S?»r*r*r>. »rfcr*r>rkrirfcr»?krkr»r>rfcr*r»r>ri I 22102058303 idi W. & Q. FOYLE, THE Booksblleri, 135 CHABiiro Crom Road, W.C. BraochM Srerywbere. Hooka Bought. EXPERIMENTAL ZOOLOGY EXPERIMENTAL ZOOLOGY BY THOMAS HUNT MORGAN PROFESSOR OF EXPERIMENTAL ZOOLOGY COLUMBIA UNIVERSITY the MACMILLAN COMPANY LONDON; MACMILLAN & CO., Ltd. 1907 All rights reserved ZOOi^O , Tj^ytJX ■■ y vie V Cv f\S,-&-(A) Gr] ^HS'- Copyright, 1907, By the MACMILLAN COMPANY. Set up and electrotyped. Published January, 19C7. WELLCOME INSTVnrnE LIBRARY Coll. welMOmec Call No. Qu KovbJootJ IPrtBS J. S. Cushing & Co. — Berwick & Smith Co. Norwood, Mass., U.8.A. PREFACE The great interest that has been shown during the last fifteen years in the study of Experimental Zoology has led to the rapid development of this branch of biology. An attempt is made in the following pages to bring together the results of this work. A series of about thirty-five lectures formed the basis for my treatment of the subject, and this will account, in part, for the way in which the matter has been handled ; many details have been omitted that an exhaustive treatment would demand ; and the plan has been to select the most typical and most instructive cases for presentation, when such a choice was possible. Neverthe- less, I believe that the reader will find a fairly full account of the subjects considered. Physiology has from the beginning made use of the method of experiment, and with notable success. Morphology has, up to the present time, followed mainly the historical and descriptive methods, although striking exceptions could be cited. While the historical study of zoology must always remain a legitimate field for activity, as human history has been a time-honored study, yet there can be little doubt that the more promising and searching method of zoological study in the future will be found in experiment. The central problem of morphology — the causes of the changes in form, or at least the determination of the condi- tions under which changes in form occur — will furnish the main theme of the present treatise. Two fields of study that properly fall under this head are, however, not considered, viz. experimental embryology and the experimental study of regeneration. Both of these subjects have in recent years received comprehensive treatment in book form, so that it did not seem desirable to go over the ground again. More- VI Preface over, their consideration would have demanded too much space to be included with the present matter in a single volume. The fascinating study of the psychical side of living phenomena also belongs to the province of experimental biology, especially comparative psychology; but this subject has quite recently been fully dealt with by Loeb and by Jennings, who have themselves been largely instrumental in developing the subject, so that further treatment would be more than superfluous. The excellent summaries and reviews of some of the topics discussed here, that have been published in recent years, have greatly facilitated my work. I need only mention Roux’s and Driesch’s analysis of the experimental method, Grafin v. Linden’s summary of the experiments on butter- flies, Herbst’s excellent treatment of the subjects of “For- mative Reiz,’’ Phillips’s very full review and literature on sex determination, and Cuenot’s, Lenhossek’s, and O. Schultze’s treatment of the same subject. I need scarcely add that, while using these and other reviews, I have made my own compilation almost exclusively from the original sources. Covering as extensive a field as I have attempted to cover, it is probable that the different subjects have received un- equal treatment, and I fear that some omissions may have been made. I trust, however, that no serious oversights or mistakes will be found. It gives me great pleasure to express here my appreciation of the generous assistance in the correction of the manuscript and proof rendered by my wife, by Professor E. B. Wilson, by Professor C. B. Davenport, and by Professor C. E. Castle. CONTENTS EXPERIMENTAL STUDY OF EVOLUTION CHAPTER I INTRODUCTION PAGE The Experimental Method • 3 CHAPTER II The Influence of External Conditions in causing Changes in the Structure of Animals . . • • • • • .12 Adaptive Responses Non-adaptive Responses 1 5 Influence of Temperature on the Coloration of Butterflies . . 15 Effect of Temperature on Caterpillars ..... 24 CHAPTER III The Influence of External Conditions in causing Changes in the Structure of Animals (^Continued) .29 Experiments on the Influence of the Food Plant ..... 29 The Influence of Light, Electricity, Centrifugal P'orce, Chemical Sub- stances, and Oxygen on the Caterpillars and Pupae of Moths and Butterflies ........... 37 The Influence of Humidity on the Characters of Moths and Butterflies 38 Experiments with Flatfish ......... 40 Experiments with Crustaceans ........ 40 Changes in Mammals and Birds . . . . . . . >41 CHAPTER IV The Inherited Effects of Changes induced by External Factors 43 CHAPTER V The Inheritance of Acquired Characters 50 Telegony ............ 59 Xenia ............. 61 Theories of Transmission of Somatic Influences ..... 61 vii VI 11 Contents CHAPTER VI PAGE Experimental Hybridizing 66 Mendel’s Law 66 Mendel’s Law and the Germ-cells 72 CPIAPTER VII Experimental Hybridizing {Continued) 82 Experiments with Mice .82 CHAPTER VIII Experiments with other Mammals and with Birds .... 99 Experiments with Guinea Pigs . 99 Experiments with Rabbits ......... 106 Experiments with Rats . . . . . . . . .110 Experiments with Cats . . . . . . . . . • H3 Data for Other Mammals and Man . . . . . . .116 Experiments with Poultry . . . . . . . . .121 Experiments with Pigeons . . . 135 CHAPTER IX Experiments with Snails, Moths, and Beetles 139 Experiments with Silkworms ........ 141 Experiments with Beetles . . . . . . . . •'^5^ Experiments with the Currant Moth . . . . . . -153 Experiments with Tephrosia . . . . . . . • *53 CHAPTER X Other Kinds of Hybridizing 156 Blended Inheritance . . . . . . . . • • *5^ Mosaic Inheritance . . . . *57 Hybridization between Linnrean Species *57 Dimorphism *60 Reversal of Symmetry . . . . *^5 Conclusions regarding Mendelian Inheritance *66 CHAPTER XI Behavior of the Germ-cells in Cross-fertilization . . . • *73 Experiments with Amphibia . . . • • • • • *73 Experiments with Echinoderms . . . • • • • • *75 Factors involved in the Entrance of the Spermatozoon . . .178 Artificial Helps to P’ertilization and to Cross- fertilization . . • *81 Polyspermy Contents IX CHAPTER XII Inbreeding Experiments with Mice and Rats Experiments with the Pomace Fly Behavior of Germ-cells in Inbreeding . CHAPTER XIII Influence of Selection . . Fluctuating or Individual Differences ....... Selection of Fluctuating Differences ....... Elimination within the Species ........ Variation and Parthenogenesis .... . . . The Results of Selection and Hybridization of Wild Elementary Species Selection under Domestication of Mutations, Saltations, Sports, and Discontinuous Varieties in General ...... CHAPTER XIV The Theory of Evolution The Analogy with Artificial Selection ...... The Influence of the Environment ...... De Vries’s Results with CEnothera ...... Variation and Mutation in Helix ....... Evolution by Means of Definite Variation Adaptation ........... The Selection Theory and the Theory of the Survival of Mutations EXPERIMENTAL STUDY OF GROWTH CHAPTER XV Introductory Normal Growth ......... Senescence .......... Length of Life in Different Species ..... Absorption of Parts by Larvm CHAPTER XVI External Factors that influence Growth Influence of Food Stimulants affecting Growth Effects of Salt on Growth Effects of Heat on Growth .... PAGE 1 86 i88 190 195 198 199 202 204 206 207 208 213 214 215 217 221 224 229 232 239 239 246 248 251 253 253 256 257 260 X Contents Effects of Light on Growth . Growth toward the Light; Phototropism Influence of Gravity on Growth . Effects of Electricity on Growth . Pressure and Contact .... The Formation of Galls CHAPTER XVII Growth and Regeneration .... EXPERIMENTAL STUDIES IN GRAFTING CHAPTER XVIII Experiments in Grafting The Union of Different Regions The Influence of the United Parts on Each Other : Formative Factors The Union of Parts of Different Species ...... Special Problems of Development EXPERIMENTAL STUDIES OF THE INFLUENCE THE ENVIRONMENT ON THE LIFE-CYCLE CHAPTER XIX Changes in the Life-cycle and Changes in the Environment The Life Histories of Some Animals ....... Influence of Food on the Life-cycle of Lepidoptera . . . . Influence of the Environment on the Time of Ripening of the Sexual Organs .... CPIAPTER XX Alternation of Sexual and Parthenogenetic For.ms Alternation of Generations in the Aphids and Phylloxerans . CHAPTER XXI Influence of External Conditions on the Life-cycle of the Lower Crustaceans CHAPTER XXII PAGE 261 261: 266 267 267 268 277 285 285 289 298 301 OF 309 309 313 315 322 322 336 Influence of External Conditions on the Life-cycle of the Ro- tifer, Hydatina Senta 346 Contents XI CHAPTER XXIII PAGE The Life-cycle of Some Hymenopterous Insects. Bees . . -35^ Other Hymenoptera 35^ EXPERIMENTAL STUDY OF THE DETERMINATION OF SEX CHAPTER XXIV Introduction: The Different Kinds of Sexual Individuals . . 363 Dioecious Species, composed of Unisexual Individuals .... 3^5 Monoecious Species, composed of Bisexual or Hermaphrodite Individuals 368 Parthenogenetic Species . . .. . • • • • • 37^ CHAPTER XXV External Factors of Sex Determination . • 37^ Influence of Food .......... 37^ Supposed Influence of Nourishment in determining Sex in Man and other Mammals .......... 3^4 Geddes’s and Thomson’s Theory of Sex ...... 387 CHAPTER XXVI Internal Factors of Sex Determination . . . ... .391 Age of Parents ........... 39^ Condition of the Germ-cells ........ 393 Vigor of the Parents .......... 393 Effects of Inbreeding . . . . . . . . . . 394 Size of the Egg ........... 394 Ratio of Nucleus to Cytoplasm ........ 396 The Extrusion of the Polar Bodies, and the Analogous Process in the Sperm-cells ........... 397 The Formation of Male and Female Producing Spermatozoa . . 401 CHAPTER XXVII The Internal Factors of Sex Determination (^Continued') . . 407 The Origin of Gynandromorphs 407 The Sex of Multiple Embryos . . . . . . . .410 The Sex of Human Twins and Double Monsters . . . . .411 Two Recent Theories of Sex Determination based on the Assumption of Male and Female Eggs . . . . . . . .413 The Reduction Process in Parthenogenetic Eggs . . . . .415 Influence of the Cytoplasm . . . . . . . . .417 Contents PAGE Conclusions : (1) The Morphological Conception of Sex Determination . 420 (2) The Physiological Conception of Sex Determination . . 421 (3) Analysis of the Results ....... 422 EXPERIMENTAL STUDY OE SECONDARY SEXUAL CHARACTERS CHAPTER XXVIII Secondary Sexual Characters . . '. . ... . . 429 Introduction ........... 429 Correlation between the Secondary Sexual Characters and the Essential Organs of Reproduction 432 CHAPTER XXIX Secondary Sexual Characters {Co7itinued') 439 Theories of the Origin of Secondary Sexual Characters . . . 439 INDEX 449 EXPERIMENTAL STUDY OF EVOLUTION EXPERIMENTAL ZOOLOGY CHAPTER I THE EXPERIMENTAL METHOD The study of Zoology by experimental methods is not a new departure, for the method of experiment has been often applied to special zoological problems. On the other hand, the recogni- tion that only by experimental methods can we hope to place the study of Zoology on a footing with the sciences of chemistry and of physics is a comparatively new conception, and one that is by no means as yet admitted by all zoologists. I do not wish to appear to disparage those studies that deal with the descrip- tive and with the historical problems of biology. They also offer a wide field for activity, and the more familiar we become with the structure and modes of development of animals, so much the better can we apply the experimental method. In fact, many of the problems of biology only become laiown to us as the result of direct observation. The wider, therefore, our general information, the greater the opportunity for experimen- tation. It is undoubtedly true that many zoologists who have spent their lives in acquiring a broad knowledge of the facts of their science fail to make use of their information by testing the very problems that their work suggests. This is owing, no doubt, to their exclusive interest in the observational and descriptive sides of biology, but also in part, I think, to the fact that the experi- mental method has not been sufficiently recognized by zoologists as the most important tool of research that scientists employ. . 3 4 Experimental Zoology Perhaps also the fact that the historical side of biology attracts such popular interest accounts, in part, for the neglect of more searching scientific methods of study. Whether the method of observation or the method of observa- tion and experiment is followed, seems to be also a question of the kind of interest aroused by living objects. If the number of collectors, naturalists, zoologists, anatomists, entomologists, ornithologists, mammalogists, conchologists, etc., be compared with the number of physiologists, physiological chemists, bac- teriologists, it will be seen that the former have an enormous advantage in numbers. It is true that a few zoologists are experimentalists, and that some physiologists do not experiment at all, but the proportion remains about the same. In other words, interest in collecting and recording the results of obser- vation and in the artistic side of nature is much more wide- spread than interest in the study of problems, or, if the interest is not lacking, the will to take the initiative in the formulation and solution of problems seems to be less cultivated in the biological sciences than the power to observe and to describe. In so far as the followers of the one or of the other method of investigation have made their selection as a matter of tempera- ment, the disproportion will probably always remain ; but in so far as the result is due to imitation, or to following the line of least resistance, or to a failure to appreciate differences in aim and method, the proportion may to some extent be altered ; and I think it will be generally admitted that at the present time there is greater need for experimental work than for descriptive and observational study. It is sometimes said that experimental study is the analytical study of problems, and this in a. sense is true, but it is only a part of the truth. It is rather the method of attacking problems that is the chief characteristic of experimental work, for is not the historical method also a study of problems? We demand in the case of a problem in experimental science that the condi- tions under which an event takes place be discovered, and that, if possible, we reproduce artificially the result by controlling the 5 The Experimental Method conditions. In fact the control of natural phenomena is the goal of experimental work. In the studies of physics and chemistry the method of experi- ment is so familiar, that we think of, their advancement as taking place by experiment alone. In biology the situation is different, and new discoveries are looked for as often in the field of observation as of experiment. This