THE ,<«?*. UNIVERSITY OF CALIFORNIA AT LOS ANGELES IFOKNV STATE NORMALSCHOOL, - LOS AKOELE3, -•- OAL. -//A THE INTERNATIONAL SCIENTIFIC SERIES VOLUME LXVII THE INTERNATIONAL SCIENTIFIC SERIES. Each book complete in One Volume, 12mo, and bound in Cloth. 1. FORMS OF WATER : A Familiar Exposition of the Origin and Phenomena of Glaciers. By J. TYNDALL, LL. D., F. R. S. With 25 Illustrations. $1.50. 2. PHYSICS AND POLITICS; Or, Thoughts on the Application of the Prin- ciples of "Natural Selection" and "Inheritance" to Political Society. By WALTER BAGEHOT. $1.50. 3. FOODS. By EDWARD SMITH, M. D., LL. B., F. B. 8. With numerous Illus- trations. $1.75. 4. MIND AND BODY : The Theories of their Relation. By ALEXANDER BAIN, LL.D. With 4 Illustrations. $1.50. 6. THE STUDY OF SOCIOLOGY. By HERBERT SPENCER. $1.50. 6. THE NEW CHEMISTRY. By Professor J. P. COOKE, of Harvard Univer- sity. With 31 Dlostrations. $2.00. 7. ON THE CONSERVATION OF ENERGY. By BALFOCH STEWART, M. A., LL.D., F.R.S. With 14 Illustrations. $1.50. 8. ANIMAL LOCOMOTION; or, Walking, Swimming, and Flying. ByJ. B. PETTIOREW, M. D., F. R. S., etc. With 130 Illustrations. $1.75. 9. RESPONSIBILITY IN MENTAL DISEASE. By HENRY MATOSLEY M D $1.50. 10. THE SCIENCE OF LAW. By Professor SHELDON AKOS. $1.75. 11. ANIMAL MECHANISM: A Treatise on Terrestrial and Agrial Locomotion. By Professor E. J. MARET. With 117 Illustrations. $1.75. 12. THE HISTORY OF THE CONFLICT BETWEEN RELIGION AND SCI- ENCE. By J. W. DRAPER, M. D. , LL. D. $1 .75. 13. THE DOCTRINE OF DESCENT AND DARWINISM. By Professor OPCAR SCHMIDT (Strasbnrg University). With 26 Illustrations. $1.50. 14. THE CHEMICAL EFFECTS OF LIGHT AND PHOTOGRAPHY. By Dr. HERMANN VOGEL (Polytechnic Academy of Berlin). Translation thoroughly revised. With 100 Illustrations. $2.00. 15. FUNGI : Their Nature, Influences. Uses, etc. By M. C. OOOKE, M. A., LL. D. Edited by the Eev. M. J. Berkeley, M. A., F. L. S. With 109 Illustrations. $1.50. 16. THE LIFE AND GROWTH OF LANGUAGE. By Professor WILLIAM DWIGHT WHITNEY, of Yale College. $1.50. New York : D. APPLETON & CO., 1, 8, & 5 Bond Street. The International Scientific Seri«»-(Continued.) 17 MONEY AND THE MECHANISM OF EXCHANGE. By W. STANLEY JEVOS8,M.A.,F.R.S. $1.75. 18 THE NATURE OP LIGHT, with a General Account of Physical Optics. By Dr. EUGENE LOMMEL. With 188 Illustrations and a Table of Spectra in Cbromo-lithography. $2.00. 19. ANIMAL PARASITES AND MESSMATES. By Monsieur VAN BENEDEN. With 83 Illustrations. $1.50. 50. FERMENTATION. By Professor ScHflrzENBEBGEB. With 28 Illustrations. $1.50. 51. THE FIVE SENSES OF MAN. By Professor BERNSTEIN. With 91 Illus- trations. $1.75. 88. THE THEORY OF SOUND IN ITS RELATION TO MUSIC. By Pro- fessor PIETRO BLASEBNA. With numerous Illustrations. $1.50. 28. STUDIES IN SPECTRUM ANALYSIS. By J. NORMAN LOCKTEB, F.R.S. With 6 Photographic Illustrations of Spectra, and numerous Engravings on Wood. $2.50. 84. A HISTORY OF THE GROWTH OF THE STEAM-ENGINE. By Pro- fessor R. II. THURSTON. With 163 Illustrations. $2.60. 88. EDUCATION AS A SCIENCE. By ALEXANDER BAIN, LL.D. $1.75. •8. STUDENTS' TEXT-BOOK OF COLOR; Or, Modern Chromatics. With Applications to Art and Industry. By Professor OGDEN N. ROOD, Colum- bia College. New edition. With 130 Illustrations. $2.00. 87. THE HUMAN SPECIES. By Professor A DE QUATBEFAGES, Membre de I'lnstitut. $2.00. 88. THE CRAYFISH : An Introduction to the Study of Zoology. By T. H. HUXLET, F. R. 8. With 82 Dlnstrations. $1.75. 89. THE ATOMIC THEORY. By Professor A. WUBTZ. Translated by E. Cleminahaw, F. C. S. $1.50. 80. ANIMAL LIFE AS AFFECTED BY THE NATURAL CONDITIONS OF EXISTENCE. ByKABLSEMPEB. With 2 Maps and 106 Woodcuts. $2.00. 81. SIGHT: An Exposition of the Principles of Monocular and Binocular Vision. By JOSEPH Li CONTE, LL. D. With 132 Illustrations. $1.50. SS. GENERAL PHYSIOLOGY OF MUSCLES AND NERVES. By ProfesFor J. ROSENTBAL. With 75 Illustrations. $1.50. W. ILLUSIONS : A Psychological Study. By JAMES SULLY. $1.50. »4. THE SUN. By C. A. YOUNG. Professor of Astronomy in the College of New Jersey. With numerous Illustrations. $2.00. Ntw York : D. APPLETON & CO., 1, 3, & 5 Bond Street. Tlt€ International Scientific terte*— (Continued.) S3. VOLCANOES: What they Are and what they Teach. By JOHN W. JUDD, K. R. s., Professor of Geology iii the Royal School of Mines. With 96 11- lustrations. $2.00. 36. SUICIDE: An Essay in Comparative Moral Statistics. By HENRY MOR- SELLI, M.D., Professor of Psychological Medicine, Royal University, Turin, $1.75. 57. THE FORMATION OF VEGETABLE MOULD, THROUGH THK AC- TION OF WORMS. With Observations on their Habits. By CHARLES DARWIN, LL. D., F. R. 8. With Illustrations. $1.50. 38. THE CONCEPTS AND THEORIES OF MODERN PHYSICS. By J. B. STALLO. $1.75. 39. THE B.EAIN AND ITS FUNCTIONS. By J. LIT*. $150. 40. MYTH AND SCIENCE. By TITO VIGNOLI. $1.50. 41. DISEASES OF MEMORY: An Essay in the Positive Psychology. By Te. RIBOT, author of " Heredity." $1.50. 42. ANTS, BEES. AND WASPS. A Record of Observations of the Habits of the Social Hymenoptera. By Sir JOHN LTTBBOCK, Bart.. F. R. S.. D. C. L., LL. D., etc. $2.00. 43. SCIENCE OF POLITICS. By SHELDON AMOS. $1.75. 44. ANIMAL INTELLIGENCE. By GEORGE J. ROMANES. $1.75. 45. MAX BEFORE METALS. By N. JOLT. Correspondent of the Institute. With 148 Illustrations. $1.75. 46. THE ORGANS OF SPEECH AND THEIR APPLICATION IN THE FORMATION OF ARTICULATE SOUNDS. By «. H. VON MEYER. Pro- fessor in Ordinary of Anatomy at the University of Ziirieh With 47 Woodcuts. $1.75. 47. FALLACIES : A View of Logic from the Practical Side. By AI.PREP SIDGWICK, B.A., Oxon. $1.75. 48. ORIGIN OF CULTIVATED PLANTS. By ALPHONSK T,E CAKDOI.I.K. $2.00. 49. JELLY-FISH, STAR-FISH, AND SEA-URCHINS. Being a Research on Primitive Nervous Systems. By GEORGE J. ROMANES. $1.75. 50. THE COMMON SENSE OF THE EXACT SCIENCES. By the late WILL- IAM KINGDON CLIFFORD. $1.50. 51. PHYSICAL EXPRESSION: Its Modes and Principles. By FRANCIS WAR- NER, M.D.. Assistant Physician, and Lecturer on Botany to the Londor Hospital, etc. With 51 Illustrations. *1.75. 52 ANTHROPOID APES. By ROBERT HARTMANN. Professor in the University of Berlin. With (53 Illustrations. $1.75. New York : D. APPLETON & CO.. 1, 8, & 5 Bond Street. < The International Scientific Seri«s- (Continued.) 68. THE MAMMALIA IN THEIR RELATION TO PRIMEVAL TIMES. By OSCAR SCHMIDT. $1.50. >! COMPARATIVE LITERATURE. By HCTCHESOX MACAULAY POBNKTT, M \., LL. D., V. L. S.. Barrister-at-Law ; Professor of Classics and English Literature, University College, Auckland, New Zealand; author of "The Historical Method," etc. fl.75. 55. EARTHQUAKES AND OTHER EARTH MOVEMENTS. By JOHN MILNE, Professor of Mining and Geology in the Imperial College of Engineering, Tokio, Japan. With 38 Figures. $1.75. 58. MICROBES, FERMENTS, AND MOULDS. By E. L. TBOUESSART. With 107 Illustration*. $1.50. 57. THE GEOGRAPHICAL AND GEOLOGICAL DISTRIBUTION OF ANI- MALS. By ANOBLO HKILPRIN. $2.00. 68. WEATHER. A Popular Exposition of the Nature of Weather Changes from Day to Day. With Dhgrams. By Hon. RALPH ABERCBOMBT. $1.75. ». ANIMAL MAGNETISM. By ALFRED BINET and CHARLES FER£, Assistant Physician at the SalpStriere. $1.50. 60. INTERNATIONAL LAW, with Materials for a Code of International Law. By LEONE LEVI, Professor of Common Law, King's College. $1.SO. 61. TUB GEOLOGICAL HISTORY OF PLANTS. With Illustrations. By Sir J. WILLIAM DAWSON, LL. D., F. R. 8. $1.75. 91. ANTHROPOLOGY. An Introduction to the Study of Man and Civilization. By EDWARD B. TTLOR, D. C. L., F. R. S. Illustrated. $2.00. 63. THE ORIGIN OF FLORAL STRUCTURES, THROUGH INSECT AND CTHER AGENCIES. By the Rev. GEORGE HENSLOW, M.A., F. L. S., F.O.S. With 88 Illustrations. $1.75. 64. ON THE SENSES. INSTINCTS, AND INTELLIGENCE OF ANIMALS, WITH SPECIAL REFERENCE TO INSECTS. By Sir JOHN LUBBOCK. With ovrr 100 Illustrations. $1.75. 65. THE PRIMITIVE FAMILY IN ITS ORIGIN AND DEVELOPMENT. By Dr. C. N. STABCKE, of the University of Copenhagen. 66. PHYSIOLOGY OF BODILY EXERCISE. By FEBNAND LAGBANQK, M. D. 67. THE COLOURS OF ANIMALS: Their Meaning and Use. By EDWABD BAGNALL POULTON, F. R.8. MIMICRY IN SOUTH AFRICAN BUTTERFLIES. KlGCRKS 1, 4, AND 5, THK FEMALES OF A SoUTH AFRICAN PafiKo, TOTALLY UNLIKE THE MAI (FlGHHB iX BltT MIMICKING RESPECTIVELY THREE SPECIF.S OF THE UNPALATABLE GENUS, DdnaiS (FlGURl 3*,- 4a, AND sa) THE FEMALE (FIGURE 2) OF A NEARLY ALUHD Papilla, IN MADAGASCAR, is NOT MIMETI AND RKSEMHLES THE HALE. DESCBIPTION OF PLATE THE figures have been copied, by kind permission of the author, and the council of the Linnean Society, from the plates accompany- ing Mr. Koland Trimen's paper, ' On some Remarkable Mimetic Analogies among African Butterflies.' ('Linn. Soc. Trans.' vol. xxvi. pp. 497, et seqq.) All figures are represented one-half of their natural size. The appearance of the under side of the wings is shown on the left hand of the four upper figures. Figure 1. — The male of Papilio merope (now called P. cenea ; the name P. merope being restricted to the West African form), from Knysna, Cape Colony. A closely allied butterfly (P. meriones), with a very similar male, is found in Madagascar. Figure 2. — The female of Papilio meriones, from Madagascar. The male is almost exactly like Figure 1. The black bar on the costal margin of the fore wing of the female probably represents the beginning of the darkening which has been carried so far in the females of the African P. merope and P. cenea. Figure 3. — First or cenea-torm of female of Papilio merope (now called P. cenea), from Knysna, Cape Colony. The female is totally unlike the male of the same species (Figure 1), but closely mimics an unpalatable butterfly, Danais echeria, prevalent in its locality. The appearance of the latter is shown in Figure Ba. The mimetic resemblance is seen to be very striking on both upper and under sides of the wings. A local variety of the Danais is also mimicked by a corresponding variety of the Papilio. Figure 4.— Second or hippocoon-form of female of Papilio merope (now called P. cenea), from Graham's Town, Cape Colony. This variety mimics the southern form of the unpalatable Danais niavius, shown in Figure 4a. Figure 5. --Third or tropJionius-tona. of female of Papilio merope (now called P. cenea), from Knysna, Cape Colony. This variety mimics the abundant and unpalatable Danais chrysippus shown in Figure 5a. In a closely allied species of Papilio from West Africa (the true Papilio merope) the male closely resembles Figure 1, while there are two mimetic varieties of female. The hippocoon-iorm is like Figure 4, except that it is larger and the white patch on the hind wing is smaller ; corresponding in both these respects to the West African variety of Danais niavius. The trophonitis-torm resembles Figure 5. There is no cenea-torm of this species. For further details see pp. 234-38. THE INTERNATIONAL SCIENTIFIC SERIES THE COLOURS OF ANIMALS THEIR MEANING AND USE, ESPECIALLY CONSIDERED IN THE CASE OF INSECTS BY EDWARD BAGNALL POULTON, M.A., F.R.S. WITH CHROMOLITHOGRAPHS FRONTISPIECE AND SIXTY-SIX FIGURES IN TEXT NEW YORK D. APPLETON AND COMPANY 1890 son Qu PEEFACB I HAVE adopted a general title, 'The Colours of Animals,' in order to indicate the contents of this volume, although the vast majority of the examples are taken from insects, and indeed almost invariably from a single order, the Lepidoptera. The examples are, however, employed merely to illustrate principles which are of wide appli cation. I have purposely abstained from multiplying in- stances when a little observation or even reflection will supply them in large numbers. For example, the ordinary Protective Resemblances of mammals and birds are barely alluded to, on this account. On the other hand, more difficult problems, such as the change of colour in arctic mammals, or the meaning of the colours of birds' eggs, are treated at far greater length. My object in both cases is the same : to stimulate observation in a subject which will amply repay investigation, from the scientific value of the results, and the never-failing interest and charm of the inquiry. viii THE COLOURS OF ANIMALS Variable Protective Eesemblance in insects is treated in considerable detail, for the reasons given above, and because much of the work is so recent that no complete account can be found outside the original memoirs. My chief object has been to demonstrate the utility of colour and marking in animals. In many cases I have attempted to prove that Natural Selection has sufficed to accouDt for the results achieved ; and I fully believe that further knowledge will prove that this principle explains the origin of all appearances except those which are due to the subordinate prin- ciple of Sexual Selection, and a few comparatively unimportant instances which are due to Isolation or to Correlation of Growth. In support of these views I have endeavoured to bring together a large amount of experimental evi- dence in favour of the theories as to the various uses of colour. 'Further experiments are still greatly needed. In the chapters on ' Sexual Selection ' I have argued in favour of Darwin's views, and have attempted to defend them against recently published attacks. At the conclusion of the volume I have brought forward a detailed classification of the various uses of colour, in which new, and, I believe, more con- venient terms are suggested. Definitions and exam- ples are also given in the classification, which is, in fact, a brief abstract of the whole book. PEEFACE IX I have to thank the Councils of various scientific societies for the courteous permission to copy figures from their respective publications. The figures in the coloured plate are copied from the plates accom- panying Mr. Eoland Trimen's paper in the ' Trans- actions of the Linnean Society,' vol. xxvi. pp. 497-522. Figures 18, 19, 20, 21, and 22 are copied from the plate accompanying Mr. E. Bowdler Sharpe's paper in the ' Proceedings of the Zoological Society,' 1873, pp. 414 et seqq. Figures 3, 4, 5, 6, 7, 8, 11, 14, 58, 60, 61, 62 are copied from the plates and woodcuts accompanying my papers in the ' Transac- tions of the Entomological Society,' 1884, 1885, 1887, and 1888. Figures 25, 26, and 27 are copied from the plate accompanying Mrs. Barber's paper in the ' Transactions of the Entomological Society,' 1874, pp. 519 et seqq. Figures 29 and 30 are copied from the plate and woodcuts accompanying my paper in the ' Philosophical Transactions of the Eoyal Society,' vol. 178 (1887), B, pp. 311-441. Figures 15, 16, 53, 54, 63, 64, 65, 66 are copied from the woodcuts and plates accompanying G. W. and E. G. Peckham's paper in the ' Occasional Papers of the Natural History Society of Wisconsin,' vol. i. (1889), Milwaukee. Figures 55 and 56 are copied from the plates accom- panying Professor Weismann's ' Studies in the Theory of Descent,' translated by Professor Meldola. Figure 10 is copied from one of the plates accompanying Dr. Wilhelm Midler's ' Siidamerikanische Nymphaliden- X THE COLOURS OF ANIMALS raupen' (' Zoologische Jahrbiicher,' J. W. Spengel, Jena, 1886). Figure 42 is copied from Vogt (' The Natural History of Animals ' : English translation : Blackie and Son). Figures 44 and 45 are copied from the plates accompanying Curtis's 'British Lepidoptera.' The remaining figures are original. Figure 17 was kindly lent me by Dr. A. K. Wallace, to whom it had been sent by Mr. Wood-Mason. In preparing the drawings of the original figures I have been greatly assisted by my wife, my sister Miss L. S. Poulton, Miss Herman Fisher, Mr. Alfred Sich, Mr. Alfred Robinson, and especially by Miss Cundell. I have almost invariably referred to original papers from which facts or conclusions have been adopted; so that any reader having access to a scientific library may easily gain possession of further details. Not wishing to overburden the book with such notes, I have abstained from referring constantly to my own papers, although most of the examples are taken from them. A list of my papers which deal with the colours of insects is therefore printed below. •Transactions Entomological Society,' London, 1884, pp. 27-60 1885, „ 281-329 1886, „ 137-179 »» » „ „ 1887, „ 281-321 * .. >. „ 1888, „ 515-606 • Philos. Trans. Royal Society,' vol. 178 (1887), B, pp. 311-441 Abstract of the above in 'Proceedings Royal Society,' 1887, vol. xlii. pp. 94-108 ' Proceedings Royal Society,' 1885, vol. xxxviii. pp. 269-315 1886, vol. xl. pp. 135-173 ' Proceedings Zoological Society,' 1887, pp. 191-274 PEEFACE Xi Short papers or notes (exclusive of those which are mere abstracts of the above) : — ' Proceedings Entomological Society,' London, 1887, pp. 1-li „ 1887, „ Ixi-lxii „ „ 1888, p. v »> »> >» »» >» PP- viii— x „ „ „ „ „ „ xxvii-xxviii „ 1889 „ xxxvii-xl ' Journal of the Victoria Institute,' 1888, vol. xxii., ' On Mimicry.' It is my pleasant duty to thank many friends for their kind assistance. I owe to Professor Meldola more than I can possibly express : his writings first induced me to enter upon this line of investigation, and I have had the benefit of his great experience and wise advice during the whole of the time that I have been at work. Nearly every subject touched upon in this volume has been discussed with him. Professor Westwood has always been most kind in helping me with the literature of the subject, with which he has so intimate an acquaintance, and in giving me the free use of the Hope collection at Oxford. Professor E. Eay Lankester has read the proof-sheets dealing with the classifications of the uses of colour, and has offered valuable suggestions. Several beautiful examples were suggested to me by Professor C. Stewart. Dr. Giinther, Mr. Eoland Trimen, Mr. Oldfield Thomas, Mr. E. Bowdler Sharpe, Mr. F. E. Beddard, Mr. W. W. Fowler, and Mr. A. H. Cocks have been very kind in answering questions upon their special subjects. Sir John Conroy has xii THE COLOURS OF ANIMALS kindly helped me in explaining the physical questions involved in the first chapter. I am especially pleased to speak of the help received from my former pupils Mr. W. Garstangand Mr. E. C. L. Perkins, who have supplied many valuable instances, which are specified in the volume, where other kind assistance is also duly acknowledged. Although I have ventured to disagree with my friend Dr. A. R. Wallace upon the subject of ' Sexual Selection,' I wish to acknowledge how very much I owe to his writings, which I have very frequently quoted. I have also made great use of the late Thomas Belt's extremely interesting and suggestive ' Naturalist in Nicaragua.' Among recent papers I wish especially to mention that by G. W. and E. G. Peckham, of Milwaukee, U.S.A. The minute observation of these authors upon the courtship of spiders of the family Attidce is a model for investigation in a subject which has never before been attacked systematically. Above all, I should wish to acknowledge, although I can never fully express, the depth of my indebted- ness to the principles which first made Biology a science, the principles enunciated by Charles Darwin. It is common enough nowadays to hear of new hypotheses which are believed (by their inventors) to explain the fact of evolution. These hypotheses are as destructive of one another as they are supposed to be of Natural Selection, which remains as the one PREFACE xiii solid foundation upon which evolution rests. I have wished to express this conviction because my name has been used as part of the support for an opposite opinion, by an anonymous writer in the ' Edinburgh Eeview.'1 In an article in which unfairness is as conspicuous as the prejudice to which it is due, I am classed as one of those ' industrious young observers ' who ' are accumulating facts telling with more or less force against pure Darwinism.' 2 On the strength of this and other almost equally strange evidence, the Eeviewer triumphantly exclaims, ' Darwin, the thanes fly from thee ! ' In view of this public mention of my name, I may perhaps be excused for making the per- sonal statement that any scientific work which I have had the opportunity of doing has been inspired by one firm purpose — the desire to support, in however small a degree, and to illustrate by new examples, those great principles which we owe to the life and writings of Charles Darwin, and especially the pre- eminent principle of Natural Selection. E. B. P. OXFORD : Dec. 28, 1889. 1 Edinburgh Review. Article V. April 1888, pp. 407-47. * p. 443. The bias of the writer appears in a most singular manner upon this page. In the short space of seventeen lines the following adjectives are divided between five writers and their works — industrious, illustrious, gifted, well-read, acute, intelligent, brilliant, thoughtful. I need hardly say that all five writers are believed by the Reviewer to oppose the theory of Natural Selection. CONTENTS CHAPTER PAGE I. THE PHYSICAL CAUSE OF AKIMAL COLOURS ... 1 II. THE USES OF COLOUR 12 III. PROTECTIVE KESEMBLANCES IN LEPIDOPTEBA ... 24 IV. PROTECTIVE RESEMBLANCES IN LEPIDOPTERA (continued) — DIMORPHISM, ETC 42 V. PROTECTIVE RESEMBLANCES IN VERTEBRATA, ETC. . . 60 VI. AGGRESSIVE RESEMBLANCES — ADVENTITIOUS PROTECTION 72 VII. VARIABLE PROTECTIVE RESEMBLANCE IN VERTEBRATA, ETC 81 VIII. VARIABLE PROTECTIVE RESEMBLANCE IN INSECTS . . 110 IX. VARIABLE PROTECTIVE RESEMBLANCE m INSECTS (con- tinued) 133 X. WARNING COLOURS 159 XI. WARNING COLOURS (continued) 189 XII. PROTECTIVE MIMICRY 216 XIII. PROTECTIVE AND AGGRESSIVE MIMICRY .... 245 XIV. THE COMBINATION OF MANY METHODS OF DEFENCE . 269 XV. COLOURS PRODUCED BY COURTSHIP . . . . 284 XVI. OTHER THEORIES OF SEXUAL COLOURING . . . 314 XVII. SUMMARY AND CLASSIFICATION 336 THE COLOURS OP ANIMALS CHAPTER I THE PHYSICAL CAUSE OF ANIMAL COLOUBS Colours due to absorption THE colours of animals are produced in various ways. By far the commonest method is the absorption of certain elements of light by means of special sub- stances which are called pigments, or colouring matters. The colour of each pigment is due to those elements of the light which it does not absorb, and which can therefore emerge and affect the eye of the spectator. Black is, of course, caused by the absorption of all the constituents of light, so that nothing reaches the eye. The colour of red pigment, like that of red glass, depends upon the fact that red is less absorbed than any other element of the light which passes through. If a sheet of red glass be placed upon white paper, the light traverses the glass, is reflected 2 THE COLOUKS OF ANIMALS from the surface of the paper, re-traverses the glass, and emerges. Similarly, in painting, bright effects are produced by covering a surface of Chinese white with the desired colour. The light passing twice through the thickness of the colour, absorption is far more complete than when only one thickness is traversed, as in a piece of red glass held up to the light. Absorption being more complete, the red colour is deeper. Animal pigments are nearly always twice traversed by the light, and therefore a very thin layer produces a considerable effect. Animal colours are therefore generally due to precisely the same optical principle which causes the colour of a wall-paper, a carpet, or a picture. Certain transparent animals are, however, for the most part coloured by light which passes but once through them, upon the same principle as the colours of a stained- glass window. The beautiful transparent blue of many pelagic animals, such as the Portuguese Man- of-war (Physalia), is caused in this way. It would be out of place to discuss the details of the causes of colour by absorption. I may, however, mention that vibrations of very different rates are started in the luminiferous ether by the sun, the electric light, &c. A certain series of these vibrations causes the effect of white light when it falls on our retina; but there are vibrations above and below this visible series — vibrations which we cannot see. We can, however, prove their existence in other ways ; THE PHYSICAL CAUSE OF ANIMAL COLOURS 3 and it is certain that some animals can see vibrations which do not affect our eyes.1 The slowest vibrations that we can see, produce the effect of red, the most rapid the effect of violet, while the intermediate vibrations cause the other well-known colours of the rainbow or the spectrum. The absorption of certain elements of light there- fore means the disappearance of ethereal vibrations with a certain speed. It is believed that these vibrations disappear because their motion has been communicated to the particles of the absorbing body. It is also believed that these particles are in a state of constant vibration, and that the vibrations of ether, which are timed to those of the body, are used up in increasing the motion of the latter. A white appearance due to light being scattered The production of white is due to a different principle, for we know that when light passes through a body without any absorption, the body is trans- parent and invisible rather than white. When all the light passes through, the body is completely in- visible. Whiteness is due to reflection of the whole visible series of vibrations, unaccompanied by the absorption of a part of them, as in the production of colours. But regular reflection, viz. reflection from 1 Sir John Lubbock, The Senses of Animals, Chapter X. (Inter- national Scientific Series). 4 THE COLOURS OT? ANIMALS a polished surface like that of a mirror, does not cause whiteness : it renders the surface itself invisible, but produces images of surrounding objects. A white appearance is produced by irregular reflection, which causes the light to be scattered or reflected in all directions. To produce such a result there must be an immense number of surfaces in an immense number of different directions. If a coloured sub- stance be reduced to powder of various degrees of fineness, the colour will diminish in intensity, and the whiteness will increase, according to the fineness of the powder ; this is because the number of re- flecting surfaces is increased, while the thickness of the grains is diminished. This will be clear from the following consideration. When a beam of light falls on a sheet of glass, a known fraction (about 4 per cent.) of the light is reflected back from the first surface : the larger portion, however, enters the glass, and, after suffering a certain amount of absorption, reaches the second surface and is again partially reflected. If the glass be powdered, the number of surfaces will be so im- mensely increased that all the light will be reflected by a small thickness of the powder. The light reflected from the second surface of each grain of coloured glass will still be coloured by absorption, but not sufficiently to produce any visible results, when the thickness of the grain is very small. Reflection is the immediate cause of whiteness, and the amount of reflection is due to the difference THE PHYSICAL CAUSE OF ANIMAL COLOURS 5 between the refractive powers (viz. the power of changing the direction of rays of light) possessed by the grains of glass and the substance, such as air or water, which lies between them. Thus the refractive powers of glass and water are much nearer than those of glass and air : hence a dry powder will reflect far more than a wet one, and will appear much whiter. To take a few familiar examples : snow is white, because of the minute globules of air which refract very differently from the crystals between which they are entangled ; ice, on the other hand, is transparent. If snow be compressed the air is driven out, and the mass becomes transparent ; if ice be powdered it be- comes white like snow. The froth of a coloured liquid is not coloured like the latter, but is white. Milk and fat are white because light is scattered from the surfaces of the countless oil globules, which refract very differently from the substance which lies between them. The surface of well-polished glass is almost invisible, because it reflects regularly, but a scratched surface is very visible, because there are surfaces in many different directions, which therefore scatter the light, while the far more numerous surfaces of ground glass scatter the light far more effectually and produce a white appearance. The white markings of animals are produced in various ways. White hairs and feathers owe their appearance, like snow, to the number of minute 2 6 THE COLOURS OF ANIMALS bubbles of gas which are contained in their inter- stices. Fat is also made use of to give a white appearance ; and the same result may be obtained by the presence of minute granules, probably akin to pigment, but differing widely from it in optical pro- perties, in that no absorption takes place. Colours due to thin plates It has been stated already that when light traverses a sheet of glass surrounded by air, a certain pro- portion of it is reflected back at the first surface and a certain proportion at the second surface. The light will be reflected in the same direction from both surfaces. It is believed that the vibrations of ether, some of which affect us as light, are in the form of undulations of different lengths ; if, therefore, the sheet of glass be sufficiently thin, some of the undula- tions reflected from the second surface will interfere with those started from the first surface. This will happen when the sheet is of such a thickness that the wave of light reflected from the second surface is half an undulation behind that reflected from the first surface ; for then the two sets of undulations will be in opposite directions, and will therefore neutralise each other. This will be quite clear if we apply the same reasoning to those visible undulations from which the name itself has been borrowed — the waves on the surface of water. If a set of ripples is started by the THE PHYSICAL CAUSE OF ANIMAL COLOURS 7 motion of an object in still water, and then another set is started from another object moved, so that the ripples succeed each other at exactly the same rate as the preceding set, and if the second set is begun when the first has advanced half a complete ripple (viz. a movement up and down), it is clear that the upward movement of the second will correspond to the downward movement of the first and vice versa, so that, if the objects are so placed that the two sets of ripples are traversing the same sheet of water, they will neutralise and destroy each other. If we compare a number of sheets of glass which are successively thinner and thinner, interference will first occur among the longest undulations of light, because half an undulation will of course require a greater distance (or thickness) than when the undula- tions are shorter. As thinner and thinner sheets are examined interference will gradually pass through the whole spectrum from red to violet, destroying sets of waves with shorter and shorter undulations. The colour seen in each case will be due to the other sets of waves which are not destroyed. The amount of reflection, and therefore of inter- ference and of colour produced, depends upon the difference between the refractive power of the thin sheet and the substance on each side of it. Such interference colours are seen in a soap- bubble, and the colours change as the bubble be- comes larger and the film thinner : they differ, too, 8 THE COLOUKS OF ANIMALS on the various parts of the bubble, because the thin- ness also varies. A bubble of melted glass may be blown thin enough to produce the same effects, which are also well seen when a thin layer of air is enclosed between two sheets of glass or between the plates of some crystals, or when a drop of oil is allowed to spread out into a thin film on the surface of water. When a substance has a laminated structure, and sufficiently thin films are enclosed between the laminae, very marked effects are seen. Thus the metallic appearance of the laminated flakes which are formed on the surface of glass which has been long buried in the earth, is accounted for. If these brilliant flakes are wetted the colour fades away, because the thin films of air between the laminas are displaced by water, with a refractive power much nearer to that of the glass, and the amount of reflected light is there- fore diminished. Interference colours due to thin films are certainly very important among animals, but the extent to which they occur is imperfectly known. The irides- cent colours of many beetles' wings are probably due to thin films of air included between layers of a horny consistence Such colours are unaltered in dried specimens. In other cases the chinks between the layers are kept open by films of less powerfully re- fractive liquids. When the tissue becomes dry the films evaporate and the colour disappears. We must suppose that the denser layers come together, THE PHYSICAL CAUSE OF ANIMAL COLOUKS 9 obliterating the chinks and excluding the air; otherwise the colours would be more brilliant than ever, because the refractive power of air is even lower than that of the liquids. The brilliant metallic appearance of many chrysalides, especially in the genus Vanessa, is caused by a large number of films of liquid enclosed between the laminae of the dense outer layer. If the pupa be kept in spirit or water the colour remains, but disappears directly it is dry, although it can be renewed any number of times by wetting. This may even occur in a living animal, for Dr. Sharp has just directed my attention to an interesting observation made by Dr. Nickerl, who found that a brilliantly golden beetle (Carabus auronitens) lost all its lustre after hybernating in captivity, but entirely regained it after drinking some water. Colours due to diffraction When white light falls upon a surface on which there are a number of fine parallel grooves the re- flected light appears coloured, the colour varying with the angle at which the light falls on the surface, and with the angle at which it is seen. This is due to the light reflected from different portions of the surface having different distances to travel before reaching the observer : and when (as occurs when the grooves are very close together) these differences amount to half a complete undulation for any particular length of vibration, interference is caused, and the vibration 10 THE COLOURS OF ANIMALS of that particular rate is wanting from the reflected light, which therefore appears coloured. Opinions differ as to the relative importance of animal colours due to thin plates and to diffraction. Many which were believed to result from the latter are in all probability due to the former. The irides- cent colours on the inner surface of many shells (mother-of-pearl) are at any rate partially caused by diffraction, for an accurate cast of the surface exhibits traces of the colours.1 The shell is, however, a laminated structure, and the colours may therefore in part be caused by thin plates. Colours due to refraction (prismatic colours) When light passes through a wedge-shaped trans- parent substance (or prism) with greater refractive power than the surrounding medium, it is bent in the same direction at both surfaces, but its different con- stituents are bent unequally. The slowest vibrations (red) are bent least, the most rapid (violet) most ; and when the substance possesses a sufficiently high re- fractive power, all the colours of white light are seen arranged like the rays of a fan in the order of their rates of vibration. Prismatic colours like those of the diamond are due to refraction. 1 Professor C. Stewart informs me that he has repeated Brewster's original experiment, upon which the above statement depends. He found that the colour was due to a thin layer of shell which had been stripped off and adhered to the surface of the wax. THE PHYSICAL CAUSE OF ANIMAL COLOURS 11 It is doubtful how far the colours of animals are caused by this principle ; but Dr. Gadow has given strong reasons for supposing that the metallic colours of birds' feathers are produced in this way,1 and there are scales on the wings-cases of certain beetles (Pachyrhynchus) which also may owe their colours to refraction. All these causes of animal colours may be conve- niently grouped under two heads — (1) pigmentary, and (2) structural. The first head includes colours caused by absorption, and the effects produced vary with the chemical nature of the substance (pigment). The second head includes the colours or appearances pro- duced in all other ways, the efficient cause being the structure of the substance rather than its chemical nature. 1 Proc Zool. Soc. 1882, pp. 409 et ae%. 12 THE COLOUKS OF ANIMALS CHAPTEE H THE USES OF COLOUR I. Non-significant colours COLOUR, as such, is not necessarily of any value to an organism. Organic substances frequently possess a chemical and physical structure which causes certain light-waves to be absorbed ; or the elements of tis- sues may be so arranged that light is scattered, or interference colours are produced. Thus blood is red, fat is white, and the external surface of the air- bladder in certain fishes has a metallic lustre, like silver. In such cases there is no reason why we should inquire as to the use or meaning of the colour in the animal economy ; the colour, as such, has no more meaning than it has in a crystal of sulphate of copper or iron. Such colours are the incidental results of chemical or physical structure, which is valuable to the organism on its own account. This argument will be still further enforced if we remember that the colours in question are, strictly speaking, not colours at all. Blood and fat are so constituted that they will be red and white, respectively, in the presence of THE USES OF COLOUR 13 light, but they cannot be said to possess these colours in their normal position, buried beneath the opaque surface of an animal. The existence of non- significant colours is, never- theless, most important, for they form the material out of which natural or sexual selection can create significant colours. Thus, the colour of blood may be made use of for ' complexion,' while fat may be em- ployed to produce white markings, as in certain insect larvae. The yellow, brown, and red fatty matters of the connective tissue are accumulated beneath the skin in patches, so as to produce patterns. All colour originally non-significant All animal colour must have been originally non- significant, for although selective agencies have found manifold uses for colour, this fact can never have accounted for its first appearance. It has, however, been shown that this first appearance presents no difficulty, for colour is always liable to occur as an incidental result. This is even true of the various substances which seem to be specially set apart for the production of colour in animals ; for pigments occur abundantly in the internal organs and tissues of many forms. The brilliant colours of some of the lower organisms are probably also non-significant. In all higher animals, however, the colours on the surface of the body have been significant for a vast period of time, so that their amount, their arrangement in 14 THE COLOURS OF ANIMALS patterns, their varying tints, and their relation to the different parts of the body, have all been determined by natural selection through innumerable generations. Because the origin of all pigments is to be found in the incidental result of the chemical and physical nature of organic compounds, it by no means follows that incidental or non-significant colours would have appeared at all on the surface of most animals. And we find as a matter of fact that such colours tend to disappear altogether, directly they cease to be useful, as in cave-dwelling animals. On the other hand, the non-significant colour of blood or of fat would persist undiminished in such forms. Colours may be destroyed by natural selection Just as natural selection may develop an appear- ance which harmonises with the surroundings, out of the material provided by non-significant colour, the same agency may lead to the disappearance of the latter when it impedes the success of an animal hi the struggle for existence. Thus the red colour of blood has disappeared in certain transparent fishes, which are thereby concealed from their enemies. Among the manifold possible variations of nature is that of a fish with colourless blood, which can, nevertheless, efficiently perform all the duties of this fluid. While such a variation would be no advantage to the great majority of vertebrates, it would be very beneficial to THE USES OF COLOUR 15 a fish which was already difficult to detect on the surface of the ocean on account of its transparency. II. Significant colours Colours may be useful in many ways, and are there- fore always liable to be turned to account in one direc- tion or another. They may be of direct physiological value to the organism, or may assist in the struggle for existence by deluding other species, or by aiding the individuals of the same species, or they may be intimately connected with courtship. 