 feet thick, with an average yield of 154 pounds per ton-about 7 per cent.— and the reduction works have a productive capacity of about 10,000 flasks annually. In conclusion, it is scarcely necessary to state that the whole of the details illustrating the subject have been worked out with the care and fulness which have characterized the author's former monograph on the Comstock lode. Whether mercury-mining in California may be in a declining state, or destined to a revival of its former prosperity at a future time, there can be no question of the high value of the record of the results hitherto obtained, which is contained in the volume it has been our pleasant task to notice. H. B. OUR BOOK SHELF. Illustrations of some of the Grasses of the Southern Punjab, being Photo-lithographs of some of the Principal Grasses found at Hissar. By William Coldstream, B.A., Bengal Civil Service. With 38 Plates and 8 pages of Introduction. (London: Thacker and Co. Calcutta: Thacker and Spink. 1889.) Except THIS work contains a series of thirty-eight photo-lithographs of the grasses used for agricultural purposes in the southern portion of the Punjab. The tract of country to which it relates lies to the west of Delhi, between the Jumna on the east and the Sutlej on the west. It constituted till recently the civil district of Hissar, which has miles, and a population of a million and a half. now been broken up. It has an area of 8500 square along the streams and canals the soil is sterile and sandy, and the crops depend upon the periodical rains. The staple cereals are Sorghum vulgare and Penicillaria spicata. In its centre is situated the great Government cattle-farm of Hissar, where for many years cattle of the finest Indian breeds have been reared by Government, principally for the supply of the ordnance and transport departments, but also to some extent for distribution through the country, with the aim of improving the commoner indigenous kinds. The Bir, or grass lands, of this great farm are of very wide extent, and in the rainy season a large number of grasses, of more or less value as fodder, grow luxuriantly over its vast parks. The farm has altogether an area of above sixty square miles, and it is mainly from this that the species figured by Mr. Coldstream are taken. The book is modelled upon the "Fodder-grasses of India," published not long ago, in two volumes, by Mr. Duthie, the director of the botanical department of Northern India, and to Mr. Duthie the author is indebted for the botanical determination of the species. He gives the native name of each plant, and a short account of the extent and manner in which it is used, and as most of them have a wide dispersion, this will be found useful in other dry sub-tropical regions. Out of thirtyseven species, the two great tropical tribes are represented, Panicea by twelve species, and Andropogone by ten, and only three species fall under Festucca, the tribe to which most of our North European pasture grasses belong. The plates are lithographed from photographs, and do not contain any dissections. Plate III., called Panicum Crusgalli, is clearly not that species, but a. form of P. colonum, another variety of which is figured on Plate II. Mr. Coldstream also has got entirely wrong with his two species of Cyperus, figured on p. 38. The left-hand figure, called Cyperus species, is evidently Cyperus Ivia, Linn., a common weed throughout India in rice-fields. The left-hand figure, labelled Cyperus Tria, is not in flower. There is no such plant known to botany; Tria is doubtless a mistake for Ivia. The figure is quite unrecognizable, but from the native name appended, "Motha," it is most likely Cyperus rotundus. J. G. B. Elementary Dynamics of Particles and Solids. By W. M. Hicks, M.A., F.R.S. (London: Macmillan and Co., 1890.) IN this excellent treatise, extending over nearly 400 pages, the author introduces to the student the principles of dynamics. Although the book is issued under the latter title, it will be found to differ considerably in its treatment from the majority of text-books on the same subject. For instance, the two subjects of statics and kinetics have been considered together, the former being regarded as a special case of the latter. Again, the discussion of force is reserved until an attempt has been made to give an idea of mass and its measurement; thus a preliminary study of momentum finds an early place. Although the mathematical acquirements of the student of these pages may be limited to a knowledge of the elements of algebra and geometry, he will be able to readily follow the methods adopted in establishing the various results. This the author has kept in view throughout his work, except in a few cases where, in the hope of rendering it useful to a larger circle of readers, he has had recourse to the trigonometrical ratios for examples which he has worked out. The volume is divided into three portions (1) rectilinear motion of a particle; (2) forces in one plane; (3) plane motion of a rigid body. One cannot read the first few chapters without observing the care taken by the writer in trying to impart to the student a correct and precise idea of the fundamental units. That this is a very important matter all will agree who have had any experience in teaching or testing students. The