difference is due to the higher stage of development that has been reached by the physical sciences, while biology is still, in large part, in the lower stages of its evolution where facts are insufficiently known. Nevertheless the amount of time still given to descriptive work is out of proportion to the present condition of development of biology. A few examples may serve to illustrate the differences be- tween the descriptive and the experimental study of zoology. The egg of an animal, if set free and fertilized, begins at once to develop. Descriptive embryology gives us the different stages through which the egg passes, but no matter how complete the description, we still know little or nothing of the causes that are operating to bring about the development. What, for instance, does the spermatozoon bring into the egg to make it develop ? What physical and chemical changes take place during cleavage ? What makes the embryo turn in at one pole ? Why do certain cells develop cilia ? These and a hundred other questions suggest themselves. Observation has failed to answer them. Another method employed in recent years has been to attempt to find certain physical or chemical changes that seem to be similar to those observed in the developing egg. Thus it has been suggested that the spermatozoon brings a ferment into the egg ; that the cleavage is due to differences of surface tension ; and that the gastrulation fe caused by osmotic pressure. Ma- chines that behave in somewhat similar ways have even been constructed to illustrate some of these changes. Interesting as the ideas derived from these sources may be, their scientific value lies only in their suggestiveness, until it can be shown 6 Experimental Zoology that the changes in the embryo are really the outcome of similar processes; and the only way in which certainty can be gained on this point is by experiment on the organisms themselves. If a ferment starts the development it is our duty to isolate it, and to introduce it hypodermically into the egg. If this could be done and the development thereby started exactly as in normal fertilization, the hypothesis becomes at least probable that a fer- ment is brought in by the spermatozoon and starts development. Similarly, for cleavage, for gastrulation, and for every stage in the development, experiment alone can give a satisfactory answer. The essence of the experimental method consists in requiring that every suggestion (or hypothesis) be put to the test of experi- ment before it is admitted to a scientific status. From this point of view the value of an hypothesis is to be judged, not by its plausibility, but by whether it meets the test of experiment. Its use is therefore primarily for the investigator, and not for the layman ; yet as a matter of fact the wildest speculations are likely to be the ones that excite most popular attention and applause. It is sometimes said that an hypothesis is useful in proportion to the number of facts it brings under one point of view. This is true for the student rather than for the investigator. Such an hypothesis may have no scientific status. It belongs rather to a system of mnemonics. To the teacher, also, hypotheses are useful in arousing the interest of his hearers, so that by exciting their imdeveloped imagination he can make his dry facts more entertaining. But let us be careful to distinguish between the forensic and the scientific value of hypotheses. Hypotheses may be useful, and have been used in various ways. They have been used, as just stated, to hold together a body of isolated facts ; as such they are in reality only fictions. They have been used in the reconstruction of supposed historical events, especially in biology in the setting up of family trees. In this case they can only claim to be more or less plausible suggestions. Hypotheses have been used to direct interest The Experimental Method 7 toward certain fields of study. As such they have often proven stimulating and have been useful in acting as a guide for others. But the hypothesis of real importance is the working hypothesis of the investigator. It is the test by means of which he tries to interpret his problem, and therefore it is essential that his hypothesis is one having a practical bearing, i.e. an hypothe- sis that can be shown to be true or false. It differs in this essential respect from purely fictitious and from metaphysical hypotheses. The working hypothesis carries along with it its dangers as well as its advantages; since, while it may lead to discoveries, it may, if it is wrong in principle, blind us to the real condi- tions. Therefore the investigator must not only be an inventor of working hypotheses, but cultivate also a skeptical state of mind toward all hypotheses — especially his own — and be ready to abandon them the moment the evidence points the other way. And herein hes one of the differences between the re- corder of observations and the experimenter. The work of the observer, if exact, is complete in itself, and stands forever as a monument to his ability, or at least to his industry; while the conclusions of the experimenter, if they are to bear fruit, must become modified with each new discovery. His results are ab- sorbed in the current of the next advance, but his consolation will be that he has had at least a share in the causal study of living things, and in helping the human race toward the control of organic phenomena. To return to our examples. The growth of animals and plants offers a wide field for experimental study. Under cer- tain conditions we see a young animal continuing to grow larger until a certain size is reached, when growth slowly ceases. Al- though the animal may live for many years longer, it has ceased to grow. What makes it grow? Why does it stop growing? We have hardly begun experimental work along these lines; yet we shall see later that there is a promising field for work in this direction. After a time old age comes on and the animal dies. We say it dies a natural death, and this seems inevitable. 8 Experimental Zoology but only because we have found that death always takes place under ordinary conditions. Suppose, however, we change the conditions ; might we not hope to prolong the duration of life ? Improbable as this may seem, there are already experiments afoot that indicate that, however difficult, the problems may not be insoluble. Why, on an average, in most animals, are equal numbers of two forms born — male and female ? Is there an internal mechanism? If so, what regulates it? Do external or inter- nal conditions determine that one egg becomes male, another female? Even if an internal mechanism exists, it might be affected by external conditions, and in any case the cause of the production of the two types must be determined. Observation has established that the evolution of animals and plants has, in all probability, taken place. But what factors are involved in the process are unknown. Only in the last few years by means of an experimental study of the subject has decided advance been made. It is sometimes stated that nature has already carried out innumerable and wonderful experiments, and that we can never hope to excel her in this power. Is it not better, therefore, to examine patiently and reverently what she has done, and in this way learn how her processes have been carried out? Let us not be blinded by rhetorical questions of this kind. No doubt nature has carried out prodigious experiments; but we can never be certain that we know how she has obtained her re- sults until we can repeat the process ourselves. What would the chemist or the physicist say if he were told that nature has already carried out experiments on a much greater scale than he can hope to accomplish, and that he should drop his ex- perimental methods and study his physics in a thunderstorm and his chemistry in a volcanic eruption ! I have brought up this point because it illustrates one side of the experimental method that is sometimes overlooked. Al- most all of the phenomena with which the biologist has to deal are so complex that he cannot determine what part each factor 9 The Experimental Method plays in the result unless he study the effects of each under different conditions that can be controlled. Little by little, in this way we can hope to gain a clearer insight into the condi- tions that, taken all together, produce the result. So much for the experimental method. It is pertinent to ask what is an experiment? If I cut an earthworm in two to see what will happen, have I performed an experiment ? Perhaps not, for the actual performance of cutting the worm in two is not the essential point. The essence of an experiment is a trial or test, and the conditions are so arranged that an answer is ex- pected. If the worm is cut in two in order to study the physio- logical behavior of the two ends, as has been done in fact, with some interesting results, or in order to see what regenerates at the two cut ends, we have a distinct purpose in view, although no formulated problem. If we proceed farther and remove a definite number of segments in order to see how many come back, and then try to determine what conditions are involved in the results, we are clearly carrying out an experiment with a more definite aim. This illustration will serve to show that the most essential feature of an experiment is the anticipation of the results of a test. The operation may be so simple, and the conditions .so little known, that to call the performance an experiment may easily expose one to the ridicule of those un- favorably inclined to the claims of experimental work. If the process is carried out with scarcely a thought as to its purpose, and in complete ignorance of all the conditions entering into the problem, it can scarcely be called an experiment at all. At most it is only a preliminary testing of possibilities. Much of the pioneer work in experimental zoology has necessarily been of this kind, and crude as such preliminary work must be, it should be looked upon only as the first step toward a further and more critical analysis. The carrying out of an experiment implies the formulation of a working hypothesis, and this usually presupposes some knowl- edge of the possible conditions that control the phenomena. The experimental work becomes more explicit and accurate the lO Experimental Zoology more we know beforehand of the possible conditions that may enter into the result. The ability of the experimenter is shown in his insight into the possible factors that may be present. His ability may be the result of a correct estimate of the possible conditions, but for the highest order of work there is demanded also great imaginative power. Good judgment and accurate observation may lead to fine work, but constructive imagination seems to be required for the highest order of original work. This does not imply that accuracy of observation is not as requi- site in original work as in descriptive and observational work, and should always be expected; but the man who sees new and overlooked combinations may open fields of research that will set to work an army of able “investigators.” The branches of biology that have made most extensive use of the experimental method are physiology, bacteriology, and physiological chemistry. The zoologist and the embryologist have also to deal with physiological problems, and already the beginning of important experimental work has been carried out in this field ; but the most distinctive problem of zoological work is the change in form that animals undergo, both in the course of their development from the egg {embryology) and in their develop- ment in time {evolution). It will be granted, I think, that these formative problems are more difficult than those relating to function with which the physiologist has concerned himself in the main; but this is all the greater reason why the experimental method should be used in their study, especially after so much purely descriptive work has been already done. The term “ morphology ” has been used in recent times to denote the study of form, as contrasted with physiology, that deals with functional changes. Morphogenesis has also been employed to signify a study of the changes in form through which organisms pass. It is mainly the experimental study of these changes in form that I propose to examine in the follow- ing pages. Experimental morphology would perhaps nearly indicate the field to be examined ; but since the line between experimental physiology and experimental morphology is often The Experimental Method 1 1 hard to draw, and since I shall not hesitate at times to enter upon the physiological side of many problems, I have chosen the somewhat broader title of Experimental Zoology to include the subjects to be treated. The principal topics to be discussed fall under the following six headings : — I. Experimental Study of Evolution. II. Experimental Study of Growth. III. Experimental Studies in Grafting. IV. Experimental Studies of the Influence of the Environ- ment on the Life-cycle. V. Experimental Study of the Determination of Sex. VI. Experimental Study of the Secondary Sexual Characters. y CHAPTER II THE INFLUENCE OF EXTERNAL CONDITIONS IN CAUSING CHANGES IN THE STRUCTURE OF ANIMALS Animals and plants are so constituted that one of their chief characteristics is that they respond to their natural environment in such a way as to insure their continued existence. These responses are in the main physiological, and therefore in large part transitory; but in some cases the response is structural, involving a temporary or even a permanent change in form or structure that persists at least so long as the external condi- tions that called it forth remain. The question arises whether these changes, directly induced by the environment, may not give origin to the more fixed characters that have become the permanent inheritance of each species. May not these have been in the first instance adaptive responses to the environ- ment? This leads to the further question of the origin of all the characters of the species, whether adaptive or non-adaptive. In this and in the following chapters the different sides of this question will be considered. ADAPTIVE RESPONSES External conditions sometimes cause adaptive structural changes in organisms. We are familiar with some effects of this sort in our own bodies. Pressure on the skin, if long con- tinued, causes it to become thicker and more capable of resist- ing the injurious effects of pressure. Sunlight tans the skin and protects it from “burning.” It is said that cold causes the fur of some mammals to become thicker, and this change better protects them against the cold. Conversely, it is said that horses and dogs lose their hair to some extent in very warm climates. 