1. The Direct Physiological Value of colour The colour of chlorophyll, which causes the green appearance of vegetation, must be intimately con- nected with the important changes which take place in this substance in the presence of light. It is well known that under these circumstances carbon dioxide (popularly called ' carbonic acid ') can be split up, and its carbon made to unite with the elements of water, forming organic substance. Although this process has been much studied it is still very imper- fectly understood. It is clear, however, that the colour of chlorophyll, involving the special absorp- tion of certain light-waves, has some direct bearing upon the changes which occur. No equally clear instance has been proved to occur in the animal kingdom, except in those few forms 16 THE COLOURS OF ANIMALS which resemble plants in possessing chlorophyll. Dr. Hickson, however, believes that among corals ' the most widely distributed colours will eventually be proved to be allied to chlorophyll . . . and perform a very simi- lar if not precisely identical physiological function.' It is much to be desired that this interesting sugges- tion, which Dr. Hickson supports by many arguments, may be thoroughly tested as soon as possible.1 In the very common association of coloured sub- stances with the important function of respiration, it is clear that the colour is not more than incidental ; while the fish with transparent blood shows that colour is not indispensable for the due performance of the function. Pigment is, however, of direct import- ance for vision : it is always present in the eyes of animals, except in the case of albinos, and it is said that even they possess the essential visual pigment associated with the termination of the optic nerve (retinal purple). The difference between the physiological importance of colour in animals and plants is well shown by the fact that a true albino variety (not merely a varie- gated example) of a green plant could not live for any length of iime. There are, however, certain cases among animals in which it is extremely probable that colour is of direct physiological value. It is well known that dark colours readily absorb radiant heat, while light 1 A Naturalist in North Cekbes (Hiokson, 1889), pp. 149-51. THE USES OF COLOUR 17 colours do so with difficulty. For this reason black clothes are most trying, and white most comfortable, in the hottest weather. Conversely, a dark surface readily parts with heat by radiation, while a white surface retains heat far more effectually. A few writers had suggested that these principles may explain the colours of certain animals, but the question was first fully entered upon in Lord Wal- singham's presidential address to the Yorkshire Naturalists' Union in 1885. 1 The predominance of dark varieties of insects and white varieties of birds and mammals in northern latitudes is connected with these facts. ' Birds and animals living through the winter naturally require to retain in their bodies a sufficient amount of heat to enable them to maintain their existence, with unreduced vitality, against the severities of the climate. Insects, on the contrary, require rapidly to take advantage of transient gleams of sunshine during the short summer season, and may be content to sink into a dormant condition so soon as they have secured the reproduction of their species ; only to be revived in some instances by a return of exceptionally favourable conditions.' It would be fatal for the temperature of one of the higher vertebrates " to sink a few degrees below the normal, except in the case of certain species, such as the dormouse, &c., which have the power of hybernating in a dormant condition; such animals were once 1 See Entomological Transactions of the Union for 1885. 18 THE COLOURS OF ANIMALS called 'warm-blooded,' but are now more correctly termed ' homothermic,' because it is the constancy of the temperature which is so important, and which must be maintained whether the surrounding medium be colder or warmer than themselves. Other animals with an inconstant temperature are now correctly called ' poikilothermic ' rather than ' cold-blooded.' Lord Walsingham's conclusions appear to be sup- ported by the fact that young dark-coloured cater- pillars, like those of the Emperor Moth (Saturnia carpini), or Tortoiseshell Butterfly (Vanessa urticce), seek the light side of a glass cylinder, and always change their position when the cylinder is turned round. The question needs further investigation, and much might be learnt by interposing various screens between such larvae and the light, thus cutting off different sets of light-waves. The most important support to the hypothesis is found in an experiment made by Lord Walsingham, in which several Lepidoptera of different colours were placed on a surface of snow exposed to bright sun- shine ; in half an hour the snow beneath the darker insects showed distinct signs of melting, but no effects were seen beneath the others. The differences were further brought out in the course of two hours, when the darkest insect of the lot, a black Geometer, the Chimney-Sweeper (Odezia charophyllata) , ' had de- cidedly won the downward race among them.' It is therefore certain that the absorption of THE USES OF COLOUR 19 radiant heat is favoured by the dark colours of northern insects, and it is in every way probable that they are benefited by the warmth received in this way. We cannot, however, as yet assert that such dark colours are not also advantageous for concealment or some other purpose. The white appearance of Arctic birds and mam- mals must be advantageous for concealment in a region so largely covered with snow, but it is very probable that advantage is also secured by checking the loss of heat through radiation. Thus Lord Walsingham's experiments and con- clusions seem to prove that colours are sometimes of direct physiological value to animals, although a great deal more work must be done before we can safely estimate the proportion which this advantage bears to others also conferred by the same colours (see also pages 92-104). 2. Protective and Aggressive Resemblance By far the most widespread use of colour is to assist an animal in escaping from its enemies or in capturing its prey ; the former is Protective, the latter Aggressive. It is probable that these were the first uses to which non-significant colours were put. The re- semblances are of various kinds ; the commonest cases are those of simple concealment. The animal passes undetected by resembling some common object which 20 THE COLOUES OF ANIMALS is of no interest to its enemies or prey respectively, or by harmonising with the general effect of its sur- roundings ; the former is Special, the latter General Resemblance, and both may be Protective or Aggressive. Among the most interesting Special Aggressive Resem- blances are the cases of Alluring Colouring, in which the animal, or some part of it, resembles an object which is attractive to its prey. 3. Protective and Aggressive Mimicry Mimicry is in reality a very important section of Special Resemblance. The animal gains advantage by a superficial resemblance to some other, and generally very different, species which is well known and dreaded because of some unpleasant quality, such as a sting or an offensive taste or smell, &c., or it may even be pro- tected from the animal it resembles : this is Protective Mimicry. When, however, the animal resembles another so as to be able to injure the latter or some other form which accompanies it or is not afraid of it, the Mimicry is Aggressive. Although, strictly speaking, Mimicry should fall under the last heading, it is so important and so different from the other examples of Special Resemblance that it is more convenient to consider it separately. In the complete classification at the end of the book it will be shown in its true position. THE USES OF COLOUR 21 4. Warning colours When an animal possesses an unpleasant attribute, it is often to its advantage to advertise the fact as publicly as possible. In this way it escapes a great deal of experimental ' tasting.' The conspicuous patterns and strongly contrasted colours which serve as the signal of danger or inedibility are known as Warning Colours. In other cases such colours or markings enable individuals of the same species easily to follow those in front to a place of safety, or assist them in keeping together when safety depends upon numbers. It is these Warning Colours which are nearly always the objects of Protective Mimicry, and it will therefore be convenient to describe the former before the latter. 5. Colours produced by Courtship Finally, in the highest animals, the vertebrata and many of the most specialised invertebrate groups, we have some evidence for the existence of an aesthetic sense. Darwin believed that this sense was brought into play in courtship, and that colours and patterns have been gradually modified by the preference of the females for the most beautiful males ; he believed that such Sexual Selection accounts for many of the most beautiful features possessed by animals, viz. those which are especially displayed during courtship. 22 THE COLOURS OF ANIMALS Although this hypothesis has been rejected by A. B. Wallace, I shall endeavour to support it by some striking observations of recent date, and by as far as possible answering the objections which have been raised, and the hypotheses which have been believed to account for the same facts. Display in courtship is probably the most recently developed of all the various uses of colour among animals, and as such, its consideration is best deferred until all the others have been described. It must not be supposed that the colours of each animal will be found to possess but a single use. Thus Protective EesemUances are often supplemented by Warning Colours or attitudes, which give the animal an extra chance of escape after its first line of defence has been broken through. It is also the general rule for the colours displayed in courtship to be hidden beneath protective tints when the animal is at rest. The colours of animals may be recapitulated as follows : I. NON-SIGNIFICANT COLOURS. II. SIGNIFICANT COLOURS. 1. Colours of Direct Physiological Value. 2. Protective and Aggressive Resemblance. 8. Protective and Aggressive Mimicry. 4. Warning Colours. 6. Colours displayed in Courtship. THE USES OF COLOUR 23 The rest of this volume will be occupied with the further consideration of the last four classes of colours. It will be remembered that the third class is but a special example of the second, which it is convenient to treat separately, and to defer until after the fourth class has been considered. 24 THE COLQUES OF ANIMALS CHAPTER m PROTECTIVE RESEMBLANCES IN LEPIDOPTERA THE first and most important use of colour is to enable an animal to conceal itself from its enemies or to approach its prey unseen.1 Special and General Resemblances These results may be achieved in one of two ways : either the animal may more or less exactly resemble some object which is of no interest to its enemies, or it may harmonise with the general artistic effect of its surroundings, so that it does not attract attention. We may therefore distinguish Special Resemblance, in which the appearance of a particular object is copied in shape and outline as well as in colour, and General Resemblance, in which the general effects of surround- ing colours are reproduced. In the latter case it is often difficult to believe, 1 This was thoroughly appreciated by Erasmus Darwin, who says : ' The colours of many animals seem adapted to their purposes of concealing themselves, either to avoid danger or to spring upon their prey.' — Zoonomia, 1794, vol. i. p. 509. PKOTECTIVE RESEMBLANCES IN LEPIDOPTERA 25 when we look at the animal by itself, that the pro- tection is effective and real. We cannot appreciate the meaning of the colours of many animals apart from their surroundings, because we do not compre- hend the complicated artistic effect of the latter. A caterpillar in the midst of green leaves may have many brilliant tints upon it, and yet may be all the better concealed because of their presence ; the ap- pearance of the foliage is really less simple than we imagine, for changes are wrought by varied lights and shadows playing upon colours which are in them- selves far from uniform. Francis Galton noticed this fact with regard to the higher animals in 1851. ' Snakes and lizards are the most brilliant of animals ; bat ah1 these, if viewed at a distance, or with an eye whose focus is adjusted, not exactly at the animal itself, but to an object more or less distant than it, become apparently of one hue and lose all their gaudiness. No more conspicuous animal can well be conceived, according to common idea, than a zebra ; but on a bright starlight night the breathing of one may be heard close by you, and yet you will be positively unable to see the animal. If the black stripes were more numerous he would be seen as a black mass ; if the white, as a white one ; but their proportion is such as exactly to match the pale tint which arid ground possesses when seen by moonlight.' l 1 Galton's South Africa, p. 187 (Minerva Library). 26 THE COLOURS OF ANIMALS We shall see that it is common for an insect to be protected by Special Resemblance at one time of its life, and by General Resemblance at another, or to be concealed at different periods of its life by different kinds of Special or General Resemblance respectively. Each of these forms of Resemblance may be Protective or Aggressive according as they are made use of to defend from attack or to assist in capture. We shall also see that Protective and Aggressive Re semblances may be either Constant or Variable ; in the latter case, the appearance is capable of adjust- ment in order to correspond with changes in the en- vironment. This, the highest form of Resemblance, will be deferred until the examples of the other form have been considered. The Larvse of Geometrse as examples of Special Protective Resemblance There is no better instance of Special Protective Resemblance than that afforded by the larvae of Geometra, * stick caterpillars ' or ' loopers,' as they are often called. These caterpillars are extremely common, and between two and three hundred species are found in this country ; but the great majority are rarely seen because of their perfect resemblance to the twigs of the plants upon which they feed. They possess only two pairs of claspers, or legs which are peculiar to the caterpillar stage, while in nearly all PEOTECTIVE KESEMBLANCES IN LEPIDOPTEEA 27 other caterpillars there are five pairs. These claspers are placed at the hind end of the body, which is long, thin, and cylindrical, and stands out like a side twig at an acute angle with the stem to which the claspers are tightly fixed. The body also often possesses little humps which resemble buds or irregu- larities of the bark. The caterpillar sits motionless FIG. 1.— The larva of Swallow-tail Moth (Ourapteryx tambucaria) ; last stage ; natural size. FIG. 2.— Twig of currant ; the general appearance much like that of fig. 1. for hours together, and the only alteration of the attitude is brought about by feeding, which generally takes place in the evening or at night. The general appearance of one of these larvse and its resemblance to a twig is shown in figs. 1 and 2, for which I am indebted to the kindness of Mr. Alfred Sich. The strain on the body during these long periods 28 THE COLOUES OF ANIMALS of absolute stillness would be far too great to be borne did not the caterpillar spin a thread of silk, •which is attached at one end to the stem, while the other end remains fixed to the head of the animal. How great the strain would be without such a support may be well understood by any one who has tried to hold out the arm straight at right angles to the body for five minutes. There is considerable tension upon the thread of silk, so that, if it be cut, the larva falls back with a jerk, making a more obtuse angle with the stem ; and it then tries to remain rigid in the new position : this is im- possible because of the strain, and after again falling backwards once or twice, and making one or two more attempts to keep firm and motionless, it is obliged to give up the twig-like position while it fixes a new sup- porting thread. In some cases the caterpillar gains support by holding a leaf or twig with one of its three pairs of true legs, or legs which will persist in the perfect insect (see fig. 3 ; also figs. 40 and 41, page 152). These pairs of legs are placed one on each of the body-rings behind the head. It is very interesting to notice how the head of these caterpillars is modified from the usual shape into one which suggests the end of a twig. It is very FIG. 3.— The larva of Peppered Moth (Am- phidasii bttularia) • last stage; natural PROTECTIVE RESEMBLANCES IN LEPIDOPTERA 29 common for the crown to be deeply notched, thus producing two humps which make a very natural end to the apparent twig. In the caterpillar of the Small Emerald Moth (Hemithea thymiaria) there are two additional humps on the body-ring (prothorax) behind the head, and the latter is bent forwards and inwards, so that the end of the caterpillar is made up of four blunt projections, forming perhaps the most suggestive of all the resem- blances to the end of a twig. In the larva of the Early Thorn Moth (Selenia illunaria) the head and first two body-rings are bent backwards at right angles to the rest of the body. The supporting thread of silk passes between the third pair of true legs, which a,re borne by a high ridge projecting from the angle. The ridge continues the line of the body, and is coloured like it, while the head and first rings are of a different colour. The whole effect is exceedingly un- caterpillar-like, and very suggestive of some eccentric vegetable growth (see fig. 4). In order that the resemblance may be complete, it is essential that the caterpillar should appear to grow out of the branch in a natural manner. The two pairs of claspers assist in producing this effect, for they partially encircle the branch, and appear to be 3 FIG. 4.— The larva of Early Thorn Moth (Selenia illunaria) ; adult ; natural size. 30 THE COLOURS OF ANIMALS continuous with it (see fig. 7, page 31). Between the two pairs there is necessarily a furrow, where the body of the larva lies along the cylindrical branch. This furrow, which, if apparent, would greatly interfere with the resemblance, is rendered inconspicuous in the fol- lowing manner. The under side of the caterpillar is somewhat flattened, so that it is in contact with a small part of the circumference of the branch, and the furrow on each side is partially filled up, at any rate in certain species, by a number of fleshy tubercles. The shadow which would betray the furrow is also neutralised by the light colour of the tubercles. The effect will be clear on comparing a, b, and c in fig. 5 : a is a section of a branch just below the point where a lateral twig comes off ; 6 a diagrammatic section of a branch and the caterpillar's body ; c the same with the addition of the tubercles, which render the outline more like that of a. I will illustrate the extraordinary degree of resem- PKOTECTIVE RESEMBLANCES IN LEPIDOPTERA 31 blance attained in Geometra, by a description of the larva of one of our most abundant species — the Brim- stone Moth (Rumia cratcegata). The appearance of the larva when seated among the twigs of its commonest food -plant — hawthorn — is shown in fig. 6. It will be observed that some of the twigs are slightly bent in the middle, and that a projection is placed on the angle; these appearances are exactly reproduced in Fio. 6.— The larva of FIG. 7.-The hind part of the larva of Brimstone Brimstone Moth (Ru- Moth (Rumia cratoegata), seen from right mia craiaegatd) ; last side, showing junction with branch • adult stage ; natural size. the larva. The hind part of the larva is represented in fig. 7 (magnified 4-5 diameters), showing the claspers and the fleshy projections which occupy the furrow between the larva and the stem. The harmony of colour is quite as perfect as the resemblance of shape. The smaller branches of the hawthorn are partially covered by a thin superficial layer of a bluish-grey colour (the cuticle), while the 32 THE COLOURS OF ANIMALS deeper layers beneath are brown or green, or mixed brown and green ; these tints become visible over a large part of the surface, owing to the breaking away of the thin layer. Hence the colour of the branches is brown or green, mottled with grey, and not only are these the exact tints of the larva, but the way in which the colours are blended is precisely similar hi the annual and the plant. The marvellous fidelity with which the details are thus reproduced, probably implies that the relation between the larva and this species of food-plant is extremely ancient. It will be shown below that this caterpillar can also adjust its colour to that of its individual surroundings, so that it would become greenish if it passed its life among young green shoots, and brown if it lived upon the older twigs. It is altogether one of the most perfectly concealed forms in existence. When, however, such 'stick caterpillars' are young, they do not sit upon the branches, but upon the leaves of their food-plant, and the twig-like attitude would then be inappropriate, for we do not see twigs projecting from leaves. In some cases the caterpillars are green (e.g. Ephyra omicronaria) , and so possess a general harmony with the colour of the surface behind them; but in other cases they are brown, and then the attitude is often modified into a different form of Special Resemblance. The caterpillar twists itself into a very irregular spiral (e.g. Ephyra pendularia, &c.), or into an exceedingly PROTECTIVE RESEMBLANCES IN LEPEDOPTERA 33 angular zigzag (e.g. Selenia illunaria; see fig. 8), thus resembling a dead and crumpled piece of leaf, or the spiral leaf-case made by other insects, or the excre- ment of birds or snails. The caterpillar of Selenia illunaria has a very similar structure and colouring at the times when it resembles such very different objects as a twig and the excrement of a bird, the whole difference being made by a modification of atti- tude alone (compare figs. 4 and 8). I have seen the larva of the Brimstone Moth twisted into a spiral, resting motionless close to the notch which it had eaten out of a leaf ; in this position it forcibly suggested the appearance of a *«». 8--The you Early Thorn (• small piece of leaf which had been accidentally torn, and had turned brown and curled up, remaining attached to the uninjured part of the leaf by one end. We may well suppose that the acquisition of a form and attitude which lend themselves so readily to the purposes of concealment, was very advantageous to the ancestral Geometra, and enabled them to spread over the vegetable world, dividing into an immense number of species, and ousting many larvae with less perfect methods of concealment. In their widening range certain Geometrce have thus come to feed upon low-growing plants which are altogether without twigs or branches. The attitude is then modified, and sug- gests some object which might be expected to occur 34 THE COLOURS OF ANIMATE upon the plant. Thus the caterpillar of the Straw Belle (Aspilates gilvaria), feeding upon such plants as yarrow and plantain, coils up the anterior part of its body into a flat spiral, with the head in the centre. Hence the attitude and the whitish colour of the larva produce a very considerable resemblance to a small bleached and empty snail-shell, which would be of no interest to any insect- eater. If the colour of the caterpillar were darker it might be mistaken for a living snail, and it is doubtful how far such a resemblance would be to its ad- vantage, in the case of birds. Another larva, that of the Large Emerald Moth (Geometra papilionaria), feeding upon catkin -bearing trees, birch and nut, resembles the catkins rather than the twigs (see fig. 9). It is short and stout, and the manner in which the body-rings succeed each other forcibly suggests the overlapping scales of a cat- kin. Some of the larva? are green and some brown, like catkins of different colours. FIG. 9. - The larva of Large Emerald Moth (Geometra papilionaria); a green va- riety ; last stage; natural Protective Resemblance to bark and lichen in Lepidoptera Certain caterpillars belonging to other groups are concealed by their resemblance to the bark of tolerably PROTECTIVE RESEMBLANCES IN LEPIDOPTERA 35 thick branches. They lie flattened and closely pressed against the bark ; while the furrow which would lead to their detection is partially filled up, and the shadow neutralised, by a row of fleshy protuberances in the caterpillars of the Eed and Crimson Under- wing Moths (Catocalida ; see figs. 38 and 39, page 151), and by hairs in the larva of the December Moth (Pcecilocampa populi). This interpretation was first offered by Meldola, and it is strongly supported by the previously mentioned fact that similar protuber- ances occur in Geometrce, and are strictly confined to the comparatively short line of contact between the larva and the branch. The lichens on the bark are very commonly resembled rather than the bark itself. This is the case with the last-named larva. The caterpillar, chrysalis, and moth of the Black Arches (Psilura monacha) are beautifully protected in this way. The black pupa is fixed in a chink in the bark by a few inconspicuous threads ; its dark colour har- monises with the shadow in the chink, while the long tufts of greyish hair project and exactly resemble the appearance of lichen. Both larva and moth are coloured so as to resemble common appearances pre- sented by lichens, and both habitually rest on lichen- covered bark. A lichen-feeding Geometer (Cleora lichenaria) is wonderfully protected in the same manner ; the larva often twists itself among the irregularities of the lichen, so that it is completely 36 THE COLOURS OF ANIMALS invisible. The moth is also similarly concealed, and rests on tree-trunks. A caterpillar which makes its surroundings resemble itself In all the examples hitherto described, and count- less others, the insect is defended by resembling its surroundings ; the very interesting caterpillar of a South American butterfly. (Anaea sp. ?), described by Wilhelm Muller, acts differently and makes its surround- ings resemble itself. It gnaws the leaf in such a manner as to leave a number of rough models of it- self attached to the mid-rib, and then sits down beside them. The cater- pillar is green above and dark beneath, although the former colour interrupts the latter at certain points and comes into contact with the mid-rib on which the insect is resting. The dark colour is not distinguishable from the deep shadow behind the leaf, and therefore the appearance is that of an elongated patch of green connected with the mid- FIG. 10. — The larva of Anaea sp. t on the mid-rib of a leaf on which are many pieces of leaf of the general appearance of the larva; third stage ; natural size ; after Wilhelm Muller. PROTECTIVE RESEMBLANCES IN LEPIDOPTERA 37 rib by two narrow stalks. The larva, in eating, leaves several pieces of leaf attached to the mid-rib by one or two stalks, which, therefore, present a very similar appearance to that of the larva itself. The concealment which is thus effected is sufficiently indicated in fig. 10. An appearance of leaf-like flatness conveyed by arrangement of colour Another very interesting example, in which the effect of shadow is gained by arrangement of colour, is afforded by the chrysalis of the Purple Emperor Butterfly (Apatura iris). The large green pupa re- sembles a leaf in the most perfect manner, mid-rib and oblique veining being represented. I showed a small twig of sallow, to which a pupa was suspended, to several friends, but it was almost invariably over- looked ; even when it was pointed out, the observer frequently failed to see any difference between it and a sallow leaf. The most extraordinary thing about this resemblance is the impression of leaf-like flatness conveyed by a chrysalis, which is in reality very far from flat. In its thickest part the pupa is 8'5 mm. across, and it is in all parts very many times thicker than a leaf. The dorsal side of the pupa forms a very thin sharp ridge for part of its length, but the slope is much more pronounced in other parts and along the whole ventral side. But exactly in these 38 THE COLOURS OF ANIMALS places, where the obvious thickness would destroy the resemblance to a leaf, the whole effect of the round- ness is neutralised by increased lightness, so disposed as just to compensate for the shadow by which alone we judge of the roundness of small objects. The degree of whiteness is produced by the relative abun- dance of white dots and a fine white marbling of the surface, which is everywhere present mingled with the green. The effect is, in fact, produced by a process exactly analogous to stippling. The degree of lightness produced in this way exactly corresponds to the angle of the slope, which, of course, determines the depth of the shadow. By this beautiful and simple method the pupa appears to be as flat as a leaf which is only a small fraction of 1 mm. in thickness. Although the effect which I have just described could not have been surpassed by the efforts of an artist, it is precisely the result which can be most readily explained by the unaided operation of natural selection. The minute white markings are present. over the whole surface, and their number and size must be subject to continual variation ; in fact, it is quite certain that no two individuals are alike in these respects. The increased protection afforded by their more appropriate distribution in certain individuals would clearly lead to the survival of the latter, while the same process continued in each generation would lead to the elaborate and beautiful form of adaptation which is now witnessed in this species. PROTECTIVE RESEMBLANCES IN LEPIDOFTERA 39 An analogous effect is produced by the larva of a Saw-fly (a plant-eating Hymenopterous insect), which rests stretched along the edge of a leaf. In this posi- tion the larva (Nematus curtispina) would be detected if it covered the notched edge of the leaf ; it has, how- ever, the habit of resting along the curved edge of the gap made by its own exertions. From the side its green ground colour is alone apparent, and it is very difficult to detect. When, however, the leaf is looked at edgeways, it would seem that the larva must be con- spicuous, because its thickness is much greater than that of a leaf. From this point of view the back of the larva is, of course, seen; along the middle line the tubular heart is more distinct than usual because of FIG. 11.— The larva of Nematus curtispina ; the transparent skin. The green must stage ; natural blood within makes the heart appear as a fine dark line against a white border on each side, which is entirely due to an accumulation of fat beneath the skin. The white band with the fine dark line down its middle produces the effect of the edge of the leaf, while the rest of the body on each side is green, shaded with dark pigment so as to appear much flatter than it really is. The appearance of the larva is shown in fig. 11. The case is also of special interest, because the colouring is chiefly derived from internal tissues or organs showing through a transparent skin. The ground colour is due to the green fluids of the body 40 THE COLOURS OF ANIMALS and the green contents of the alimentary canal. The dark shading is the only part of the appearance caused in the usual way by superficially placed pigment. Nearly all the colours of this animal are non- significant in many other insects. Eeply to objection that methods of concealment would certainly be detected It has been sometimes objected that these methods of concealment cannot be intended as a means of defence, because insect-eating animals would be sharp-sighted enough to penetrate the disguise. Of course, the progressive improvement in the means of concealment has been attended by a corresponding increase in the keenness of foes, so that no species can wholly escape. But so long as a well-concealed form remains motionless, it is easy to prove by experiment that enemies are often unable to recog- nise it. Thus I have found that the insect-eating, wood-haunting Green Lizard (Lacerta viridis) will generally fail to detect a ' stick caterpillar ' in its position of rest, although it is seized and greedily devoured directly it moves. The marvellous resem- blance of Cleora lichenaria (see p. 35) even deceived one of these lizards after the larva had moved more than once. The instant the caterpillar became rigid the lizard appeared puzzled, and seemed unable to realise that the apparent piece of lichen was good to PROTECTIVE RESEMBLANCES IN LEPIDOPTERA 41 eat. After a few moments, however, the lizard was satisfied, and ate the caterpillar with the keenest relish. Furthermore, the fact that all well-concealed forms are good for food, and are eagerly chased and devoured by insectivorous animals, while unpalatable forms are conspicuously coloured, points strongly towards the conclusion that the object of concealment is defence from enemies. 42 THE COLOUKS OF ANIMALS CHAPTER IV PROTECTIVE RESEMBLANCES IN LEPIDOPTERA (continued), DIMORPHISM, ETC. General Protective Resemblance and changes of colour corresponding to changes in the surroundings ALL the examples hitherto described illustrate Special Protective Resemblance. A good instance of General Resemblance is afforded by the large and common caterpillar of the Privet Hawk Moth (Sphinx ligustri). Although the caterpillar looks so conspicuous, it harmonises very well with its food-plant, and is some- times difficult to find. The purple stripes increase the protection by breaking up the large green surface of the caterpillar into smaller areas. This cater- pillar also affords a good example of a rapid change of colour corresponding to a change of environment. When full grown it descends to the ground and hurries about in search of a spot to bury in, and, being very large and bright green, it would be ex- ceedingly conspicuous against the brown earth if it retained the usual colour. But just before it descends the back begins to turn brown, and becomes finally PEOTECTIVE RESEMBLANCES— DIMORPHISM, ETC. 43 dark brown, so that the caterpillar harmonises well with the colour of its new surroundings. The sig- nificance of this change was first pointed out by Professor Meldola. Other nearly allied caterpillars feeding upon trees, such as willow or poplar, which grow in damp situations where the ground is covered with green vegetation, do not turn brown to anything like the same extent. A very interesting instance of exactly the opposite change at a corresponding period is afforded by the caterpillar of the August Thorn Moth (Ennomos angu- larid), a brown ' stick caterpillar,' protected by a very perfect Special Eesemblance to the dark twigs of the elm on which it feeds. When full-fed it constructs a very loose cocoon of elm leaves, so loose and open that it is easily seen within, and its brown body would be conspicuous against the background of green leaves. But at the same time the dark brown colour of its surface entirely disappears, and the animal is tinted by its green blood, which is seen through the trans- parent skin ; it is thus well concealed by General Resemblance to its new surroundings. Another exceedingly interesting case of the same kind of change is witnessed in the caterpillar of the Miller Moth (Acronycta leporina), which sits motionless on the under side of the leaves of the birch and alder, and is covered with very long beautiful hair which is brilliantly white, and bends over on all sides so as to touch the leaf, forming a wide margin round the 44 THE COLOURS OF ANIMALS caterpillar. Hence all we can see is an oval convex mass of a substance resembling white cotton wool, an appearance very suggestive of a cocoon (see fig. 12). The caterpillar's body is almost invisible ; but on looking carefully we can just make out a dim curved shape beneath the white covering, just as a caterpillar or chrysalis appears through the walls of its cocoon : PIG. 12. -The larva of Miller Moth (Aero- FIG. 13. -Thelarva of Miller Moth (Aero- nyctaleporina), at rest on a birch leaf ; nycta leporina), wandering about on adult ; natural size. bark before forming cocoon ; natural size. furthermore, the larva is very short and thick, and thus resembles the contracted state of a caterpillar before turning to a chrysalis. This perfect Special Eesem- blance is kept up until the caterpillar is full-fed, when it wanders over the bark and finally burrows in it. But a cocoon is a motionless object, and the resemblance, if continued, would be fatal, for it would attract attention. But as soon as the larva PEOTECTIVE RESEMBLANCES— DIMOKPHISM, ETC. 45 is mature, the hairs become black and the body of a much darker tint, and the animal is then well pro- tected by General Eesemblance to the dark surface over which it moves (see fig. 13). Although the bark of large birch trees is chiefly white, the caterpillar is, upon the whole, better con- cealed by becoming dark-coloured. It lives on small birches and alders with dark bark, as well as on large birches, and in the latter case it probably wanders among the wide dark chinks rather than over the smooth wide expanses, for it would certainly burrow in the former rather than the latter. Just before pupation the colours of caterpillars nearly always become dull, and it is in every way probable that such incidental changes have been seized upon by natural selection, and have been ren- dered advantageous to the species. Such alterations of colour are entirely different from those which will be described below, in which an animal can modify its appearance into correspondence with its individual surroundings. The larva of the Privet Hawk Moth almost invariably wanders over the earth when it has come down from its food-plant ; but if it were to descend upon turf, the brown colour would still be assumed, although green would conceal it more effec- tually. The change to brown is, however, far safer for the average caterpillar, and is beneficial to the species on the whole, although it must lead to some in- dividual failures. In the far higher form of Variable 46 THE COLOURS OF ANIMALS Protective Resemblance, which will be described in Chapters VII., VIII., and IX., the individual can adjust its appearance to any of the various environ- ments it is h'kely to meet with in nature. The consideration of changes in colour very naturally leads to the subject of Dimorphism. Dimorphism in Lepidopterous larvae It has been already mentioned that the caterpillars of the Large Emerald Moth are sometimes green and sometimes brown. The same is true of many larvae, and in some of the Mocha Moths (Ephyrida) the chrysalides are the same colour as the larvae from which they develop. These colours have nothing to do with sex, and the appearance of the perfect insect does not seem to be influenced in any way by the larval dimor- phism. It is noteworthy that both colours of dimorphic larvae are invariably of protective value : they are, in fact, nearly always the two chief tints of nature — green and brown. If we breed from moths developed from the green larvae of, e.g., the Large Emerald, the larvae in the next generation are chiefly green, and after several generations there is little doubt that the brown form would become excessively rare ; so also the green form would disappear if we bred