most deplorable state of ignorance sometimes exhibited by them, in giving their results in all manner of absurd units, should encourage both teacher and author to make a special effort when dealing with the question of units, fundamental or otherwise. As the subject of statics is included, an opportunity has been taken of introducing the method of drawing stress diagrams for loaded framework; this will be valuable to engineering students. Notwithstanding that the writer has forbidden himself the use of the integral calculus, he has been able to establish (in some cases very neatly) many useful results in the two chapters on centre of gravity and moment of inertia, which should be read with care. Neatness in method characterizes the book throughout 2 A and an unusually large number of examples will be found at the end of each chapter. The work is based on a series of lectures delivered by the author at the Firth College, Sheffield, and many details for which time can generally be found at the lecture table have in this case found their way into the book. These will help to lessen the individual difficulties of students, and their views of the subject will be enlarged thereby. There can be little doubt that the text-book will have a deservedly favourable reception. G. A. B. Catalogue of the Fossil Reptilia and Amphibia in the British Museum (Natural History). Part III., containing the Order Chelonia. By Richard Lydekker, B.A, F.G.S., &c. (London: Printed by Order of the Trustees, 1889.) MR. LYDEKKER is to be congratulated on having added one more to the valuable series of catalogues of the palæontological collections in the British Museum which he has compiled during the last few years. Like his previous catalogues, the present work indicates an enormous amount of careful and accurate work, which, however, is of such a special kind that it cannot easily be summarized in a short review. The extreme difficulty of correlating the fossil forms of Chelonia with the recent, on account of the fragby the fact that, out of the 52 genera and 131 species mentary character of many of the remains, is indicated or varieties described, the author has only been able to place with certainty 18 genera and 10 species amongst existing forms. The classification adopted is to a great extent that followed by Mr. Boulenger in his catalogue of recent Chelonians. The work is illustrated by 53 woodcuts, and abundant references to the bibliography of the group are given. It must be added, as stated in the preface, that "the collection which forms the subject of this Catalogue is particularly rich in Chelonians from the Purbeck Beds of Swanage, the Cretaceous of England and Holland, the Eocene Tertiaries of Warwick, Sheppey, Hampshire, the Isle of Wight, and the older Pliocene of the Siwaliks of India." The last-named beds have yielded the largest tortoise known (Testudo [Colossochelys] atlas of Falconer), the carapace of which measures about six feet in length. 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.] Systems of "Russian Transliteration." As one who takes an interest in the Russian tongue, quite apart from the value of the scientific papers published in that language, I may perhaps be allowed to express my regret that the author of " A Uniform System of Russian Transliteration," published in your issue of February 27 (p. 397), has departed every point where it is possible to do so from the system of transliteration which has been in use in England for about a century, and which has, moreover, the advantage of being almost identical with that current in France. in almost A system of transliteration may be founded on one of two bases-namely, the empirical, in which little or no account is taken of the sound of the letters in the foreign language, and the rational; in the latter the letters of the foreign language are, where possible, represented by letters or groups of letters which have as nearly as may be the same sound as the original. For instance, B in Russian would be represented by B in English, these two having the same sound. It seems to me that the latter is the most cor.venient system, and the one which ought to be generally adopted; the author of this new "uniform system," however, has chosen the other course. If the author of the "uniform system" had been contented with tabulating the system of transliteration which has been so long in use, he would have earned the gratitude of those devoted to literature, as well as of those who cultivate science. As it is, I am afraid he has merely given the world of art and letters an opportunity for gibes at what they are sometimes pleased to call the narrowmindedness and pedantry of scientific men. I may, perhaps, be permitted to give a few examples of the defects of the new system; r in Russian has three sounds, one nearly resembling the English g, another very like h, and a third guttural sound, to which there is nothing analogous in our tongue. The author proposes to get over this by transliterating r by gh!! The eminent chemist Hemilian thus becomes masked as Ghemilian, whilst Gustavson appears as Ghustavson, and a well-known political character, Gortchakoff, is altered to Ghorchakov'. For comparison, I give these names, and a few others, as transliterated in accordance with the two systems : New system. Ghemilian. Ghustavson'. Ghorchakov'. Aleksyeev'. Ghrighor'ev'. Oghloblin. Mendelyeev'. Khrushchov. Mikhail. Zhukovskič. Geographical names are even more weird; for example, it becomes somewhat difficult to recognize under the disguise of Nizhni Novgorod and Volgha, the town of Nijni Novgorod and the River Volga. Such words as 66 'Journal" and "Chemie,' when occurring in titles, can be at once recognized; this can scarcely be said of them if the new system of transliteration is used, as they become "zhurnal" and Khimii" respectively. It is much to be regretted that the Royal Society, the Linnean Society, and the Geological Society should have pledged themselves to adopt this novel " system of transliteration," instead of adhering to the one which has been so long in use. Fellow of the Royal Society, I feel very great regret that the Council are going to adopt this system in their publications, as it will seriously detract from the value of their supplementary Catalogue of Scientific Papers" now in the press, at all events as far as Russian literature is concerned. 46 As a to two No protest of mine, however, can be half so forcible as the unconscious sarcasm of the author himself, in his paper, where he says that "an expression of grateful thanks is due Russians "who have assisted in the arrangement of the system." The names of the Russians are then given, and if my readers will take the trouble to study them by the light of the table for transliteration by the new system, he will see how they express their appreciation of the author's labours by carefully avoiding every one of the novelties he has introduced. CHARLES E. GROVES, Editor of the Journal of the Chemical Society. Burlington House, March 17. HAVING in view the increasing importance of Russian to literary and scientific men, it becomes very desirable to have a uniform system of transliteration, such as that recently proposed in your columns. But, in order to be useful, everyone must agree to conform to it, nor should any such system be adopted off-hand without full discussion of any points which may seem susceptible of improvement. It seems to me objectionable to indicate the semi-vowels (1 and b) by a simple, and to omit them altogether at the end of a word. They really correspond, to a certain extent, to our e (mute); and I would suggest that it would be better to indicate them by a full letter—perhaps & for one and é for the other. March II. W. F. KIRBY. ONE or two points in the criticisms on this subject call for some notice before the publication of a more detailed account of the system. As regards Mr. Kirby's suggestion, the transliteration of the semi-vowels was discussed, but it was not thought advisable to exaggerate their importance by using two letters for them, especially as their use is becoming discontinued in Russia. When recommending a uniform system, we did not imagine that Mr. Groves or anyone else would infer that this was intended to limit the right of Russians who dwell in England or who write in English to spell their names as they please ; we have not asked Messrs. Kelly to apply it to all Russian names in the Post Office Directory or the Court Guide; we should never think of altering such names in ordinary correspondence. Even in catalogues and records, for which this system is intended, the familiar form should of course be quoted with a cross reference, as recommended by us in the clause dealing with proper names. Mr. Groves asks why we have not tabulated "the system which has been in use in England for about a century." Our efforts began with an attempt to discover such a system, and resulted in the tabulation of a large number of systems, including that employed by Mr. Groves in the Journal of the Chemical Society; since, however, no two authors agree in the English symbols intended to represent either the sounds or letters of Russian words, we endeavoured to frame a system combining as far as possible the features of those already in use in England and America. We are much obliged to Mr. Groves for supplying further illustrations of the desirability of using gh for r; the letter has, of course, more than the three sounds to which he limits it. The uniformity of "the system which has been so long in use may be illustrated by the following examples, in which we confine ourselves to the names of chemists, and to the words quoted by Mr. Groves : Consulting the "Imperial Gazetteer," Lippincott's "Gazetteer," and Keith Johnston's "Atlas" alone, we find Nijni, Nijnei, Nishnii, Nizhnee, Nijnii, and Nischnii-Novgorod. One journal is given in Bolton's "Catalogue of Chemical Journals" as Zhurnal russkova khimicheskova i fizicheskova; in the Geological Record as Jurnal rosskoi chimitcheskago i phizitcheskago; and in Scudder's "Catalogue of Serials " as Zhurnal; russkoye khimitcheskoye i fizitcheskoye. unintelligible. Hence it is difficult to see why Nizhnii and Zhurnal should be In the Royal Society Catalogue, the Geological Record, and Chemical Society's Journal, the same name is spelt Jeremejew, Jeremejeff, Jereméeff. Which of these words represents the pronunciation ? In the Chemical Society's Journal, Wroblewski and Flawitzky correspond to the Wroblevsky and Flavitzsky of Armstrong and Groves' "Organic Chemistry." The same journal frequently quotes