12 The Influence of External Conditions 13 A number of Arctic animals become white in winter. This change seems to be in part due directly to the cold, for it has been found if these animals are transferred to warmer climates they show less marked changes on the approach of winter. Flounders and some other fish and some amphibians become lighter in color on a light background and darker on a dark background. The most remarkable case of this sort is that of the pupas of certain butterflies. If the pupation takes place on a light background, the chrysalids are lighter ; and if on a dark back- ground, they are darker. Experiments by Poulton ^ have shown that this effect is produced directly through the skin and not through the ocelli. Poulton thought in one case that even the color of the silk in which the caterpillar incloses itself is influ- enced by the color of the background, but this has been shown not to be the case. It is popularly supposed that the African chameleon becomes green in green surroundings, and brown in a dark environment, but this is probably not true ; at least it has been shown in another lizard, Anolis, that can also change from green to brown and the reverse, that the animal is as a rule green if warm and brown if cold. The effects are produced by change in the pig- ment of the dermal pigment cells. In the cold the black pigment spreads out over the surface and conceals the stationary green pigment. In the warmth the black pigment migrating inward exposes the green. Light has a somewhat different effect. In both the African chameleon and in Anolis a strong light acts like a low temperature, causing the black pigment to migrate to the surface, and a faint light or darkness acts like a high temperature, causing the black pigment to wander inward, so that the animal becomes green. Other lizards give reverse effects in light. Parker and Starratt have shown for Anolis that when both light and heat act together, the results are as follows : ‘ For details of the experiment, see Poulton, “The Colours of Animals,” p. no, 1890; “Further Experiment upon the Colour Relation,” etc., Trans. Ent. Soc., p. 293, London, 1893; “An Inquiry into the Cause and E.xtent of a Special Colour Relation,” Proc. Roy. Soc., Vol. XII, 1887. 14 Experimental Zoology At a low temperature, io° C., Anolis changes from green to brown, irrespective of illumination. At a high temperature, 40° to 45° C., it turns from brown to green, irrespective of illumination. Thus heat is the controlling factor at these extremes. Between these extreme temperatures there is a range from 25° to 35° through which light is the controlling factor, although heat is not without its influence, as shown by the rate of the change. Parker and Starratt have discovered the astounding fact that the effects of the illumination may be produced when an area of the skin no larger than a square millimeter is exposed, the rest of the animal being in the dark. The changes that have just been described, except perhaps the last ones, seem to be of benefit to the animal, either in directly protecting it from the agent that brings about the result, as in the effects of pressure, cold, sunlight, etc., or in more effectually concealing the animal from its enemies. These responses are said to be adaptive, but it is remarkable how rare are adaptive structural responses, when we recall the fact that adaptation of the organism to its surroundings is one of its most characteristic properties. The poverty of adaptive structural response does not encourage one to look to external agents as having brought about directly the structural adaptation of organisms to exter- nal conditions, even if it could be shown that such influences are inherited. There are, on the other hand, many cases of physiological responses that are adaptive; in fact, nearly all functional changes are directly beneficial to the organism. Animals re- spond in a most remarkable way to poisons. If certain alka- loids are injected in ever increasing doses, the animal becomes immune to a dose which if given in the first instance would have been fatal. In the case of the poisonous ptomaines pro- duced by bacteria, the animal produces a counter poison, an antitoxin, that nullifies the effects of the ptomaine. In a num- ber of bacterial diseases the animal becomes more or less im- mune after the first attack. Equally striking are the adaptive responses shown by animals 15 The Injiuence of External Conditions to temperatures higher or lower than those to which they are normally subjected. If the change of temperature is gradual, the organism may become adapted to a temperature that would have been fatal if met at once. Somewhat similar results have been found by subjecting ma- rine animals to water containing less salt. If the change is gradual, the animal will become adapted to the new density. The extent to which the process may be carried differs greatly in different species. In some cases the animal may be gradually transferred even to fresh water. There are also other animals that may pass at once from salt to fresh water without serious injury. Salmon and shad leave the ocean to migrate up the rivers, and other fish do the same thing. Conversely, the young salmon migrates back to the sea. The number of fish that will stand as great a change as this is, however, limited, although many oceanic species will live in water much less salt than that of the sea. NON-ADAPTIVE RESPONSES In contrast to the few cases of adaptive structural responses to the environment there are quite a number of cases in which definite structural responses occur that are not adaptive. One of the interesting points connected with these responses is that the differences effected by changes in the environment have been shown in some cases to resemble the kind of differences that separate species from each other; but whether species have really originated, either directly or indirectly, in this way, must be carefully considered later. Influence of Temperature on the Coloration of Butterflies The earliest experimenters on the influence of temperature on butterflies were Dorfmeister (1864) and Weismann (1875). Earlier naturalists, who were familiar with seasonal dimor- phism in butterflies, had supposed, it is true, that differences in temperature might be responsible for the differences in color that characterize the summer and the winter broods; but it 1 6 Experimental Zoology required the experimental work of Dorfmeister and of Weis- mann to show that this supposition was correct. Weismann showed for Vanessa levana-prorsa that when a pupa, destined to give rise to the summer form, is kept at a low temperature, it may produce the winter form, V. levana, or a type transitional between the summer and the winter forms. He also succeeded, by raising the temperature, in changing the winter pupa so that it gave rise to the summer type of butterfly. More recently Merrifield, Standfuss, Fischer, Grafin von Linden, and others have carried out extensive experiments on the effects of temperature. The butterflies and moths used for this work are usually those having summer and winter broods, that differ in color and often in size, and even in the shape of the wings. In other cases, however, similar changes have been brought about in forms that do not show seasonal dimor- phism. It has been found that not only the summer form can be changed into the winter form, and vice versa, but in certain cases the type may be changed by cold so that it resembles northern varieties of the same species, and by heat to resem- ble southern varieties. Temperatures that are only somewhat higher or lower than normal produce the southern and northern types respectively, while much higher or lower temperatures produce effects that are rarely or never found in nature. These latter changes are sometimes called aberrations. We may first examine a few examples of these effects given by Standfuss. The effect of heat on Vanessa cardui is shown in Fig. 5. The colors are much lighter above than those of the normal butterfly. The black bands are much reduced. Similar changes are ob- servable on the under side of the wing. The pupa had been kept for 60 hours at 36-37° C., and then at normal temperatures for six to seven days, when the butterfly emerged. The effect of cold on Vanessa cardui is shown in Fig. 6. The color is darker and the white spots are reddish in tint as seen especially on the under side. The pupae had been kept for 33 days in an ice chest, then for five days in a cellar (-f 13° C.), and lastly for nine days at room temperature. Fig. I. Showing the iilrtuence of heat and cold on the pupie of some butter- flies. Vanessa antiopa: Fig. i. effect of heat; of cold, Figs. 2 and 3. V^anessa cardui: Fig. 5, effect of heat; Fig. 6. effect of cold. Vanessa atalanta : Fig. 7. effect of heat; Fig. 8, effect of cold. Fig. 4. Vanessa pollychloros. aberratio dixeyi. effect of cold. (After Standfuss.) The Influence of External Conditions ij The effect of heat on Vanessa atalanta is shown in Fig. 7. The blue spots on the outer edge of the wings are so much re- duced that often two small flecks alone remain. The red cross- band of the fore-wings is spread out. In the black, that is somewhat brownish, red-brown shading appears near the base of the wing. The large white spot at the anterior border of the fore-wing and the neighboring five scattered white spots show a reduction, and the latter may even disappear. In all of these aspects the type approaches a variety from the Canary Islands. The pupae were kept 172 hours at 37° C., then three to four days at 24° C. The effect of cold on Vanessa atalanta is shown in Fig. 8. The ground color of the upper surface is bluish black. The white spots are larger, and the whole of the outer tip of the fore- wing is lighter. The red cross-band is reduced and broken. The under surface of both wings are much changed, as the figures show. The influence of heat and of cold on Vanessa antiopa is shown in Figs. I and 2. The effect of heat (Fig. i) is to make “dusty” the brown ground color of the upper surface, especially on the hind wing, that may become nearly black. The blue spots of the marginal row are reduced to half their normal size and are more violet in color. The characteristic yellow margin is dusted with brown. The pupae had been kept 60 hours at 37° C., and then for 12 days at 24° C. The influence of cold is shown in Fig. 2. The blue spots are much larger, and the ground color of the wings is darkened. The marginal yellow band is reduced and richly dusted with black scales. The pupae had been kept for 44 days in an ice chest, and then for 15 to 19 days at normal temperatures. Another individual produced by cold (29 to 34 days) is shown in Fig. 3. It approaches Vanessa polychloros. Fig, 4, in a number of points, Fischer has obtained some very aberrant forms of Vanessa antiopa by heat and by cold. Some of those produced by cold are shown in Figs. 1-8 as seen from above, and in Figs. 9-1 1 as seen from below. The former set (Figs. 1-8) is arranged to 1 8 Experimental Zoology show how the broadening of the yellow margin begins in the posterior wings and finally extends forward to the same extent on the anterior wings. The change is in an antero-posterior direction. The conditions under which these butterflies had been kept were as follows: — In the first experiment 20 pupae of Vanessa antiopa, about 12 hours old, were kept for six hours at a temperature of 14° C. and then four hours in a temperature decreasing from 14° C. to 0° C. After this they were put three times daily for a short period in a temperature of —3° for 18 days. They were then keptinthe cellar (14° C.) and finally at room temperature (22°C.). Six pupae died, the remaining 14 began to emerge after 10 to 12 days. Figure 7 shows one of these butterflies which is the aberration known as hygiaea; another is shown in Fig. 6, in which the blue spots and the dark border have completely dis- appeared, but the yellow border does not extend inward so far as in the last case. Three other individuals (Figs. 2, 3, 4, 5) show transitional forms in some of which traces of the blue spots could be seen. The last butterfly of this series to emerge was like the normal antiopa with the blue spots even larger than the normal, but less sharply defined. Thus under identi- cal external conditions quite a range of colors result; but of course the caterpillars themselves had probably not lived under identical conditions, nor had they been subjected to the cold at precisely the same stage in their pupation. In a second experiment the conditions were the same, except that the pupae were brought into the extreme cold (three times daily) for only 14 days. Five of these showed the aberration artermis (like that in Fig. i, which developed, however, under different conditions).^ The five individuals showed also aber- rations with the hygiaea characters. Two individuals were like Fig. 4 ; the others were like the intermediate forms. In a third experiment the conditions were the same as before, except that the pupae were brought into the extreme cold for only six days. They produced one normal butterfly; one transi- ‘ The butterfly had been kept at a temperature below 0° and 6° C. Fig. 2. Vanessa antiopa : Figs. i-8 showing effects of a low temperature on pupa. (After Fischer.) 19 20 Experimental Zoology tional form like Fig. 3 ; one form, the aberration arlermis, with many groups of golden scales on the brown ground color and the black border, and even within the blue spots ; the rest for the most part belonged to the aberration hygiaea, one of these being as extreme a form as that shown in Fig. 8. Fig. 3. Vanessa antiopa, under surface : Figs. 9-1 1 showing influence of cold. (After Fischer.) The under side of the wings of these butterflies is also , changed, but not always in the same way as the upper side. As shown in Fig. 10, not only does the yellow border of the wing extend inward, as it does also on the upper side, but at the same time the black scales spread toward the periphery, darkening the broader yellow band, as shown in Fig. 10. This figure shows the under side of the aberration hygiaea. In the butterfly shown in m Fig. II the whole under side is black, the yellow border having disappeared. This is the under side of the butterfly shown in Fig. 8 that had the broadest development of the yellow band on the upper side. Some further details may now be considered. Merri- field did not employ ver)^ ex- treme temperatures. The pupae were either iced (33° F.) or only cooled (39° to 57° F.). For “ forcing,” temperatures not higher than 70° to 80° F., or even 90° F., were employed. He found The Injiuence of External Conditions 21 in general that cooled or iced pupae gave dark and much-spotted moths, while “forced” pupae were pale and spotless or with reduced spots. He states that the markings are affected by- long- continued exposure, especially during the early pupal period, but the color is chiefly affected during the penultimate period. Standfuss has experimented on as many as 7000 individuals in all. He determined that the period in the life of the pupa when it is most sensitive to heat is at the beginning, although it is best not to try to produce the effect too soon after the cater- pillar has become a pupa. If exposed soon to extreme cold, the pupae die; but if exposed soon to heat the best results, i.e. the most divergent forms, are obtained, although the mortality is high. If the pupa is exposed as soon as it can stand it to extreme cold and then to heat, the heat-type is produced. In other words, the cold delays the development so that the heat produces the greater effect; for as soon as the temperature