from the brown varieties. But in nature both forms are common, and therefore it is certain that both must be advantageous to the species, PROTECTIVE RESEMBLANCES-DIMORPHISM, ETC. 47 or one of them would quickly disappear. I -believe that it is a benefit to the species that some of its larvae should resemble brown and others green catkins, instead of all of them resembling either brown or green. In the former case the foes have a wider range of objects for which they may mistake the larvse, and the search must occupy more time, for equivalent results, than in the case of other species which are not dimorphic. Dimorphism is also valuable in another way : the widening range of a species may carry it into coun- tries in which one of its forms may be especially well concealed, while in other countries the other form may be more protected. Thus a dimorphic species is more fully provided against emergencies than one with only a single form. To take an ex- ample : the green colour of the young caterpillars of the Convolvulus Hawk Moth (Sphinx convolvuli) some- times persists, and is sometimes replaced by brown in the later stages. In Europe the latter form pre- dominates, because the creeping food-plant (Con- volvulus arvensis) is so small that it is safer for a large caterpillar to resemble the earth beneath rather than the small leaves on its surface. In the Canary Islands and Madeira, where the larva feeds on many large- leaved species of Convolvulus, the green form pre- dominates, for it is far better protected than the other against a continuous green background. This result appears to have been brought about by 48 THE COLOURS OF ANIMALS the ordinary operation of natural selection, leading to the extermination of the less protected variety. I have experimented with all the dimorphic larvae men- tioned above, and could not find any trace of suscep- tibility to the influence of surroundings, so as to lead to the production of the appropriate form. When such susceptibility is present, of course the dimorphism has a far higher protective value. The description of such cases is reserved for a future chapter. Occasionally the two forms of a dimorphic species appear at different times and correspond to the tints which successively predominate in the surroundings. At one time I thought the brown form of the Large Emerald caterpillar might appear rather later than the other, when the green catkins had been replaced by brown ; but further examination did not confirm the observations which pointed in this direction. Dr. Alexander Wallace, of Colchester, has, however, found that the moths of Bombyx cynthia which are the first to emerge from the pupae possess, as a rule, an olive- green ground colour, while those which emerge in September are generally of a yellow tint. These colours harmonise with the appearance of the Ailanthus leaflets at corresponding times of the year. Dimorphism in the Perfect Insect Dimorphism is also met with in perfect insects, and it is especially frequent in the females (see page PROTECTIVE RESEMBLANCES— DIMORPHISM, ETC. 4U 302 for an example of a dimorphic male among spiders). Its meaning is obscure, but one of the two forms is generally much rarer than the other, and probably the older. The facts seem to point towards the replacement of an older by a younger form, because the latter is more attractive to the opposite sex, or because it is better concealed, or because the appearance is accompanied by other benefits to the species. The dark variety of the female Silver-washed Fritillary (Argynnis paphia, var. ralezina), and the white variety of the female Clouded Yellow (Colias edma, var. helice), are examples of dimorphism among British butterflies. I exclude that form of dimor- phism, or polymorphism, which is caused by one sex ' mimicking ' two or more species which are specially protected (for a good example see pp. 234-38). An extremely important form of di- or poly- morphism occurs among the females of the social Hymenoptera. In this case, however, the different forms are specially fitted for certain duties, and the consequent division of labour is beneficial to the society and therefore to the species. Seasonal Dimorphism Finally, a species which passes through two or more cycles of development in a year, viz. one that is ' double ' or ' treble-brooded,' is often characterised by * seasonal dimorphism/ in which the first brood is 50 THE COLOURS OF ANIMALS different in appearance, and often in size, from the later ones. Professor Weismann has investigated this question, and he finds that while the later broods can be readily made, by the application of ice in the pupal stage, to assume the form of the first or winter gene- ration, the latter cannot be made to assume the form of the summer brood by the application of warmth. He infers that such species were single-brooded in the short summers which succeeded the Glacial Period, and that the appearance was that of the present winter form. As the summers became longer, other newer generations with a different appearance were added (summer broods), but the species always tends easily to revert to the more ancient form. An important part of the evidence consists in the proof that such species are now single-brooded hi the northern part of their range, and that the one form is that of the winter brood of more southern localities.1 I have given a very brief sketch of dimorphism, hardly alluding to polymorphism, which is only an extension of the same principle. Although the subject is only touched upon, enough has been said to show that there are many distinct kinds of dimorphism, some of which are very obscure. By far the most important kind of di- or polymorphism remains to be described below (see Chapters VIII. and IX.), in which 1 See Studies in the Theory of Descent, by August Weismann. English translation by Professor Meldola. PROTECTIVE RESEMBLANCES— DIMORPHISM, ETC. 51 each individual has two or more appearances, as it were, at its command, and can develop that one which is most suited to its own peculiar surroundings. A reason for the wonderful concealment of Lepidopterous larvae In the remarkable abundance and variety of methods by which concealment is effected in Lepi- dopterous larvae, we probably see a result of their peculiarly defenceless condition. A larva is a soft- walled cylindrical tube which owes its firmness, and indeed the maintenance of its shape, to the fact that it contains fluid under pressure, which is exerted by the sides of the body. This construction is extremely dangerous, for a slight wound entails great loss of blood, while a moderate injury must prove fatal. Hence larvae are so coloured as to avoid detection or to warn of some unpleasant attribute, the object in both cases being the same — to leave the larva un- touched, a touch being practically fatal (see also pp. 175-76). The concealment of Pupae Protective Resemblance, either Special or General, is seen in nearly all exposed pupae, but most chrysalides are buried in the earth or protected by cocoons. The cocoons are often sufficient defence, because the silk is very unpleasant in the mouth ; but such protection 52 THE COLOUES OF ANIMALS only applies in the warmer weather when there is an abundance of insect food. In the winter, insectivo- rous animals are pinched by hunger, and would devour the pupa in spite of the cocoon. We therefore find that all cocoons which contain pupae during the winter are well concealed, either spun between leaves which fall off and become brown, or hidden under bark or moss, or constructed on the surface of bark with a colour and texture which renders them extremely difficult to detect. It is very common for particles of the bark to be gnawed off by the larva and fixed on to the outside of the cocoon. It will be shown below that many larvae can also control the colour of their cocoons. Protective Resemblances in Butterflies and Moths The perfect insect is also commonly defended by very efficient methods of concealment. The under sides of the wings of butterflies are generally coloured like the surface on which the "insect habitually rests, and they are the only parts seen during repose. We can form some idea of the perfection of this conceal- ment when we remember the entire disappearance of common butterflies in dull weather. Many of them creep far down among thickly set leaves, while others rest freely exposed upon surfaces which harmonise with their colours. Perhaps the most perfect concealment attained by any butterfly is seen in the genus Kallima, found PKOTECTIVE RESEMBLANCES— DIMORPHISM, ETC. 53 in India, the Malay Archipelago, and Africa. The way in which the insect is concealed has been described by Wallace in his ' Malay Archipelago,' and also in the 'Essays on Natural Selection.' The tip of the fore-wing is pointed like the apex of a leaf, and the hind-wing has a short tail like a leaf-stalk, while the outline of the folded wings between these extremities is exactly like that of a withered and somewhat shrivelled leaf. At rest the wings are held upright over the back, the head and antennae are concealed between them, while the tails touch the branch to which the insect clings by its almost invisible legs. Along the supposed leaf runs a distinct mark like a mid-rib, with oblique veining on either side. But dead and withered leaves are not all alike ; they may be almost any shade of brown, grey, or yellow, while they are often attacked by fungi of dif- ferent colours and in different places. Similarly the under sides of the wings of the butterfly are excessively variable, the different colours and markings only agreeing in that they all represent some familiar ap- pearance presented by withered or decayed leaves. Dead leaves are often pierced by insect larvae, and a detail of great interest is added to the disguise in the semblance of a small hole. The scales are absent from both sides of a certain spot on each fore- wing, which is therefore only covered by the thin transparent wing membrane. These spots come opposite to each other in the position pf rest, and the effect produced 4 54 THE COLOUES OF ANIMALS is exactly that of a hole, for the two membranes are BO transparent that they are completely invisible. The size of the apparent hole varies very greatly in the numerous specimens of Kallima inachis, in the Hope Collection at Oxford. The upper sides of the wings, concealed during rest, are dark, with a deep orange bar across the fore-wings. I have heard a naturalist, who is ac- quainted with the Indian species (Kallima inachis) in its natural surroundings, object to the interpretation afforded by Mr. Wallace, on the ground that he has often seen the butterfly displaying the conspicuous upper sides of its wings when settled, and has seen it resting on inappropriate surfaces. I do not think that this objection is fatal ; for butterflies only dis- play their brilliant tints during the short pauses between the successive flights, when they are on the alert and can evade their enemies by wariness and by the swiftness of their flight. Our own beautiful Eed Admiral (Vanessa atalanta), Peacock (V. Jo), and Small Tortoiseshell (V. urticce) similarly display their bril- liant colours when pausing on a flower or even on the ground. But during prolonged rest, when the insects are often semi- torpid and would be easily captured if detected, the wing?, are invariably held so that the sombre tints of the under sides are alone visible. Hence the display of bright colours by the Indian Kallima is no argument against the protective value of the leaf-like appearance of the under sides ; for the PKOTECTIVE RESEMBLANCES— DIMORPHISM, ETC. 55 latter acts as a disguise when it is most necessary, for the butterfly to be concealed. It appears that the Malayan species (Kallima paralekta) is more cautious during the brief pauses between the flights ; for Mr. Wallace states that it frequents dry woods and thickets, and that it invariably settles on bushes with dry or dead leaves. He never saw one of these butterflies settle upon a flower or green leaf. A recent paper by Mr. .S. B. J. Skertchly ! en- tirely supports Mr. Wallace's statements. The author calls attention to the fact that leaf-mimicking butter- flies, of several genera in addition to Kallima, settle in an entirely different manner from that of other butterflies. While the latter gradually slacken their speed and settle deliberately, the leaf butterflies ' fly rapidly along, as if late for an appointment, suddenly pitch, close their wings, and become leaves. It is generally done so rapidly that the insect seems to vanish.' Certain English moths are also protected by their resemblance to dead leaves. One of the most beautiful examples is afforded by the Herald Moth (Gonoptera libatrix), which suggests the appearance of a decayed red leaf sprinkled with a few white spots of fungoid growth ; the irregularly toothed margin of the wings adds to the effect. The bright eyes of the moth might expose the deception, but they are covered during rest by a tuft of hair which springs from the base of 1 Ann. and Mag. Nat. Hist. Sept. 1889, pp. 209 et seg. 56 THE COLOURS OF ANIMALS the antennae (see fig. 14). When the moth is about to fly the antennae are brought forward, and the same action raises the tufts and uncovers the eyes. The moth appears in the autumn and lives through the winter, so that the resemblance to dead leaves is peculiarly appropriate. The Angle-shades (Phlogophora meticulosa) is also beautifully concealed by resembling a withered and crumpled leaf. The colours of the Yellow Underwing (Triplicena pronuba), as seen during flight, strongly suggest the appearance of a yel- low leaf whirled along by the wind and then suddenly dropping. The sudden swift rise and rapid descent are very unlike the flight of a moth. When at rest, it hides deeply amid thick foliage or among dead leaves on the ground ; it is extremely difficult to detect, and instantly rises when disturbed. The Rev. Joseph Greene has pointed out that the various shades of yellow and brown are especially characteristic of autumn moths, while grey and silvery tints predominate in the winter species ; such tints FIG. 14.— The base of left antenna of Herald Moth (Gonoptera libatrix), show- ing the tuft of hair which covers the eyes of the moth in the position of rest ; x 24-5 diameters. PROTECTIVE RESEMBLANCES— DIMORPHISM, ETC. 57 harmonise with those that are most characteristic in the corresponding seasons. The Buff-tip Moth (Pyg&ra bucephala) is very perfectly concealed by resembling a broken piece of decayed and lichen-covered stick. The cylindrical shape is produced by the wings being rolled round the body. A friend l has raised the objection that the moth resembles a piece of stick cut cleanly at both ends, an object which is never seen in nature. The reply is that the purple and grey colour of the sides of the moth, together with the pale yellow tint of the parts which suggest the broken ends, present a most perfect resemblance to wood in which decay has induced that peculiar texture in which the tissue breaks shortly and sharply, as if cut, on the application of slight pressure or the force of an insignificant blow. The excreta of birds are also very commonly re- sembled by moths as well as by caterpillars. This is the case with the little Chinese Character (Cilix spinula}? and with many grey and white Geometers which rest on the upper sides of leaves with their wings extended as if ' set.' In this position they forcibly suggest the appearance of birds' excrement which has fallen on to a leaf from a great height, and has therefore become flattened into a wide patch. In spite of a faithful resemblance to such an object, these moths possess very great beauty. » Dr. C. M. Chadwick, of Leeds. * Arthur Sidgwick, Journ. of the Rugby School Nat. Hist. Soc. 58 THE COLOUKS OF ANIMALS The appearance of splinters of wood is also often suggested by moths such as the ' Sharks ' (Cucullia) or Goat Moth (Cossus). Others resemble the surfaces of rock upon which they habitually rest (Bryophila, many Geometers, &c.). I have merely given a few striking instances of resemblance to objects which are of no interest to insect-eating animals. Numerous other examples might have been added, but my object is merely to illustrate from the Lepidoptera a principle of colour- ing which is of extremely wide application, viz. its use in aiding an organism to escape from enemies by the method of concealment. Abundant examples of this principle will be recognised by every one interested in natural history, among other orders of insects as well as the Lepidoptera, among vertebrate animals no less than among the invertebrate sub- kingdoms. Protective Resemblances in other Insects and in Spiders In the other orders of insects, the Orthoptera (locusts, grasshoppers, &c.) will be found to include the most beautiful examples of Protective Eesemblance. The tropical ' leaf insects ' and ' walking-stick insects ' belong to this order. The latter hold their limbs irregularly, so that the resemblance to a dead branch with lateral twigs is rendered all the more perfect. PEOTECTIVE RESEMBLANCES-DIMORPHISM, ETC. 59 Spiders are often protectively coloured ; many ex- cellent examples are given by Elizabeth G. Peckham.1 One of the most remarkable is Ccerostris mitralis, from Madagascar, which sits motionless on a branch and resembles a woody knot. Its appearance is shown in fig. 15. A common Wisconsin spider, Epeira prompta, FIG. 15.— Cceroslris mitralis in proaie; FIG. IS.—Speir from Peckham ; after Vinson. lichen - covered" tree"- trunk ; from Peckham. generally rests on the branches of cedar bushes, and closely resembles lichen (see fig. 16). Spiders are especially relished by insectivorous animals, so there is every reason for the faithfulness of these resem- blances. In many other cases, however, the resem- blance is chiefly aggressive, enabling the spider to approach its prey. 1 Occasional Papers of tJie Natural History Society of Wisconsin, vol. i. 1889, Milwaukee, pp. 61 et seq. 60 THE COLOURS OF ANIMALS CHAPTEE V PROTECTIVE RESEMBLANCES IN VEETE- BE AT A, ETC. MANY of the lower Vertebrata have the power of rapidly modifying their colour according to the en- vironment, and these will be described in a future chapter. Such a power appears to be possessed by few reptiles, and by no bird or mammal. Protective Resemblances among Reptiles Our two English snakes are well concealed by their colours ; the olive-green Grass Snake (Tropi- donotus natrix) harmonising well with the grassy banks which it chiefly frequents; while the brown viper (Pelias berus) is difficult to detect upon the dry heaths where it is most commonly found. Our lizards are also well protected in the same manner. Protective Resemblances among Birds Wallace has directed attention to the protective colours of female birds which build open nests, and he points out that the males are similarly protected when PROTECTIVE RESEMBLANCES IN VERTEBRATA 61 they undertake the duties of incubation. The same necessity does not apply to species which construct covered nests or build in holes. The Colours and Markings of Birds' Eggs The protective value of the tints and markings of eggs are of great interest, and have not been sufficiently investigated. The fact that eggs are protectively coloured was fully recognised by Erasmus Darwin, who places them under ' colours adapted to the purpose of concealment.' He says ' the eggs of birds are so coloured as to resemble the colour of the adjacent objects and their interstices. The eggs of hedge-birds are greenish with dark spots; those of crows and magpies, which are seen from beneath through wicker nests, are white with dark spots ; and those of larks and partridges are russet or brown, like their nests or situations.' l This description of the eggs of crows and magpies is incorrect. The eggs of crows are greenish with umber markings; those of magpies pale greenish with dark markings. It is probable that Erasmus Darwin correctly explained the appearance of the eggs of the wood-pigeon (see p. 62), and inadvertently illustrated this principle of colouring by erroneous instances. The special men- tion of the interstices between the parts of sur- rounding objects, as well as the objects themselves, is 1 Zoonomia, 1794, vol. i. p. 510. 62 THE COLOURS OF ANIMALS of great interest; protective colouring can never be fully understood until this principle is taken into account. In order to make out the true meaning of the colours of eggs they must be observed in their natural surroundings, and must be looked at from all points of view and at varying distances. It is very probable that the bright blue colour of certain eggs will be explicable under these conditions. The fact that concealed eggs are almost invariably white strongly confirms the conclusion that the colours of expos-ed eggs are of value to the species, and are maintained by the operation of natural selection. Certain exposed eggs may, however, be white, as in the wood-pigeon, but in these cases the eggs are pro- tected from enemies beneath ; for the holes in the loosely constructed nest through which they are seen cannot be distinguished from others through which the bright sky appears. The whiteness of eggs hidden in holes or in covered nests, or buried under leaves, is of a very different nature, for it is due to the cessation of natural selec- tion, perhaps aided by reversion to the ancestral colour, which is still preserved in the eggs of reptiles. All useful characters are kept up to a high pitch of efficiency by the continual elimination of the unfittest, and as soon as such elimination ceases, the level of efficiency must fall. This interpretation is confirmed by the fact that the eggs of certain species which now PROTECTIVE RESEMBLANCES IN VERTEBRATA 63 nest in the dark still retain traces of patterns which are well developed upon the eggs of their nearest allies with other habits. Thus the egg of the puffin, which nests in a burrow, would be called white at a little distance, but closer examination reveals the presence of very faint spots, which are distributed as in the very distinctly marked egg of the razor-bill.1 Certain other species still lay strongly marked eggs in the dark, and in their case the change of habit presumably took place at a comparatively recent date. Such a conclusion can be tested by an investigation of the habits of closely allied species. Although white must have been the ancestral colour of birds' eggs, it has probably been re-acquired in species which nest in the dark. It would be very difficult to believe that such a habit has persisted continuously since the time when all birds' eggs were white. The strongest confirmation of this explanation of the whiteness of hidden eggs is, however, to be found in the colours of the eggs in the various breeds of domestic fowls. If the gradual disappearance of colour is due to the cessation of natural selection, we must expect it to occur, however the cessation has been brought about. Natural selection cannot operate to preserve the colour of eggs laid in the dark, and it 1 This interesting example attracted my attention while looking over a collection of eggs in the possession of my friend, Mr. E. H. Greenly. Mr. H. Seebohm informs me that he has no doubt about the validity of this interpretation, which was suggested in hiu work on British Birds, 1885, vol. iii. p. 367. 64 THE COLOUES OF ANIMALS is equally inoperative when enemies are artificially excluded from eggs laid in open nests. And the eggs laid in our poultry-yards afford conclusive evidence that colour disappears as surely under the latter condition as under the former. The brown colour must be a very important protection to the eggs of the ancestor of our domestic breeds, the Asiatic jungle fowl (Gallus bankiva) ; while a white appearance would greatly add to the danger of discovery by egg- eating animals. But there is no such difference be- tween the value of white and brown in confinement, and we accordingly find that the colour is disappear- ing. Certain fowls lay white eggs, and the tint of those which still lay coloured eggs varies considerably, ' the Cochins laying buff-coloured eggs, the Malays a paler variable buff, and Games a still paler buff. It would appear that dark-coloured eggs characterise the breeds which have lately come from the East, or are still closely allied to those now living there.' ' Erasmus Darwin further suggested that the colours of eggs, in common with other protective colours, may be due to the effect of the imagination of the female.2 This suggestion has been still further elaborated by A. H. S. Lucas,3 but no real proof of it is brought forward in his paper. That eggs resemble their sur- 1 Darwin, Variation of Animals and Plants under Domesti atiou, 1875, vol. i. p. 261. 2 Loc. cit. p. 611. » Royal Society of Victoria, 1887, pp. 52-60. PROTECTIVE RESEMBLANCES IN VERTEBRATA 65 roundings is explicable by the operation of natural selection, while the gradual loss of colour when natural selection ceases to operate, is in opposition to Mr. Lucas's hypothesis, which assumes that the colour of the shell is determined by the influence of surround- ing tints upon the retina of the female. If the rusty spots on the eggs of birds of prey are due, as Mr. Lucas supposes, to the sight of blood, eggs laid in the dark should still be affected by the memory of some colour either predominant in the surroundings, or of especial interest to the female. The hypothesis might easily have been tested by surrounding birds with un- usual colours and observing the tints of their eggs : and, had this been done, I believe that the paper would not have been written. Since the last paragraph was printed, a letter from the Rev. F. F. Grensted has been communicated to ' Nature ' by Mr. A. E. Wallace.1 The writer believes that the colour of the egg of the red-backed shrike varies with the tint of the lining material of the nest. Mr. E. B. Titchener maintains that there is not suffi- cient evidence for this opinion. At one time Mr. Titchener believed that Variable Protective Resem- blance was exhibited by the eggs of the yellow- hammer and spotted fly- catcher as well as by those of the red-backed shrike. Further observation con- vinced him that the evidence was insufficient.2 1 Nature, vol. 41, Nov. 21, 1889, p. 53. 2 Ibid., Dec. 12, 1889, pp. 129-30. 66 THE COLOURS OF ANIMALS The strongest argument used by Mr. Lucas is the fact that cuckoos at first sight appear to have the power of adjusting the colour of their eggs to those of the birds which are so successfully imposed upon. It seems to be certain, however, that the cuckoo carries her egg to the nest in her beak ; for there are numerous instances of the cuckoo's egg having been found in a nest which the bird itself could not possibly enter, and Mr. Lucas gives examples of the same kind from Australia. The cuckoo has therefore the chance of seeing the colour of her egg, and of carrying it to the appropriate nest. It is also possible that different individuals lay eggs of a different shade, and deposit them in the nests of species with eggs of a correspond- ing appearance. The whole relation of the cuckoo to the birds it deludes is most interesting, but very difficult to decide satisfactorily, because of the extreme shyness of the bird. I do not think, however, that the facts which .are now at our disposal afford sufficient justification for the opinion that the female cuckoo can control the colour of her eggs. I have discussed the colours of birds' eggs at some length in the hope that those readers who are interested in the subject may be induced to observe for them- selves, and assist in obtaining a far more complete knowledge of the meaning of the colour and marking of eggs than we at present possess. I know of no more inspiring subject than the colours of birds' eggs. The most superficial glance PBOTECTIVE RESEMBLANCES IN VERTEBRATA 67 over a collection of eggs reveals hosts of interesting problems which require solution. I look forward to the time when any description of colour and marking will be considered incomplete unless supplemented by an account of their meaning and importance in the life of the species. Protective Resemblances among Mammalia Among the Mammalia it would be hardly possible to meet with a better example of protective colouring and attitude than that of the hare as it sits motion- less, exactly resembling a lump of brown earth, for which indeed it is frequently mistaken. The dark brown or grey colours of all our smaller quadrupeds are also highly protective. The change of colour in northern mammals in the winter will be described in Chapter VH Protective Resemblances among Fish The power of colour adjustment is very widely distributed among fish and Amphibia, and will receive attention in a later chapter. I will therefore only say a few words about the Protective Eesemblance of the former. Professor Stewart has shown me a beautiful ex- ample in the Australian Sea Horse (Phyllopteryx eques), a fish which is covered with numerous cutaneous 68 THE COLOUKS OF ANIMATE appendages, most of which are supported by a bony core. The appendages are flat, and are alternately banded with dusky brown and orange, exactly resembling the form and colour of the sea-weed to which the fish clings with its tail. There also are many bony spines without the flat folds of skin, and these are doubtless defensive. The general arrangement of colour on porpoises, most fish, &c., has been well explained by Wallace. Looking down on the dark back of a fish it is almost invisible, while, to an enemy looking up from below, the light under-surface would be equally invisible against the light of the clouds and sky.'1 The white colour of one side of such fish as the sole, turbot, &c. (Pleuronectidce), viz. the side which is in contact with the sand or mud, cannot be ex- plained in this way. In such a case we see the dis- appearance of colour in consequence of the cessation of natural selection, as in the white eggs laid in the dark, while the white bellies of many fish may be compared to the whiteness of the eggs of the wood- pigeon, an appearance produced by the operation of natural selection. It has been already pointed out that natural selection may not only remove the pigment from an animal, but may oven replace the red blood of a vertebrate by a colourless fluid. The transparency 1 Tropical Nature, p. 171. PROTECTIVE RESEMBLANCES IN VERTEBRATA 69 of the surface swimming fish, Leptocephalus, is in- creased in this way.1 Protective Resemblances among Marine Animals Before leaving this part of the subject I must allude to Protective Eesemblances among marine animals. Although large numbers of isolated cases are understood, the principles of colouring in marine forms have been very incompletely worked out. The difficulties are far greater than in land animals, be- cause it is often nearly impossible to observe the species in their natural environment, and it has been already shown that this is essential if we are fully to understand the meaning of all details in their appearance and attitudes. It is, however, very satis- factory to know that the whole subject of the colouring of marine mollusca is being undertaken in a syste- matic manner by Mr. W. Garstang,2 assisted by all the appliances of the laboratory of the Marine Biological Association at Plymouth. Protective Resemblances among Marine Mollusca E. S. Morse has shown, contrary to Darwin's opinion,3 that the appearances of many mollusca are such as to afford concealment. An extremely 1 E. Ray Lankester, « On the Distribution of Haemoglobin,' Proc. Boy. Soc. No. 140, 1873. 2 Journ. Mar. Biol. Assoc., New Series, vol. i. No. 2, Oct. 1839, pp. 173 et seq. 3 Descent of Man, vol. i. p. 316. 70 THE COLOURS OF ANIMALS interesting example was brought before me by Mr. Garstang, viz. that of the Opisthobranch mollusc, Hermcea, which is transparent and therefore invisible, except for the ' hepatic ' canals, which simulate in form and colour the reddish weed (Griffithsia) on which the animal usually lives. Mr. Garstang finds that the colour is purely adventitious, being due to the food undergoing digestion (see pp. 79, 80). Another English species of Hermcea is green, and lives on green weeds. Mr. H. L. Osborn has published a very interesting note on the resemblance between the colour of a coral on the North American coast and the mollusc which habitually lives upon it.1 He states that Dr. E. B. Wilson, working in 1879 in Dr. Brooks's laboratory at Beaufort, N.C., found an orange-yellow coral (Leptogorgia virgulata) invariably attended by a gastro- pod of the same Colour (Ovulum uniplicaturri), which was never seen apart from the coral. Dr. Wilson's coral occurred in shallow water. In 1884, Mr. Osborn, also working at Beaufort, found a Leptogorgia in ten fathoms of water, of the same general habit as L. vir- gulata, but of a deep rose colour, almost purple. The ground-colour was mottled with white round the open- ings of the polypes. A large number of molluscs were found on the coral, and these had red-brown shells, with the surrounding skin of a deep rose colour mottled with white. Except for this difference in colour the molluscs exactly resembled 0. uniplicatum, and Mr. Osborn con- 1 H. L. Osborn, Science, New York, 1885, vi. pp. 9-10. PROTECTIVE RESEMBLANCES IN VERTEBRATA 71 eiders that they were undoubtedly of the same species. When placed in an aquarium the molluscs always sought their own corals, but if red- molluscs and yellow corals were put together the former took no notice of the latter. It is very interesting to find that Mr. Garstang notices a similar association between species allied to the above at Plymouth. A reddish coral (Gorgonia verrucosd) is attended by a gastropod (Ovula patula) , adapted in form and colour for concealment on the stems of the Gorgonia. It might be argued that these are cases of Pro- tective Mimicry, inasmuch as one animal resembles a portion of another for the purpose of protection. Similar examples are to be found in certain parasites which resemble the colour of the hair or skin of the animal they infest. Protective Mimicry, how- ever, leads one animal to be mistaken for another, and thus to live upon the reputation of the latter. Protective Kesemblance simply renders an animal difficult to detect. Animals defended in the former manner are almost invariably conspicuous ; in the latter they are admirably concealed. Mr. Garstang tells me that Gorgonia is exempt from the attacks of fishes, so that the molluscs gain additional advantage by resembling an inedible form. Inasmuch as con- cealment appears to be the chief object of the form and colour, the example falls under Protective Eessmblance, although it leads in the direction of true Mimicry. 72 THE COLOUKS OF AKIMALS CHAPTER VI AGGRESSIVE RESEMBLANCES— ADVENTITIOUS PROTECTION Aggressive Resemblances PASSING now to Aggressive Resemblances, the appear- ance of the larger Carnivora harmonises well with their surroundings, enabling them to approach their prey. The colours of snakes, lizards, and frogs are doubtless Aggressive as well as Protective. Certain carnivorous insects, such as the Mantidae, are well- concealed by their colour ; and this, although chiefly Protective, is probably also of value in enabling them to creep upon their prey. Aggressive, like Protective, colouring may be either Special or General. Alluring Colouration Special Aggressive Resemblance sometimes does more than hide an animal from its prey ; it may even attract the latter by simulating the appearance of some object which is of special interest or value to it. Such appearances have been called Alluring Coloura- AGGRESSIVE RESEMBLANCES, ETC. 73 tion by Wallace, and they are some of the most in- teresting of all forms of Aggressive Eesemblance. An Asiatic lizard, Phrynocephalus mystaceus, is a good example. Its general surface resembles the sand on which it is found, while the fold of skin at each angle of the mouth is of a red colour, and is produced into a flower-like shape exactly resembling a little red flower which grows in the sand. Insects, attracted by what they believe to be flowers, approach the mouth of the lizard, and are of course captured. Professor C. Stewart kindly brought this instance before me, and showed me a specimen of the lizard in the Museum of the Royal College of Surgeons. ' Similar examples are to be found among fishes. The Angler, or Fishing Frog (Lophius piscatorius) , pos- sesses a lure in shape of long slender filaments, the foremost and longest of which has a flattened and divided "extremity. The fish stirs up the mud so as to conceal itself, and waves these filaments about : small fish are attracted by the lure, mistaking it for worms writhing about in the muddy water ; they approach and are instantly engulphed in the enormous mouth of the Angler. This interesting habit has been known since the days of Aristotle. Certain deep-sea forms allied to Lophius behave in a similar manner, but as the depths of the sea are dark, they have a special ' phosphorescent organ, which probably illu- minates the play of the tentacles which serve to allure other creatures.' In some of these fish (certain 74 THE COLOURS OF ANIMALS species of Ceratias) the foremost tentacle bears a luminous organ which is suspended as a lure in front of the mouth.1 The prey are attracted by the light into this convenient position for effecting their capture. An Indian Mantis (Hymenopus Ucornis) feeds upon other insects which it attracts by its flower-like PIG. n.—Bymenopui bicomis in active pupa stage. shape and pink colour. The apparent petals are the flattened legs of the insect. The appearance of the Mantis in the active pupa stage is shown in fig. 17.- The figure has been copied from a drawing sent by 1 A. Giinther, Challenger Reports, vol. xxii. p. 52, and Introduc- tion, p. xxx. The function of the luminous organ was first sug- gested by Liitken. AGGRESSIVE RESEMBLANCES, ETC. 75 Mr. Wood-Mason to Mr. Wallace, who kindly lent it to me. Another beautiful example of Alluring Colouring was discovered by Mr. H. 0. Forbes in Java. Butter- flies are often attracted by the excreta of birds, and a spider (Ornithoscatoides decipiens) takes advantage of this fact to secure its prey. The resemblance to a bird's dropping on a leaf is carried out with extraordi- nary detail. Such excreta consist of a * central and denser portion, of a pure white chalk-like colour, streaked here and there with black, and surrounded by a thin border of the dried up more fluid part, which, as the leaf is rarely horizontal, often runs for a little way towards the margin ' and there evaporates, forming a rather thicker extremity. The margin is represented by a film-like web, with a thickened part to represent the fluid which has run to the edge or apex of the leaf ; the central mass is represented by the spider itself with white abdomen and black legs, lying on its back in the middle of the web, and held in position by the spines on its anterior legs which are thrust under the film.1 The whole combination of habits, form, and colouring afford a wonderful example of what natural selection can accomplish. In such a case there is no necessity for calling in the aid of any other principle, for the addition of each new feature and the improvement of every detail would at once 1 H. 0. Forbes : A Naturalist's Wanderings in the Eastern Archi- pelago, pp. 63-65. 