the name Markownikoff and ff, while in the examples of Mr. Groves it is also reprewhere the same Russian letter (and sound) is denoted both by sented by ; here, of course, and in similar cases, the name comes through a German channel. Mr. Groves transliterates a few names; since, however, in his "rational system one Russian letter has more than one English equivalent (v, ff), and one English letter (e) has more than one Russian equivalent, while the sound is not correctly represented (o, é), it is obvious that this is neither "rational' nor a system (it does not profess to be "empirical"; perhaps Mr. Groves will now call it the "graphic method"). Since, moreover, the system recommended by Mr. Groves is not used by him in the Chemical Society's Journal, we hope that he may yet see his way to adopting the one which has now been accepted by so many of the leading English Societies. H. A. M. J. W. G. "Like to Like "- -a Fundamental Principle in Bionomics. THE following letter has been intrusted to me for seeing through the press, and therefore I deem it desirable to state that it does not constitute the writer's reply to Mr. Wallace's criticism of his paper on "Divergent Evolution." This reply, as previously stated (NATURE, vol. xl. p. 645), will be published by him on some future occasion. I FOLLOW Prof. Lankester in the use of bionomics to designate the science treating of the relations of species to species. If the theory of evolution is true, bionomics should treat of the origin, not only of species, but of genera, and the higher groups in which the organic world now exists. In his very suggestive review of "Darwinism," by Mr. A. R. Wallace, in NATURE of October 10, 1889 (p. 566), Prof. Lankester refers to "his (Mr. Wallace's) theory of the importance of the principle of like to like' in the segregation of varieties, and the consequent development of new species." Prof. Lankester has here alluded to a principle which I consider more fundamental than natural selection, in that it not only explains whatever influence natural selection has in the formation of new species, but also indicates combinations of causes that may produce new species without the aid of diversity of natural selection. The form of like to like which Mr. Wallace discusses is "the constant preference of animals for their like, even in the case of slightly different varieties of the same species," which is considered not as an independent cause of divergence, but as producing isolation which facilitates the action of natural selection. If he had recognized this principle, which he calls selective association, as capable of producing in one phase of its action sexual and social segregation, and in another phase sexual and social selection, he would perhaps have seen that its power to produce divergence does not depend on its being aided by natural selection. Mr. Wallace's view is very clearly expressed in the following passages, though I find other passages which lead me to think that the chief reason he does not recognize segregation as the fundamental principle in divergence is that he has not observed its relations to the principle of like to like. He says :-"A great body of facts on the one hand, and some weighty arguments on the other, alike prove that specific characters have been, and could only have been, developed and fixed by natural selection because of their utility" ("Darwinism," p. 142). "Most writers on the subject consider the isolation of a portion of a species a very important factor in the formation of new species, while others maintain it to be absolutely essential. This latter view has arisen from an exaggerated opinion as to the power of intercrossing to keep down any variety or incipient species, and merge it in the parent stock' ("Darwinism," p. 144). For I think we shall reach a more consistent and complete apprehension of the subject by starting with the fundamental laws of heredity, and refusing to admit any assumption that is opposed to these principles, till sufficient reasons have been given. Laws which have been established by thousands of years of experiment in domesticating plants and animals, should be, it seems to me, consistently applied to the general theory of evolution. example, if in the case of domesticated animals, "it is only by isolation and pure breeding that any specially desired qualities can be increased by selection" (see Darwinism," p. 99), why is not the same condition equally essential in the formation of natural varieties and species? If in our experiments we find that careful selection of divergent variations of one stock does not result in increasingly divergent varieties unless free crossing be ween the varieties is prevented, why should it be considered an exaggeration to hold that in wild species "the power of intercrossing to keep down any variety or incipient species, and merge it in the parent stock," is the same. Experience shows that segregation, which is the bringing of like to like in groups that are prevented from crossing, is the fundamental principle in the divergence of the various forms of a given stock, rather than selection, which is like to like through the prevention of certain forms from propagating; and I think we introduce confusion, perplexity, and a network of inconsistencies into our