is raised the development goes forward rapidly. To get the best effect from cold, Standfuss found it advantageous to expose first to low temperatures, then for 5 to 10 days to moderate cold (11° to 14° C.), and lastly bring the pupae to room temperature. In his earlier work Fischer subjected the pupae for several days to extreme temperatures of heat or of cold, but later he found it better to subject the pupae two or three times a day to the extreme temperatures during a period of ii to 20 days. This method applies particularly to extreme temperatures. For a low temperature he used 0° C. or - 3° C. or even — 8° to - 20° C. The high temperature was from 35° to 46° C. The more ex- treme the temperature that the pupae will stand the greater the effect produced. Thus temperatures between 0° and — 20° gave greater effects than those between 0° and -|- 10°. Tempera- tures of 42° to 46° gave more striking results than those of to 41° C. The most important result obtained by Fischer was to show that the same aberrations are obtained by extreme heat and by extreme cold. The result may seem puzzling, but it must be re- membered that the coagulation of proteids, which is probably 2 2 Experimental Zoology one of the factors in the results, can be caused artificially either by a high or by a low temperature. In a mixture of proteids we might expect some slight differences in the results of coagu- lation, even if the principal changes are the same, and it is not improbable that such minor differences do exist. An excellent general review of the effects on the colors has recently been given by Grafin Marie von Linden. She points out that in the Vanessa series, a higher temperature makes the red or yellow deeper or more fiery. The dark background suffers a reduction. Cold gives the reverse, a brightening of the general dark ground color, the yellow expanding at the cost of the red. There is also a lightening of the red and increase of white scales. Extreme heat and cold, as 'stated above, give remarkably similar results. The black spots on the border run together, so that the peripheral dark spots are lost. The dark border zone becomes clearer (in some forms only at the tip). Despite this peripheral clearing up, extreme temperatures cause an increase of dark pigment elsewhere. It may be said, there- fore, that extremes of heat and of cold do not give specific effects, but produce the same physiological change. As a result of these changes the differences between related species sometimes seem to disappear to a greater or less extent. The nearer the forms experimented upon, the more alike are their aberrations. This result led Fischer to the conclusion that extreme heat and cold cause an atavistic return to the primi- tive type of all the Vanessas, i.e. a return to the stem-form from which they have come. He attempts to explain this result on the grounds that during the development of the color the butter- fly passes in its ontogeny through phylogenetic stages. Cold and heat cause an arrest of development, so that an ancestral stage emerges. Standfuss, on the other hand, looks upon the changes as some- thing new, and points out certain contradictions to Fischer’s idea that the aberrations are atavistic. For instance, the males are much less prone to atavism than the females, and yet pro- duce a much greater number of aberrations. He thinks it im- The Injiiience of External Conditions 23 probable that the original form of the Vanessas was darker than the present forms, to judge by related groups. Uncertain as arguments, pro and con, based on these phylogenetic specu- lations necessarily are, there is fortunately experimental evi- dence that shows how little ground there is for Fischer’s argu- ment. Grafin Marie von Linden has examined the developing pigment in the wing itself. She finds that the lighter colors develop first and that in the younger stages the red and yellow tones occupy a greater area than they do later. This is the reverse of Fischer’s primary assumption of sequence. In pupae exposed to freezing and to heat it seems that disturbances in the development of the color occur. Standfuss observed in these cir- cumstances that the color develops later than in the normal ; and von Linden finds that the black color appears relatively earlier than the others, and at times even before the red and yellow. Grafin von Linden has also shown that changes in certain colors similar to these shown by the pupae can be produced in vitro. Extracts of the red color, when heated, become fiery red or more red-brown in color ; while on ice the red and the yellow- red tone remain constant. It is likely, therefore, that some of the effects of high and low temperature can be explained en- tirely as due to the direct influence of the temperature on the chemical composition of the pigment. Not only have the summer and winter forms been changed by changing the temperature, but in two cases, in which sexual dimorphism exists, it has been possible to change the female coloration into that of the male. For example, the colors of the female of Parnassius apollo can be changed into those of the male. In the female of Rhodocera rhammi the white colora- tion of the wings can be changed by warmth into the intensely yellow color of the male. It has been suggested that while ordinary temperatures suffice to cause the development of the deeper color in the male, it requires a higher temperature than that ordinarily met with to cause the same change in the tissues of the female. Standfuss has pointed out that while in many cases a lower temperature may cause a darker color and a higher 24 Experimental Zoology temperature a lighter one, this is not invariably the rule, as seen in certain dimorphic forms. Thus, in Lythria rotaria the butterfly that hatches in the spring (from over-wintering pupae) is darker than the summer type (the second generation, variety L. purpuraria). On the other hand, Vanessa levana (also from over- wintering pupae) is lighter than the summer variety, V. prorsa, A lower temperature does not therefore always produce darker colors and a higher temperature lighter ones, although this is the general rule, but the reaction depends also on the nature of the organism. In these experiments with butterflies the more conspicuous result is the change in color ; but it should not be forgotten that changes in size also often occur, and even constant changes in the shape of the wings have been observed, the outline of the wing in some cases being quite different from that of the normal animal. Effect of Temperature on Caterpillars Standfuss has shown that by rearing caterpillars at a higher temperature (20°-2 5° C.) than normal, the characters of the moths may be affected.^ Changes in the shape of the wings are sometimes caused in this way. The color is also altered to some extent. The most constant change, however, is in the size. Standfuss gives the following rule: “The more the feeding period of the caterpillar is shortened by raising the temperature the smaller the butterflies. Lasiocampa quercifolia, for example, had its weight decreased in this way to one seventh of the normal. On the other hand, if the time of feed- ing (or in other words the time of the caterpillar stage) is not, or only very slightly, shortened, despite the higher temperature, there is an increase in size, which in the case of Arctia fasceata may be as much as half again the entire normal weight.” Standfuss points out that results similar to these are found in nature where the size of certain forms appears to be connected with the time of year at which the pupas appear. Species * Merri6eld also has made observation on the effects of temperature on caterpillars. The hijiuence of External Conditions 25 whose growth occurs in late autumn (with lower temperature and decreasing vegetation) and is completed in the autumn so that they winter as pupae, are generally smaller in the first gen- eration and larger in the second or summer generation; for the caterpillars of the latter are born in summer and have plenty to eat. Conversely, species which winter over as small cater- pillars and develop further during the favorable conditions of spring produce larger butterflies in the first than in the second generation. LITERATURE, CHAPTER II Bachmetjew, P. Ueber die Temperatur der Insecten nach Beobachtun- gen in Bulgarien. Zeit. f. wissen. Zool. LXVI. 1899. Der kritische Punkt der Insecten und das Entstehen von Schmetter- lings-aberration. 111. Zeit. f. Entomol. V. 1900. Experimentelle entomologische Studien. i. Band., Temperaturver- haltnisse bei Insecten. Leipzig. 1901. Kalometrische Messungen an Schmetterlingspuppen. Zeit. f. wissen. Zool. LXXI. 1902. Bateson, W. On Variation in the Colour of Cocoons of Eriogaster Trans. Entomol. Soc. London. 1892. On Variation in the Colour of Cocoons, Pupae, and Larvae; Further Experiments. Ibid. 1892. Brucke, E. Ueber den Farbenwechsel der Chamaeleonen. Sitz. Kais. Akad. Wiss. Wien, math.-naturw. Cl. VII. 1851. Untersuchungen ueber den Farbenwechsel des Afrikanischen Chama- leons. Denkschr. Kais. Akad. Wiss. Wien, math.-naturw. Cl. IV. 1852. Carlton, F. C. The Colour Changes in the Skin of the so-called Florida Chameleon, Anolis Carolinensis Cuv. Proc. of the Amer. Acad, of Arts and Sci. XXXIX. 1903. Chlodkovsky, N. Sur quelques Variations Artificielles du Papillon de rOrtie (Vanessa urticae). Ann. Soc. Entomol. de France, LXX. 1901. Dixey, F. a. On the Phylogenetic Significance of the Wing marking in certain Genera of the Nymphalidae. Trans. Entomol. Soc. Lon- don. 1891. Mr. Merrifield’s Experiments in Temperature-Variation as Bearing on the Theory of Heredity. Trans. Entomol. Soc. London. 1894. Doremeister, G. Ueber die Einwirkung verschiedener wahrend der Ent- wickelungsperiode angewandter Warmegrade auf die Farbung und Zeichnung der Schmetterlinge. Mitt. d. naturw. Ver Steiermark. II. 1864. Ueber den Einfluss der temperatur bei der Erzeugung von Schmetter- linge varietaten. Ibid. 1879. Edwards, W. H. An Abstract of Weismann’s paper on “The Season Dimorphism” of butterflies, to which is appended a statement of 26 Experimental Zoology some experiments made upon Papilio ajax. Canad. Entomol. 1875-1876. EfiFects of Cold applied to the Chrysalides of Butterflies. Amer. Ento- mologist, III. 1880. Farkas, K. Beitrage zur Energetik der Ontogenese. Arch. ges. Physio- logie, XCVIII. 1903. Fickert, C. Kunstliche Kalteabartungen von Schmetterlingen. Jahres- hefte d. ver. f. vaterland. Naturkunde in Wurtemburg, LIII. 1897. Fischer, E. Transmutation der schmetterlinge infolge Temperaturan- derungen. Experimentelle Untersuchungen ueber die Phylogenese der Vanessen. Berlin. 1894. Neue experimentelle Untersuchungen und Betrachungen ueber das Wesen und die Ursachen der Aberrationen in der Faltergruppe Va- nessa. Berlin. ’ 1896. Zwei sonderbare Aberrationen von Vanessa antiopa und eine neue Methode zur Erzeugung von Kalteaberrationen. Illustr. Wochen- schr. f. Entomol. 1897. Beitrage zur experimentellen Lepidopterologie. 111. Zeitschr. f. En- tomol. II. 1897. III. 1898. IV. 1899. V. 1900. Experimentelle kritische Untersuchungen uber das prozentuale Auf- treten der durch tiefe Kalte erzeugten Vanessa aberrationen. Soc. Entomol. XIII. 1899. Experimentelle Untersuchungen ueber die Vererbung erworbenerEigen- schaften. Allg. Zeitschr. f. Entomol. VI. 1901. Lepidopterologische Experimentalforschungen. 111. Zeitschr. f. En- tomol. VI. 1901, und VIII. 1903. Weitere Untersuchungen ueber die Vererbung erworbener Eigen- schaften. Allg. Zeitschr. f. Entomol. VII. 1902. Flemming, W. Ueber den Einfluss des Lichts auf die Pigmentirung der Salamanderlarve. Arch. f. mikr. Anat. XLVIII. 1897. Weitere Bemerkungen ueber den Einfluss von Licht und Temperatur auf die Farbung der Salamanderlarve. Arch. f. mikr. Anat. XLVIII. 1897. Frings, C. Experimente mit erniedrigter Temperatur im Jahre 1898. Soc. Entomol. XIV. 1899. Experimente mit erniedrigter Temperatur im Jahre 1899. Soc. En- tomol. XV. 1900. Temperaturversuche im Jahre 1900. Soc. Entomol. XVI. 1901. Berichte ueber Temperaturexperimente im Jahre 1901. Soc. En- tomol. XVI. 1902. Gauckler, H. Einfluss hoher Temperatur auf den Organismus von In- sekten. Entomol. Nachr. XII. 1886. Experimente mit niedrigen Temperaturen an Vanessenpuppen. Iris, II. 1896. Keller, R. Ueber den Farbenwechsel des Chameleons und einiger anderer Reptilien. Arch. f. ges. Physiol. LXI. 1895. Krodel, E. Durch Einwirkung niederer Temperaturen auf das Puppen- stadium erzielte Aberrationen der Lycaena-Arten : Corydon Poda und Damon Schiff. Allg. Zeitschr. f. Entomol. IX. 1904. Linden, M. v. Versuche ueber den Einfluss aiisserer Verhaltnisse auf die Gestaltung der Schmetterlinge. 111. Zeitschr. f. Entomol. IV. 1898. 27 The Influence of External Co7iditions Morphologische und physiologisch chemische Untersuchungen ueber die Pigmente der Lepidopteren, I. Die gelben und roten Farb- stoffe der Vanessen. Arch. f. d. Ges Physiol. XCVIII. 1903- Die Ergebnisse der experimentellen Lepidopterologie. Biol. Centralb. xxiv. 1904. _ Merrifield, F. Report of Progress in Pedigree Moth-Breeding. Trans. Entomol. Soc. London. 1888. Incidental Observations in Pedigree Moth-Breeding. Ibid. 1889. ^ Systematic Temperature Experiments on Some Lepidoptera in all Stages. Ibid. 1890. Conspicuous Effects on the Markings and Colouring of Lepidoptera. Ibid. 1891. The Effects of Artificial Temperature on the Colouring of Several Species of Lepidoptera with an Account of Some Experiments on the Effects of Light. Ibid. 1892. The Effects of Temperature in the Pupa Stage on the Colouring of Pieris napi, Vanessa atalanta, Chrysophanus phloeas, and Ephyra puctaria. Ibid. 1893. Temperature Experiments in 1893 on Several Species of Vanessa and Other Lepidoptera. Ibid. 1904. Recent examples of the effect on Lepidoptera of extreme Temperature applied in the Pupal Stage. Proceedings of the South London En- tomol. and Nat. Hist. Soc. 1897. Parker, G. H., and Starratt, S. A. The Effect of Heat on the Color Changes in the Skin of Anolis Carolinensis (Cuv.). Proc. of the Amer. Acad, of Arts and Sci. XL. 1904. Pictet, A. LTnfluence des changements de nourriture sur les chenilles et sur la fonnation in sexe de leurs papillons. Compt. Rend. Soc. de phys. et d’hist. Nat. Geneve, XIX. 1902. L’influence des changements de nourriture des chenilles et sur le developpement de leurs papillons, I. Compt. Rend. Soc. Hebret. sc. Nat. Geneve. Arch. sc. phys. et nat. Variations des papillons provenant des changements d’alimentalion de leurs chenilles et de I’humiditd. Arch. sc. phys. et nat. Geneve. XVI. 1903. Notes complementaires sur les variations des papillons provenant de I’humiditd. Soc. de Phys. et d’hist. Nat. Bull. d. stances, I. 1903. PouLTON, E. B. The Essential Nature of the Coloring of Phytophagous LarvEe (and their Pupae), with an Account of Sortie Experiments upon the Relation between the Color of such Larvae and that of their Food Plants. Proc. of the Roy. Soc. 1885. A Further Enquiry into a Special Color-Relation between the Larva of Sinerinthus ocellatus and its Food Plants. Proc. of the Roy. Soc. 1886. An Inquiry into the Cause and Extent of a special Color-Relation between certain exposed Lepidopterous Pupae and the Surfaces which immediately surround them. Proc. of the Roy. Soc. XLII. 1887. The Experimental Proof of the Protective Value of Color and Mark- ings in Insects in reference to their Vertebrate Enemies. Proc. of the Zool. Soc. of London, 1887. Further Experiments upon the Color-Relation between certain Lepi- 28 Experimental Zoology dopterous Larvae and their Surroundings. Trans. Entomol. Soc. London. 