76 THE COLOURS OF ANIMALS give an advantage to the possessor in the constant struggle for food. Adventitious Protection Before proceeding to describe the power of Vari- able Protective Eesemblance possessed by many animals, it is necessary to point out that effects similar to those described above may be gained by means which supplement the acquisition of any special colour and form, or which may entirely replace these methods of producing concealment. Many animals cover themselves with objects which are prevalent in their surroundings and are of no interest to their enemies. Sometimes the meaning of this habit is concealment alone, but in other cases objects of great strength are selected and bound firmly together so as to form a resistant armour. Many Lepidopterous larvae live hi cases made of the fragments of the substance upon which they live. The cases of the larvae of Clothes Moths are only too well known ; those of the Psychidce are made of leaf or brown grass stems. The larva of the Essex Emerald Moth (Geometra smaragdaria) covers itself with a loose case made of fragments of leaves spun together with silk. The cocoons of Lepidoptera are frequently concealed by containing fragments of wood or bark gnawed off the surface on which the cocoon is constructed (Centra, Cilix, Hemerophila, &c.). Birds' AGGRESSIVE RESEMBLANCES, ETC. 77 nests are often similarly concealed ; the lichen- covered nest of the chaffinch is an obvious instance. - The well-known cases of Caddice-worms (Trichop- tera) are partly for concealment and partly for defence ; they are built of grains of sand, small shells (often alive), vegetable fragments, — in fact, of any suitable objects which are abundant at the bottom of the stream in which they happen to be. Some of the best examples are to be found among marine animals. Certain sea-urchins cover themselves so completely with pebbles, bits of rock, shell, &c., that one can see nothing but a little heap of stones. Many marine mollusca have the samp habits, accumulating sand upon the surface of the shell or allowing a dense growth of algae to cover them.1 The best example of the kind was shown me by Professor C. Stewart, and is all the more interesting because of the transition observed in the habits of different species of the same genus, Xenophora. Many of these gastro- pods include pieces of shell, rock, coral, &c., in the edge of the growing shell. The effect is probably to obscure the junction between the shell and the surface on which it rests, and thus to assist in rendering the organism difficult of detection. Thus the growth of the shell may be traced by a spiral line of included fragments (X. calculifera) . In X. Solaris the habit is only maintained during the early stages of growth, 1 E. S. Morse: Proc. Boston Soc. Nat. Hist, yol, xiv. April 5, 1871, p. 7. 5 78 THE COLOUKS OF ANIMATE and the spiral line of fragments extends for a certain distance, and is then suddenly replaced by spines which are doubtless of value as a defence. In X. cerea and X. solaroides the size of the adventitious particles is so great as to nearly conceal the shell, while in X. conchyliophora nothing can be seen but a heap of fragments. Specimens of Xenophora and of the crabs mentioned below are to be seen in the Museum of the Eoyal College of Surgeons, as part of a beautiful series intended by Professor Stewart to illustrate the various uses of the colours of animals. The tube of certain well-known marine worms (Terebellida) is constructed of sand-grains cemented together. One of the most interesting examples of adventi- tious protection is afforded by certain crabs (Steno- rhynchus, Inachus, Pisa, Mala) , which fasten pieces of sea-weed, &c., on their bodies and limbs. Bateson has watched the process in Stenorhynchus and Inachus. 1 The crab takes a piece of weed in his two chelae, and, neither snatching nor biting it, deliberately tears it across, as a man tears paper with his hands. He then puts one end of it into his mouth, and, after chewing it up, presumably to soften it, takes it out in the chelae and rubs it firmly on his head or legs until it is caught by the peculiar curved hairs which cover th m. If the piece of weed is not caught by the hairs, the crab puts it back in his mouth and chews it up again. The whole proceeding is most AGGRESSIVE EESEMBLANCES, ETC. 79 human and purposeful. Many substances, as hydroids, sponges, Polyzoa, and weeds of many kinds and co1 our s, are thus used, but these various substances are nearly always symmetrically placed on corre- sponding parts of the body, and particularly long, plume-like pieces are fixed on the head, sticking up from it. ... Not only are all these complicated processes gone through at night as well as by day, but a Stenorhynchus if cleaned and deprived of sight will immediately begin to clothe itself again, with the same care and precision as before.' 1 Bateson states that Stenorhynchus does not betray any disposition to remain in an environment which harmonizes with its dress. Adventitious Colouring The protective colouring of many animals may be due to the food in some part of the digestive tract, seen through the transparent body. This is impor- tant in many transparent caterpillars, such as the Noctuce, and probably in many marine organisms. If a larva, such as that of the Angle-shades (Phlogophora meticulosd), be fed on the orange- coloured marginal florets of the marigold, the passage of food along the alimentary canal can be distinctly traced by the pro- gressive change in the colour of the caterpillar. The green colour of the blood of most larvae is 1 Journ. Mar. Biol. Assoc., New Series, vol. i. No. 2, Oct. 1889, pp. 213-14. 80 THE COLOURS OF ANIMALS adventitious in origin, having been derived from the chlorophyll of the leaves ; it is, however, much modi- fied in constitution by the time it reaches the blood. The green colouring matter passes from the blood into the cells of the surface of the body in many caterpillars, but is re-dissolved in the blood of the chrysalis. It is then made use of, in certain species, to tinge the eggs, and, after this, is absorbed into the body of the young larvae which afterwards hatch from them, protecting them with a green colour before they have had time to acquire fresh chlorophyll from the leaves. The passage of an adventitious colouring matter on into a second generation is a very remark- able phenomenon. There does not, however, seem to be any doubt about its occurrence in certain species (e.g. Smerinthus ocellatus) ,l and I have a good deal of unpublished evidence on the subject. 1 See ' Proceedings of Physiological Society,' pp. xxv and xxvi, in Journal of Physiology, vol. viii. 1887. CHAPTER VH VARIABLE PROTECTIVE RESEMBLANCE IN VERTEBRATA, ETC. PROTECTIVE Resemblance in its highest and most per- fect form must not be fixed, but capable of adjustment, so that the animal is brought into correspondence with each of the various tints which successively form its environment, as it moves about. An active and wide-ranging animal will be benefited by the power of resembling the tints of many different environments, and by that of changing its colour rapidly. More sluggish animals only require the power of bringing their appearance into harmony with a single environment, although the capability of adjustment is still of great value, because the environments of the different individuals vary, at any rate to a slight extent. Thus a moth lays some of its eggs upon one tree of a certain shade of green, and others upon another with leaves of a rather different shade ; so that the caterpillars would not have same environment, and would gain, by possessing the power to adapt their colours. At the same time there would be no im- 82 THE COLOUES OF ANIMALS perative necessity for the change to be extremely rapid. In other cases it will be of advantage to the animal to possess the power of changing twice in its life, once for the peculiar surroundings of the caterpillar, and once for the peculiar surroundings of the chrysalis. There is indeed some ground for the belief that in certain cases the colours of the perfect insect also may be adjusted to correspond with the peculiar environ- ment. Variable Protective Resemblance in Fishes Instances of the power of rapid adjustment are very common, although most people are not aware of them. Nearly all fishermen know that the trout caught in a stream with a gravelly or sandy bottom are light-coloured, while those caught in a muddy stream are dark. It is also well known that the same fish will soon change in colour when it passes from one kind of background to the other. Thus Mr. E. D. Y. Pode tells me that all the trout in a stream near Ivy Bridge have become unusually light ever since the pollution of the stream by white china clay. The same facts are true of many other freshwater and sea fishes. The interior of a minnow-can is painted white in order that the bait may become light- coloured and thus conspicuous in the dark water where the pike or perch is likely to be found. The change of colour occupies an appreciable time, and the fisher- VAEIABLE RESEMBLANCE IN VERTEBRATA, ETC. 83 man knows that he stands an extra chance of catch- ing his fish while the bait remains unadapted to its environment. This experience serves to prove in a practical way that the power of changing the colour is essentially protective. Variable Protective Resemblance in Amphibia Other animals possess the same power. The Com- mon Frog (Rana temporaria) can change its tints to a considerable extent. Thus Sir Joseph Lister states that ' a frog caught in a recess in a black rock was itself almost black; but after it had been kept for about an hour on white flagstones in the sun, was found to be dusky yellow with dark spots here and there. It was then placed again in the hollow of the rock, and in a quarter of an hour had resumed its former darkness. These effects are independent of changes of temperature . . .' l The Green Tree Frog (Hyla arborea), so common in the South of Europe, is bright green when seated among green leaves, but becomes dark-brown when resting on the earth or among brown leaves. It is very interesting to notice that when this frog turns brown, irregular spots become conspicuous upon its skin, spots which evi- dently correspond to those upon the Common Frog (Rana), but which are invisible when the green tint is assumed. 1 Lister : Phil. Trans., 1858, vol. 148, p. 628. 84 THE COLOURS OF ANIMALS Variable Protective Resemblance in Reptiles The power of Variable Protective Eesemblance is therefore present among fish and Amphibia, but the most remarkable and well-known example is afforded by a reptile, the Chamaeleon. The rapidity with which the change of colour takes place, and the wide range of tints which the animal has at its command, have caused this lizard to be regarded as a type of everything changeable. But the same power is also present in certain of the South American lizards- belonging to the family Iguanidce, and it is probable that Variable Protective Eesemblance is much more common than has been generally supposed. The changes of colour depend upon the eye The physiological mechanism by means of which these rapid changes of colour are effected has been investigated by Lister in this country, by Briicke in Germany, and by Pouchet in France. At first sight it appears likely that the light may directly determine the distribution of colouring matter in the pigment cells in or immediately beneath the skin. Each of the various surroundings of an animal would, accord- ing to its colour, reflect light of a certain constitution, and it might well be supposed that each kind of re- flected light would produce a different effect upon the VARIABLE RESEMBLANCE IN VERTEBRATA, ETC. 85 pigment cells. It is, however, now well known that the action is extremely indirect; certain kinds of reflected light act as specific stimuli to the eye of the animal, and differing nervous impulses pass from this organ along the optic nerve to the brain. The brain being thus indirectly stimulated in a peculiar manner by various kinds of reflected light, originates different impulses, which pass from it along the nerves distributed to the skin, and cause varying states of concentration of the pigment in the cells. The highest powers of the microscope, assisted by ah1 the varied methods of histology, have failed to detect the con- nection between the nerves and the pigment cells in the skin, and yet such connection appears to be ren- dered certain by the fact that light falling on the e$B modifies the distribution of the pigment granules. The pigment cells in the skin are often of various colours, and are 'arranged in layers, so that very different effects may be produced by concentration in certain cells, leading to the appearance of those of another colour, or to a combined effect due to the colours of two or more kinds of cells. Blind animals cannot vary their colour protectively It has been shown by experiment that blinded frogs have no power of altering their colour so as to correspond with surrounding tints. The same facts also have been proved in a most interesting manner 86 THE COLOURS OF ANIMALS by the observation of living animals in their natural surroundings. Thus Pouchet noticed that one single plaice out of a large number upon a light sandy surface was dark-coloured, and thus unlike its sur- roundings. Examination showed that this individual was blind, and therefore unable to respond to the stimulus of reflected light.1 Another very interesting example of the same kind was brought under my notice by my friend, Mr. H. Nicoll. This gentleman had observed that in addition to the light- coloured trout usually seen in a chalk stream in Hampshire (a tributary of the Test), very dark individuals are occasionally met with. He was puzzled for a long time, but the fact that the dark fish could never be induced to rise to a fly finally led him to examine them, when he found that they were invariably blind, the crystalline lens being opaque. Sometimes the fish were blind in one eye, but this did not affect their colour. The darkness appears to come on gradually with increasing blindness, for the depth of the tint varies in different individuals, and some- times only part of the body (e.g. the tail) is affected. The blindness probably comes on with age, for the dark fish are always large, generally between one and two pounds in weight. 1 Quoted by Semper, Animal Life, International Scientific Series, pp. 95-9ti. VAEIABLE RESEMBLANCE IN VEETEBRATA, ETC. 87 The power of varying the colour essentially protective The protective value of the change of colour in normal trout was especially well seen when contrasted with these blind individuals. As it has been some- times asserted that protection is not the meaning of resemblance to the environment, I was anxious to observe so striking a contrast for myself. Mr. Nicoll kindly gave me the opportunity of seeing the fish in his stream, and I can in every way confirm his state- ment that a person unaccustomed to the observation of animals would certainly fail to detect any trout except the black ones. No one who had the opportunity of comparing the changing colours of the normal fish — ever harmonising with their surroundings,— with the unvarying conspicuous darkness of the blind in- dividuals, could hesitate for a moment in admitting that concealment is the one object of the adjustment of colour. The change of colour may also be voluntary or may follow from mental excitement. Thus the colours of fish often become much brighter while they are feeding.1 The food of blind trout It may be objected that the dark fish still continue to live in the same stream with the more perfectly 1 For a carious change of colour in the conger, see Bateson, Ix. pp. 214-15. 88 THP COLOURS OF ANIMALS concealed individuals. This is sufficiently explained by the facts that the waters are carefully preserved, and that the blindness only comes on when the fish are large, and are therefore exposed to the attacks of comparatively few enemies. Mr. Nicoll informed me that the black fish were usually in very poor condition, and I was very anxious to ascertain the kind of food which was still accessible to them. We therefore caught two fine specimens which were in fair condition. I opened them and found their stomachs quite full of caddice-worms, cases and all, together with a few fresh-water shrimps (Gammarus), These animals are doubtless hunted by scent and touch, while the insects on the surface of the water can only be obtained by sight.1 Loss of power of varying colour in a chameleon before death The changes which took place in a chamaeleon in my possession probably show the dependence of the power of adjustment upon the state of the nervous system. In the summer, while the lizard was healthy and had an abundance and variety of insect food, it was dark-coloured by day, when it rested on some dark branches or walked about in its shaded cage. Placed upon a leafy branch in strong light it became 1 Certain fish habitually seek their food by the olfactory and tactile senses. See Bateson, I.e. p. 214. VAKIABLE RESEMBLANCE IN VERTEBRATA, ETC. 89 yellowish-green in a very short time, the change beginning in a few seconds. At night when it was asleep it became light and straw-coloured. In the winter it died, probably on account of the scarcity and monotony of the only insect diet which could be obtained for it. For many days before its death it became almost black and lost all power of changing its colour. Its weakened nervous system either ceased to respond to the influence of light, or was unable to pro- duce any effect upon the pigment cells, which were thus paralysed, with their pigment permanently diffused. Green frogs also generally become dark before they die. Explanation of darkness of blind animals Some authorities have maintained that an animal of a kind which possesses the power of altering its colour should, when blind, become light- instead of dark-coloured. When the skin is light-coloured the pigment in the cells is strongly contracted, so that the coloured surface contributed by each cell occupies but a small space, and produces but little effect ; when the skin is dark the coloured parts of the cells are relaxed, and stretch out into the long branching processes, so that each dark surface becomes as large as possible. The latter is evidently the condition of rest, while concen- tration is the state of activity. It is therefore to be expected that when the coloured parts of cells are cut off from all stimuli, they will be permanently 90 THE COLOURS OF ANIMALS relaxed, so that the skin will be dark ; and we have seen that such a result actually occurs. When a muscle is cut off from nervous stimuli it also enters a condition of permanent rest or relaxation. Thus when the face is paralysed on one side, the muscles are relaxed and unable to balance the contraction of those on the other, so that the face is drawn over towards this latter side. The contractions of a muscle cell and those which take place in a pigment cell are not essentially different ; the former are far more specialised and powerful, but both of them exhibit manifestations of that contractile power which is possessed by the simplest cells. It is therefore of interest that both should behave in a similar manner when cut off from the nervous system which provides the stimuli under which both normally contract. In 1858, Sir Joseph Lister showed that the coloured part of a pigment cell contracts independently of the cell itself. Cells are now recognised as composed of a network containing a glassy substance in its meshes ; pigment granules are only contained in the network, and, as this contracts, it carries them inwards from the long branching processes towards the centre of the ceU. Loss of colour in cave-dwelling animals On the other hand it has been argued that the Proteus, a blind amphibian living in the underground rivers of Carniola, and Carinthia, is light-coloured, VARIABLE RESEMBLANCE IN VERTEBRATA, ETC. 91 and that other blind animals living in dark caverns are often white. In the majority of cases this result is undoubtedly due to the gradual disappearance of the useless pigment, and not to the excessive con- traction of the structures in which it is usually con- tained. Just as the useless eye has become rudi- mentary in these animals, so has the useless colour gradually disappeared from the skin. The energy necessary for the production and maintenance of such structures has been diverted, either wholly or in part, to other and more useful ends. In the Proteus, however, the degeneration is as yet incomplete, for the skin still retains pigment cells. An individual now in my possession has gradually become much darker since its removal from the cave at Adelsberg. It is probable that this result has followed from the direct effect of light upon the skin ; for it is known that superficial pigment cells are sensitive to light, although the changes of colour thus induced differ from those caused indirectly through the nervous system, in the absence of any harmony with the colours of the environment. The skin of the Proteus is probably extremely sensitive to light. By day the animal in my laboratory always lies concealed beneath a plate at the bottom of the aquarium, while ife comes out every night and swims freely about. As the eyes are very degenerate and buried beneath the surface, it appears certain that the difference between light and darkness is appreciated by the skin. W. Bateson has shown that blind 92 THE COLOURS OF ANIMALS shrimps and prawns bury themselves in the sand by day and swim about at night, exactly like the uninjured animals.1 The seasonal change of colour in northern mammals The well-known fact that many northern quad- rupeds become white in winter has given rise to a great deal of discussion as to the manner in which the change is brought about. Some have maintained that the animals simply acquire a new coat of white hair which conceals the darker fur beneath, while the long hairs of the summer coat are believed to be shed. Others believe that these latter actually change and become white, and that, although an abundance of new hairs also appear, nothing is shed. Most ob- servers agree that the white hair is shed at the close of winter : this is of course independently neces- sary, in order to reduce the thickness of the winter coat. I shall bring forward what appears to be conclusive evidence that the latter view is the right one, at any rate for certain species. But however the change is brought about, it will be rightly considered in this part of the subject, if it can be proved that it is called up either directly or indirectly by the stimulus pro- vided by the external conditions, and is not merely a contemporaneous change, harmonising with those in 1 L.c. p. 212. VARIABLE RESEMBLANCE IN VERTEBRATA, ETC. 93 the surroundings. The simplest view to take of the matter would be to suppose that natural selection has favoured an extra growth of hair of a white colour for the winter season, so that if an animal were trans- ported to the equator, a similar change would take place at a corresponding time. If the change was thus merely contemporaneous and without any actual physiological relation to the surroundings, it would require discussion in the previous chapters, for it would be precisely parallel to the darkening of the larva of the Privet Hawk Moth, which takes place whether it will descend upon brown earth with which it will harmonise, or green turf against which it will be conspicuous (see pp. 42-43) , It is possible that the change of certain purely Arctic animals is of this kind ; but it must be remembered that many such animals range southward into districts where the white coat would be conspicuous in whiter, so that the higher power of Variable Protective Resemblance would be very beneficial. The question is, however, one of evidence, and I shall show that in certain species the change in colour is physiologically associated with the conditions, like the change in the colour of a fish which depends on the reflected light entering its eye. A discussion of the probable nature of the physiological association is better deferred until after considering the evidence. 94 THE COLOURS OF ANIMALS Sudden change of colour determined by sudden exposure to extreme cold A classical experiment made by Sir J. Boss, con- siderably over fifty years ago, seems decisive on the above-mentioned point, as far as the species ex- perimented upon is concerned. A Hudson's Bay Lemming kept in the cabin, and thus shielded from the low temperature, retained its summer coat through the winter : ' It was accordingly placed on deck in a cage on February 1, and next morning, after having been exposed to a temperature of 30° below zero, the fur on the cheeks and a patch on each shoulder had become perfectly white. On the following day the patches on each shoulder had extended considerably, and the posterior part of the body and flanks had turned to a dirty white ; during the next four days the change continued but slowly, and at the end of a week it was entirely white, with the exception of a dark band across the shoulders, prolonged posteriorly down the middle of the back. . . .' No further change took place, and the lemming died of the cold on February 18, the thermometer having been between 30° and 40° below zero every night. 'On examining the skin it appeared that all the white parts of the fur were longer than the unchanged portions, and that the ends of the fur only were white so far as they exceeded in length the dark-coloured fur; and by VARIABLE RESEMBLANCE IN VERTEBRATA, ETC. 95 removing these white tips with a pair of scissors it again appeared in its dark summer dress, but slightly changed in colour, and precisely the same length as before the experiment.' l This experiment conclusively proves, — (1) that the external condition itself provides the cause which brings the appropriate change in colour; for the animal did not change until subjected to the con- dition ; 2 (2) that in all probability the cause is a lowered temperature acting upon the skin ; (3) that the existing dark hairs become white at the tips ; for we cannot well believe that a fresh growth could have overtopped the existing hair in a single night ; (4) that the whitening hairs grow suddenly and rapidly. Nature of the change of colour in the American Hare The same conclusions are also supported by some extremely careful observations conducted by F. H. Welch upon the American Hare (Lepus Americanus) in New Brunswick.3 In the latter district the animal keeps its winter coat till May, when it is gradually shed, the change being complete in June. The winter coat gradually develops in October and November, and 1 Sir J. Ross : Appendix to Second Voyage, Nat. Hist. p. xiv. 1835. 2 This conclusion is also supported by the fact that such changes occur earlier when the winter is exceptionally early. Concerning the Alpine Hare, see Tschudi, Thierleben der Alpenwelt, p. 300. » Proc. Zool. Soc. 1869, p. 228. 96 THE COLOURS OF ANIMALS is retained from December till the end of April. The appearance of the back and sides in summer is ' glis- tening fawn-colour interspersed with black, especially over the vertebral ridge.' The colour is conferred by long thick hairs (the pile) covering a woolly under- growth of a slaty colour. Early in October the first changes appear ; the whiskers become white at the tip or in some part of the shaft, and a few of the longer hairs on the back also become white at the tip or throughout. At this time there is no addition to the summer coat, only a change in the colour of existing hairs. The changes advance during November, and on separating the fur a new growth of stiff white hairs is seen over the sides and back : these grow rapidly, while the long hairs of the summer coat also grow and become white very quickly as soon as the new hairs appear on the surface. * The shaft of the hair of the new growth is invariably white, a circumstance which renders it easily distin- guished from the autumnal hair in process of change.' This change is most frequent at the tip, proceeding downwards, but it sometimes begins in the middle, and occasionally at the base. ' The whiskers, which apparently do not lengthen but merely alter in colour, will demonstrate each variety. ' Thus the whiter hue would appear to be brought about by a change of colour in the pile of the autumnal coat, combined with a new hybernal white crop, the latter undoubtedly playing no small part in the VARIABLE RESEMBLANCE IN VERTEBKATA, ETC. 97 colouring process and in the thickening of the fur. There is no indication of shedding : an increase in length ensues over the whole body.' There is considerable individual difference in the time of change : it sometimes commences before the first fall of snow, indicating that the stimulus is the fall of temperature affecting the skin rather than the colour affecting the eyes. Great differences are seen when the same species is followed into other localities. ' On the seaboard it (the winter change) is postponed in comparison with inland districts in the same lati- tudes.' In Hudson's Bay Territory it changes early and carries the winter coat till June, while no change of colour takes place in the winter in the southern parts of the United States. An individual kept in a warm barn at St. John's, New Brunswick, retained the summer colours. The consideration of the Hudson's Bay Lemming and the American Hare lead to the conclusion that all species in which the northern change does not occur in the southern individuals, possess the power of Vari- able Eesemblance. It is possible that the change is merely contemporaneous when it occurs uniformly in all individuals of the species, and it is at any rate probable that it would soon become so, because the extreme complexity of the mechanism by which Vari- able Resemblance is brought about would need the constant operation of natural selection to keep it in 9.8 THE COLOUES OF ANIMALS a state of efficiency. This consideration is better deferred until after the probable nature of the mechanism has been discussed. The physical cause of the change of colour It is now necessary to inquire into the actual physical cause of the change in appearance. It has already been explained that the dark colour depends upon absorption, while the whiteness depends upon scattering of light. The former is occasioned by pigment granules, the latter by included gas bubbles. When the latter are sufficiently abundant, the hair becomes white in spite of the pigment ; if then the gas were absorbed the dark colour would be restored.1 It appears to be well authenticated that in certain cases patches of human hair have become white during some nervous attack, again becoming dark at its cessation. Such changes can be explained by the evolution of gas (probably carbon di-oxide) at the base of the hair, and its subsequent absorption (probably by some alkaline fluid). It is therefore probable that the nervous system can so modify the processes taking place in the cells at the base of the hair as to cause 1 This explanation only applies to the existing dark hairs which become white. It is very improbable that any pigment exists in the new hairs which make up the great part of the winter coat. Hence, in testing the explanation offered above, the hairs must be selected with the greatest care, and the investigation should be conducted in connection with an experiment like that of Sir J. Boss (see pp. 94-96). VAEIABLE RESEMBLANCE IN VERTEBRATA, ETC. 99 the formation of gas bubbles. The many recorded cases of hair turning white in a few hours as the result of some strong nervous shock are to be explained in the same manner. . That the change in the long autumnal hairs of Lepus Americanus is due to the appearance of large numbers of bubbles is rendered probable by an exami- nation of Welch's figures and descriptions. He speaks of the white part of a hair being much broader than the coloured part, and containing additional rows of 'cells.' His 'cells' appear to be bubbles of gas, and he draws them with the characteristic dark borders. It must be remembered that the dark parts of a hair also contain bubbles, although in smaller amount. The change in the hair is indirectly caused by the change of temperature It is extremely improbable (to say the least) that such changes as the evolution of bubbles, and above all, the growth in length, are the direct result of a lowered temperature on the hair itself. That they are indirect results, through the nervous system, is in every way probable, and is furthermore in harmony with certain well-known facts concerning the regula- tion of temperature. The direct tendency of cold is clearly to diminish the activity of those processes upon which the pro- 100 THE COLOUKS OF ANIMALS duction of heat depends, just as it would tend to dimmish rather than promote the growth of hair. This direct effect is obvious in animals which are un- injured by variations of temperature. ' The body of a cold-blooded animal behaves in this respect like a mixture of dead substances in a chemist's retort : heat promotes and cold retards chemical action in both cases.' But the higher vertebrates are warm- blooded (homothermic), and such direct effects of cold would be fatal. ' In these animals there is obviously a mechanism of some kind counteracting, and indeed overcoming, those more direct effects which alone obtain in cold-blooded animals.' The influence of cold upon the nerves of the skin constitutes a stimulus to that part of the central nervous system which regulates the production of heat : thus cold indirectly increases the amount of heat, and the temperature of the body remains constant. I may mention that the amount of heat produced in the body at any one time may be gauged by the amount of oxygen ab- sorbed in respiration.1 It is in every way probable that such changes in colour as that of Sir J. Boss' Lemming and the American Hare are also indirectly caused by the cold, which we may suppose acts as a stimulus to that part 1 For a further account of the regulation of temperature see Professor MichaeT Foster's Physiology, from which the quoted sentences are taken. I owe the correct understanding of the physical cause of the change of colour to a conversation with Professor Foster. VAEIABLE KESEMBLANCE IN VERTEBEATA, ETC. 101 of the nervous system which presides over the nu- tritive and chemical changes involved in the growth of hair and the appearance of the bubbles. Probable variation in susceptibility to stimulus of cold in different districts In the northern part of an animal's range, natural selection would favour great delicacy in the adjustment of the mechanism by which such changes are produced, so that the winter coat would be ready in time to harmonise with the mantle of snow. Conversely, extreme delicacy would be a disadvantage in the southern part of the range, if the climate were such that the snow did not lie on the ground for any great part of the winter. There is abundant evidence of variations in the delicacy of adjustment, upon which natural selection could operate. Mr. F. E. Beddard has directed my attention to three Arctic Foxes (Cam's lagopus) from Iceland, which have been in the Zoological Gardens since 1887. One l of these turns perfectly white every winter, while the other two remain dark. The stoat always becomes white in the alpine districts of Scotland,' frequently in the north of Eng- 1 When I examined this fox on October 14, 1889, the change in colour was nearly complete : there was, however, a grey patch of hair on the back which was certainly moulting. It is possible, therefore, that the change is effected in an entirely different manner in this species. 6 102 THE COLOURS OF ANIMALS land, occasionally in the Midlands, and Mr. Couch has seen two white stoats in Cornwall.1 It would be extremely interesting to take a number of Scotch stoats to Cornwall and an equal number of Cornish stoats to Scotland, in order to test whether the southern individuals are less susceptible to change than the northern. It is likely that the great differ- ence is not wholly to be explained by the relation of northern to southern temperature, but at any rate partially by the fact that the change is disadvantageous in most parts of England ; for it would render the animal conspicuous against the prevalent tints of a midland or southern winter. Of course, any such disadvantage implies that natural selection would gradually blunt the susceptibility of the apparatus by which the change is produced. The rare cases of a change of colour in Cornwall are probably examples of a formerly beneficial susceptibility, as yet unaltered by natural selection. Loss pf susceptibility to stimulus of cold in animals which remain white all the year Such a nervous mechanism as that to which I have alluded, would be of the highest intricacy and com- plexity, and would speedily lose its efficiency unless constantly preserved by natural selection. Thus certain Arctic animals which remain on the snow 1 Boll : British Quadrupeds, second edition, pp. 196-201. VAEIABLE RESEMBLANCE IN VEETEBEATA, ETC. 103 nearly all the year retain the white coat permanently, and there is no need for the mechanism by which the change is produced. And yet in certain species we may feel sure that such a mechanism existed under former conditions. Thus the Arctic Hare (Lepus glacialis) usually remains white all the summer ; oc- casionally, however, it becomes greyish, the change of colour being limited to the points of the hair : the young are born grey, but change to white at their first winter (Welch). The latter change appears to be independent of cold, for Sir J. Boss speaks of a young hare turning white as early as those running wild, although in a temperature not much below freezing. This observation forms an in- teresting contrast with the behaviour of species pos- sessing an efficient nervous mechanism (the Hudson's Bay Lemming and the American Hare) when shielded from a low temperature. The white winter coat chiefly for concealment, but may also help to retain heat Certain northern animals, especially those fre- quenting trees, do not become white in winter : this is true of the Glutton (Gulo luscus). Occasionally dark winter individuals occur in species which as a rule change their colour regularly : thus, a black Arctic Fox is well known, but its rarity (Sir J. Eoss found three individuals out of fifty white ones) prob- 104 THE COLOTJKS OF ANIMALS ably indicates that it is, as we should expect, at a disadvantage, and that it will disappear. Mr. Wallace considers that the dark colour of arboreal northern animals, which is clearly for concealment, disproves the theory that the white colour is of value in retain- ing animal heat. But it does not follow that such benefits are wholly non-existent, because they must be dispensed with under the pressure of a stronger necessity. Mr. Wallace's argument shows that con- cealment is the paramount necessity ; but this does not disprove the opinion that other advantages also may be conferred by one particular mode in which concealment is attained. The seasonal change of colour in northern birds The same convincing evidence as to the nature of the .change, and the manner in which it is brought about, has not yet been brought forward hi the case of birds. Mr. A. H. Cocks, who has had a very wide experience of northern animals, believes that it is at least partially due to a change in the autumnal feathers. He writes : ' I have some specimens of Lagopus (various species) showing brown feathers with white tips, and in one species, at any rate, the converse.' Mr. E. Bowdler Sharpe does not however think that the evidence of a winter change in existing feathers is sufficient.1 He has nevertheless proved 1 H. Seebohm thinks it ' possible that the white winter feathers VAEIABLE KESEMELANCE IN VERTEBRATA, ETC. 105 that other changes of colour do occur, as will be seen in the following passage from his most interesting paper.1 ' Let any one who doubts the possibility of mark- ings such as those on the Greenland Falcon becoming gradually changed without an intermediate moult, study the changes exhibited by the common Sparrow Hawk in its progress towards maturity. The general characteristic of the species of Accipiter is to have a striped plumage when young and a barred dress when old. But it is not- generally known that this is effected by a gradual change in the markings of the feather, and not by an actual moult. On the first appearance of the feathers from the downy covering of the nestling, the markings on the chest are longi- tudinal drops (fig. 18) of a pale rufous-brown colour. The gradual dissolution and breaking up into three bars is shown in fig. 19. Hence, when the bars are perfectly developed a shade of darker brown over- (of Ptarmigan) gradually change colour in spring, only those being moulted which have been injured in winter.'— British Birds, vol. ii. p. 427, n. 1 Proc. Zool. Spc. 1873, pp. 414 et sec[. 