exposition of the subject, whenever we assume that the latter is the fundamental factor, and especially when we assume that it can produce divergence without the cooperation of any cause of segregation dividing the forms that propagate into two or more groups of similars, or when we assume that segregation and divergence cannot be produced without the aid of diverse forms of selection in the different groups. The theory of divergence through segregation states the principle through which natural selection becomes a factor promoting sometimes the stability and sometimes the transformation of types, but never producing divergent transformation except as it co-operates with some form of isolation in producing segregation; and it maintains that, whenever variations whose ancestors have freely intergenerated are from any combination of causes subjected to persistent and cumulative forms of segregation, divergence more or less pronounced must be the result. The laws of heredity on which this principle rests may be given in the three following statements: (1) Unlike to unlike, or the removal of segregating influences, is a principle that results either in extinction through failure to propagate, or in the breaking down of divergences through free crossing. (2) Like to like, when the individuals of each intergenerating group represent the average character of the group, is a principle through which the stability of existing types is promoted. (3) Like to like, when the individuals of each group represent other than the average character of the group, is a principle through which the transformation of types is effected. In my paper on "Divergent Evolution" (Linn. Soc. Journ., Zoology, vol. xx. pp. 189-274), I pointed out that sexual and social instincts often conspire together to bring like to like in groups that do not cross, and that in such cases there will be divergence even when there is no diversity of natural selection in the different groups, as, for example, when the different groups Occupy the same area, and are guided by the same habits in their use of the environment. There is reason to believe that under such circumstances divergence often arises somewhat in the following way. Local segregation of a partial nature results in some diversity of colour or in some peculiar development of accessory plumes, and through the principle of social segregation, which leads animals to prefer to associate with those whose appearance has become familiar to them, the variation is prevented from being submerged by intercrossing. There next ari es a double process of sexual and social selection, whereby both the peculiar external character and the internal instinct that leads those thus characterized to associate together are intensified. The instinct is intensified, because any member of the community that is deficient in the desire to keep with companions of that kind will stray away and fail of breeding with the rest. This process I call social selection. The peculiarity of colour or plumage is preserved and accumulated, because any individual deficient in the characteristic is less likely to succeed in pairing and leaving progeny. This latter process is sexual selection. It can hardly be questioned that both these principles are operative in producing permanent varieties and initial species; and in the circumstances I have supposed, I do not see how the process can be attributed to natural selection. Varieties thus segregated may often develop divergent habits in their use of the environment, resulting in divergent forms of natural selection, and producing additional changes; but so long as their habits of using the environment remain unchanged, their divergencies cannot be due to natural selection. 66 66 means Mr. Wallace's very interesting section on "Colour as a Means of Recognition," taken in connection with the section on Selective Association," already referred to, and another on Sexual Characters due to Natural Selection," offers an explanation of "the curious fact that p ominent differences of colour often distinguish species otherwise very closely allied to each other" (p. 226). His exposition differs from mine in that he denies the influence of sexual selection, and attributes the whole process to natural selection, on the ground that of easy recognition must be of vital importance" (p. 217). The reasoning, however, seems to me to be defective, because the general necessity for means of easy recognition is taken as equivalent to the necessity for a specialization of recognition marks that shall enable the different varieties to avoid crossing. In the cases I am considering, there is, however, no advantage in the separate breeding of the different varieties, and even in cases where there is such an advantage (as there would be if the variety had habits enabling it to escape from competition with the parent stock, but only partially preventing it from crossing with the same), it does not appear how this advantage can prevent the individual that is defective in the special colouring from following and associating with those that are more clearly marked. The significant part of the process in the development of recognition marks must be in the failure of such individuals to secure mates, which is sexual selection; or in the unwillingness of the |