1892. The Experimental Proof that the Colours of Certain Lepidopterous Larvae are largely due to Modified Plant Pigments. Nature XLVIII. 1893. Reaumur, M. de. Memoires pour servir h I’histoire des insectes. Tome II. 1737- Reichenau, W. V. Die Ziichtung des Nesselfalters (Vanessa urticae L.); ein Beweis fur den direken Einfluss des Klimas. Kosmos, V. 1882. Ruhmer, R. G. W. Die Uebergange von Araschnia levana L. Zu var. prorsa L. und die bei der Zucht anzuwende Kaltemenge. En- tomol. Nachr. XXIV. 1898. Standfuss, M. Handbuch fur Sammler der europaischen Grosschmet- terlinge. 1891. Ueber der Variation und Aberration des Falter stadiums bei der Schmetterlingen mit ausblicken auf die Entstehung der Arten. Leipzig. 1894. Weitere Mitteilungen ueber den Einfluss extremer Temperaturen auf Schmetterlingspuppen. Entomol. Zeitschr. 1895. Handbuch der palaarktischen Grosschmetterlinge. Jena. 1896. Experimentelle zoologische Studien mit Lepidopteren. Neue Denk- schrift d. Allg. Schweiz. Gesellsch. f. d. gesamte Naturw. 1898. Gesamtbild der bis Ende, 1898, an Lepidopteren vorgenommenen Tern peratur- und Hybridationsexperimente. Insektenborse, XVI. 1899. Urech, F. Experimentelle Ergebnisse der Schniirung von noch weichen Puppen der Vanessa urticae quer ueber die Fliigelchen. Zobl. Anz. XX. 1897. Ergebnisse von Temperaturesperimenten von Vanessa io, L. 111. Zeitschr. f. Entomol. 1898. Varigny, H. de. Experimental Evolution. Lectures. 1892. Venus, C. Pr. Ueber Varietatenzucht. Korresp. Bl. des entomol. Ve- reins. Iris z. Dresden, I. 1888. WeiSmann, a. Studien zur Deszendenztheorie, I. Ueber den Saisondi- morphismus der Schmetterlinge. Leipzig. 1875. Neue Versuche zum Saisondimorphismus der ScWetterlinge. Zobl. Jahrb. Abt. f. Syst. VIII. 1895. CHAPTER III THE INFLUENCE OF EXTERNAL CONDITIONS IN CAUSING CHANGES IN THE STRUCTURE OF ANIMALS ((Continued) Experiments on the Influence of the Food Plant Pictet has studied the influence of the food of the caterpillar on the color, the size, and other characters of the butterfly. As a rule the caterpillars of each species are found on a par- ticular plant, and cannot be induced to eat the leaves of a differ- ent one, or only with great difficulty. A few species, however, are polyphagous, i.e. they feed on a number of different plants. For example, the caterpillars of the Arctiidse feed upon all sorts of herbaceous plants; many Noctuidse consume indifferently several species of Compositae ; Papiho macchaon lives on dif- ferent Umbelliferae ; Ocneria dispar, Porthesia chrysorrhaea, and Bombyx neustria are found on nearly all kinds of trees. ^ Occa- sionally caterpillars are found on plants that are not those nor- mal for them, and the question has often been asked whether the aberrant types of butterflies sometimes met with may not have arisen in consequence of a change in the food plant. This question Pictet has studied experimentally. In captivity certain caterpillars adapt themselves readily to very different kinds of food. As a rule a caterpillar will feed on the flowers of its natural plant, and even the fruit may be used, as in the case of Cossus cossus, which will eat pieces of apple instead of the wood and the bark of the tree. It appears that in nature, also, certain species have recently extended their dietary. Thus Lasiocampa quercus was known at the time of Linnaeus to feed on the oak tree (as its ^ These examples are given by Pictet. 29 30 Experimental Zoology name implies) and on the leaves of certain shrubs. It is now found on a number of other trees. Certain aberrations have been traced directly to the food of the caterpillar. The caterpillars of Ellopia prosapiaria living on the pine become reddish butterflies ; but if they, occur on the fir, they give rise to the aberration prasinaria, which is green. An analogous case is that of Cidaria variata, whose caterpillars living on the fir give rise to a form that is gray, but on the pine produce a variety obeliscata, which is brownish red. In certain countries where a variety almost entirely replaces the parent species it is not uncommon to find the caterpillar on a different plant. For example, Lasiocampa quercus fives on different food plants in different countries. In Scotland, where the variety callunae dominates, the caterpillars five on the heather ; in the South, where the variety roboris is found, the caterpillars five on the oak, Quercus robur. In changing the food, Pictet often made use of plants that were very different from the natural one. Thus the oak was often replaced by the esparcette,^ by the dandelion, by the lettuce or by the pimprenelle.^ In other cases it was replaced by the walnut, neflier,^ and the sorbier.^ In some cases a food plant closely related to the normal was the only one that could be substituted ; thus the European Evonymus europaeus (spindle tree) can be replaced by E. japonicus. Certain species refuse aU food but their natural one. Other species can be induced to take a different nourishment, but only after much persever- ance. Others, such as Lasiocampa quercus, when young ac- commodate themselves to nearly all kinds of vegetation, but once full size, many plants are rejected. The acceptation of a new kind of food is transmitted by heredity, and individuals whose parents have become accustomed to a strange food will consume the same food with greater facility. Some of the effects that Pictet obtained by changes in the food are as fol- lows : — * Onobrychis saliva, or holy clover. ^ Mespilus germanica, or mespit. * Poterium sanguisorba, or salad burnct. ■* Sorbus ancuparia. The Influence of External Conditions 31 The typical male and female moth of Ocneria dispar are shown in Fig. i and Fig. 2. In the male the wings are gray, ash, or dusky, with four zigzag black lines. The female is whitish gray or shghtly yellow, with the same grayish white Fig. 4. Ocneria dispar : male, Fig. i ; and female, Fig. 2. First generation fed on walnut leaves : male. Fig. 3; female. Fig. 4. Second generation fed on walnut : male. Fig. 5 ; female. Fig. 6. First generation on walnut, second on oak : male. Fig. 7; female. Fig. 8. First generation on walnut, second on oak, third on wal- nut, fourth on walnut : male. Fig. 9. (After Pictet.) pattern as .the male, but more marked. The normal food is oak or birch. The young caterpillars can be made to eat the leaves of the walnut (Juglans regia), at first with difficulty, but the subsequent generations eat it with avidity. The male of the first generation is shown in Fig. 3. The wings are pale yellow, the 32 Experime7ital Zoology central lines have partly disappeared, and the rest of the pattern is less marked. The female of this first generation is shown in Fig. 4. The wings are slightly transparent, with rarely a darker mark on the upper surface. The insect is almost white. The caterpillars of the next generation were also fed on the walnut. The male is shown in Fig. 5.. The wings are whitish, the marginal band has partly disappeared, and the transverse .lines are but slightly visible. The female is shown in Fig. 6. The wings are transparent ; the V and the fifth of the marginal points alone remain of the markings. The figures show a decided decrease in the size of the indi- vidual, which is accompanied with a constitutional weakness. It was not possible to obtain eggs from this generation. Pictet resorted, therefore, to the device of feeding the first generation on the walnut, the second on the oak, and then the third and even the fourth on the walnut. When the first generation was fed on the walnut and the second on the oak, the male moth appeared as in Fig. 7, and the female as in Fig. 8. The fig- ures show that the effects, produced by feeding the first genera- tion on the walnut, persist during the second generation when returned to the normal food. In another experiment the caterpillars were fed first on wal- nut, then on oak, and then on walnut. The male was much like that produced by two generations of walnut, as were the females also. Another lot was fed walnut (ist gen.), oak (2d gen.), walnut (3d gen.), walnut (4th gen.). The male type is shown in Fig. 9. The wings are gray ash, or dusky; the marginal points strongly marked, as well as most of the trans- verse lines. The marginal band is accentuated in the four wings. The females have white wings — sometimes yellowish. As in the male, the lines and the marginal band of the lower wings is much accentuated. Thus while the greatest change is effected by walnut, oak, walnut, there is a return to the darker type if walnut is used again for the fourth generation. Pictet interprets this as a return to the ancestral type; the result of becoming accustomed to the walnut, and states that he has ob- The Injiueiice of External Conditions 33 tained similar results in other experiments. It is not entirely clear from his figure that the darker form in the fourth gener- ation is really a return to the ancestral type — at least the possibility of a different interpretation must be kept open. Another experiment showed that when the first generation was fed on walnut, and .the second and the third on oak, the effects of the walnut were still apparent in the last generations. When the first generation was fed on walnut, the second on the oak, and the third on flowers of different kinds (rose and peony), the last seemed to accentuate the effects of the walnut and “tend to cause to disappear those of the normal food. To the same extent as the walnut the flowers appear to be a poor alimentation.” The caterpillars of Ocneria were also fed on other plants, the results being in some respects like the pre- ceding cases, in others different. There are two known aberrant forms of the moth Psilura monacha. The typical male form is shown in Fig. 10, and the female in Fig. ii. The aberrant form nigra, male and female, is shown in Figs. 12 and 13, and the form, eremita, male and female, in Figs. 14 and 15. The caterpillars are found on oak, birch, and conifers. Caterpillars fed on oak and on birch produced moths of all three types, in the proportion of 58 per cent type-form; 24 per cent ab. nigra, and 18 per cent ab. eremita; fed on walnut the proportions were 38 per cent type-form; 23 per cent ab. nigra; 30 per cent ab. eremita. The moths were smaller than the aver- age. The proportion of the aberrant forms is greater from the walnut than from the normal food. In the second generation, nourished on the walnut, the females of the type form assume the characters of the typical males. In all, Pictet’s experiments with feeding included 21 species, comprising 4695 individuals. In nearly all cases some effects of a change from the normal type could be directly traced to the food.^ ' For somewhat similar experiments see Romanes, “Darwin and After Dar- win,” II, pp. 217-218. D 34 Experimental Zoology Pictet has also tried the effects of insufficient nourishment, using smaller amounts of the normal food. In the case of Pieris crataegi, lack of food during the last days of the life of the caterpillar causes the wings to become lighter in color and more transparent. If all food is removed from the full-grown Fig. 5. Psiluria monacha: typical male, Fig. lo; and female, Fig. ii. Aber- rant form P. ingra : male. Fig. 12; female. Fig. 13. Aberrant form P. eremita : male. Fig. 14; female. Fig. 15. (After Pictet.) caterpillar of Vanessa urticae, it becomes a chrysalid in a few days after the last moult. The butterflies are dwarfs, but show no changes in coloration. On the other hand, if the very young caterpillars are kept on a regime that becomes less sufficient every day, they become pup® even before the last moult, and produce dwarf butterflies that have the aberrant characters much accentuated. The Influence of External Conditions 35 It is not difficult to starve caterpillars, but it is difficult to make them take more food than they do normally; yet by indirect means this can be done by giving them food containing an excessive amount of nutritive substances. Reviewing the results of the effects produced by changes in food, Pictet points out that, in general, the variations produced are either in the direction of lighter (albinistic) or darker (mel- anistic) shades. If the kinds of food that produce the former effect are compared with those that produce the latter effect, it will be seen that the lighter colors result from feeding on plants which, owing to their anatomical structure, present obstacles to mastication — either as a result of a thick epidermis or of the pres- ence of crystals or of hairs, etc. In consequence of the imperfect nutrition of the caterpillar, the butterfly is less pigmented. On the other hand, the melanistic or black variations result from those plants that offer no obstacles to mastication or nutrition. The caterpillars develop rapidly, acquire a size greater than the normal, and the butterfly shows a greater development of pig- ment. The influence of different food plants in causing an albinistic or melanistic change in the moths can be traced, according to Pictet, to the length of the pupal stage, and this in turn to that of the caterpillar stage, which again is due to the amount of nourishment received. Thus, those food plants that give in- sufficient nourishment prolong the caterpillar stage ; but this leads to a shortening of the pupal stage, and the albinistic effect is produced, because the dark pigments do not have time to develop extensively. Conversely, an abundant nourishment shortens the caterpillar stage ; but this causes a prolongation of the pupal stage during which the pigments have time to develop even more fully than normally, and the melanistic variation is the result. Not only the colors of the moth or butterfly are affected by the character of the food, but the colors of the caterpillars them- selves may sometimes be affected, although to a less degree. Pictet finds, in fact, that there is a correlation between the pig- 36 Experimental Zoology mentation of the caterpillar and of the adult. The caterpillar of the death’s head moth (Acherontia atropos) is ordinarily yellow, but has also a black type. Dasychira pudibanda is white, but has a green and gray variety, etc. The color of many caterpillars is much influenced by the food in the digestive tract that can be seen through the semi- transparent skin. Poulton has also shown that certain of the epidermal pigments are derived from the chlorophyll absorbed with the food. He separated into three groups caterpillars of Agrotis pronuba: the first was fed with the green leaves of cabbage, the second with the yellow etiolated leaves. Both lots developed their normal yellow-brown color. A third group was fed on leaves deprived of the green and of the yellow coloring matter, and only the brown pigments developed. Standfuss has produced within a few hours color changes in the caterpillar of Eupithecia absinthiata. It becomes “a lemon-yellow when fed on the yellow bunches of