106 THE COLOUES OF ANIMALS spreads the upper margin, gradually eclipsing the rufous-brown shade, which remains the evidence of the previous plumage (fig. 20). Hence are shown two successive stages of the development of the dark brown shade which at last removes all traces of the reddish tint (figs. 21 and 22).' If the winter change does not occur in the autumnal feathers, it by no means follows that the power of Variable Eesemblance is absent. The growth of new white feathers may be indirectly due to the cold, acting through the medium of the nervous system. This is, however, very far from being proved ; for it does not appear to be certain that there is a single species becoming white in winter which retains its dark colour at this time of the year in the southern- most part of its range. Mr. A. C. Billups, of Niagara, Ontario, tells me that during an exceptionally mild winter, about seven or eight years ago, neither the ' snow bird ' nor the American Hare acquired the winter dress. Hence the power of Variable Resemblance appears to be possessed by certain birda. VAEIABLE EESEMBLANCE IN VERTEBKATA, ETC. 107 Variable Resemblance of northern animals most nearly related to that of certain insects The acquisition of a special winter covering as a response to the stimulus of cold is most nearly related to the Variable Eesemblance exhibited by many cater- pillars and chrysalides. It will be shown in the next chapter that these latter changes are similar to the above in that the stimulus (of reflected light) acts upon the skin ; that the results are in all probability indirect, and take place through the part of the nervous system which regulates the production of colour ; finally, that far greater time is required for the accomplishment of the change than in those animals in which the stimulus acts upon the eye, and in which existing pigments are arranged instead of new substances elaborated. Rapid adjustment of Colour in certain invertebrate animals Certain invertebrate animals, however, possess the power of rapidly adjusting their colour to that of their surroundings. It is well-known in Crustacea, and is probably very common among them. The power has been proved to depend upon the eye as among the vertebrates. Some cuttle-fish also can modify their colours in the same manner, with remarkable pre- 108 THE COLOURS OF ANIMALS cision and rapidity. The resemblance between certain individuals of Ovulum and the rose-coloured coral, and between other individuals and the yellow coral (see pp. 70-71), is probably due to the existence of a power of adjustment ; but this suggestion needs experimental verification. A fact mentioned by Morse is even more convincing : he states that individuals of the same molluscan species occupying different stations are differently coloured, and he quotes from Dr. A. A. Gould the observation that the colour of all the shells found in the sandy harbour of Provincetown is remarkably light.1 There is no evidence as to whether the change in colour, if produced at all, takes place rapidly or slowly ; but the latter is the more probable in these animals. Professor Stewart found four or five bright red individuals of the Nudibranchiate mollusc Archidoris tuberculata in a mass of bright red sponge (Hymenia- cidon sanguined) upon which they were feeding.2 The colour was very different from that of individuals taken upon another sponge (Halichondria) . The ob- servation strongly suggests the existence of a power of Variable Protective Eesemblance, although it is possible that the colour of the food may be made use of. It is very likely that Variable Eesemblance will be found to occur far more generally than has been hitherto supposed. 1 Proc. Boston Soc. Nat. Hist. vol. xiv. April 5, 1871. 2 W. Garstang, l.c. p. 177. VARIABLE RESEMBLANCE IN VERTEBRATA, ETC. 109 Variable Resemblance, Protective and Aggressive As in so many other cases, the specialised form of concealment by the organism resembling its surround- ings treated of in this chapter may be either Protective or Aggressive ; it may enable an animal to escape its enemies or to approach its prey unseen. Frequently it may be turned to both uses by a single animal. Thus the green tree frog is probably aided in cap- turing the insects on which it feeds because of its close resemblance to the leaves around it ; but it is also protected in the same manner from the animals which prey upon it. Thus Mr. E. A. Minchin tells me, from his experience in India, that tree frogs are sought for with especial eagerness by snakes, which greatly prefer them to others. It is probable that this power when possessed by a vertebrate animal nearly always bears a double meaning, although a con- sideration of the different instances will show that it is especially Protective in some and especially Aggres- sive in others. In the next chapter we shall meet with a large number of cases briefly alluded to at the beginning of this chapter, in which the power is in many respects different, and possesses an entirely Protective meaning. 110 THE COLOUES OF ANIMALS CHAPTEE VIII VARIABLE PROTECTIVE RESEMBLANCE IN INSECTS No insect is known to possess the power of rapidly adjusting its colour to the tints of its surroundings, and it has not long been known that any power of adjustment exists. There is still a great deal to be done in finding out the extent to which the power is present, and in further investigating the physiological processes which are involved in its operation. Up to the present time the Lepidoptera (butterflies and moths) alone have been made the subjects of in- quiry, and we know nothing of other insects in this respect. Many caterpillars and chrysalides have been proved to be capable of adjusting their colours to those of the surroundings, and it is also known that certain cater- pillars can construct cocoons of different colours, so as to harmonise with the environment. The latter extremely interesting example of Variable Protective Resemblance has been very insufficiently investigated. It is also probable that a relatively small number of per- fect insects possess the same power ; but in this case VARIABLE PROTECTIVE RESEMBLANCE IN INSECTS 111 no experimental researches have been conducted. I will first consider the chrysalides, because they were first found to possess the power, and because it has been more completely investigated in them than in the other cases. Variable Protective Eesemoiance in Lepidopterous Pupae. The capability of adjusting the colour to that of the surroundings is only present in exposed chry- salides, and has not been found hitherto among the pupae of Heterocera (or moths), nearly all of which are either buried in the earth or concealed in opaque cocoons. In both cases the chrysalides are generally reddish-brown hi colour, the shade varying greatly in different species. The dark colour is of pro- FIG. 23. — The pupa of Swallow- FIG. 24. — The pupa of Swallow- tailed Moth, showing colour tailed Moth (Uropteryx sambu- assuined when the larva has caia) ; the usual dark colour been placed on white paper assumed in cocoon ; natural before pupation. size. tective value when the chrysalis is accidentally ex- posed upon the surface of the earth. Since this last paragraph was written, I have found that the chrysalis of the Swallow-tailed Moth 112 THE COLOURS OF ANIMALS (Uropteryx samlwcata) becomes light-coloured when the caterpillar has been placed upon white paper shortly before pupation (see fig. 23). The chrysalis is usually dark (see fig. 24), and is contained in a cocoon which is formed of the brown fragments of leaves or twigs spun together with threads of silk. The cocoon, which is suspended from the food-plant and swings freely, is so loose and open in texture that the enclosed pupa is easily seen, and is in fact as exposed as that of many butterflies. The chrysalides of butterflies are generally freely exposed, and many species have been proved to possess the power of adjusting the pupal colour to that of the adjacent surface. Such pupse are often suspended head downwards from a boss of silk, to which the hooks at the posterior end are affixed ; or they are fre- quently attached horizontally, or in a vertical position with the head upwards, by similar posterior hooks and a strong silken girdle, which is fixed on either side to the supporting surface, and which sinks into a groove across the back of the pupa. The group which in- cludes the Tortoiseshell and Peacock Butterflies adopts the former mode of suspension; that to which the * Garden Whites ' belong adopts the latter. VARIABLE PROTECTIVE RESEMBLANCE IN INSECTS 1 13 The History of the Discovery of Variable Protective Resemblance in the pupae of Butterflies In 1867 Mr. T. W. Wood exhibited to the Entomo- logical Society of London l a number of chrysalides of the Swallow-tailed Butterfly (Papilio machaori), and of the large and small Garden White Butterflies (Pieris brassicce and P. rapes), which corresponded in colour to the surfaces to which they were attached. Dark pupse had been found on tarred fences and in subdued light ; light ones on light surfaces ; while green leaves were shown to produce green chrysalides, at any rate in certain cases. Mr. Wood's inclusion of the chrysalis of the Swallow-tail, with which he states that he was imperfectly acquainted, was most unfortunate, and doubtless prevented his suggestive paper from gaining the success it deserved. It is quite true that this chrysalis appears in two forms, being sometimes green and sometimes dark grey ; but, without sufficient evi- dence, it was unwise, although most natural, to assume that these colours could be adjusted to green or dark surroundings respectively. I have since tested the chrysalis, and as far as my experiments (which were with small numbers) are conclusive, they show that it has no power of adjustment.2 In the discussion which followed Mr. Wood's paper, Mr. Bond stated that ' he 1 Proc. Ent. Soc. 1867, pp. xcix.-ci. • Phil. Trans vol. 178 (1887), B. p. 406-408. 114 THE COLOURS OF ANIMALS had had thousands of pupae of Papilio machaon, and had often had the brown variety of pupa on a green ground colour, whilst in some seasons he had obtained no brown specimens at all.' In spite of this unfortunate mistake, Mr. Wood adduced quite sufficient evidence concerning the Garden Whites to show that the subject was worth investigation. But the great example and the great principles of Darwin had not penetrated far into the mass of naturalists ; and distinguished entomologists preferred the expression of an adverse opinion, to making an easy experiment upon one of our commonest insects. Mr. Wood also stated that the chrysalis of the Large Tortoiseshell Butterfly (Vanessa polychloros) was coloured like a withered elm-leaf, when suspended among the foliage of the elm on which its caterpillar feeds. Its colour was then light reddish-brown with a cluster of metallic silvery spots, but when suspended from a wall, the metallic spots were not produced, and the pupa was of a mottled greyish colour. This observation led Mr. Wood to conclude ' that by the proper use of gilded surfaces the gilded chrysalides of Vanessa, and perhaps of other genera, would be ob- tained ' ; and he added, ' I hope to be able to try the experiment next season.' If this intention had been carried out, such startling results would have been obtained that opposition would have broken down before them, and the combined researches of many VAEIABLE PROTECTIVE RESEMBLANCE IN INSECTS 115 naturalists would have been brought to bear upon the subject. The experiment, however, was not made till nineteen years later, when I was led to do as Mr. Wood had proposed, although unaware at the time of his suggestion. Nevertheless, during these nineteen years, gradual confirmation of Mr. Wood's central position was afforded. In 1873 Professor Meldola supported the observations upon the chrysalides of the ' Garden Whites.' He compared large numbers of individuals, and found that the pupse upon black fences were darker than those upon walls.1 In 1874 a paper by Mrs. M. E. Barber, and commu- nicated by Mr. Darwin to the Entomological Society of London, was printed in the Transactions of that society.2 'Mrs. Barber had experimented with a com- mon South African Swallow- tailed Butterfly (Papilio nireus), and had found the chrysalis wonderfully sensi- tive to the colours of its environment. When the pupae were attached among the deep green leaves of 1 Zool. Soc. Proc. 1873, p. 153. 2 1874, p. 519. FIG. 25.— The pnpa of Papttio nireus attached to orange tree ; natural 116 THE COLOUES OF ANIMALS the food-plant, Orange, they were of a similar colour (see fig. 25) ; when fixed to dead branches covered with withered, pale yellowish-green leaves, they re- sembled the latter (see fig. 26). One of the cater- pillars ' affixed itself to the wooden frame of the case, FIG. 26.— The pupa of PapHio niretu attached to plant (Veprii lanceo- lata) with withered yellowish-green 7. — The pupa of Papilio nireus attached to woodwork. and then became a yellowish pupa of the same colour as the wooden frame' (see fig. 27). The case was made partly of purplish-brown brick and partly of wood, and one of the pupae, attached close to the junction, was believed by Mrs. Barber to have assumed VARIABLE PROTECTIVE RESEMBLANCE IN INSECTS 117 both colours, that of the brick upon its back and that of the wood upon its under surface. My experiments upon the chrysalis of the Small Tortoiseshell Butterfly, to be described below, do not support this conclusion, and it is a common thing for the colours of pupae to differ greatly in the dorsal and ventral regions. Mrs. Barber also tried the effect of scarlet cloth, but little if any influence was exerted. Mr. Mansel Weale also showed that the colour of certain other South African pupae can be modified,1 and Mr. Roland Trimen made some experiments upon another African Swallow-tail 2 (Papilio demoleus, common at Cape Town), confirmatory of Mrs. Barber's observations. He covered the sides of the cage with bands of many colours, and found that green, yellow, and reddish-brown tints were resembled by the pupae, while black made them rather darker. Bright red and blue had no effect. The larvae did not exercise any choice, but fixed themselves indiscriminately to colours which their pupae could resemble and those which they could not. In the natural condition the latter would not exist, for the pupae can imitate all the colours of their normal environments. Finally, Fritz Miiller experimented upon a South American Swallow-tail (Papilio poly damns)? and found Trans. Ent. Soc. Land. 1877, pp. 271, 275. 2 Described in a letter to me, published in my paper already referred to, p. 316. * Eosmos, vol. 12, p. 448. 118 THE COLOURS OF ANIMALS that its pupae, although appearing in two forms, dark and green, like those of our own Swallow-tail, also resemble the latter in having no power of adjusting their colours to the surroundings. Theories as to the manner in which the colours of such pupae are determined These observations and experiments had been made when I began to work at the subject in 1886 : they appeared to prove that the power certainly exists, but nothing was really known as to the manner in which the adjustment is effected. Mr. T. W. Wood's original suggestion, that 'the skin of the pupa is photographically sensitive for a few hours only after the caterpillar's skin has been shed,' was accepted by most of those who had worked at the subject. And yet the suggestion rested upon no shadow of proof ; it depended upon a tempting but overstrained analogy to the darkening of the sensitive photographic plate under the action of light. But the analogy was unreal, for, as Professor Meldola stated in the discussion which followed Mrs. Barber's paper, ' the action of light upon the sensitive skin of a pupa has no analogy with its action on any known photo- graphic chemical. No known substance retains per- manently the colour reflected on it by adjacent objects.' The supposed ' photographic sensitiveness ' of chrysalides was one of those deceptively feasible sug- VARIABLE PROTECTIVE RESEMBLANCE IN INSECTS 119 gestions which are not tested because of their apparent probability. It would have been very easy to transfer a freshly formed pupa from one colour to another which is known to produce an opposite effect upon it ; and yet if this simple experiment had been made the theory would have collapsed, for the pupa would have been found to resemble the first colour and not the second. Furthermore, Mr. Wood's suggestion raised the difficulty that chrysalides which had become ex- posed in the course of a dark night would have no opportunity of resembling the surrounding surfaces, for the pupal colours deepen very quickly into their permanent condition. In working at the subject I determined to pay especial attention to such ques- tions. Experiments upon the chrysalis of the Peacock Butterfly I began work with the common Peacock Butterfly (Vanessa lo), of which the chrysalis appears in two forms, being commonly dark grey (see fig. 28), but more rarely, bright yellowish-green (see fig. 29) : both forms are gilded, especially the latter. The gilding cannot be represented in the woodcuts. Only six caterpillars could be obtained, and these were placed in glass cylinders surrounded by yellowish-green tissue paper. Five of them became chrysalides of the corresponding colour; the sixth was removed imme- diately after the caterpillar skin had been thrown 120 THE COLOUES OF ANIMALS off, and was placed in a dark box lined with black paper, but it subsequently deepened into a green PIG. 28.— The pupa of Peacock FIG. 29.— The pupa of Peacock Butterfly ; dark form ; natural Butterfly ; light yellowish-green size. form. pupa exactly like the others. Obviously the sur- roundings had exercised their influence before the pupa was removed. Experiments upon the chrysalis of the Small Tortoise- shell Butterfly Being unable to obtain more larvae of the Peacock, I worked upon the allied Small Tortoiseshell Butterfly (Vanessa urticce), which can be obtained in immense numbers. In the experiments conducted in 1886, over 700 chrysalides of this species were obtained and their colours recorded. Green surroundings were first employed in the hope that a green form of pupa, unknown in the natural state, might be obtained. The results were, however, highly irregular, and there seemed to be no susceptibility to the colour. The pupas were, however, somewhat darker than usual, and this VAEIABLE PROTECTIVE RESEMBLANCE IN INSECTS 121 result suggested a trial of black surroundings, from which the strongest effects were at once witnessed : the pupae were as a rule extremely dark, with only the smallest trace, and often no trace at all, of the golden spots which are so conspicuous in the lighter forms. These results suggested the use of white sur- roundings, which appeared likely to produce the most opposite effects. The colours of nearly 150 chrysalides obtained under such conditions were very surprising. Not only was the black colouring matter as a rule absent, so that the pupae were light-coloured, but there was often an immense development of the golden spots, so that in many cases the whole surface of the pupae glittered with an apparent metallic lustre. So remarkable was the appearance that a physicist, to whom I showed the chrysalides, suggested that I had played him a trick and had covered them with gold-leaf. These remarkable results led to the use of a gilt background as even more likely to produce and in- tensify the glittering appearance. By this reasoning I was led to make the experiment which had been suggested by Mr. Wood nineteen years before. The results quite justified the reasoning, for a much higher percentage of gilded chrysalides, and still more remark- able individual instances, were obtained among the pupae which were treated in this way. The following table shows the results of some ex- periments in which the above-mentioned colours were employed : — 122 THE COLOUKS OF ANIMALS Dark Light (1) (2) (3) (3) (3) (4) (5) •„->> i| 'L If g2-3 a| | Degrees of colour *% •8-S *£ - g -Si ^ g-§ "" 0 s ? I js"S 11 ii 0 '-£ *!^ ,§» slS .£fS « |g "il "£>„„ S^« £ 1 ||| 0 c3 •all &1| al H 3 o" SKI guw t> Totals Green surroundings Black 2 11 8 29 27 25 22 14 1 2 8 = 3 = 105 White — 7 21 37 44 25 11 = 145 Gilt „ — 1 2 7 16 27 14 - 67 356 In order to realise these results it must be remembered that the appearance represented by (4) or (5) is very rarely seen in nature, except when the pupa is dis- eased. By far the commonest varieties met with are those represented by (3), which are therefore called normal forms. Special advantages of the Small Tortoiseshell for purposes of experiment From the results expressed in the tabular form given above, it was clear that this species was very susceptible to surrounding colours, and that black and gold produced the most opposite effects. Another advantage in the use of this species is the fact that the caterpillars live in companies, each of which VARIABLE PROTECTIVE RESEMBLANCE IN INSECTS 123 develops from the eggs laid by a single butterfly. Hence by keeping the companies separate, the varying hereditary tendencies, due to different parentage, are eliminated ; for a company of moderate size would contain over one hundred larvae, and would therefore furnish the material for several experiments. Con- sidering also the abtuadance of the species, I determined to employ it for the investigation of the process by which the change of colour is effected. The pupae darker when crowded together Very early in the investigation a possible source of error was detected. It seemed probable, when many individuals were collected together at the sus- ceptible period, which will be shown to occur towards the end of larval life, that each of them would be affected by that part of the surroundings which was constituted by the black skins of its neighbours. It was therefore necessary to take into account the relative positions of the pupae, and, in the most careful experiments, to place only a single individual in each coloured case. Experiment soon showed that these precautions were necessary. Many of the darker pupae, shown in the table to be produced by white and gilt surroundings, were proved to have been influenced 'by mutual proximity, so that the results would have been even more striking if this source of error had been allowed for. 124 THE COLOUKS OF ANIMALS The period during which the colours of pupae are determined A very large number of experiments and the closest and most frequent observations were devoted to the determination of the time during which these organisms are sensitive to surrounding colours. It was first necessary to observe everything that happens to a caterpillar between the cessation of feeding and the change into a chrysalis, for I felt sure that the time of susceptibility lay somewhere within these limits. When one of these caterpillars is full-fed, it descends from its food-plant (nettle) and wanders about in search of some suitable surface upon which to pass the pupal period. This is stage i., and its length varies greatly, according to the proximity of suitable surfaces. Then the caterpillar, having found the surface, rests motionless upon it, generally in a somewhat curved position. This is stage ii., and it is also variable in length, but fifteen hours may be accepted as a fair average of the time spent in this position. Finally the caterpillar hangs, head down- wards, suspended by its last pair of claspers (larval legs), which are attached to a boss of silk spun at the close of the second stage. This is stage iii., which lasts for about eighteen hours, at the end of which time the skin splits along the back behind the head, and the chrysalis is exposed by the skin being worked VAKIABLE PROTECTIVE EESEMBLANCE IN INSECTS 125 up towards the boss of silk. Then the tail of the chrysalis is withdrawn from the interior of the skin and is forced up the outside of the latter, until it comes in contact with the boss of silk. Contact immediately causes some of the numerous hooks on the end of the chrysalis to be entangled in the silk. During this apparently perilous operation the chrysa- lis is suspended to the larval skin, although different opinions obtain as to the exact method of its attach- ment. The sight is extremely interesting and beautiful, and the operation is almost always performed with precision and success. As soon as the pupa is firmly attached to the silk, it endeavours, by the most violent movements, to get rid of the skin, and generally succeeds in detaching it. Exact determination of the period of susceptibility The whole of the period before pupation, including the three stages, may be estimated at about thirty-six hours. Even if the caterpillars were susceptible during stage i., no effective results could be obtained ; for they are then wandering over surfaces of various colours, of which few can be the same as that which will form the environment of the chrysalis. Many experiments were conducted with the object of as- certaining the exact period of susceptibility. Larvae were exposed to one colour during stages i. and ii., and then transferred to another colour for stage iii., 7 126 THE COLOURS OF ANIMALS •while other larvse were exposed to each of the colours for all three stages : the effects were then compared. The results of the largest experiment of the kind are given below : — ! Degrees of colour Dark Light ,2 (as before) (1) W (3) (») (3) (*) (6) -5 In black surroundings for all three stages . . . — 1 — 6 — 1 — = 7 Transferred from black into gilt for stage iii. . . — — — 1 5 3 — = 9 Transferred from gilt into black for stage iii. . . — — — — 6 9 — = 15 In gilt surroundings for all three stages . . . — — — 5 7 8 = 20 I 51 This analysis speaks for itself. Stages ii. and iii. are both sensitive, but stage iii. is much less sensitive than the other. This is proved by the fact that the larvae which had been exposed to gilt surroundings during stage ii. and to black afterwards, were lighter thm those which had been exposed to black during stage ii. and to gilt afterwards. In other words, the coloured surroundings, both gilt and black, produced more effect during stage ii. than iii. ; but both stages are sensitive, because the black and gilt surroundings produced still greater effects when they operated for the whole period before pupation. It must be ob- served that the caterpillars, in the experiment sum- marised above, tended as a whole to produce the lighter forms of chrysalides, so that the black did VARIABLE PROTECTIVE RESEMBLANCE IN INSECTS 127 not cause nearly such strong effects as the gilt sur- roundings. The tendency was evidently hereditary and shared by all the caterpillars of the company, so that we have a striking example of the errors which were eliminated by keeping the companies separate. It is almost unnecessary to point out how com- pletely the old theory of ' photographically sensitive ' chrysalides is broken down by these experiments. Not only is the adjustment of the pupal colours to their surroundings due to larval susceptibility, but the larva itself has ceased to be highly sensitive many hours before pupation takes place. And this is to be expected, for during the latter part of stage iii. rapid changes are going on beneath its surface, and the developing pupa is becoming loosened from the larval skin which encloses it like a shell. Putting together the results of all the experiments, it is probable that in this species the influence of surrounding colours operates upon the larva during the twenty hours immediately preceding the last twelve hours of the larval state. Hence stage ii. is the great period of susceptibility, and this is probably the true meaning of the hours during which the caterpillar rests motion- less on the surface upon which it will pupate ; while stage iii. has other meanings connected with the rapid pupal development which is taking place. 128 THE COLOURS OF ANIMALS Determination of the part affected by surrounding colours Having thus defined the time of susceptibility, the next question was to ascertain the organ or part of the larva which is sensitive. At first it appeared likely that the larvae might be influenced through their eyes (ocelli), of which they have six on each side of the head. Hence in many experiments the eyes of some of the larvae were covered with an innocuous opaque black varnish, and they, together with an equal number of normal larvae from the same company, were placed in gilt or white surroundings. The pupae from both sets of larvae were, however, always equally light- coloured. It then seemed possible, although highly improbable, that the varnish itself might act as a stimulus similar to that caused by white or gilt surroundings, and therefore the experiment was re- peated with black surroundings in darkness ; but the pupae of the two sets were again almost identical, so that it appeared certain that the eyes can have nothing to do with the influence. It then seemed possible that the large branching bristles, with which the larvae are covered, might con- tain some organ which was affected by surround- ing colours, but experiments in which half of the larvae were deprived of their bristles showed con- clusively that the sensitive organs must have some VARIABLE PROTECTIVE RESEMBLANCE IN INSECTS 129 other position, for the pupse from both sets of larvae were identical. I was thus driven to the conclusion that the general surface of the skin of the caterpillar is sensi- tive to colour during stage ii. and part of stage iii. In order to test this conclusion I wished to subject the body of the same larva to two conflicting colours, such as black and gold, producing the most opposite effects upon the pupa. Such an experiment, if successfully carried out, would decide some important points. If the part of the body containing the head was not more sensitive than the other part, a valuable confirmation of the blinding experiments would be afforded. Mrs. Barber's suggestion that parti- coloured pupae may be produced by the influence of two colours would be tested in a very complete manner ; if parti- coloured pupae were obtained it seemed probable that the light acts directly upon the skin, but if they could not be obtained it seemed more probable that the light influences the termination of nerves in the skin, and that the pupal colours are produced through the medium of the nervous system. The practical difficulties in the way of such an ex- periment were very great, for the conflicting colours could only be applied during stage iii., when the larva is motionless and may be disturbed with impunity. If, on the other hand, a larva be disturbed in stage ii. it begins to walk about and thus renders the experiment impossible. The only way to obtain 130 THE COLOUES OF ANIMALS satisfactory results in spite of the slight susceptibility of stage iii. was to employ large numbers of larvaa, and to pay careful attention to minute differences of pupal colour as well as to the time during which the conflicting colours had been applied. The experiments were conducted in two ways. In the first the larvae were induced to suspend them- selves from sheets of clear glass, by placing them in wide shallow glass boxes so that the ascent to the glass roof was easily accomplished.. As soon as sus- pension (stage iii.) had taken place, each larva was covered with a cardboard tube divided into two chambers by a horizontal partition which was fixed rather below the middle. There was a central hole in the partition just large enough to admit the body of the larva. The tube was fixed to the glass sheet with glue; the upper chamber was lined with one colour, e.g. gilt, and the lower chamber with the opposite colour, e.g. black, with which the outside of the cylinder was also covered, in case the larva should stretch its head beyond the lower edge. The parti- tion was fixed at such a height that the larval head and rather less than half of the total surface of skin were contained in the lower chamber, while rather more than half of the skin surface was contained in the upper chamber. The arrangement is shown in section in fig. 30. The second method of conducting the conflicting colour experiments was superior in the more equal illu- VARIABLE PROTECTIVE RESEMBLANCE IN INSECTS 131 mination of the upper and lower colours. The bottom of a shallow wooden box was covered with alternate areas of black and gilt paper, and partitions were fixed along the lines where the two colours came into contact. Each par- tition was gilt to- wards the gilt surface and black towards the black, and was c perforated close to the bottom of the box with holes which would just admit the body of a larva. The box was then placed in a vertical position towards a strong light, so that the partitions became horizontal shelves, while the black and the gilt sur- faces were uppermost alternately. As soon as a larva was suspended to a glass sheet, the boss of silk was carefully scraped off and was pinned on the upper colour above one of the holes, so that the head and first five body-rings passed through the hole on to the colour beneath, which tended to produce opposite effects. Other larvae were similarly fixed between the shelves upon one colour only, so as to afford a comparison with the results of the conflicting colours. FIG. 30.— The larva of Small Tortoiseshell Butter- fly suspended in a tube of which the upper compartment is lined with gilt, the lower with black ; x 2. s. Boss of silk. b. Black, c. Card- board, g. Gilt. 132 THE COLOURS OF ANIMALS A careful comparison of all the pupse obtained in the conflicting colour experiments showed that, when the illumination of the two surfaces was equal, the effective results were produced by that colour to which the larger area of skin had been exposed, whether the head formed part of that area or not. Parti-coloured pupae were never obtained. It there- fore appears to be certain that the skin of the larva is influenced by surrounding colours during the sensi- tive period, and it is also probable that the effects are wrought through the medium of the nervous system. This latter conclusion receives further confirmation from other observations which will be described in the next chapter (see pp. 142- ^1G). CHAPTER IX VARIABLE PROTECTIVE RESEMBLANCE IN INSECTS (continued) The meaning of the metallic appearance of pupae APART from the general physiological significance of the results described in the last chapter, they are of extreme interest in giving us a possible clue to the meaning of the remarkable metallic appearance of the pupae of many butterflies. This wonderful appear- ance has given the name chrysalis to the second stage of Lepidopterous metamorphosis, although rela- tively few pupae are really entitled to bear it. But some pupae which deserve the name are very common, and probably have attracted attention ever since men began to look with interest on the world around them. Not only did the alchemists believe that in the appear- ance of these animals they received encouragement for the successful issue of the projects which were always before them, but we find that Aristotle, writing more than 2,200 years ago, mentions chrysalis as a word which was generally used * in his time, and which had therefore been invented as descriptive of the 134 THE COLOURS OF ANIMALS golden appearance at a still earlier period. There can be no doubt of this, for Aristotle's word is xpva-a\\is, identical with our own ; nor can there be any doubt as to the stage of insect life to which Aristotle was refer- ring, for his language is precise and descriptive. In fact, if a naturalist wished to convey to any one igno- rant of the changes undergone in insect metamorphosis a short and simple but perfectly accurate account of the two first stages of a Lepidopterous insect, he could not do better than use the very words of Aristotle : ' Caterpillars take food at first, but afterwards they cease to take it and become quiescent, being generally called chrysalides ; ' ! or again in another passage : 'Afterwards the caterpillars, having grown, become quiescent, change their shape, and are called chrysa- lides.' 3 Mr. T. W. Wood suggested that the metallic appearance was so essentially unlike anything usually found in the organic kingdoms, that it acted as a protection to the organisms possessing it. Others have thought that it has the value of a warning colour, indicating an unpleasant taste (see Chapter X.). It is probable that the appearance sometimes bears this meaning now, but it is unlikely that such was its original use ; for the fact that metallic colours can be called up or dismissed by the appropriate sur- 1 o? re yap Ka.fj.vai \afi.&avovffi rii vpurov Tptxpfyv, yueri ravra ovKtrt i', d\A.' aKivtiTi^ovfflv al KaAov/j.ei/ai \nr& Tivtav xpWAAfSes. iiero 5£ -ravra (at Ka.fj.ifui] av|7j0«i ical T^V (iop1)v, ical KaXovvrai VARIABLE PROTECTIVE RESEMBLANCE IN INSECTS 135 roundings shows that they are essentially protective, and as far removed as possible from conspicuous warning colours, the object of which is to render their possessors unlike the environment. What can be the object in nature which the glittering pupae resemble ? It is obvious that metals are not sufficiently abun- dant on the surface of the earth to afford models for successful imitation, and there is the same objection to certain metallic sulphides which otherwise would answer the purpose admirably. A consideration of the darker non-glittering va- rieties of the same species helps us to an explana- tion. These certainly resemble the grey surface of weathered rocks, and the whole shape of such pupae, with their angular projections and tubercles, com- bines with their colour to produce a most perfect Protective Kesemblance to rough dark surfaces of rock. In fact, did we not delude the larvae by offering them flat mineral surfaces in our walls and sides of houses, the protection would be so complete that we should hardly ever find the chrysalides ; and, as a matter of fact, they are rarely seen except in such situations. In England we very rarely see a brightly metallic pupa because in our moist climate exposed rock- surfaces quickly weather and become lichen-covered. If, however, the bright appearance of many recently fractured rocks were retained, as they are in drier countries, they would cause the production of a similar 136 THE COLOUES OF ANIMALS appearance in the pupae of those larvae which sought them. Although metallic surfaces are not conspicuous in nature, there is a very abundant glittering mineral •which is quite common enough to offer a surface against which the larvae might often suspend them- selves. I refer to the mineral mica, the substance forming the glittering flakes which are so well-known in common granite. Furthermore, any recently broken rock contains bright and glittering surfaces, although they may not be so brilliant as mica, and the bright spots of the pupae would thus be of pro- tective value against almost any freshly exposed mineral surface. Hence we see that the pupae would occur as dark or glittering forms, as the surrounding mineral sur- faces are dark or glittering : they appear in two different varieties which are respectively in harmony with the two conditions of the mineral surfaces they resemble — the dark and weathered, and the bright and freshly exposed condition. It may be that this adaptation to mineral sur- roundings arose when the widespread green tints of the vegetable kingdom contributed less to the total appearance of land-surfaces ; or the adaptation may have followed the habit of feeding upon herbaceous plants which withered away in the hot season, changing from green to brown during the time when the insect was in the chrysalis state and could VARIABLE PROTECTIVE RESEMBLANCE IN INSECTS 137 undergo no corresponding change of colour. How- ever the adaptation arose in the ancestor of all butter- flies which now emerge- from gilded chrysalides, it is probable that it took place in some hot dry country, where mineral surfaces did not weather quickly but remained glittering for long periods of time. The manner in which golden chrysalides are adapted for concealment on plants In the origin and gradual progress of our modern aggressive forms of vegetation, less and less of the land-surface has been formed by mineral substances, until the green colour of foliage and the brown colour of stems and of withered leaves have become the pre- dominant tints of nature and the most feasible models for Protective Eesemblance. It is therefore interesting to note how the species with gilded pupse have adapted themselves to the change. The chrysalis of the Peacock Butterfly (Vanessa lo) still retains the dark variety, which is formed when pupation takes place upon dark rock surfaces ; but the golden form has been replaced by a green variety, which is produced when the chrysalis is suspended from the leaves of its food-plant. The green variety still retains the metallic appearance, and exhibits it to a much greater extent than the dark variety. During the summer of 1888 I found that the green form is produced by the surroundings which cause 138 THE COLOURS OF ANIMALS the appearance of the gilded form of the Small Tortoiseshell chrysalis, viz. by a gilt and by a white environment. The chrysalis of the Eed Admiral Butterfly (Vanessa Atalanta) has no green variety, but it appears, like the Small Tortoiseshell, as a dark or a glittering form resembling the two conditions of rock-surfaces upon which it often pupates, hanging suspended without any attempt at concealment except such as is afforded by its very perfect colour-harmony with the surroundings. I have shown that this species also is susceptible, and that either variety of pupa is produced by the appropriate environment. But this chrysalis is very commonly found attached to the food-plant, and when this is the case it hangs suspended in a tent formed of leaves carefully spun together by the caterpillar, so that it is concealed from view. The larva also often has the habit of partially biting through the leaf-stalk or stem, so that the leaves of its retreat hang down and wither. The dead brown leaver thus afford a far more harmonious background for the dark pupa, if by any chance it becomes exposed to view. The Small Tortoiseshell has neither the green variety of the Peacock nor the protective habit of the Eed Admiral, and therefore it almost invariably seeks mineral surroundings for the pupal period, and very rarely becomes a chrysalis on its food-plant. In 1886 I only found three such pupse suspended to VARIABLE PROTECTIVE RESEMBLANCE IN INSECTS 139 the food-plant, although I examined the nettle-beds where many hundreds of caterpillars had been feeding and had left for pupation. All these three pupae were dead, being filled with the parasitic larvae of Ichneumon flies. In 1888 I found many more pupae upon the food-plant, but a very high percentage of these had been killed by parasites, and the hurrying on of pupation which occurred in the other cases and pre- vented the larvae from wandering in a normal manner may, I think, be attributed to the state of health induced by that extraordinarily wet season. The colours of certain dimorphic pupae cannot be adjusted to the surroundings I have already mentioned that I experimented upon the pupae of the Swallow-tailed Butterfly (Papilio machaori), and found that they were not sus- ceptible to the influence of surrounding colours. This is also true of the small family of Mocha moths (Ephyrida) which have freely exposed pupae, fixed like those of many butterflies by a silken girdle and boss, and often appearing in two varieties, green and brown. The caterpillars of the same species are also of two colours, and always produce pupae of corre- sponding tints (see page 46). 