Solidago, green on non-flowering leaves of the same plant, rose on the ‘ buttons ’ of Statice armeria, white on the umbels of Pimpinella saxifraga, brown on the bouquets of Artemisia vulgaris, and a delicate blue on the small balls of Succisa pratensis.” Similar results are known for other forms. Pictet finds also that the colors in certain caterpillars that hibernate tend to disappear, but reappear again a few days after feeding begins again in the spring. The colors of some species change with each moult. Difference in the color, and even in the markings of caterpillars, occur when they are fed on food other than the normal. In some cases the effect is direct, as explained above; in other cases the effects are indirect, i.e. not due to the food in the digestive tract or to the pigments directly absorbed and present in the skin or in the blood. In the case of Ocneria dispar the food plants that produce lighter colors in the moths have the same effects on the caterpillar. Similarly for the dark or mel- anistic changes. Moreover, the effects may last over for one or two generations after the caterpillar is again fed on its normal food. The Injitience of External Conditions 37 The Influence of Light, Electricity, Centrifugal Force, Chemical Substances, and Oxygen on the Caterpillars and Pupa of Moths and Butterflies It has been shown by Grafin von Linden that the size, the colors, and the markings of butterflies may be altered by sub- jecting the caterpillars or the pupas to several kinds of external conditions other than temperature. In general, however, the results are similar to the effects produced by higher and lower temperatures. Light. — Vanessa urticae and V. io were used. Some of the caterpillars were kept, with their food, in red, or green, or blue light. Others were kept in the dark. The principal changes were in the ground color of the wings of the butterflies. This was intensest and brightest in red light, dusky in green, and paler in blue and in the dark. The butterflies of V. urticae were largest from the caterpillars reared in the blue, and of V. io in the dark. The changes in the markings were very slight. Electric Shocks. — The pupae of V. urticae, in a fresh but dry condition, were put into an iron box through which an electric current was passed of sufficient strength to cause a pricking sensation when applied to the hand. In other cases one elec- trode was applied at the wing axis and the other at the tip of the wing case. The resulting butterflies were brightly colored. The black border of the wing was broader, and the tip of the wing was sometimes dark. The blue and yellow scales of the sides of the wing were little developed. Centrifuge. — After the pup£e had become hard they were subjected to a centrifugal force for ten minutes each day. The effects were in general similar to those caused by light. Chemical Substances. — Contrary to the results of Standfuss, who found changes in the food had no influence on the color ! of the moth, except in one case, when salt was added to the food of Callema, Grafin von Linden found that certain substances given with the food produced distinct effects. Some effects, for [ the most part slight, were obtained with (i) defibrinated blood; I I 38 Experimental Zoology (2) iron albuminate (officinal solution: 4 parts metallic iron to 1000 water); (3) argonin silver casein (5 per cent); (4) sugar; (5) lupulin ; (6) capsicum ; (7) morphine (i per cent) ; (8)atropin (i per cent). The iron and the silver compounds by exciting hunger tended to produce larger forms. The largest butterflies came from caterpillars fed with argonin silver casein and the smallest from those given morphine with their food. The younger the caterpillars, when the experiment began, the better the results. Darkening of the ground color of the wings was found after argonin and morphine. Also much red was present. Capsi- cum also gave a dark background. The color was strongly developed after iron, sugar, and lupulin. The markings were affected in much the same way as by cold and warmth. Re- duction of the black or of the blue flecks occurred in some cases. Pure Oxygen. — Pupas kept in pure oxygen produced in the usual time butterflies normally colored and marked. Young caterpillars kept in oxygen took food for only one day, wandered about the next day, and died on the third or fourth day. The moist atmosphere rather than the oxygen may have been respon- sible for the early death of the caterpillars. The Influence of Humidity on the Characters 0} Moths and Butterflies Humidity is also supposed to have an effect on the coloration of butterflies. Marshall thinks that in the Transvaal, where a dry and a wet season alternate, as do summer and winter in northern climates, the seasonal dimorphism of certain butterflies is the direct result of the effects of moisture and of dryness, in the same way that cold and heat cause the seasonal dimor- phism of northern species. Similar results have been described by Doherty for Melanites leda of India. The differences fol- low the dry and the wet seasons. Pictet has shown that effects of this kind can be artificially produced in certain Eu- ropean species. Moisture on the leaves produces great mortality 39 The Influence of External Conditions amongst the young caterpillars, but while the moths may be small their markings show no variations. On the other hand, although the fully formed caterpillars better resist the effects of moisture, aberrations in the color are produced. Thus when the caterpillars of Vanessa urticae were fed for 8 to lo days on leaves kept constantly wet, changes in the markings of the butterflies were produced. Similar results were brought about in Vanessa polychloros and in Hybernia defoliaria. In Ocne- ria dispar variations were induced in the first generation, but disappeared in the second, owing, Pictet thinks, to the cater- pillars having adjusted themselves to the change, in the same way as they do to changes in their food. The variations in- duced in the first generation are very much like those pro- duced by esparcette and dandelion, especially in respect to the males, and these plants contain a great deal of water in their tissues. A fine spray of water at ordinary temperatures was con- stantly applied for 36 to 48 hours to caterpillars of Vanessa urticas that had suspended themselves preparatory to changing into chrysalids. A marked effect was produced on the but- terfly, a yellow band appearing across both fore and hind wings. The blue spots became gray or violet, and the border of the wings was clear. The variation resembles the variety polaris of this butterfly. Humidity also acts on the chrysalid’s stage, causing aberra- tions in a number of moths and butterflies. If the chrysalids are kept moist at a warm temperature, 30° to 35° C., very little effect is produced, because the development is so much has- tened that the chrysalids pass through the critical stages be- fore the protecting waxy covering is worn away. But if the chrysalids are kept cool while the moisture is applied, the effects are much more marked; for now the development is so much delayed that the water has time to penetrate the protecting coat and affect the critical stages. 40 Experimental Zoology Experiments with Flatfish There are some experiments made by Cunningham with young flatfish which also show the effect of an external agent, viz., light, on the development of color. Very young fish were put into aquaria lighted from below. As the young fish. under- went their metamorphosis, the pigment gradually disappeared on one side, as it does under normal conditions, i.e. as when they are lighted from above only. If, however, the fish are still illuminated from below, the pigment begins to come back again on the lower surface. The markings are similar in all respects to those on the upper side of the animal. The result shows that the lower side of flatfish in their natu- ral environment is white, because it is not exposed to fight ; but whether the result shows, as Cunningham believes, that the lower side has become white in the course of generations, be- cause it has been turned away from the fight, is not shown conclusively by the experiment. Experiments with Crustaceans There are a few cases, in other groups, where it has been shown that external agents produce changes in form. Bran- chippus ferox, inhabiting salt and fresh water, shows small differences in the length of the ovigerous sac, in the form of the segments of the body, in the length of the lobes that terminate the abdomen, and in the disposition of the abdominal bristles. Daphnia degenerata of salt water is only a degenerate variety of Daphnia magna of fresh water. A species related to Daphnia — Moina rectivoctris — oc- curs in one form in fresh water and in another in brackish. The two differ in general points of structure due to the indi- viduals becoming sexually mature before the final structural changes are completed. The oft- cited case of Artemia may also be mentioned here. Schmankewitsch has described the slow transformation of Artemia salina into A. milhausenii, as the lagoon in which the The Influence of External Conditions 41 c former was contained became more concentrated by evapo- ration. Experimentally he procured the same result by con- centration. Furthermore Schmankewitsch claimed that by gradually diluting the brackish water in which Artemia salina lives he obtained a form having the characters of the genus Branchippus. These conclusions have been seriously questioned by Bateson and by Samter and Heymons/ who have shown that Artemia salina is subject to great individual variability and that there is no close connection between the different variations and the concentration of the water in which they occur. Especially doubtful is Schmankewitsch’s comparison with Branchippus, whose diagnostic features he seems to have imperfectly under- stood. Samter and Heymons find nevertheless that the salt content of the water has some influence on the form of Artemia salina, although in different pools of the same concentration a large range of variability exists. They think that other fac- tors than concentration probably also affect the result. Changes in Mammals and Birds The Porto Santo rabbits, so fully described by Darwin, fur- nish another instance of influence of the environment. It is said that these rabbits originated from a single pregnant female that produced a litter on board ship in the year 1418 or 1419. Set free on the island of Porto Santo, the rabbits increased rapidly and soon became a pest. Darwin examined these rabbits and found that, compared with domesticated rabbits of average size, the Porto Santo rabbits had lost three inches in length and almost half the weight of the body. In other points also they differed — in the skull, and especially in color. “But here we meet with a singular circumstance: In June, 1861, I examined two of these rabbits recently sent to the Zoological Gardens, and their tails and ears were colored ' Schimkewitsch {Biolog. Cenlralhlatt, XXVI, 1906) states that Anikin (1889) and Butschinsky (1901) have obtained results contradictory to those of Schman- kewitsch. I have not seen their papers. 42 Experimental Zoology as just described {i.e. like the feral rabbits of Porto Santo); but when one of their dead bodies was sent to me in February, 1865, the ears were plainly edged, and the upper surface of the tail was covered with blackish gray fur and the whole body was much less red, so that under the English climate this in- dividual rabbit had resumed the proper color of its fur in rather less than four years.” ^ Darwin thinks that “from the direct action of a humid cli- mate and poor pasture the horse rapidly decreases in size in the Falkland Islands.” There is a great deal of evidence to indicate that cHmate acts directly in bringing about a change in the hair of animals, and not only is the thickness of pelt affected, but also the character of the hairs. Geoffroy St. Hilaire states that horses that have lived for several years in deep coal mines come to have a vel- vety hair, somewhat like that of the mole. Wild ducks are said to change under the influence of do- mestication. The collar around the neck of the mallard be- comes broader and more irregular. White feathers appear in the wings. The birds also increase in size, etc. Certain foods affect the color of the feather of birds. Hemp seeds cause the color of bullfinches to become darker.^ Wal- lace states that the natives of the Amazonian region feed the common green parrot with the fat of siluroid fishes, which causes it to become variegated with red and yellow feathers. How far these effects, produced directly by the environment, may become inherited will be examined in the next chapter. ‘ “Animals and Plants,” Chap. V. 2 Romanes, in his book on “Danvin and After Darann,” further cases of this sort, p. 217. II, gives some CHAPTER IV THE INHERITED EFFECTS OF CHANGES INDUCED BY EXTERNAL FACTORS It has been pointed out that in butterflies the changes brought about by higher and lower temperatures give rise to forms that resemble southern and northern varieties of the same butter- flies. The question at once arises whether species may not have originated in this way. Fischer found that when a dark moth, produced by cold, was paired with another similar moth, the offsprings were also dark. His experiments extended, however, only to the first generation, and consequently the cold may have acted directly on the germ-cells of the parents. Highly important as this observation is in showing that the undeveloped germ-cells may be affected in the same way as the somatic tissues of the pupa, so that even under altered conditions the effect persists, yet the result as it stands does not conclusively show that permanent racial or specific changes have been produced in this way. Were such forms bred for several generations at a warmer temperature, it is possible that they would return again to their original condition. The new type might persist only so long as, or a little longer than, the external conditions are the same as those that produced it. The change may represent only an extreme fluctuating variation that has been caused by an ex- ternal factor. The results do not appear, from this point of view, to belong to the kind of changes by which new species are made. Nevertheless the question still remains an open one as to whether changed external conditions may not at times cause more permanent effects. Some observations of Standfuss seem to show that such may be the case, although 43 44 Experimental Zoology the evidence is not satisfactory in all respects. He states that certain aberrations, that occur when the external factors have produced a sudden divergence from the parent form, are in- herited. He shows by numerous experiments that these aber- rant forms when crossed with the parent type do not give intermediate varieties, but that the offspring correspond to one or to the other parent. Thus the new type is not lost by inter- crossing; but if adaptive, might perpetuate itself and form the beginning of new species. The aberrations seem to resemble in these respects mutations. If they are really produced by a change in the environment, as Standfuss believes, the results throw most important light on the question of the origin of mutations — a question which at the present time is one of the most pressing questions of the theory of organic evolution. Standfuss believes that effects of these kinds are inherited, and that new species may be evolved at the limits of the range of forms through the effect of external agents. It is not quite clear to me, how he supposes such results come about, since his own experiments seem to show that the effects of moderate changes in temperature are only temporary. He appears to believe, however, that the effects, if long enough carried out, become in large part fixed. In plants the effects of the environment on the form, growth, and time of flowering have long been known; and many ex- periments have been made by transporting plants from one locality to another. It has been shown that many alpine plants will flourish in the valley, and often great changes in the char- acter of the plants have been noted. Naegeli has carried out elaborate experiments of this sort. He collected in the Botani- cal Garden of Munich 2500 varieties of mountain plants, and for several years made observations on the effects of changes in the environment. The changes are seen at once, showing that no permanent effects have been produced by the alpine climate. Other botanists, however, claim that for a time at least the effect of the original habitation may be seen. 45 The Inherited Effects of Changes Pictet, as has been shown in the preceding chapter, has also obtained direct evidence of the inherited effects of food. It should be noticed in this case that the young caterpillar fed on the leaves of a new food plant produces changes of a certain type in its somatic cells that appear in the butterfly. The caterpillars of these butterflies, reared on their natural food plant, produce butterflies that also show evidence of the effect caused in the preceding generation. It is probable that this influ- ence was directly induced in the germ-cells of the first genera- tion, so that the effects were not inherited through the soma, but were directly produced. The point, however, of special interest in these cases, aside from the question of inheritance, is that the influences that induce certain changes in the somatic cells of the caterpillar affect the germ-cells of that caterpillar in the same way, so that when they develop they, too, give the same results. The effect is weakened, it is true, in the second genera- tion; but this may be due to the counteracting influence of the normal food. The results show that the influence of the envi- ronment may persist for one or more generations in another environment. Recently de Vries has dealt with the same subject, and has carried out certain experiments that bear on the question. He believes that in general the effect of the environment produces only the fluctuating variations seen in plants. The full effects may not appear at once, but may be accumulated, through sev- eral generations, and hence would seem to be inherited. He speaks of these changes as acquired characters, and believes, as I have said, that individual variations are simply acquired characters due to differences in the environment. We are to understand, however, that de Vries means that these characters are acquired either by the somatic cells or by the germ-cells, but independently of each other, i.e. the effects acquired by the somatic cells are not supposed to affect the germ-cells except indirectly, namely, by affecting the nourishment of the germ- cells. The question arises whether these new characters are inherited 46 Experimental Zoology when the environment that has caused them is altered. De Vries answers this question in the following way: If the seeds of the best-nourished ordinary plants are selected and planted under the most favorable conditions, the average plants of the next generation will not only be as vigorous as the former, but their seeds themselves will have a higher average of size. If this process is continued, the average of the race will be in- creased in the direction of selection, and in time the race, i.e. all the individuals, may be brought to and maintained near the highest plane to which fluctuating variation ever reaches. There is a counteracting principle that must also be con- sidered, viz., what Galton has called regression toward medi- ocrity. If variations that depart from the average of the type are united, their descendant will tend in some degree to return to the average or to mediocrity, as Galton has shown. If, however, we pick out in each generation those that are most above the average, the average of the descendants in each gen- eration rises, despite the regression, and in time the average may be brought near to the highest point to which individual varia- tions reach. In other words, the character that appeared in the first generation as a somatic variation of one individual has become temporarily transferred to the seeds; but unless the same favoring external conditions are vigorously maintained, there will be a return to mediocrity. This fastening, as it were, of the individual variation upon the race is due, de Vries thinks, to the action of the nourish- ment on the germ-cells. If a plant is well nourished, it produces larger seeds that are better supplied with nourishment. Hence on an average they will give rise to more vigorous plants, and these in turn will produce an average of larger or at least of better-nourished seeds, and the plants from these will be still stronger. Tlius by slow degrees the seeds acquire the same sort of characteris- tics as those shown by the best-nourished plants in any one generation. The curious thing about this is that the seeds do not respond completely in the first generation, but it takes sev- The Inherited Effects of Changes 47 cral generations to produce seeds all of which contain nearly the maximum possible food supply. Suppose when this condition is attained we reverse the pro- ' cess and plant the seeds under poorer conditions. The first crop from such seeds will still be above the original average, but not as much above as though they had grown under favor- able conditions. Their seeds again will contain less nourish- ment, and the second generation will be still smaller on the aver- age, until finally the race reaches- the lowest level shown by fluctuating variations. The inheritance of the acquired character in these cases is brought about in rather a peculiar way. The favorable exter- nal conditions, for instance, act favorably on the body-cells of the plant, and the germ-cells are therefore well nourished and store up more food : not all to the same extent, but the average is higher than in the preceding generation. The next genera- tion thus gets a better start, and if the plants are better nourished than the average, a similar advance is again made. We must not forget that in each generation some of the seeds may be as good as the best, but others are not as good. It is the average that is improved, and not the best individual seed. It may appear from the case just given that the food condi- tions determine only the size and vigor of the plant. This is not the whole matter, for even the characters of certain parts may be changed in a plus or a minus direction, provided they are correlated with the condition of nourishment of the plant. De Vries cites the case of the poppy, Papaver somniferum. In this plant the stamens may be changed into supernumerary or accessory carpels by changing the external conditions. The number of these carpels may be as great as 150 or more. From the seeds of flowers with an average crown of carpels, plants may be reared having many or few of these organs. The more favorable the conditions, the more numerous the carpels, and vice versa. Poorly nourished plants may have only one or two rudimentary or accessory carpels. If we select in each generation the most vigorous plants (hav- 48 Experimental Zoology ing, therefore, a high average of carpels), and keep them under favorable conditions, the average of carpels of the race can be temporarily increased. Conversely, if we select the less vig- orous plants, and put them under unfavorable conditions, the accessory carpels can be made to disappear. This connection between individual vigor and the forma- tion of carpels is not an absolute one, for there are means by which vigorous plants can be obtained without accessory car- pels. If very young plants are transplanted, and then put under favorable conditions, vigorous plants result without or with few carpels. The interpretation that de Vries gives of this result is that the flower buds were not laid down at the time of trans- planting, but develop soon afterward, when the conditions are temporarily unfavorable. Later, when the favorable conditions begin to act, the rest of the plant responds, but it is too late to affect the carpel formation. It will be observed that we are dealing here with rather a special case, viz., that of nourishment alone. Somatic cells and body- cells alike are affected in the same way. If along with this condition of nourishment there are certain correlated changes, such as the formation of accessory carpels, the prin- ciple remains the same. The point of special interest is that the effects may be accumulated only slowly by the seeds, so that it takeg several generations to produce the best average effects. The effect, once produced, may persist in part through several generations subjected to the reverse conditions. The results are not unlike those in the butterflies, in which the effects of temperature or of feeding are marked in the first generation, and then decline if the external conditions that produced them are changed. It is difficult to reach any probable conclusion from the evi- dence given in the preceding pages in regard to the inherited effects of the influence of the environment. Possibly we are dealing with two distinct problems. In most cases the effects on the body-cells and on the germ-cells are only temporar}’, and persist only as long as, or a little longer than, do the condi- 49 The Inherited Effects of Changes tions that called them forth. The result is only a special case of the inheritance of fluctuating variations. , On the other hand, in some of the cases of aberrations given by Standfuss, it ap- pears that a permanent change has been brought about by the environment which persists afterward, independently of the influences that caused it. What is still more interesting is that the new characters may be transmitted to the offspring formed by a cross between the new and the parent type. We shall consider this question at greater length in a later chapter. LITERATURE, CHAPTERS III AND IV Bateson, W. Materials for the Study of Variation. Pages 96-101. London. 1894. CufiNOT, L. L’influence du milieu sur les animaux. Cunningham, J. T. An Experiment concerning the Absence of Color from the Lower Side of Flatfishes. Zool. Anz. 1891. The Evolution of Flatfishes. Nat. Sc. I. 1892. Researches on the Coloration of the Skins of Flatfishes. Jour. Mar. Biol. Assoc. III. 1893. Cunningham, J. T., and MacMunn, C. A. On the Coloration of the Skins of Fishes, especially of Pleuronectidae. Phil. Trans. CLXXXIV. B. 1893. 1 Darwin, C. Animals and Plants under Domestication. Dewitz. Zur Verwandlungderinsectenlarven. Zool. Anz. XXVIII. 1904. i Gauckler, H. Untersuchungen ueber beschleunigte Entwickelung ueber- winternder Schmetterlingspuppen. 111. Zeitsch. f. Entomologie, 1 IV. 1899. 1 Pictet, A. Des Variations des Papillons provenant des changements d’alimentation de leurs chenilles et de I’humiditd. Comptes rendus du Congr&s intern, d. Zool. Session d. Berne. 1904. < Influence de TAlimentation et de I’Humiditd sur la Variation des Papil- lons. Mem. Soc. Phys. et Hist. Natur. de Geneve, XXXV. 1905. PoucHET, G. Des changements de coloration sous I’influence des nerfs. Jour. Anat. et Physiol, norm, et pathol. XII. 1876. Samter, M., and Heymons, R. Variationen bei Artemia salina Leach, und ihre Abhangigkeit von aeusseren Einfluessen. Abh. d. k. Preuss. Akad. d. Wissen. Anhang. 1902. ‘ Thilenius, G. Die Farbenwechsel von Varanus griseus, Uromastix acanthinurus und Agame inermis. Morph. Arb. VII. 1897. DE Vries, H.^ Die Mutationstheorie Versuche und Beobachtungen ueber die Entstehung der Arten im Pflanzenreiche. I. Die Ent- stehung der Arten durch Mutation. Leipzig. 1901. Die Mutationen und die Mutationsperioden bei der Entstehung der Arten. Vortraz. Vers. Deutsch. Naturf. u. Arzte zu Hamburg. 1901. E CHAPTER V THE INHERITANCE OF ACQUIRED CHARACTERS In the preceding chapters we have seen that external factors may cause definite changes in the form, color, markings, etc., of animals, as well as certain changes in plants. The question arose whether these effects are transmitted to the offspring of the next generation. An examination of the evidence seemed to show that in most cases the effects are inherited only in so far as the germ-cells are also affected by the external factors. It has been long recognized that internal factors also may cause changes in animals. The use of an organ may increase its size, and also its effectiveness, even when the change in form is so slight as to escape notice. Here, also, the question arises whether these effects are inherited. Disease or injury may bring about changes in an organ, and again the question has been raised as to whether the effects are transmitted to the offspring. The phrase “the inheritance of acquired characters” is used to include supposed cases of inheritance of these different kinds of effects; and it is customary to use this term to include also cases like those described in the last chapter, in which the e.x- fernal agents act directly on the germ-cells, as well as on the body-cells. In this chapter I shall use the term in a restricted sense, and include under it only those cases in which the body- cells are first affected and are then supposed to transmit their influence to the germ-cells. It is not necessary to consider at length the historical origin of the idea that acquired characters are inherited. It is well known that Lamarck based his theory of evolution on this 5° The Inheritance of Acquired Characters 51 assumption, not only in the restricted sense as defined above, but also in the sense that external conditions may directly affect the whole organism, i.e. germ-cells as well as the body- cells. He made no such sharp distinction, however, between these two sides of the question as we find it convenient to make nowa- days. Let us examine critically the experimental evidence on which the theory of the inheritance of acquired characters, in the restricted sense, rests, for, could it be shown that changes acquired through use are transmitted to the next generation, we might seem to be able to explain how many of the complicated adaptations and coordinations of animals have arisen. Amongst modern writers the first to seriously question the truth of the generally admitted doctrine that acquired charac- ters are inherited was August Weismann. In his essay “On Heredity,” published in 1883, and in two subsequent papers “On the Supposed Botanical Proofs of the Transmission of Acquired Characters” (1888), and “The Supposed Transmis- sion of Mutilations” (1888), Weismann challenged the ac- cepted point of view. His conclusions have become common i knowledge, so that it will not be necessary to go over the ground I again. While many zoologists and botanists were convinced I by Weismann’s argument, which seemed to show that the evi- I dence fails to support the view that acquired characters are i inherited, a few zoologists have always insisted nevertheless i that such characters are inherited, and a few investigators have \ brought forward experimental evidence which they believe ^ shows convincingly that changes brought about in the body ii may be transmitted to the germ-cells. ’ It is this experimental s evidence that I propose especially to consider. ■ Darwin believed that acquired characters are inherited, and i in the “Origin of Species,” in the “Animals and Plants under Domestication,” and in the “Descent of Man” he accounts for a many adaptations in this way. Herbert Spencer, especially, l| although not always with discrimination, used this hypothesis to c account for many structures and habits of animals and plants. 