140 THE COLOURS OF ANIMALS Variable Protective Resemblance in the pupae of the Pieridae. The susceptibility of the two species of Garden White Butterflies (Pieris brassica and P. rapes) was also investigated in the same season (1886), and the results of previous observers were confirmed and extended. Many colours were employed, and it was found that the light reflected from yellow and orange surroundings was very potent in producing bright green varieties of the chrysalides of both species. It is therefore probable that when the light reflected from green leaves produces this effect in nature, the yellow and orange constituents of the light form the stimuli. When, therefore, these constituents are made use of nearly alone, they produce still more marked effects. Black and white backgrounds caused the pupae of both species to become dark and light respectively, and all other colours except yellow and orange produced more or less dark pupae. Experiments were made upon P. rapce, to ascertain the susceptible period, the larvae being transferred as in the case of the Small Tortoiseshell. The results were as in the latter : the larva is sensitive and not the pupa, and the time of chief susceptibility is during stage ii. A few larvae of P. rapce were blinded, but the chry- salides were similar to those produced by normal larvae. VARIABLE PROTECTIVE RESEMBLANCE IN INSECTS 141 Further experiments on the same subject During the summer of 1888 I conducted further experiments upon the same subject. The results are as yet imperfectly worked out, and are unpublished, but I will shortly mention the chief conclusions. Other glittering metallic surfaces, such as silver or tin, do not produce anything like so striking an effect as gold upon the pupaB of the Small Tortoiseshell. It seems probable that the yellow light reflected from the gold is effective in preventing the formation of pigment, and in thus producing the gilded chrysalides, just as the yellow light also prevents the appearance of pigment and produces the bright green pupae among the Pierida. Two new species also were investigated, and proved to be sensitive. The pupae of the Silver-washed Fritillary (Argynnis paphia) can be rendered dark or light in colour, although the metallic spots do not seem to be affected. The pupae of the Large Tortoise- shell (Vanessa polycTiloros) were also rendered dark brown without metallic spots, or light reddish-brown with the spots, by the use of appropriate surroundings. The metallic spots could not be extended over the pupal surface as in the case of the Small Tortoise- shell. 142 THE COLOUKS OF ANIMALS Confirmatory results obtained by other workers It is also interesting to record that many of these results have been since confirmed by independent workers. Mr. G. C. Griffiths worked at the chrysalis of the Small Garden White (Pieris rapce), and con- firmed my results in many important respects.1 The Eev. J. W. B. Bell and Mr. Pembrey have worked at the pupae of the Small Tortoiseshell and Peacock, and the former also at the pupse of the Large Tortoise- shell.2 Their results are, on the whole, confirmatory of those described above. Variable Protective Resemblance in the colours of cocoons It has been already mentioned that the colour of the cocoon in certain species can be adjusted to the environment. I obtained proof of this in 1886, at the suggestion of Mr. W. H. Harwood of Colchester, who had observed that the colour of the cocoon of the Emperor Moth (Saturnia carpini) varied, and seemed to suit its environment. I found that caterpillars of this species spun very dark brown cocoons in a black calico bag (see fig. 31), while white cocoons were Trans. Ent. Soc. Lond. 1888, pp. 247 et seq. The Midland Naturalist, Dec. 1889, pp. 289-90. VARIABLE PROTECTIVE RESEMBLANCE IN INSECTS 143 spun in white surroundings in a strong light (see fig. 82). » In this case it seems almost impossible for the surrounding colours to influence directly the colour of the cocoon. It is necessary to assume the existence FIG. 31.— The cocoon of Emperor Moth PIG. 32.— The cocoon of Emperor Moth, (S. carpini), spun in a black calico spun on a white surface in strong bag ; natural size, although an excep- light ; natural size, tionally small cocoon. of a complex nervous circle as a medium through which the stimulus of colour can make itself felt. If this conclusion be correct it is probable that the colours of the pupa and larva are adjusted in the same manner. The observation upon S. carpini has been con- 1 Proc. Hoy. Soc. vol. xlii. p. 108. I have since found that the fact must have been known previously, for it is quoted in Mr. A. R. Wallace's Tropical Nature. I do not yet know the name of the naturalist who made the observation. 144 THE COLOURS OF ANIMALS firmed, and has been extended to other species. Thus Kev. W. J. H. Newman showed that the cocoons of the Small Eggar Moth (Eriogaster lanestris) are creamy white when spun upon white paper (see fig. 33) , dark FIG. 33. — The cocoon of Small Eggar PIG. 34. — The cocoon of Small Eggar Moth (E. lanestris), spun upon white spun among green leaves, paper ; natural size. brown when constructed among leaves (see fig. 34) -1 These cocoons are so compact and smooth that they re- semble birds' eggs : a fact which explains the name of the moth. In constructing the cocoon the caterpillar leaves a few holes, which are doubtless of importance in permitting a free exchange of air. The fact that light reflected from green leaves is here the stimulus for the production of a dark colour is readily intelligible when we remember that the moth does not emerge till the following February at the earliest, while the in- sect often remains in the pupal state for one or two years longer. The leaves in contact with the cocoon soon die and turn brown, and after this change the dark colour is highly protective. It is also of especial importance for the cocoon to be well concealed during the winter months, when insect-eating animals are 1 /Yoc. Ent. Soc. Land. 1887, pp. 1. li. VARIABLE PROTECTIVE RESEMBLANCE IN INSECTS 145 pressed for food, and are obliged to search for it with extreme care. I have also shown that the cocoon of the Green Silver Lines Moth (Halias prasinand) can be modified in colour like that of the Small Eggar.1 This species also passes the winter in the chrysalis state, when the brown colour is highly protective. One of my caterpillars had begun to spin a brown cocoon upon an oak leaf (see fig. 35). I then removed the caterpillar to a white box ; it remained motionless for several hours and then FIG. 36. — White cocoon spun by same caterpillar when transferred from oak leaf to white paper; natural Fio. 35. — Brown oocoon begun by caterpillar of Green Silver lines Moth (fi.prasinana)\ifOQ an oak leaf ; natural size. FIG. 37, — The brown oocoon of Green Silver Lines spun upon an oak leaf. spun a white cocoon (see fig. 36). The brown cocoon of the same species is shown in fig. 37. Remembering the experiments upon the Small Tortoiseshell, it is very probable that the colour of the cocoon was deter- Proc. Ent. Soc. Lond. 1887, pp. 1. li. 146 THE COLOUES OF ANIMALS mined during the time when the caterpillar was motionless in the box. Still later in 1888 Dr. E. G. Lynam sent me some cocoons of the Gold-tail Moth (Liparis auriflua) which had been also modified in a similar manner, and I found that the same power is possessed by the cater- pillar of the Brimstone Moth (Rumia crat&gatd) .l In this latter case a green tissue-paper background pro- duced brown cocoons like those spun upon green leaves. It is probable that this power of adjusting the colour of the cocoon is very common among species which spin in exposed situations. It may also be expected to occur in those Hymenoptera with similar habits. The investigation of the physiological pro- cesses involved in the adjustment would be of extreme interest. Last year (1888) I obtained a large number of Small Eggar caterpillars, intending to begin such an investigation, but nearly all of them died just before reaching maturity. It is to be hoped that many species will now be tested in order to ascertain whether this form of susceptibility is present. Variable Protective Resemblance in Lepidopterous larvae It now remains to briefly consider the power of colour-adjustment possessed by certain caterpillars. Naturalists have long known that in certain species 1 Proc. Ent. Soc. Land. 1888, p. xxviii. VARIABLE PEOTECTIVE KESEMBLANCE IN INSECTS 147 the colour of the caterpillars may vary according to the colour of the plant upon which they are found. This is especially true of caterpillars feeding upon brightly coloured parts of the plant, such as the anthers or petals. At the same time there has been, until recent years, hardly any systematic investigation of these interesting facts. Professor E. Meldola's editorial notes to his translation of Dr. Weismann's ' Studies in the Theory of Descent ' (the essay on ' The Origin of the Markings of Caterpillars ') contain many instances of this kind, together with most sug- gestive remarks upon them, which first induced me to work at the subject. At a still earlier date the same writer had brought together all the scattered examples of this kind, including the power of adjusting the colours of pupae, and had drawn attention to the general principles involved.1 Experiments upon the larva of the Eyed Hawk Moth (Smerinthus ocellatus) The instance which Professor Meldola chiefly con- siders in his editorial notes is that of the caterpillar of the Eyed Hawk Moth (Smerinthus ocellatus), which is of a whitish-green colour when it is found upon apple and certain kinds of willow, and of a bright yellowish-green when found upon other species or varieties of willow. ' The colours are on the whole 1 Proc. Zool. Soc. 1873, p. 153. 148 THE COLOURS OF ANIMALS protective; the larva resembles the under side of a rolled- up leaf, and when the food-plant bears leaves with white and downy under sides (apple, Salix viminalu, &c.) the larva is usually whitish ; while it is generally yellowish-green upon trees of which the leaves have green under sides (Salix triandra, S. baby- lonica, S. rubra, S.fragilis, &c.). I remember, when a boy, finding the two varieties of larva, and being much astonished at the difference between them. I began working at the species in 1884, and have bred large numbers of the larvae for every season since that year. Only the results of the earlier experiments are published.1 The eggs of each female moth were kept separate, and the caterpillars of each batch were fed upon a variety of food-plants, and manifested de- cided differences in their shade of green. At the same time remarkable exceptions occasionally occurred : sometimes, also, when collecting I have found bright green individuals upon apple. Blinding experiments like those upon the Small Tortoiseshell led to negative results. These experiments were very laborious, for a caterpillar changes its skin four times, and with it the covering to its eyes and the opaque varnish. Hence, before each change of skin the caterpillars were sepa- rated from the food, and, after changing it, were re- blinded before being restored. Before this investigation had been begun, it waa believed that such variability in caterpillars was due to 1 Proc. Boy. Soc. vol. xxxviii. p. 269 ; vol. xl. p. 135. VARIABLE PROTECTIVE RESEMBLANCE IN INSECTS 149 the direct chemical effect of different kinds of leaves upon them after being eaten, and it was therefore called phytophagic variability. Many special experi- ments were directed toward the solution of this question. Thus, leaves were sewn together, so that the caterpillars were exposed to the colour of the upper or of the under side alone, although they ate the same leaf in both cases. In other instances the ' bloom ' was rubbed off the under sides of some leaves (Salix fragilis, incorrectly described as Triandra in my papers) , while others were left normal. The results proved that the caterpillars are affected by the colour of the leaves and not by the leaves as food. Com- parison with the experiments on pupae renders it most probable that reflected light influences the skin. Experiments upon the larvae of other Sphingidse Professor Meldola had also quoted the instance of the larva of the Privet Hawk Moth (Sphinx ligustri), which is of a much brighter green when found upon privet than when found upon lilac. Larvae of this species, from the same batch of eggs, were fed upon the two plants, and the above quoted observation was confirmed. The larvae of the Lime Hawk Moth (Smerinthus tilice) were similarly modified, being made unusually light green by the use of variegated elm and a lime having leaves with very white and downy under sides. 8 ISO THE COLOUKS OF ANIMALS Experiments upon the larvae of Geometrae and Noctuae The experiments were then extended to many other dark-coloured larvae (chiefly Geometree). The method of experiment was as follows : a larva which resembles the twigs or bark of its food-plant was selected, and was surrounded by the leaves upon which it fed, and by white or green surfaces. No brown twig or anything dark-coloured was allowed to come near it during its whole life. Under these circum- stances the larvae, in the majority of the species selected for experiment, became very light brown or light grey in colour, and quite unlike the darker larvae of the same kinds which were produced when an abundance of dark twigs had been mixed with the leaves of the food-plant.1 The results were certainly protective, for the lighter larvas were far less conspicuous en the green leaves and stems than the darker ones would have been. At the same time it must be admitted that the resemblance of the darker forms to the dark branches 1 These experiments have been successfully applied to the follow- ing Gcometrce : — Crocallis elinguaria (for two seasons), Ennomos angularia, E. htnaria, Boarmia rhomboidaria (this species was inves- tigated by my friend and pupil, Mr. E. C. L. Perkins, B.A., of Jesus College, Oxford), B. roboraria ; and to one of the Noctuce, Catocala sponsa, Since this note was written, I have found, during the past summer (1889), that the Geometer Heterophylla abruptaria, and the Noctuas Catocala electa and C. elocata, are also sensitive, the first and last named to a marked degree. VAEIABLE PKOTECTIVE RESEMBLANCE IN INSECTS 151 is much stronger than that of the light varieties to the leaves (see figs. 38 and 39). Two species, how- ever, are already known in which the green stems and FIG 38. The larva of a continental Noctua (Calocala elocata) with the colour adjusted to that of the dark twigs mixed with its food-plant ; nearly full-fed ; two-thirds natural size. FIG. 39. — The colour of a larva of the same species when only preen twigs and leaves were supplied to it. The food-plant in both cases was black poplar (Populut nigra). leaves cause the production of green larvae, so that the concealment is very perfect. And we may be quite sure tha,t there are many other species with equal powers. Experiments upon the larvae of the Brimstone Moth Lord Walsingham first pointed out to me that the larvse of the Brimstone Moth (Eumia cratcegata) vary from brown to green, and through all intermediate shades. I found that when brown objects were 152 THE COLOUKS OF ANIMALS entirely excluded the larvae became greenish-brown, brownish-green, or sometimes of a decided green colour, and thus harmonised well with the leaves and young green twigs of the hawthorn. In the presence of dark twigs they became dark brown like so many other larvae.1 Experiments upon the larvae of the Peppered Moth The second instance is even more remarkable, and has only been observed during the present year (1889). i I obtained some hundreds of eggs from a single wild female of the Pep- pered Moth (Am- phidasis betularid), and the caterpillars which hatched were treated as in the other experiments. The larvae reared among green leaves and shoots became FIG. 41.— The larva of Peppered Moth surrounded by bright green (866 abundant dark twigs as well as leaves. fig. 40) iciihout ex- ception, while the others in nearly all cases assumed the colour of the dark-brown twigs, which were mixed 1 Report of the British Association, 1887, p. 756 ; also Nature, vol. 36, p. 594. FIG. 40.-The larva of Peppered Moth (A. betularia) surrounded by green twigs and leaves ; full-fed ; half natural size. VAKIABLE PROTECTIVE RESEMBLANCE IN INSECTS 153 with the leaves upon which they fed (see fig. 41) : about one or two per cent., however, took their colour from the latter. The food-plants were the same in both experiments. The change of colour is not due to the food seen through a transparent skin Some authorities have supposed that the change of colour under such circumstances is a comparatively simple thing, that the younger green leaves eaten and seen in the alimentary canal through the more or less transparent tissues cause a brighter appearance, while the older leaves produce in the same manner a darker appearance. This cause of colour is certainly efficient in many transparent caterpillars (see p. 79), such as some of the Noctua, but it does not account for any of the results obtained in my experiments. As a precaution against such an error, I reversed the surroundings of a few larvae of most of the species experimented upon. The new conditions were main- tained for some days, during which the contents of the alimentary canal must have been changed many times, but no perceptible effect was produced. This result also serves to show that the influences act very slowly, and that the processes of adjustment are totally dif- ferent from those which cause the rapid changes of colour considered in Chapter VII. 154 THE COLOUKS OF The difference between slow and rapid adjustment of colour The essential difference between the two kinds of adjustment is that, in the one case, the pigmented part of certain cells contracts in obedience to nervous stimuli, and thus alters the general appearance ; while in the other case the coloured part is actually built up of the appropriate tint, or loses its colour alto- gether and becomes transparent in obedience to the same stimuli. The frog or fish has a series of ready- made screens which can be shifted to suit the environ- ment ; the insect has the power of building up an appropriate screen. In many cases, however, the green colour of caterpillars is due to the ready-made colour of the blood, which becomes effective when pigment is removed from the superficial cells, but which disappears when the latter are rendered opaque. Here, however, the superficial cells form the screen which has to be built up or from which the colour must be dismissed ; and in certain species even the colour of the blood is entirely changed in the passage from a green to a dark variety or vice versa. Hence it is to be expected that the changes occur- ring in an insect will occupy a considerable time as compared with those which take place in a frog. Another difference between the two processes is that VARIABLE PROTECTIVE RESEMBLANCE IN INSECTS 155 the stimulus from the environment falls upon the eye in the one case and probably upon the surface of the skin in the other. Variable Protective Resemblance in insects is no explanation of the origin of colour Many authorities have believed that, in the results of these experiments upon the colours of insects, we see an explanation of the origin of colour, by the direct influence of environment accumulated through many generations. This is a very tempting conclusion, and one which for a time appeared to me to be satisfactory. But as soon as there was clear evidence that the medium of the nervous system is necessary, the results were seen to be indirect, and to have needed the most astonishing adaptations on the part of the organism before the colour of the environment could exercise any influence upon it. It might still be maintaine'd that the existing colours and markings of certain caterpillars are at any rate in part due to the accumulation through heredity of the indirect influence of environment, working by means of the nervous system. To this it may be replied that the whole use and meaning of the power of adjustment depends upon its freedom during the life of the individual ; any hereditary bias towards the colours of ancestors would at once destroy the utility of the power, which is essentially an adaptation to the 156 THE COLOURS OF ANIMALS fact that different individuals will probably meet with different environments. As long ago as 1873, Pro- fessor Meldola argued that this power of adjustment is adaptive and to be explained by the operation of natural selection.1 Comparison between the varying effects of green leaves upon the different stages of an insect strongly supports the view that the results are due to adaptation. Thus the caterpillar of the Brimstone Moth remains upon its food-plant for a few weeks in the summer when the leaves are green, and green leaves cause the larva to become green and to lose the dark pigment. But the chrysalis remains among the leaves in winter when they have become brown, and green leaves cause the caterpillar to spin a dark cocoon. Hence precisely opposite effects are produced by the operation of the same force, the nature of the effects having been determined by adaptation. Furthermore, there is no positive evidence for any of these effects becoming hereditary. I have carried on some of my experiments for more than one genera- tion, always carefully noting the effects produced in the parents, and have never been able to detect any resulting hereditary tendencies, even when the previous generation had been powerfully influenced. When therefore we meet with a dimorphic species which is not influenced by its environment, so as to produce the appropriate form, I do not believe that we ' Proc. Zool. Soc. 1873, p. 153. VAKIABLE PROTECTIVE RESEMBLANCE IN INSECTS 157 are witnessing the results of a power of adjustment which existed in the past but is now lost. I think, on the other hand, that variability or dimorphism pre- ceded the power of adjustment in all cases. I have already shown that these appearances possess a pro- tective value even when they cannot be adjusted (see pp. 46-48). When Variable Protective Eesemblance is present, but acts somewhat uncertainly (as in the larva of the Eyed Hawk Moth), it is probable that the power has been only recently acquired and is still imper- fect. This conclusion is supported by the fact that the closely allied caterpillar of the Convolvulus Hawk Moth (Sphinx convolvuli) has no power of adjustment, although it is completely dimorphic 1 (see pp. 47, 48). Before finally leaving this part of the subject I will briefly allude to facts which render it probable that certain perfect insects possess the power of Variable Protective Eesemblance. Variable Protective Resemblance probable in certain Moths The colour of certain insects varies with the pre- vailing tint of the locality in which they occur. The best instance known to me is that of one of the Geometrce, the Annulet Moth (Gnophos obscurata). This moth is light-coloured in chalky localities (e.g. on the chalk at Lulworth), but darker when the pre- 1 Trans. Ent. Soc. Land. 1888, pp. 552-553. 158 THE COLOUES OF ANIMALS vailing tint of the earth is dark, as in peaty districts. It is improbable that these are local races, and the only other interpretation is that the colours can be varied as the result of a stimulus. No experimental proof of this has been as yet afforded. If the view adopted here be correct, it will be of extreme interest to define the susceptible period ; it will most probably be found at the close of larval life. I have treated this part of the subject at some length and have discussed many details. I have done so because the inquiry is new, and will not be found in other books on the colours of animals ; l and also because I hope that some of my readers may be in- duced to carry on investigations for themselves in a field which is easily entered, and in which further help is especially necessary. 1 Since this sentence was written, Mr. A. B. Wallace's most inter- esting volume, Darwinism, has appeared. A short account of Variable Protective Resemblance in insects will be found in it. CHAPTEE X WARNING COLOURS WE now come to a class of colours with a meaning precisely opposite to that of the large class we have just been considering. The object of the latter is to conceal the possessor from its enemies, the object of the former is to render it as conspicuous as possible. As in other classes of colour, the most familiar and striking illustrations are to be found among insects.1 The sharp contrast between most Protective or Aggres- sive Resemblances and Warning Colours It must have been obvious to any one interested in natural history that the insects met with during a walk in summer may be arranged in two great groups : those which are extremely difficult to find and excite our wonder by the perfect manner in which they are concealed, and those which at once attract our attention by their startling colours and conspicuous attitudes, the effect being often greatly increased by the habit 1 Many of the facts and conclusions in this chapter are taken from my paper in the Proc. Zool. Soc. 1887, p. 191. 160 THE COLOURS OF ANIMALS of living in companies. These two groups form, perhaps, the sharpest contrast in nature. We assume, almost as a matter of course, that the latter are protected in some other way, that if captured they would prove to be of little value, or even posi- tively nauseous or dangerous. The value of Warning Colours At first sight the existence of this group seems to be a difficulty in the way of the general applicability of the theory of natural selection. Warning Colours appear to benefit the would-be enemies rather than the conspicuous forms themselves, and the origin and growth of a character intended solely for the advan- tage of some other species cannot be explained by the theory of natural selection. But the conspicuous animal is greatly benefited by its Warning Colours. If it resembled its surroundings like the members of the other class, it would be liable to a great deal of accidental or experimental tasting, and there would be nothing about it to impress the memory of an enemy, and thus to prevent the continual destruction of individuals. The object of Warning Colours is to assist the education of enemies, enabling them to easily learn and remember the animals which are to be avoided. The great advantage conferred upon the conspicuous species is obvious when it is remembered that such an easy and successful education means an WARNING COLOUES 161 education involving only a small sacrifice of life. It must not be supposed that nauseous properties are necessarily attended by Warning Colours ; there are very many instances in which they are accompanied by Protective Kesemblances and habits. The common cockroach is a familiar example of this latter asso- ciation. Warning Colours in Mammalia The highest vertebrate animals are rarely protected by the possession of the qualities which-are most com- monly attended by Warning Colours, viz. an unplea- sant taste or smell. There is, however, at least one mammal of which this is certainly true. This ex- ample is brought forward hi Belt's most interesting book, ' The Naturalist in Nicaragua.' l Thus he tells us that at night ' the skunk goes leisurely along, holding up his white tail as a danger- flag for none to come within range of his nauseous artillery.' He also alludes to the fetid fluid which these animals ' discharge with too sure an aim at any assailant.' He describes the large white tail as laid over against the black and white body, producing a very conspicuous effect in the dusk, so that the animal ' is not likely to be pounced upon by any of the Carnivora, mistaking it for other night-roaming animals.' The conspicuous appearance of the skunk is shown in fig. 42. 1 Second edition, 1888, pp. 174, 249, 250, 320, 321. See also Mr. A. E. Wallace's Darwinism, 1st edition, p. 233. 162 THE COLOURS OF ANIMALS I know of no instance of this kind among birds, but it is probable that the gaudy and strongly-contrasted colours of certain tropical species may be found to be accompanied by some nauseous property and to be of warning significance. RG. 42. — The Brazilian Skunk (Afephiti* suffocant) : showing the conspicuous black and white appearance of the animal which serves as a warning to its enemies. The brilliant and conspicuous colours of many powerful birds are, I think, to be explained as a result of the free scope given to sexual selection (see pp. 311-12). WAENING- COLOURS 163 Warning Colours in Reptiles Warning characters are not uncommon among poisonous reptiles. The various species of Coral Snake (Elaps), occurring in tropical America, are ex- tremely venomous, and are highly conspicuous, their bodies being alternately banded with bright red and black, and often with yellow.1 It is extremely in- teresting to observe that the deadly Eattlesnake (Crotalus) warns an intruder of its presence by sound instead of by sight, like the Coral Snake. The Cobra is protectively coloured, but, if attacked, it expands the hood with the conspicuous eye-like marks, and thus endeavours to terrify its enemy by the startling appearance. The majority of poisonous snakes, how- ever, depend entirely upon Protective Resemblance together with the use of their fangs. This, for ex- ample, is the case with our common Viper. It is, however, an advantage to some snakes to acquire warning characters and to live on their repu- tation for being poisonous; for although an animal bitten by one of them would probably die, the effects are never immediately fatal, and there would be plenty of time for the snake itself to be killed. Again, the snake possesses only a limited supply of poison at any one time, and if this had been recently drawn upon 1 See also A. B. Wallace's Essays on Natural Selection, 1875, p. 101. 164 THE COLOUES OF ANIMALS for purposes of defence or for killing prey, the snake would be comparatively harmless. Hence it .would be to the advantage of certain snakes to advertise publicly the fact that they are dangerous, retaining the poison to use if necessary ; and others would gain by concealing themselves by Protective Eesemblance, while they also would use their poison fangs if detected and attacked. The question is not whether one of these methods is better than the other, but whether either of them is better than an intermediate con- dition ; so that we can well understand why one group of poisonous snakes should adopt one method, while the other method is made use of by another group. Warning Colours in Amphibia Among the Amphibia a beautiful example has been afforded by Mr. Belt's acute powers of observation.1 ' In the woods around Santo Domingo there are many frogs. Some are green or brown, and imitate green or dead leaves, and live amongst foliage. Others are dull earth-coloured, and hide in holes and under logs. All these come out only at night to feed, and they are all preyed upon by snakes and birds. In contrast with these obscurely coloured species another little" frog hops about in the daytime, dressed in a bright livery of red and blue. He cannot be mistaken for any other, and his flaming vest and blue stockings ' Loc. tit. p. 32L WARNING COLOURS 165 show that he does not court concealment. He is very abundant in the damp woods, and I was convinced he was uneatable so soon as I made his acquaintance and saw the happy sense of security with which he hopped about. I took a few specimens home with me and tried my fowls and ducks with them, but none would touch them. At last, by throwing down pieces of meat, for which there was a great competition amongst them, I managed to entice a young duck into snatching up one of the little frogs. Instead of swal- lowing it, however, it instantly threw it out of its mouth, and went about jerking its head, as if trying to throw off some unpleasant taste.' It is also ex- tremely probable that the well-known European Sala- mander (Salamandra maculosd), so conspicuous with its irregular yellow blotches on a black ground, pos- sesses some unpleasant attribute. I do not think, however, that there is any direct evidence for this, like that obtained by Mr. Belt in the case of the Nicaraguan frog. Warning Colours in Marine Animals Many fish are poisonous, and many possess for- midable defensive spines, but I do not know that any attempt has been made to connect these characters with a conspicuous appearance. It is very probable, however, that such a connection exists in many cases.1 1 Mr. Garstang suggests that the weever-fish (Trachinus viper a) is an example of Warning Colouration. It possesses a pair of in- 166 THE COLOUES OF ANIMALS "Warning Colours are probably wide-spread among marine organisms. Mr. Garstang had suspected that the bright colours of certain compound Ascidians were of warning significance, because these helpless animals are thus rendered extremely conspicuous, and because some of them emit a most unpleasant odour. He now finds that fish invariably refuse them : although some- times tasted or even swallowed, they are never retained. The bright colours of many sea-antmones and sponges are probably to be explained in the same way. Evi- dence in favour of this conclusion is given on pp. 200-204. ( Warning Colours in Caterpillars : the history of their discovery Warning Colours are greatly developed in insects, and an account of the first recognition of this prin- ciple among caterpillars is of great historical interest. When Darwin was investigating the bright colours of animals, and was elaborating his theory of their ex- planation as of use in courtship, he came across the brilliant colours of certain caterpillars, and saw at tensely poisonous spines on its gill-covers, and is rendered conspicuous by a deep black first dorsal fin. The body of the fish is completely buried in the sand, which it resembles in colour, the black fin alone being seen. Mr. Garstang thinks that this conspicuous character prevents such fish as gurnards from mistaking the weever for the dragonet (Callionymus lyra), which is similar in size and habits. He has frequently found the dragonet in the stomachs of gurnards, but the weever never. WARNING- COLOURS 167 once that they were a difficulty in the way of the theory. For caterpillars are undeveloped organisms ; they have been described as ' embryos leading an in- dependent life,' and there is no way of distinguishing the sexes by external colour or structure (except in a few instances). Here, therefore, we meet with bril- liant colours, often rendering the possessors con- spicuous, which cannot be of any use in courtship. Seeing, therefore, that the bright colours must be of use in some other way, Darwin drew the attention of Wallace to the subject, and asked whether he could suggest any explanation. Wallace accordingly thought over the subject, and considered it as part of the wider question of the varied uses (other than sexual) of brilliant and startling colour, in other stages of insect-life, and in numerous instances scattered over the whole animal kingdom ; and he finally ventured to predict that birds and other enemies would be found to refuse such conspicuous caterpillars if offered to them. He believed, in fact, that such larvae are pro- tected by possessing a nauseous taste or smell, or some other property which renders them unfit for food. Conversely Wallace argued that inconspicuous cater- pillars would be eaten and relished whenever they were detected. It is most inspiring to read the letter in which the great founder of modern biology accepted this fruitful suggestion. ' . . . You are the man to apply to in a difficulty. 168 THE COLOUKS OF ANIMALS I never heard anything more ingenious than your suggestion, and I hope you may be able to prove it true. That is a splendid fact about the white moths ; it warms one's very blood to see a theory thus almost proved to be true.'1 Very soon after the suggestion was made public 2 it received confirmation by experiments conducted by Mr. J. Jenner Weir 3 and Mr. A. G. Butler.4 At a later date experiments of the same kind were made by Professor Weismann,5 and still later by myself.6 It was found that while birds devoured with eagerness the well-concealed caterpillars, they refused those with conspicuous colours; it was also found that other insect-eating animals, such as frogs, lizards, and spiders, refused larvae with warning colours, or did so after first tasting them. Examples of Warning Colours among Caterpillars A very common example of a caterpillar with warning colours is afforded by the larva of the Cur- rant Moth or Magpie Moth (Abraxas grossulariata) , which is excessively abundant in gardens (see fig. 43) 1 Life and Letters of Charles Darwin, 1887, vol. iii. p. 94. 2 Proc. Ent. Soc. Lond. Ser. 3, v. p. Ixxx. 1867. » Trans. Ent. Soc. Lond. 1869, Part i. April. * Ibid. p. 27. • Studies in the Theory of Descent, Part ii. pp. 336-340. English translation by Professor B. Meldola. 6 Proc. Zool. Soc. 1887, p. 191. This paper contains an account of all previous work on the same subject. WAENING COLOURS 169 The caterpillar is extremely conspicuous, being of a cream colour with orange and black markings. Al- though it belongs to the group of well- concealed ' stick-caterpillars ' (Geome- tm), of which several instances have been considered in Chapter III., it makes no attempt to hold itself in any of the attitudes characteristic of its group (compare fig. 43 with figs. 1, 2, 3, 4, 6, 8, and 9). All observers agree that birds, lizards, frogs, and spiders either refuse this species altogether, or exhibit signs of the most intense disgust after tasting it. FIG. 43. — The larva of Magpie Moth (A. grosnulariata), showing Warning Colouring ; full- led ; natural size. PIG . 44.— The larva of Buff-tip Moth (P. Bucephala ), showing Warning Colouring; full-fed ; natural size ; from Curtis. FIG. 45.— The larva of Cinnabar Moth (E. Jaeobcece), showing Warning Colouring ; full-fed; natural size ; from Curtis. 170 THE COLOURS OF ANIMALS The caterpillar of the Buff-tip Moth (Pygaera bu- cephala), fig. 44, and the Cinnabar Moth (Euchelia jacobaa), fig. 45, are also extremely abundant, and are good examples of the association of Warning Colours with a nauseous taste. Both of them are gregarious, living in large companies, so that their conspicuous appearance is greatly intensified. The colours of the first-named larva are black, yellow, and orange. It feeds on oak, elm, lime, birch, hazel, &c., and the large bare branches which attest its appetite are very familiar sights in autumn. The second caterpillar is coloured by alternate black and yellow rings ; it feeds upon ragwort in the summer. There is plenty of experimental evidence for the unpleasant taste of both caterpillars. The conspicuous gregarious caterpillars of the Large Garden White Butterfly (Pieris brassicce), which are only too well known in cabbage gardens in the autumn, are also protected in the same manner. Many other instances will be found in the papers already referred to. A caterpillar may be freely exposed rather than conspicuous In some cases the warning of an unpleasant quality is conveyed by the caterpillar being freely exposed, while its colours, although sober, do not harmonise with those of the food-plant. This may WARNING COLOURS 171 be true of gregarious species, such as the dark larvae of the Peacock or Small Tortoiseshell butterflies, which feed freely exposed on the tops of nettles, and which are known to be refused by some insect-eating animals.1 The various unpleasant qualities possessed by caterpillars with Warning Colours Other unpleasant attributes, as well as that of a nauseous taste, may be associated with Warning Colours. A strongly smelling or irritant fluid may be discharged from special glands on the approach of an enemy. Glands of this kind occur on the back of many common caterpillars, such as the brilliantly coloured ' Palmer worm ' (larva of Porthesia auriflua), or the onspicuous ' Hop-dog ' (larva of Orgyia pudilunda). The larvae of some common gregarious saw-flies (Hymenoptera), such as Croesus septentrionalis, which completely denudes the branches of birch trees, have a number of odoriferous glands along the middle of the ventral surface. When disturbed, the body is turned forward over the head, and the glands are everted so that their secretion escapes into the air. The meaning of the gregarious habit is very 1 The gregarious habit may render an insect so conspicuous that it is unnecessary for it to acquire bright colours. The ' warning ' significance of the gregarious habit was first suggested by Fritz Miiller (Kosmos, Dec. 1877). An abstract of this paper has been published by Professor Meldola (Proc. Ent. Soc. Lond. 1878, pp. vi. and vii.) 172 THE COLOURS OF ANIMALS clear in this and parallel cases ; for when many indi- viduals combine to discharge an unpleasant odour, they become surrounded by an atmosphere which acts as a most effective barrier. Irritating hairs possessed by certain larvae Again, caterpillars may be protected by possessing irritating hairs. This is the" case with the ' Palmer worms ' mentioned above, which are thus doubly pro- tected. Many people have discovered this fact to their cost after handling these pretty black, red, and white caterpillars, which are so abundant and freely exposed on our hawthorn hedges in early summer. When the face or neck is touched by the hands, which are covered with minute barbed hairs shed by the caterpillar, an intensely irritating rash soon makes its appearance. The same effect is produced, as I shall always remember, if an old cocoon, in which the hairs are interwoven, be pulled to pieces with the fingers. These caterpillars were nearly always refused, but Mr. Butler records that they were in one case eaten with- out hesitation by a young sky-lark, which soon after- wards died with symptoms which may have been due to the irritating hairs. . One of my lizards also seized a larva, but relinquished it after biting it for sometime. The lizard was evidently greatly irritated by the hairs in its mouth. Many other hairy caterpillars also produce a rash : thus, the larvae of the Fox Moth WARNING COLOURS 173 (Lasiocampa rw&t),0ak Eggar (L. quercus), and Drinker (Odonestis potatorid), have this effect on the skin of the hands if they are held for a long time, and they would certainly act rapidly upon the delicate skin of the mouth. All thre.e caterpillars are fairly con- spicuous, and there is experimental evidence that the two latter are disliked. It will be shown that the hairs are sometimes arranged in tempting tufts, which invite an enemy to seize the caterpillar at a point which does not injure the latter, while it causes the former the greatest dis- comfort. The hairs of nearly all caterpillars are probably more or less unpleasant in the mouth. Delicate and sensitive animals, such as the marmoset, although ex- cessively fond of insects, cannot be induced to touch any hairy larva. Birds appear to eat them more readily than other animals, but they have peculiar advantages in their power of rubbing off the hairs. The association of hairs with a conspicuous appearance Sir John Lubbock l has tabulated the appearance of the larvae of all British butterflies and the larger moths, and he thus shows in a most convincing manner the general association of hairs or spines with conspicuous warning colours. His conclusion is as follows : ' Thus summing up the caterpillars, both of 1 Trails. Ent. Soc. 1878, pp. 239, et sej. 9 174 THE COLOURS OF ANIMALS the butterflies and moths, out of the eighty-eight spiny and hairy species tabulated only one is green (L. syUlla), and even this may not be protectively coloured, since it has yellow warts and white lateral lines. On the other hand, a very great majority of the "alack and brown caterpillars, as well as those more or less marked with blue and red, are either hairy or spiny, or have some special protection.' ) Sir John Lubbock, however, fully recognises that hairs may contribute towards tb£ Protective Eesem- blance of certain species, examples of which have been already given (see page 35). Professor Meldola suggests that a probable original meaning of the hairy covering was protection from injury after falling from the food-plant. Warning Colours in other stages of metamorphosis in Lepidoptera Lepidoptera of many species are protected by Warning Colours and unpleasant attributes, in other stages in addition to that of the larva ; and the same method of defence is also adopted in other orders of insects. The chrysalis of the Magpie Moth, which is black with yellow bands, and exposed to view in a very slight cocoon, is nauseous like the larva, and the slow- flying moth itself, with white wings rendered conspicu- ous by yellow markings and black spots, is defended in the same manner. When captured it makes no attempt WARNING COLOURS 175 to escape, but ' feigns death.' The conspicuous and sluggish day -flying black and red Burnet Moths (Zygcena) and Cinnabar Moth (Euchelia Jacobcece) are also nauseous, and so is the gaudy Garden Tiger Moth (Arctia caja). Many white moths, or black and white moths, have also been refused by insect-eating animals with every sign of disgust. Consideration of the later stages of species with unpalatable larv» A comparison of the means of defence and palata- bility in the three stages of metamorphosis, in species of which the larvae are known to be nauseous, proved to be extremely interesting, and much more work is needed in the same direction. In the first place the com- parison showed that when the later stages are nauseous the larva was also nauseous in all cases. The Tiger Moth is probably an exception, for the caterpillar may be defended by its hairs instead of by taste, and the chrysalis seems to be palatable. The Leopard Moth (Zeuzera cesculi] is another exception. Such cases are probably very rare, and it is clear that this method of defence, among Lepidoptera, nearly always arose in the larval stage. The larval stage is exposed to more danger and is more helpless than any other : the imago can escape by flight, and the pupa, if exposed, may render its Protective Resemblance complete by entire quiescence, and it is usually effectually protected 176 THE COLOURS OF ANIMALS in other ways. But the larva must feed, and at the same time is sluggish in its movements, defenceless, and when palatable is more relished than any other stage, for it does not possess the hard investment of the one or the scaly covering of the other. Hence it is that the great needs of the larva have been so fre- quently met in this way; but as soon as the un- pleasant quality has appeared it will tend to pass on by simple continuity into the other stages. If these latter are hard pressed, there is always the possibility that such qualities may be made the starting-point of a similar method of defence for them also. But the disagreeable properties may also pass on into stages which hold their own successfully by elaborate and perfect Protective Eesemblances, and then such quali- ties, unattended by Warning Colours, are entirely use- less to the stage, but may be important as a latent possibility for the future. It must be remembered that an unpleasant attribute must always appear in advance of the warning colouring. An example is afforded by the Buff-tip Moth (Pygcera Imcephala), which is beautifully protected, during rest, by resem- bling a piece of rotten lichen -covered stick (see page 57), but which nevertheless retains something of the unpleasant taste by which its caterpillar is effectually protected. WARNING COLOUKS 177 The metallic appearance of certain pupae may be of value as a warning At this point it is of interest to consider the cases in which the metallic appearance of a chrysalis may act as a warning. Dr. Fritz Miiller tells me that the brilliant metallic pupae of the South American butter- fly, Mechanitis lysimnia, hang in groups from the leaves of their food-plant (Solanum). The butterfly of this species is certainly distasteful, for the genus is mimicked by butterflies of other families. This fact, and the gregarious habit of the pupae, render it nearly certain that the glittering appearance has a warning significance. The same is probably true of the pupa of the abundant Indian butterfly (Euplcsa core), which Mr. E. A. Minchin tells me possesses a brilliant silvery appearance, and is so conspicuous that it can be seen from a great distance. This butterfly also belongs to a group protected by an unpleasant taste or smell, and there is little doubt that the metallic appearance of the pupa has a warning meaning. Warning Colours in other orders of Insects Passing now to the other orders of insects, highly conspicuous and abundant beetles (Coleoptera), such as the black and red ' soldiers and sailors ' (Tele- phorus], the red and -black ladybirds (Coccinella), and the red and blue-black Chrysomela populi, have been 178 THE COLOURS OF ANIMALS shown to be extremely nauseous, and the two latter emit a very unpleasant smell. The sting possessed by the females of so many Hymenoptcra is obviously an unpleasant attribute, . rendering the insect disagreeable or even dangerous to eat. We find accordingly that stinging insects are often rendered conspicuous by warning colours, of which the contrasted dark and yellow bands of the Common Wasp, the Hornet, and of many Humble Bees, furnish examples. Warning Colours are also to be found in other orders, but it is unnecessary to give further examples. They will be recognised in numbers in any country walk during the summer, although the experimental proof of the co-existence of some unpleasant attribute is still wanting in a large proportion of the cases. Warning Colours can only be safely adopted by a small proportion of the Insects in any country The acquisition of an unpleasant taste or smell, together with a conspicuous appearance, is so simple a mode of protection, and yet apparently so absolutely complete, that it seems remarkable that more species have not availed themselves of it. What can be the principle which works in antagonism to such a mode of protection? Thinking over this subject, as the result of a lecture upon the facts and conclusions already described,1 it appeared probable that such an 1 Delivered at the Royal Institution in the spring of 1886. WARNING COLOUES 179 antagonistic principle would be found in the too com- plete success of the method itself. If a very common insect, forming the chief food of some animal, gained protection in this way, the latter might be forced to devour the unpalatable food in order to avoid starva- tion. And the same result might readily be brought about if a scarce and hard-pressed form adopted the same line, and became dominant, after ousting many species which were important as food. If once an insect- eating species were driven to eat any such insect in spite of the unpleasant taste, it would gradu- ally come to devour it with relish, and the insect would be in great danger of extermination, because of its conspicuous appearance. If this reasoning be correct, it is clear that the mode of defence is by no means perfect, and that it depends for its success upon the existence of relatively abundant palatable forms ; in other words, its employ- ment must be strictlv limited. Absence of Warning Colours in the seasons when Insect life is scarce A very interesting fact in support of this argument is the entire disappearance of all insects with Warning Colours during the seasons when insect life is scarce, and when insect-eating animals are hard pressed for food. And yet, if it were safe to rely on such a mode of defence, the Warning Colours would be especially 180 THE COLOURS OF ANIMALS conspicuous at these times, when all the tints of nature are sombre and form a background against which the Warning Colours would stand out in startling contrast. Certain species, which are defended in this way, pass the winter in the brightly coloured stage of metamor- phosis ; but they conceal themselves as completely as possible under loose bark or among dead leaves, &c. This is true of the common ladybird, and I have noticed that they begin to hide comparatively early in the autumn, when the insects are rapidly diminishing in numbers, but before the beginning of the cold weather. It is therefore probable that they hide in order to conceal their bright colours and not to escape the cold.- It is also known that ladybirds are eaten by the Green Tree frog in winter, when other insect food is scarce, and also by hungry birds, although they are intensely disliked and are refused (at any rate by the frogs) if other food can be obtained. Experimental proof that Insect-eating animals, if hungry, will eat unpalatable species This conclusion was tested as completely as pos- sible by offering conspicuous unpalatable insects of many species to animals from which all other food was withheld. Under these circumstances the insects were eaten, although often after many attempts, and evidently with the most intense disgust. WARNING COLOUES 181 Natural selection has enabled certain animals to eat unpalatable Insects with apparent relish Naturalists have always recognised that an insect may be distasteful to one animal, but palatable to another. It is, however, very probable that these differences have been acquired comparatively recently, and have arisen out of the competition for food. In most cases the change of habit has not become so far confirmed that the previously distasteful food is eaten with avidity and pleasure. Again, when we find that the taste of an insect is recognised as nauseous by a standard wide enough to include mites and spiders as well as birds, lizards, and frogs, it appears probable that any difference of opinion is due to an altogether exceptional immunity conferred upon certain species by natural selection. Nauseous qualities probably do not affect Insect parasites It is probable, however, that this argument does not apply to insect parasites, which are not in the position to gratify their tastes, but must make the best of the larva in which the parent has deposited her eggs. It is clear that even the most nauseous forms must suffer greatly from the attacks of enemies, for the average number of individuals in each species appears to remain constant. It is likely that the 182 THE COLOURS OF ANIMALS numbers are kept down by special liability to the attacks of insect parasites, in one or more stages. Thus the larva of the Large Garden White (Pieris brassicte) is known to be nauseous, but the im- munity from attack which it enjoys by no means extends to its insect foes. In the autumn of 1888 I collected some hundreds of these larvae in order to experiment upon the colours of their pupaa. I ob- tained 109 pupae, while 424 mature larvae died from the presence of the parasitic grubs of Ichneumon flies The likes and dislikes of Insect-eating animals are purely relative It may be taken as proved that the continued spread of some distasteful form and the correspond- ing diminution of edible species would lead to the former becoming the prey of insect-eating animals ; for a point would ultimately be reached, as it was reached in many of my experiments, when hunger would become a stronger stimulus than those lesser prejudices in which a species can very well afford to indulge while palatable food is abundant. These pre- judices having been overcome in confinement, there is nothing in the conditions of natural life which could prevent the same result from being reached, as doubt- less it has been reached again and again. The com- parison of all experiments of this kind ever made with insects will show that the likes and dislikes of insect- WARNING COLOURS 183 eating animals are purely relative, and are manifested to a marked extent when they are offered a variety of insects, and when obviously nauseous species are excluded from the list. Butterflies and moths are freely eaten by lizards, but they are not enjoyed like houseflies or many cater- pillars. This is probably because the former are such dusty and unsatisfactory things to eat, with such a small proportion of body in which the nutriment and taste is contained, and so large an expanse of dry membranous wings with their scaly covering. In this respect butterflies contrast unfavourably with moths, and the latter are certainly greatly preferred by lizards and especially birds. The latter have special advantages in being able to pick off the wings before eating the body. In the excessive abundance of insect-eating ani- mals and the keen competition for food which takes place, we see the conditions which must render the acquisition of an unpleasant taste together with Warn- ing Colours an exceedingly hazardous mode of defence, if assumed by more than a small proportion of the insects of a country. For in so great a press of com- petition among the innumerable insect-eaters, we may feel sure that some at least would be sufficiently enterprising to make the best of food which at least has the advantage of being easily seen and caught. The great principle of Warning Colours has de- servedly taken a most important place among the 184 THE COLOUES OF ANIMALS principles which deal with the infinitely complex and ever-changing relations which obtain between the most widely separated, no less than between the most closely allied, members of the organic kingdom. But, nevertheless, this principle carries with it its own compensating principle, which will come into opera- tion precisely as the former advances to the possession of undue influence. The education of enemies assisted by the fact that Warning Colours and patterns often resemble each other It is probably unnecessary for the young insect- eating animal actually to make trial of every species of nauseous insect in its locality, in order to be equipped with an efficient stock of experiences with which to conduct its later life. Such an education would be somewhat dearly bought ; it would be un- pleasant to the insect-eater and destructive to the insect. Since, however, the same colours are em- ployed again and again by unpalatable or dangerous insects of very different groups, and since the patterns are also frequently repeated, it is obvious that a com- paratively few- unpleasant experiences would be suffi- cient to create a prejudice against any insect with a colour or pattern at all resembling the nauseous forms which had already produced BO deep an impression upon the memory. This conclusion was drawn from the careful com- parison of the colours and patterns of all insects which WAKNING COLOURS 185 have been experimentally shown to be distasteful. The colours which produce the greatest contrast, and therefore the greatest effect, upon the eye of an insect- eating Vertebrate, are black and white, and next to this black (or some very dark colour) and yellow, orange, or red ; and it was found that nearly all unpalatable or dangerous insects were coloured with these tints. The advantage gained by the acquisition of such colours is twofold : they afford the combina- tions which are most conspicuous to an enemy, and their number being small, while nauseous forms are numerous, the continual repetition of the same com- bination becomes a necessity, and this also facilitates the education of enemies. There are similarly a few eminently conspicuous and simple patterns. These are — alternating rings of different colours, and alternating longitudinal stripes, both especially suited to the cylindrical body-form, such as that of caterpillars, &'c. ; spots upon a back- ground with a contrasted colour, especially suited to a wide expanse such as the wings of Lepidoptera. Every insect which has been proved to be distasteful was found to possess one or other of the above patterns or some combination of them. Here also the frequent similarity of the patterns of nauseous insects is primarily due to the fact that there is only a limited number of appropriate patterns, but also because the repetition is in itself advantageous and has therefore been encouraged by natural selection. 186 THE COLOURS OF ANIMALS Some of the advantages of true mimicry (to be more fully described below) also follow, when a group of insects is rendered conspicuous by the same colours and patterns, and when certain members of the group are noted for the possession of especially unpleasant attributes. Thus it is more than probable that the species marked by alternate rings of black and yellow (including the chrysalis of the Magpie Moth and the caterpillar of the Cinnabar Moth), gain considerable advantages from the justly respected appearance of Hornets and "Wasps. It must not be forgotten, however, that the latter also are probably benefited, although to a much smaller extent, by the greater publicity which follows from the resemblance. The causes which have determined the resemblance between Warning Colours in different Insects Hence the causes which determine the frequent repetition of the same colours and markings in dis- tasteful forms are as follows: (1) The fact that a limited number of colours and patterns are especially efficient in attracting the attention of enemies, and in thus facilitating their education; (2) the fact that the education of enemies is also rendered easy by requiring them to learn only a small number of pat- terns and colours ; (3) the great additional advantage conferred by trading upon the reputation of a well- known and much-feared or much -disliked insect. WAENING COLOURS 187 The warning appearance acquired by any insect is also largely determined by the character of its previous appearance, which formed the material upon which at any rate the first steps of the change were built. In some cases we can successfully read the history of past changes, and can point to certain parts of a warning appearance which are remnants of a previous mode of defence by means of Protective Eesemblance. Thus the orange rings of the caterpillar of the Cinnabar Moth harmonise well with the flowers of its food - plant, ragwort.1 The acquisition, or perhaps only the greater prominence, of the strongly contrasted black bands, and above all, the gregarious habits, are the later developments which have followed the acquisition of an unpleasant taste. Again, the caterpillars of the Mullein Moth (Cucullia verbasci), which are so abundant and con- spicuous on various species of Mullein (Verbascum) , are even now difficult to detect when resting among the dark and yellow sessile flowers studded upon the surface of the 1 First noticed by T. W. Wood : Insects in Disguise, Student. isea FIG. 46.— Two larvae of Mullein Moth (Cucullia verbasci) on the spike of the Mullein ; small in last stage ; natural size. 188 THE COLOUES OF AOTVLAL9 thick green spike (see fig. 46). ! The conspicuous appearance chiefly depends upon the gregarious habits, and upon the fact that the larvse com- monly rest on the upper surface of the large leaves, which form a background against which the larval PIG. 47.— Larva of Mullein Moth on leaf of Mullein ; full-fed ; natural siza colours stand out with startling distinctness (see fig. 47). It has been proved that these larvae possess an unpleasant although not an extremely nauseous taste, so that here also we have evidence that the change from a palatable well-concealed form is only recent and is as yet incomplete. * This observation w&s ^o^^p^.tiDicfttod to Professor ^kltddolci by Mr. Thomas Eedle. CHAPTEE XI WARNING COLOURS (continued) Sexual colouring may be made use of for warning purposes In addition to the modes of producing a warning appearance which we have hitherto considered, and which are almost universal in this country, there is another method which is very conspicuous in the tropics. In certain groups of mature insects, and especially in butterflies, the beautiful colours and patterns which have been produced by courtship, appear to have been made use of as an indication of some unpleasant quality. The differences between Sexual and Warning Colours The tints used in and produced by courtship are as a rule easily distinguished from Warning Colours, even when both occur in sexually mature insects. The former rarely usurp the whole surface of an in- sect, and they are carefully concealed during repose. Thus the upper sides of the upper wings of most moths, and the under sides of both wings in butter- 190 THE COLOUES OF ANIMALS flies, are generally protectively coloured, and hide the bright colours of other parts when the insect is at rest. If the parts exposed during rest are conspicu- ously coloured it is clear that they chiefly possess a warning significance. I say ' chiefly,' because it is probable that the appearance of the mature individuals of any species, however much it may be specialised for other ends, possesses a sexual significance, and appeals as an adornment to the modified taste of the individuals of the same species. We have a rough criterion of the extent to which the taste has been modified when we compare the appearances which have other addi- tional meanings with those which possess a sexual value alone, and which are concealed except during flight and are especially displayed in courtship. Warning Colours are also displayed during the slug- gish flight of a nauseous species, but the insects with purely ornamental colours are swift and wary when upon the wing. But quite apart from these considerations, the Warning Colours can be distinguished by the subordi- nation of every other feature to that of conspicuous- ness. Crude patterns and startling strongly contrasted colours are eminently characteristic of a warning appearance, while the colours and patterns produced by courtship include everything that is most beautiful in insects. The two kinds of appearance differ as an advertisement differs from a beautiful picture: the one attracts attention, the other excites admiration. WARNING COLOURS 191 The transition from Sexual to Warning Colours The two groups nevertheless run into each other, and a beautiful transition is afforded by the insects in which sexually produced colours and patterns are made use of for warning purposes. When this is the case the colours spread on to the parts which are ex- posed during rest, and the flight becomes sluggish, so that they are displayed as completely as possible. These are the insects which are the principal models of mimicry in tropical countries, and Bates's classical paper, in which an intelligible theory of mimicry was first brought forward, deals with the groups which are found in the Amazon valley, and with the forms which resemble them and share the advantages conferred by their well-known and nauseous qualities. The evi- dence for the existence of such qualities is better considered under the next heading, viz. Mimicry. Resemblance between such Warning Colours in different species The members of each of these groups resemble one another to a marked extent ; far more so than the species of other groups without Warning Colours. Thus the advantage of facilitating the education of enemies is gained by them, although it has arisen in a manner different from that already described in other unpalatable insects (see pp. 184-86). 192 THE COLOUKS OF ANIMALS The similarity has arisen from the fact that the species in each group are closely related, so that natural selection has maintained an initial resem- blance, instead of causing convergence, as it has done with more distantly related species. Hence repetition of the same appearance may be produced by a pre- vented divergence, as in these cases, or by the actual convergence of forms originally unlike, as in the former cases. The convergent forms are more perfectly con- spicuous, more ideally warning, because they have been further modified from their original appearance ; while the forms in which divergence has been arrested have merely adopted, with comparatively slight modi- fication, an appearance which was produced by the operation of other principles, but which is sufficiently well known for the purpose. These interesting conclusions have gradually grown out of the observations of many naturalists. The arrested divergence, sometimes aided by actual convergence, has produced such remarkable resem- blances between certain species of unpalatable insects, that Bates speaks of the wonderful fact that such species mimic each other. Wallace at first looked upon these mysterious resemblances as due to some unknown cause connected with locality, for the similar species are nearly always found together. The difficulty was at length explained by Fritz WARNING COLOURS 193 Miiller.1 This eminent naturalist suggested that both species were benefited by the resemblance, because the number of individuals which must be sacrificed to the inexperience of young birds and other enemies would be made up by both of them instead of by each independently. This fruitful suggestion was at once accepted by Wallace.2 The mathematical aspects of the subject were accurately worked out by Mr. Blakis- ton and Mr. Alexander, of Tokio, Japan.3 The next step was taken by Professor Meldola,4 who extended Fritz Miiller's explanation of these comparatively rare cases of close resemblance, to the general similarity which obtains throughout whole groups of unpalatable and conspicuous species. ' The prevalence of one type of marking and colouring throughout immense numbers of species in protected groups, such as the tawny species of Danais, the barred Heliconias, the blue-black Euplceas, and the fibrous Acrceas, is perfectly intelligible in the light of the new hypothesis.' This list comprises the whole of the large groups of butterflies alluded to in the last few pages. The species belonging to them are very familiar in every collection of tropical butterflies, while some of them are even abundant in temperate climates. Until re- 1 Ituna and Thyridia, Kosmos, May 1879, p. 100, translated by Meldola, Proc. Ent. Soc. Lond. 1879, p. xx. 2 Nature, vol. xxvi. p. 86. * Ibid., vol. xxix. p. 405. * Ann. and Mag. Nat. Hist. December 1882. 194 THE COLOUES OF ANIMALS cently, the British butterflies did not include an ex- ample, but a large and handsome American Danaid (Danais archippus) seems to be gradually extending its range into every country where the food-pla.nts (Asclepiads) of its larva are to be found. Several individuals have been caught in this country of late FIG. 48.— The North American Dnnais arch'ppus, which has now spread into this country ; upper side ; iiiUI natural size. Pl&. 49. — Danriis archippus, showing the conspicuous colours on the under sides of the wings. years, and there is no doubt that it will thoroughly establish itself if it can meet with a sufficient supply of larval food, and can withstand the ceaseless energy of collectors. It is far larger than any of our native butterflies. It is shown half the natural size in fig. 48, while fig. 49 gives the appearance of the under side. The latter figure shows that the insect WAKNING COLOUES 195 must be as conspicuous at rest as it is on the wing, a fact which is characteristic of those groups of butterflies which are specially defended by being un- palatable. In North America Danais archippus is mimicked by Limenitis misippus, a butterfly belonging to a very different group. Although the general resemblance between the species in each of these groups is doubtless due to arrested divergence, there is one very interesting case which is probably to be explained by convergence of groups which were formerly unlike. The Danaids, which are found in the same localities as the Heli- conias of tropical America, have taken the peculiar appearance of the latter, in the arrangement of the colours, and in the long narrow form of the wings. These Heliconoid Danaids are therefore distinguished from all the other members of the group. It is quite obvious that both Heliconian and Danaids are benefited by the fact that the insect-eating animals of the region they inhabit have to learn but a single mode of flight, shape of wing, and general arrangement of colours. Although these resemblances, produced by conver- gence or by arrested divergence, are transitional into and often contain an element of true Mimicry, they must be distinguished from the latter. In true Mimicry, the mimicking species are without unplea- sant attributes, and are sheltered under the reputation of abundant and well-known forms in which such attributes are strongly marked. In the resemblances 196 THE COLOUES OF ANIMALS considered here, all the similar forms possess unplea- sant attributes although they may possess them in different degrees. The remarkable likeness between many of the species of Burnet Moths (Zygaenidae) is probably due to arrested divergence. They are all conspicuous black and red moths, and some of them are known to be nauseous. The unpleasant qualities may be concentrated in special parts, which are so placed and coloured as to attract the attention of enemies In certain cases the warning appearance is of a different kind. The organism possesses a highly conspicuous feature to which the attention of an enemy is directed ; if seized, the structure breaks off without harm to the animal, but with very unpleasant results to the enemy. It is probable that this method of defence will be found to be wide-spread. The defensive value of ' tussocks ' Only recently, in the summer of 1887, this ex- planation of the beautiful flat-topped tufts of fine hairs, called 'tussocks,' which occur on certain cater- pillars, was shown to me by the results of an experi- ment.1 The tufts are often light-coloured, and are 1 Trans. Ent. Soc. Lond. 1888, pp. 589-91. WARNING COLOURS 197 generally placed on an intensely black ground colour, which shows them up and makes them appear to project more than is actually the case. In the well- known ' H'op-dog ' or ' Tussock ' caterpillar (Orgyia pudibunda) there are four ' tussocks,' each upon a separate ring, and the furrows between these rings are of the most intense velvety black. They are con- cealed until the caterpillar is irritated, when the body is curved in a vertical direction, so that the * tussocks ' diverge, and the furrows appear as black semilunar areas separating them and rendering them con- spicuous. The hairs of the ' tussocks ' are so fine and so closely packed that the tuft does not appear to be made up of hairs at all, but to be rather a fleshy pro- jection from the back of the caterpillar, and a most convenient part for an enemy to seize. Fine as the hairs are, they nevertheless bristle with minute lateral branches, and would certainly be most unpleasant if brought into contact with the skin of the mouth. If seized by an enemy, the fine hairs come out in im- mense numbers, and produce such an effect upon the skin of the mouth that the caterpillar escapes unhurt. The following experiment suggested the explanation which has just been given. A caterpillar of the Common Vapourer Moth (Orgyia antiqua) was intro- duced into a lizard's cage, and when attacked, instantly assumed the defensive attitude, with the head tucked in and the ' tussocks ' separated and rendered as 10 198 THE COLOURS OF ANIMALS prominent as possible. An unwary lizard seized the apparently convenient projection ; most of the ' tus- sock ' came out in its mouth, and the caterpillar was not troubled further. The lizard spent a long and evidently most uncomfortable time in trying to get rid of its mouthful of hairs. On another occasion a full-grown ' Hop-dog ' was offered to a hungry adult Lacerta viridis, but the lizard knew the danger, and kept trying to find some part of the body which could be safely seized. The caterpillar remained motionless in the defensive attitude during the whole attack, which lasted several minutes. In this attitude the ' tussocks ' were held in the most tempting manner, while all other parts of the body bristled with sharp stiff spines. This ex- perienced lizard finally seized the back of the larva a long way behind the ' tussocks,' evidently looking upon the bristles as the lesser evil. Although killed the caterpillar was not swallowed, and it had only been seized after many attempts and the closest examina- tion. It is quite clear that the hairy covering would have saved it from any except a very hungry enemy. Evidence that Insect-eating Animals learn by experience When we compare the behaviour of these two lizards we find strong evidence for the opinion that insect-eating animals learn by experience. I have, however, come across more direct and convincing WAENING COLOUES 199 proofs of this conclusion. Thus, the chamseleon which has been previously referred to had just been imported into this country when I received it, and it had probably never seen a common hive-bee in its life. I put a living bee in the cage, and the lizard immediately began to watch it, and, as soon as it had settled, captured it with a dexterous shot of its long tongue. As the tongue was being withdrawn with the bee adhering to the sticky pad at its extremity, the chamaeleon was stung and immediately showed signs of discomfort, throwing its head from side to side, and thus jerking the bee off. For many months after this I put bees into the cage at irregular intervals; but the chamseleon's education in this direction was complete, the single experience was sufficient, and no other bee was touched. Similar highly-coloured and specially defended features occur in certain Marine Animals A very similar example from an entirely different group of animals has been recently brought forward independently by Professor W. A. Herdman l and Mr. Garstang. These naturalists suggest that the brightly- coloured dorsal papillae of Eolids (Nudibranchiate gastropods) have the same meaning as the ' tussocks ' of Orgyia, being far more conspicuous than the rest of the body, easily detached, and often reproduced by 1 Report of British Association at Neivcastle, 1889. 200 THE COLOURS OF ANIMALS growth in two or three days. The nematocysts at the tips of the papillae would convey a lesson to the enemy similar to that taught by the fine hairs of the 'tussocks.' Mr. Garstang has now tested this suggestion by experiment, and he finds that fish will not attack the Eolids under normal conditions. He therefore threw one of them (the orange variety of Cavolina Farrani) into the tank containing young pollack (Gadus polla- chius), which generally swallow any object while it is descending to the bottom. The Eolid was swallowed and rejected after a second or two by two fish, which then shook their heads as if experiencing discomfort. Similar movements were made when the fish were induced to seize the specially defended tentacles of sea- anemones, and when they attempted to swallow the Poly cirrus, described on p. 201. Mr. Garstang then found that the Eolid causes a distinct, though faint, tingling sensation when placed on the tongue ; while larger species (Facelina coronata and Eolis Alderi) produce much more marked effects. The protectively coloured Opisthobranch, Hermeea (see p. 70), has well-developed defensive papillae, and Mr. Garstang finds that whenever a shadow passes over it, the head is at once retracted, and the papillae rendered very prominent. This behaviour is exactly similar to that described in the larvae of Orgyia (see pp. 197-98). The reaction under the stimulus of light is associated with the unusually large eyes of the genus. WARNING COLOURS 201 Mr. Garstang has still more recently come across an instance of the same kind in a bright red marine worm, one of the Terebellidce (Poly cirrus aurantiacus), which, unlike the rest of its family, has dispensed with the protection of a tube, and creeps about in the crevices of stones and among the roots of Laminaria. It has an immense number of long, slender- tentacles, and when touched, coils itself up in the middle of them. The tentacles break off very easily, and evidently possess some unpleasant attribute. When the animal is irri- tated the tentacles become brilliantly phosphorescent, so that they are conspicuous by night as well as day. Mr. Garstang obtained experimental evidence of the validity of this interpretation. He placed a specimen in one of the fish-tanks in the Plymouth Laboratory : only one pollack ventured to seize the worm, but ejected it immediately, and would not touch it again. Another fish made three vigorous attempts to swallow it, but finally left it. Another, a very voracious rock-fish, ac- tually swallowed it, but immediately afterwards began to work its jaws about as if experiencing discomfort. Mr. Garstang then cut the head and tentacles away from the body and threw both pieces into the pollack tank : the tentacles were untouched, but a fight took place over the body, which was torn into several pieces and swallowed with great relish. Mr. Garstang has kindly allowed me to describe these interesting experiments, which have only just been made, and have not, as yet, been published elsewhere. 202 THE COLOUKS OF ANIMALS Adventitious warning colours Under this head we may include a few very in- teresting cases in which palatable animals make use of others which are specially defended and conspicuous in order to gain protection. Such a method of defence bears the same relation to Warning Colours as the examples of Adventitious Protection and Colouring bear to true Protective Eesemblance (see pp. 76-80). A mollusc which encourages a dense growth of algae upon its shell is defended by Adventitious Pro- tective Eesemblance ; if, however, the algae were brilliantly coloured and nauseous or poisonous, the example would fall under Adventitious Warning Colouration. Professor Eomanes l brings forward examples of a most interesting association of crabs with sea- anemones. He quotes from Mobius 2 the remarkable case of * two crabs belonging to different genera which have the habit of firmly grasping a sea-anemone in each claw, and carrying them about ; ' also from P. H. Gosse 3 the fact that when the sea-anemone (Adamsia palliata) is removed from its position upon the shell of the hermit crab (Pagurus Prideauxii), which in- variably carries it, the crab 'always took it up in 1 Animal Intelligence, International Science Series, pp. 233-34. * Beitrage zur Meeresfauna der Insel Mauritius. • Zoologist, June 1859, pp. 6580-6584. WARNING COLOURS 203 its ckws and held it against the shell " for the space of ten minutes at a time, until fairly attached by a good strong base." ' This fact seems to indicate that the crab detaches and refixes its anemone when it changes its shell in the course of growth. Eomanes, however, quotes from Dr. E. Ball * the statement ' that when the common Sagartia parasitica is attached to a stone, and a hermit crab is placed in its vicinity, the anemone will leave the stone and attach itself to the hermit's shell.' Mr. Garstang tells me that at Plymouth there are two species of hermit crab associated with two distinct species of anemone : the Pagurussmd Actinia mentioned on p. 202, and P. bernhardus, which bears Adamsia Rondeletii. He finds that hermit crabs are eaten with great relish by fish ; they are, in fact, much used as bait by fishermen. Hence the association with the inedible Actinians must be of great service. When the hermit crabs are young and small they are obliged to live in shells without anemones, and Mr. Garstang has often found them, shells and all, in the stomachs of gurnards and other fish. He has never found the larger crabs with shells suited for Actinians in the stomachs of fish. Another hermit crab at Plymouth (Pagurus cua- nensis) is always found in shells covered with a bright orange-red sponge (Suberites domuncula). Mr. Gar- stang finds that sponges are intensely disliked by fish ; 1 Critic, March 24, 1860. 204 THE COLOUES OF ANIMALS the smell alone is generally sufficient to repel them. Many Crustacea are known to live in the canal systems of sponges, and are thus protected. The significance of this association is to be found in the fact that Crustacea are the animals most relished and sought after by fishes, and that sponges are extremely re- pugnant to the latter. Such cases as these are some indication of the severity of the struggle for existence among marine forms of life. Very interesting evidence of this is to be found in Bateson's notes on the protective habits of shrimps and prawns.1 He states that the wrasse will find a shrimp if the least bit be exposed, in spite of its protective colouration. If, however, ' the sand be fine, a shrimp will bury itself absolutely.' We can well understand the immense advantage which would be gained by a much persecuted crustacean if it associated with some animal repugnant to its foes. Colours and markings which direct the attention of an enemy to some non-vital part, but which are not attended by unpleasant qualities From cases such as those which have been just described we pass, by a very natural transition, to colours and markings which attract the attention of an enemy to some non-vital part after the animal 1 Journ. Mar. Biol. Ass., New Series, vol. i. no. 2, Oct. 1889, pp. 211 et seq. WARNING COLOUES 205 has been discovered. The part seized by the enemy breaks away, and thus gives the animal another chance of escape. The cases differ, however, from the pre- ceding ones in that the enemy is in no way injured by its mistake. In correspondence with this difference such features are associated with Protective Kesem- blances leading to concealment, and are not them- selves highly conspicuous, or are only conspicuous when the animal is thoroughly on the alert. The object of these characters is to direct attack to some unimportant part after all other methods of defence have failed, after disguise has been penetrated or speed surpassed. This is probably one of the meanings of the brightly-coloured wings of butterflies, in addition to their more obvious use in courtship. When the insect is flying they form a conspicuous mark easily seized by an enemy, and yet readily tearing without much injury to the insect. On this account we generally find the wings torn and notched when an insect has been long on the wing. In the spring of 1888 I caught a large number of Clouded Yellow Butterflies (Colias edusa) in Madeira. The limited number of species in the island (there are only about a dozen), and the abundance of small omnivorous lizards and of insect-eating birds, lead to the keenest pursuit of the butterflies, and I noticed that the hind wings of a considerable pro- portion of the Clouded Yellows were notched just 206 THE COLOURS OF ANIMALS behind the body. The notches generally corresponded on both hind wings, the insect having been seized at the instant when the wings came together in flight, or during one of its short pauses upon a flower. From the position of the injury it is clear that the enemies were aware of the situation of the body and attempted to seize it, but that they had been frustrated by the swift and wary butterfly with its bright yellow wings extending behind the short body, and offering an apparently convenient point for seizure.1 The bright yellow black-bordered under wings of the moths of the genus Tryphcena (Yellow Underwings) also possess this among their other meanings, as Mr. Jenner Weir has pointed out : they are exposed during flight, and their colours are far brighter than any other part of the insect. It is also very common to find the margin of these wings notched in captured specimens, and this is often the case when all other parts are fresh and perfect. The red and black under wings of the genus Catocala (including the Bed and Crimson Underwings) are perhaps useful in the same way. A still more interesting and obvious character of this kind is to be found in markings which actually suggest the presence of a vital part, such as an enemy would be likely to seize. On one occasion I intro- 1 Skertchly has found such mutilations not uncommon among Bornean butterflies : he also notices the correspondence of the injury on the two sides. The wings are not torn in this way by flying through thick branches ; Skertchly states that even the most fragile butterflies can pass unharmed through dense undergrowth. WARNING COLOURS 207 duced a Small Heath Butterfly (Coenonympha pam- philus) into a lizard's cage. It was at once obvious that the lizard was greatly interested in the large eye- like mark on the under side of the fore wing: it examined this mark intently, and several times at- tempted to seize the butterfly at this spot. The observation seems to point to, at any rate, one use of the eye-like markings which are common on the under sides of the wings of butterflies. A very perfect and elaborate example of the same kind is witnessed in the Hair streak Butterflies (Thecla). Each hind wing in these butterflies is furnished with a ' tail,' which in certain species is long, thin, and apparently knobbed at the end. When the butterfly is resting on a flower the wings are closed and the hind wings are kept in constant motion, so that the ' tails ' continually pass and re-pass each other. This movement, together with their appearance, causes the ' tails ' to bear the strongest likeness to the antennae of a butterfly ; the real antennae being held so as not to attract attention. Close to the base of the supposed antennae an eye-like mark, in the most appropriate position, exists in many species. The effect of the marking and movement is to produce the deceptive appearance of a head at the wrong end of the body. The body is short and does not extend as far as the supposed head, so that the insect is uninjured when it is seized. This interesting fact of the resemblance of the 208 THE COLOUKS OF ANIMALS tails and adjacent parts to a head in Thecla has been long known : it was first observed by Dr. Arnold in the case of a foreign species (Thecla larbas), and was confirmed by Dr. Forsstrona in other species. The fact is quoted by Kirby and Spence under ' Means of Defence of Insects,' but the interpretation offered, that the insects ' perhaps thus perplex or alarm their assailants,' hardly expresses the true significance of the character.1 The same fact was independently discovered by Mr. E. C. L. Perkins in 1888, and this keen naturalist at once perceived the meaning of the character — to divert the attention of an enemy towards a non-vital part. The discovery is of especial interest because it was made upon an English species (Thecla W-album), and because Mr. Perkins tested his explanation by finding that this part had been torn in a considerable proportion of the butterflies. The observation renders it extremely probable that the slender ' tails ' which occur in the same position in many ' Blues ' (Polyommatus), and the bright colours and eye-like spots which are often associated with them, have a similar meaning. The 'Blues,' when resting on a flower, have the same habit of moving the hind wings, as I have often ob- served in our common English species which are without ' tails.' The movement is such as would render the ' tails ' prominent and antenna-like if 1 Kirby and Spence, People's Edition, 1867, p. 423. WARNING COLOURS 209 they were present ; and it may, therefore, have per- sisted from a time when the butterflies possessed these appendages. . Similar features in Reptiles A similar interpretation applies to the tails of lizards, which break off the instant an attempt is made to capture the animal by seizing this part. The tail is, of course, the first part which the pursuer has the chance of seizing. The great length of the tail, and the rapidity with which it is renewed after being shed, also support this interpretation. Similar features in Mammals It is very possible that the well-known peculiarity of the tail of the dormouse is to be explained in the same manner. The large bushy tail of the squirrel may possess a similar meaning (among others), for an enemy in pursuit would be liable to get only a mouthful of fur. In the north of Europe the squirrels which frequent the birches are black, while those on the pines are brown : both varieties, which are probably protective, become greyish in winter, and thus har- monise with the frosted bark. But the tails of both retain their summer colour, and would be thus more conspicuous. This fact was pointed out to me by my friend Mr. H. Balfour.1 1 On the other hand a seasonal change in the colour of the tail 210 THE COLOUBS OF ANIMALS Similar features in Mollusca Semper ' has shown that certain freely exposed and active snails in the Philippines (Helicarion) have the same power of readily parting with their tails, and this is also true of a snail in the West Indies (Stenopus). The tail, or rather hinder part of the foot, which the animal sheds when it is seized and after- wards renews, is more conspicuous than the rest of the body. Semper found that the tails had been shed in about ten per cent, of the individuals of a species (Helicarion gutta) very common in the north-east of Luzon. Recognition Markings A special kind of marking is often of great value in attracting the attention of individuals of the same species, instead of attracting the attention of enemies. From its obvious relation to the latter form of marking it is best included under the division of Warning Colours. Mr. A. E. Wallace has directed attention to the importance of Eecognition Markings, and an account of them will be found in his recently published seems to be not uncommon in certain localities in this country. The tail becomes cream-coloured at the end of summer, but resumes its ordinary appearance at the beginning of winter. Smaller differences appear to be general, the summer fur on the tail being coarser and more uniformly red than the winter fur. —Bell's British Quadrupeds, 2nd edition, p. 279. 1 Semper : Animal Life, International Scientific Series, pp. 395 WARNING COLOURS 211 volume, ' Darwinism ' (pp. 217-27) -1 Such characters may be of use in aiding a species to escape from its enemies. Thus gregarious mammals, ' while they keep together, are generally safe from attack, but a solitary straggler becomes an easy prey to the enemy ; it is therefore of the highest importance that in such a case the wanderer should have every facility for dis- covering its companions with certainty at any distance within the range of vision ' (loc. cit. p. 217). Kecog- nition Markings would be especially useful * at a dis- tance or during rapid motion in the dusk of twilight, or in partial cover.' Recognition Markings in Mammals A very beautiful and familiar illustration is given by Mr. Wallace — the white upturned tail of the rabbit, by which the young and inexperienced, or the least wary individuals, are shown the way to the burrow by those in front. It is very interesting to compare this marking with that of the skunk, which has been al- ready described as possessing a very conspicuous white tail. In the latter case the tail is held so that the slow-moving animal is always conspicuous, and appeals to the imagination and memory of its enemies ; the tail of the rabbit only becomes conspicuous when it is needed by other individuals of the same species, and 1 The principle of Recognition Markings is set forth in a work by the late Alfred Tylor, Colouration in Animals and Plants, 1886, p. 30. 212 THE COLOURS OF ANIMALS when the animal is already alarmed and in full retreat for a place of security. In this way Mr. Wallace explains the conspicuous markings often present on gregarious ruminants, which are nevertheless protectively coloured in other respects. The remarkable differences in the length and form of the horns of different species are explained in a similar manner. Recognition Markings in Birds Mr. Wallace also shows that such characters are especially numerous and suggestive among birds. ' Recognition Marks during flight are very important for all birds which congregate in flocks or which migrate together ; and it is essential that, while being as conspicuous as possible, the marks shall not inter- fere with the general protective tints of the species when at rest. Hence they usually consist of well- contrasted markings on the wings and tail, which are concealed during repose, but become fully visible when the bird takes flight ' (loc. cit. p. 222). Recognition of Birds' eggs may be aided by variation in certain species It is very probable that the great variation in the colours and markings of birds' eggs, which are laid close together in immense numbers, may possess this WABNING COLOURS 213 significance, enabling each bird to know its own eggs. I owe this suggestive interpretation to my friend Mr. Francis Gotch : it is greatly to be hoped that experimental confirmation may be forthcoming. The suggestion could be easily tested by altering the positions of the eggs and modifying their appearance by painting. Mr. Gotch's hypothesis was framed after seeing a large number of the eggs of the guillemot in their natural surroundings. It appears to be a more feasible explanation than that offered by Mr. Wallace. ' The wonderful range of colour and marking in the eggs of the guillemot may be imputed to the inaccessible rocks on which it breeds, giving it com- plete protection from enemies ' (loc. cit. p. 214). Recognition Markings in Insects Turning to insects, I do not believe with Mr. Wallace that colours and markings generally are to be explained in this way, although many instances of undoubted recognition characters will probably be found among them. In fact, a very interesting example only recently came before me. It has been already mentioned that Lepidopterous larvaB are especially subject to the attacks of parasitic insects (Hymenoptera and Diptera), which lay their eggs in or upon them. It is of the highest importance for a parasite to know whether a larva is already * occupied,' and also to ensure that other parasites 214 THE COLOURS OF ANIMALS shall recognise and avoid the larvae in which it has laid its eggs. When the eggs are laid within the body of a caterpillar the skin is pierced, and a small amount of blood exudes and generally forms a black clot. These black spots are probably recognised by other parasites, and the larva is consequently avoided. Although the spots would disappear after a change of skin, the parasites generally lay their eggs at about the same period of larval growth, and would be warned over a considerable part of this period. Al- though these are not, properly speaking, Eecognition Markings, we shall see that they form the foundation on which such characters have arisen. Other parasites (among the Hymenoptera), such as those of the genus Paniscus, lay eggs upon the body of their prey. The eggs are pear-shaped, and are firmly fixed by the stalk, which is knobbed at the end. So tightly do the eggs adhere that the caterpillar can change its skin without removing them (see pp. 275-77). Several eggs are fixed upon a large caterpillar, two or three upon a small one, although the number varies greatly in different individuals. These external eggs are black and shining, and they are very conspicuous against the colour of the caterpillar, which is generally green. When Professor Weismann was staying with me in the summer of 1887, 1 showed him a larva of the Puss Moth (Cerura vinuLa) to which several eggs were attached. This led to a discussion as to the meaning of the colour, in WAKNING COLOURS 215 the course of which we both independently arrived at the opinion that it is adapted to serve as a warning to other parasites that the larva is already ' occupied.' The eggs, being black, somewhat resemble the scars caused by the introduction of internal eggs, so that the species which deposit such eggs may be warned off, as well as those of the genus Paniscus. 216 THE COLOURS OF ANIMALS CHAPTEE XH . PROTECTIVE MIMICBY WE now approach one of the most interesting aspects of our subject, and one that has played an important part in the history of evolution and of natural selection. History of the subject The fact that certain butterflies belonging to widely separate groups, but inhabiting the same localities, possess the most remarkable superficial resemblance, has been known for a very long time. An interesting quotation from Boisduval's ' Species General des Lepidopteres ' (pp. 372, 373) is given by Mr. Eoland Trimen at the head of his paper on Mimicry among African butterflies.1 Boisduval's sen- tence, written in 1836, refers to an African Swallow-, tailed Butterfly, which still remains the most remark- able instance of Mimicry known in the _worldl ' C'est une chose bien remarquable que de voir la nature creer a cote les uns des autres 1' Euplcea Niavius, le 1 Linn. Soc. Trans, xxvi. p. 497. PEOTECTIVE MIMICRY 217 Diadema dubia, et le Papilio Westermanni, trois Lepidopteres qui se ressemblent presque completement par le port, le dessin, et la couleur, quoique apparte- nant a des genres fort eloignes et de tribus differentes.' From 1836, and the even earlier dates at which these remarkable resemblances had been noticed, until 1862, no attempt at explanation had been made ; but in that year Mr. Bates's classical paper appeared.1 In this admirable~essay the author showed the advan- tage which must necessarily be gained by a palatable form, hard pressed by enemies, if it sheltered itself under the reputation of some conspicuous species well known to be inedible. Only three years before, Darwin, writing to Asa Gray, had said : ' I cannot possibly believe that a false theory would explain so many classes of facts as I think it certainly does explain. On these grounds I drop my anchor, and believe that the difficulties will slowly disappear.' 2 One great difficulty which had so long been a puzzle to naturalists was therefore satisfactorily explained by the new theory, within a few years of Darwin's prediction. It is most delightful to read of the interest and enthusiasm with which Bates's paper was received by Darwin. ' In my opinion it is one of the most re- markable and admirable papers I ever read in my 1 Contributions to an Insect Fauna of the Amazons Valley. Linn. Soc. Trans, vol. xxiii. 2 Life and Letters, vol. ii. p. 217. 218 THE COLOURS OF ANIMALS life,' he writes. ' I am rejoiced that I passed over the whole subject in the " Origin," for I should have made a precious mess of it. You have most clearly stated and solved a most wonderful problem. Your paper is too good to be largely appreciated by the mob of naturalists without souls ; but rely on it that it will have lasting value, and I cordially congratulate you on your first great work.' * This was Darwin's opinion of a theory which is often lightly criticised or even con- demned by many biologists who offer nothing in its place. The relation of the theory of Mimicry to Evolution Mr. Bates's paper afforded a twofold support to the arguments in the ' Origin of Species,' at a very criti- cal time in the history of these opinions. In the first place it showed that an important class of facts was unintelligible upon any theory except that of evolution. The proof of this is best given in Dar- win's own words, also quoted by Mr. Francis Darwin.2 'By what means, it may be asked, have so many butterflies of the Amazonian region acquired then- deceptive dress ? Most naturalists will answer that they were thus clothed from the hour of their creation — an answer which will generally be so far triumph- ant that it can be met only by long-drawn argu- ments ; but it is made at the expense of putting an 1 Life and Letters, vol. ii. pp. 391-93. 1 Ibid., vol ii. pp. 391-92. PEOTECTIVE MIMICEY 219 effectual bar to all further inquiry. In this particular case, moreover, the creationist will meet with special difficulties; for many of the mimicking forms of Leptalis can be shown by a graduated series to be merely varieties of one species ; other mimickers are undoubtedly distinct species, or even distinct genera. So again, some of the mimicked forms can be shown to be merely varieties, but the greater number must be ranked as distinct species. Hence the creationist will have to admit that some of these forms have become imitatory by means of the laws of variation, whilst others he must look at as separately created under their present guise ; he will further have to admit that some have been created in imitation of forms not themselves created as we now see them, but due to the laws of variation ! Professor Agassiz, indeed, would think nothing of this difficulty ; for he believes that not only each species and each variety, but that groups of individuals, though identically the same, when inhabiting distinct countries, have been all separately created in due proportional numbers to the wants of each land. Not many naturalists will be content thus to believe that varieties and individuals have been turned out all ready made, almost as a manufacturer turns out toys, according to the tempo- rary demand of the market.' l But Mr. Bates's theory was equally important in 1 From a review of Bates's paper by Charles Darwin. — Natural History Revieiv, 1863, p. 219. 220 THE COLOURS OF ANIMALS another respect. It not only supported the doctrine of evolution, but it afforded strong confirmation of the theory of natural selection, by which Darwin explained how it was that evolution took place. Every step in the gradually increasing change of the mimicking in the direction of specially protected form, would have been an advantage in the struggle for existence, while the elements out of which the resemblance was built exist in the individual variability of the species, a variability which is hereditary. The transition from Warning to Mimetic appearances It will have been observed that Mimicry has already been mentioned in the pages on Warning Colours, and that a gradual transition may be traced from the one principle to the other. And yet Mimicry itself was explained long before many of the conclusions con- cerning Warning Colours which have been described. In this, as in so many other cases, the steps by which the subject is best approached are almost exactly oppo- site to the historical steps by which it was gradually understood. The transition from warning to mimetic forms may be shortly recapitulated. 1. The existence of Warning colours, attitudes, &c. in species which possess some quality unpleasant to the enemies of their class : recognised by Bates in butter- flies which are mimicked by others (loc. cit. 1862) ; the PROTECTIVE MIMICRY 9.21 principle especially supported and extended by Wal- lace; also greatly supported by Trimen, Belt, and many others. 2. The tendency for the species in each specially protected group of butterflies to resemble each other (by convergence or arrested divergence) more closely than those in other groups not similarly protected, thus suffering a smaller amount of destruction while their enemies are being educated to avoid them ; sug- gested by Meldola (' Ann. and Mag. Nat. Hist.' Dec. 1882) as an extension of the principle discovered by Fritz Miiller and described in the next paragraph. 3. The tendency for the members of distantly re- lated groups of specially protected butterflies to resemble each other, thus gaining the advantages described above : discovered by Fritz Miiller (' Proc. Ent. Soc. Lond.' 1879, p. xx.). The fact of the resemblance was first observed by Bates (loc. cit.). 4. An extension of the same principle to all the groups of such specially protected animals : in these the same colours and patterns occur again and again, and advantage is gained by the fact that the types of appearance are those which produce most effect upon the sight of an enemy, as well as by the fact that only a few different types have to be learnt. Certain types of colour and pattern are eminently advantageous for animals in which the special protection is imperfect, because they are so thoroughly advertised by other animals in which the protection is complete and 11 222 THE COLOURS OF ANIMALS inspires great dread (Poulton, « Proc. Zool. Soc.' March 1887). 5. The latter cases naturally lead to those of true Mimicry, in which a group of animals in the same habitat, characterised by a certain type of colour and pattern, are in part specially protected to an eminent degree (the mimicked), and in part entirely without the special protection (the mimickers), so that the latter live entirely upon the reputation of the former. Discovered by Bates in Tropical America, (loc. cit. 1862), then by Wallace in Tropical Asia and Malaya (loc. cit. 1866), and by Trinien in South Africa (loc. cit. 1870). Cases to which the term Mimicry is best applied The term Protective Mimicry is best applied to the deceptive appearance of the unprotected forms in the last class only. Instances of such true Mimicry, in which the resemblance deceptively suggests the pre- sence of some positively unpleasant quality, are so common and striking that we need some name for them ; and it is in every way best to retain the historic term. An additional advantage is that the word Mimicry implies the deception and unreality which is so obvious in the last class of cases de- scribed above. For this reason it is best to include all the other classes and the protected forms in the fifth class, unc|er Warning Colours ; for their object PROTECTIVE MIMICRY 223 is to warn an enemy, as effectually as possible, of real danger or unpleasantness ; while the object of the unprotected forms in the fifth class is to suggest the presence of some unpleasant attribute which has no existence in fact. The transition from Warning to Mimetic colours which occurs in the fourth class is no objection to this arrangement ; for we cannot escape transition in any classification of the uses of colour in animals. Some authorities have failed to make any distinction between Mimicry and the other forms of Protective and Aggressive Resemblance ; but such an arrangement would confound together cases in which appearance is used for concealment, and those in which it is made use of in order to attract attention.1 1 S. B. J. Skertchly has recently (Ann. and Mag. Nat. Hist., Series vi. vol. iii. pp. 477 et aeq.) urged (a) that 'protective resemblance copies stationary objects, mimicry simulates moving ones.' He accord- ingly maintains that the former is a defence against enemies which attack butterflies at rest, the latter against those which attack them on the wing. He further argues that the attacks of birds constitute the only real danger to an insect on the wing ; (6) that certain ob- servers (Skertchly, Pryer, Scudder) agree in considering these attacks to be of very little importance ; and (c) that therefore Mimicry, to- gether with the shyness of moving objects exhibited by all butterflies on the wing, ' are habits acquired long since, which have survived the necessity that gave them birth.' He maintains that this argument is supported by (d) the ' law that the amount of apprehended danger is measurable by the efforts taken to avoid it,' inasmuch as ex- amples of Mimicry are far rarer than examples of Protective Resem- blance. To this^e may reply: (a) the alleged contrast between Pro- tective Resemblance and Mimicry is only a usual consequence of the real difference between them (see pp. 222-23, also p. 71). Further- 224 THE COLOURS OE ANIMALS Convenience of the term Mimicry Mr. Bates's term has been criticised because it is generally used to describe voluntary actions, whereas the Mimicry alluded to in these pages is of course un- conscious, and has been gradually produced by the operation of natural selection.1 This use of the word more, the alleged contrast frequently breaks down ; thus, a dead leaf driven by the wind (see p. 56), or a piece of stick swinging by a thread, are not uncommonly resembled ; while the conspicuous appearance of Mimetic and Warning Colours are most certainly of value during rest as well as during flight (see pp. 189-95). Against the evidence offered by Skertchly (b) may be put the observations of other naturalists (see pp. 228-30; many other examples might have been recorded). We must also remember that very little of the destruction of life which we know takes place is actually witnessed by us. Against (c) may be urged the fact that characters begin to decline directly they be- come useless, while certain mimetic resemblances are perhaps more wonderfully elaborate and perfect than anything else in the animal kingdom. Every naturalist will agree with (d), but it is really de- structive of the argument based upon it. The danger to most mimetic species must indeed be great if measured by the efforts taken to avoid it. Some of the marvellous results of such ' efforts ' are de- scribed in this volume. Skertchly's argument only applies to the numerical ratio between the examples of Mimicry and Protective Resemblance ; and this ratio is readily explicable on other grounds. It has been shown that Warning Colours can only be adopted safely by a small proportion of the insect fauna in any country (see pp. 178-80), and also that mimetic individuals must be far rarer than the nauseous forms they resemble (see pp. 243-44 ; see also p. 231). 1 After this sentence had been printed I came across a most extraordinary statement of the theory of Mimicry by Skertchly, I.e. This theory ' presupposes (a) that danger is universal ; (b) that some butterflies escape danger by secreting a nauseous fluid ; (c) that other butterflies noticed this immunity ; (d) that they copied it.' The opinions expressed in the words I have italicised will hardly be accepted by a PROTECTIVE MIMICRY 225 is, however, well known, and is not likely to mislead anyone ; and in addition to its historical accuracy the word is more convenient than any other, as Mr. Wallace has pointed out. Thus we obtain the con- venient series of words — mimic, mimicry, mimetic, mimicker, mimicked, mimicking. Various degrees of affinity between mimicking and mimicked species The various examples of Mimicry may be divided according to the affinity of the forms which resemble each other. Thus a species may mimic another closely allied species, or one of a widely separated family, of a distinct order, class, or even sub-kingdom. Mr. Bates first explained the Mimicry of butterflies and moths, so that in this case the divergence between the species, although generally very great, is not nearly so large as that of other examples. The former, however, in- cludes some of the most striking examples known, and will be first described. The Butterflies which afford models for Mimicry In giving some account of Mimicry among butter- flies, it is first necessary to speak of the models which are most generally copied in all the warmer parts of single naturalist. I imagine that even the American Neo-Lamarckians do not follow their founder so far as to believe that the volition of an animal could account for all the details of mimetic resemblance. 226 THE COLOUKS OF ANIMALS the world. These models almost invariably belong 'to the two great families Danaidce (including Euploea, Danais, and Hestia) and Acrceidce, while the Heliconidte of Tropical America are also mimicked. It has been already pointed out that the Danaida, which inhabit the region of which the Heliconidce are characteristic, have adopted the appearance of the latter, and may therefore be called Heliconoid Danaidce.1 Proof that the mimicked Butterflies are specially pro- tected by a nauseous taste or smell It is of the greatest importance to prove that these butterflies are specially protected in some unusual and exceptionally complete manner, so that resemblance to them would be advantageous. All observers speak of their slow flight, gaudy colours, and abundance. Thus Mr. Trimen's descriptions of the Danaidce and Acrceidce are equally true of the Heliconidce, and of all other butterflies or moths which are the objects of Mimicry. ' The slow flight, the conspicuous colours, the complete disregard of concealment, no less than the great abundance of individuals, are characteristics indicating unmistakably that these butterflies are favoured races, enjoying advantages and immunities above their fellows.' 2 The colours of the under sides 1 Bates called them ' Danaoid Heliconidce,' but Trimen pointed out that the transposed words more truly express the relationship (loc. cit. p. 499). 2 Loc. cit. p. 498. PEOTECTIVE MIMICRY 227 of the wings are the same as those of the upper sides, or at any rate are equally conspicuous. A peculiar and frequently unpleasant smell has been noticed by all observers who have studied these groups. It is pro- bable that the same means of defence is present in the other stages, and this has been proved in certain cases (e.g. Acraa horta, Trimen). The unpleasant smell frequently resides in a clear yellow fluid which exudes on the slightest pressure. Mr. Trimen l has also called attention to the fact that the conspicuous butterflies and moths which possess such qualities have a remarkably elastic struc- ture, and can endure very severe pressure without injury. The wings are so flexible that they can be bent and distorted without breaking the nervures. The insects can in this way often recover from the mistaken attacks of insect-eating animals. Skertchly also maintains, from his experience in Borneo, that nauseous properties are accompanied by strong vitality. There is, unfortunately, too little direct experi- mental proof of the unpalatability of the specially protected groups which are the chief models of Mimi • cry. When, however, all the observations are brought together they constitute a fair amount of evidence, and there can be no doubt about the results of future experiments. Mr. Bates mentions the glands near the anus 1 Loc. cit. pp. 498, 499. 228 THE COLOUES OF, ANIMALS of certain Heliconid