52 Experimental Zoology I do not propose to consider in detail the cases that Darwin and Spencer have brought forward (most of them will not bear critical examination, as Weismann has so ably shown), but, as has been said, I shall consider rather those cases, most of them recent, in which attempts have been made by direct experiment to show that acquired characters are inherited. The work that has attracted most attention is that of Brown- Sequard. A full statement of Brown- Sequard’s experiments and results is given by Romanes in his book “Darwin and After Darwin,” Vol. II, Chap. IV. The experiments were made with guinea pigs. Epilepsy was induced by operations on some part of the nervous system. The young of these animals sometimes developed epilepsy, or some of its effects, in the same part of the body as that affected in the parent. The details of the experiments are as follows : The parents became epi- leptic after injury to the spinal cord or by section of the sciatic nerve. The “epileptiform habit” does not supervene until some time after the operation and lasts only “for some weeks or months.” The convulsions “never occur spontaneously, but only as a result of irritating a small area of skin behind the ear on the same side of the body as that on which the sciatic nerve had been divided.” ^ The attack lasts only a few minutes, and during this time the animal is unconscious and convulsed. The habit is only rarely transmitted to the young, but as the disease occurs only in guinea pigs whose parents have been made epileptic by an operation of the sort described above, and never in the young of guinea pigs that have been operated on in other ways, there seems to be here something more than a coincidence. Another series of experiments consisted in cutting the cervi- cal sympathetic nerve. This operation causes a change in the shape of the ear, and a similar change is said to appear in the young. By cutting the cervical sympathetic nerve or by re- moving the superior cervical ganglion, the eyelids partially close, and this closure was also seen in the young. ^ Romanes, “Darwin and After Darwin,” II, 1895. The Inheritance of Acquired Characters 53 If a particular spot of the restiform body of the brain is injured, a marked protrusion of the eyeball quickly follows. The progeny of parents thus affected show also an abnormal protrusion. Romanes also observed this, but he found that the young show less protrusion than do the parents, and since the amount of protrusion of the eyeball is variable in normal guinea pigs, Romanes is not certain that there is anything more than a coincidence in the cases that he observed. An injury to the restiform body may also cause dry gangrene (and haematoma) in the ears. This disease may appear either several weeks after the operation or even later. It affects, Romanes says, usually the upper parts of both ears and may gradually “eat its way down” until two' thirds of the tissues of the ears are affected. In the offspring from animals of this sort . a morbid condition of the ears may arise at any time in their ,.i lives, even after they have become full grown. The disease > does not go so far as in the parents, and “almost always affects II the middle third of the ears.” Romanes points out that this i particular disease never appears amongst guinea pigs unless • their own or their parents’ restiform bodies have been injured. 1 Furthermore, he tested the possibility that the results are due tto contagion by inoculating “corresponding parts of the ears of normal guinea pigs by first scarifying those parts and then rubbing them with the diseased surfaces of the ears of muti- ' lated guinea pigs.” The disease was not communicated in this way. Brown-Sequard found that after cutting the sciatic nerve (or this and the crural also) the leg became anaesthetic, and the •guinea pigs would sometimes eat off two or three of their hind loes. In the offspring of these animals he found sometimes an absence of toes, or only a part of one or more of the toes might be missing. The inheritance occurred in only one or two per cent of cases. Romanes, who repeated the operation Ithrough six successive generations, never obtained any results. Another outcome of injury to the sciatic nerve is to induce i ‘morbid states of the skin and hair of the ndck and face in 54 Experimental Zoology animals.” This result is also said by Brown-Sequard to be inherited. I have given somewhat fully these remarkable results of ' Brown-Sequard because the experiments appear to have been f carried out with such care, and the results are given in such V detail that it seems that they must be accepted as establishing * the inheritance of acquired characters. t Moreover, similar results have been obtained by other in- | vestigators. They have been corroborated in part by Ober- j Steiner.^ Westphal has produced epilepsy by striking the heads | of the animals with a hammer, and has found that the young V are often epileptic.^ Still more important are the experiments | of Romanes. His conclusions, it is true, are much more can- : tious, and his statements more guarded than those of Brown- ' Sequard; yet on the whole they confirm Brown-Sequard’s claims. Weismann has attempted to discredit these results on the ^ ground that we are still ignorant of the cause of epilepsy. The | possibility that it is a bacterial disease must be admitted, he | claims, and, if this is the case, the bacteria themselves may be I transmitted to the young during their uterine existence. It is | supposed, in fact, that other diseases may be inherited by direct | contagion through the germ- cells of the father or the mother, j This objection is, however, purely formal, and as long as we do | not know that epilepsy is a bacterial disease that is contagious v in the way supposed, the objection may raise a doubt but cannot 'f set aside the results.^ There are some quite recent experiments that may have a '■ very direct bearing on the questions here raised. In a paper j by Charrin, Delamare, and Moussu, the inherited effects of i, injury are described. The liver or the kidneys of pregnant rab- 1 bits and guinea pigs were injured, which caused these organs | * Oesterreichischc medicinischc Jahrbiicher , p. 179, 1875. J ’ See Weismann, “Essays,” Vol. I, p. 323. ? ® It has also been suggested by some more recent authors that epilepsy occurs * as the result of a weakening of the general condition either direct or inherited ^ This view will not e.xplain the localized inheritance claimed by Brown-S^quar . , The Inheritance of Acquired Characters 55 to become diseased. The offspring sometimes showed defects in the corresponding organs. The authors suggest that some substance may be set free from the diseased organ which may be carried in the blood, and by diffusion get into the blood of the embryo, and directly affect the development of the corre- . spending organ. They attempted to test this hypothesis by in- jecting into the blood of a pregnant animal extracts from the ;1 diseased kidney of another animal, and while the authors do I not appear so certain that similar effects are here also pro- I duced in the organs of the embryo, yet this seemed to be i the case. It will be observed that this transmission of an acquired ! character, if it really occurs, appears to be different from that of transmission through the egg, for it is the developing organ : itself that is acted upon. These results may possibly have a bearing on Brown-Sequard’s work, since they seem to show that if an organ of an adult ani- I mal (that is viviparous) is diseased, the sarhe organ of the young :imay develop abnormally (or become diseased). Whether the ttwo cases are really the same we do not know. In fact, it > would be extremely hazardous to conclude even that they may ibe, until we know in what way epilepsy is caused — whether Iby a physical defect in the nervous system, or by bacteria, or fby some other means. It should not be overlooked that the .epilepsy was not present in the young when born, but developed later. There is urgent need that experiments of this sort be carried out on an extensive scale. Many other cases of mutilations have been cited to show that acquired characters are inherited. Defects in the parents are said sometimes to reappear in the young, and it is linf erred that in some way the two things are connected. ■We are apt to overlook the fact that thousands of injuries are not inherited, and that a malformation appearing in the child in the same organ that had been injured in the mother or father ‘ may chance to occur at any time, and would be certain to arrest 56 Experhnental Zoology . there was any direct connection between the conditions in the parent and in the offspring. On the other hand, if the results of Charrin, Delamare, and Moussu are confirmed, there is a chance that diseased organs at least may affect the young in utero. This would, of course, only allow the inheritance of acquired characters from the mother, and not from the father. This difference might give us a chance to test the view that the effects are produced in the embryo in utero and not in the germ-cells. It is interesting to note that Brown-Sequard found that epilepsy is more often transmitted through the mother than through the father. In the case of mutilations, the injury may have been inflicted years before, and the wounds have completely healed before the young are conceived ; yet cases of this sort have often been cited to show inherited effects. In such cases it is difficult to see how such effects could become transmitted, especially through the male. Weismann has given in the “Essays ” referred to above an interesting and very full discussion of the supposed cases of inheritance of acquired characters; He himself carried out some experiments with mice. For four generations the tails of mice were cut off. Of the 901 mice born during this time not one had a short tail, and careful measurements showed that there was no shortening at all of the tail of the offspring. This experiment was, it is true, almost needless, since it is cus- tomary to cut off the tails of certain breeds of sheep and the ears of dogs without these breeds ever having become tailless or earless, and circumcision has been practiced in man for cen- turies without any appreciable effect. Since in these cases the organs operated upon had healed over before the next generation was born, there would be little chance of the injury directly affecting the embryos. Moreover, even if such effects are inherited, it might not follow that the tails would be shortened, but at most only diseased in some way. Recently Nussbaum has commented on this side of the ques- The inheritance of Acquired Characters 57 tion. He claims that we should not expect the inheritance of short tails, because even if such an influence were transmitted to the egg, the young embryo would promptly regenerate its missing portions. Unfortunately there is no evidence that they can regenerate in the embryo. It has been claimed that the removal of or injury to an organ is a different process from the modification of an organ caused by some external condition. There are, however, cases in which an organ has been greatly modified for several generations and no inherited effect has been produced. The feet of Chinese women are compressed so that their form is greatly changed, and in the higher classes this has been kept up for generations, yet no effect has been produced on the feet of the children. Certain races of Indians are known to have changed the shape of the heads of their children by compressing them between boards, yet the effect does not seem to have been inherited.' Tight lacing may change very greatly the shape of the ribs and to some extent the position of the viscera, yet no inherited effect has been produced. We may now consider the cases of the supposed inherited effect of use and of disuse. We are familiar with the decrease of a part in size and in function through disuse, in the case of muscles, glands, bones, etc., and also with the enlargement of organs through use. It is especially this class of facts that impressed Lamarck, and which led him to assume that these effects are inherited. The rudimentary eyes of animals living in the dark are often cited as having resulted from disuse. The long neck of the giraffe and the long tongue of the ant-eater, of humming birds, and of woodpeckers are examples, that La- marck has given, of effects produced by use; and many natu- ralists since Lamarck’s time have cited these and other similar cases as bearing on the question. But neither Lamarck nor his successors have been able to demonstrate by experiment that the effects of use and disuse of this kind are inherited by the next generation, and until this proof is forthcoming we must regard their view as purely speculative. It is surprising 58 Experimental Zoology that no experimental proof of this kind has been furnished, because it would seem that it ought not to be difficult to make crucial experiments that would settle the question at once. It is true that many of the latter-day Lamarckians claim that effects of this sort are only very slowly brought about, and that we should not expect to observe any results that are meas- urable in the course of a few generations. It seems to me that by taking this position the Lamarckians distinctly avoid the real issue; for it is not evident why such effects, if produced at all, should not appear at once in the embryo, since they appear at once in the adult. It is evident that if the effects of use and disuse were inherited, many of the adaptations of organisms to their surroundings could be quickly attained; for the parts most used would be strengthened in successive generations, those less used would soon decrease in proportion to their use, and the comphcated adjustments of the different parts could be accounted for. The habits and instincts of animals would also be made to con- form to the needs of the animal. These benefits are so obvious that it is small wonder that the theory of the inheritance of ac- quired character has always had its adherents. But the plausi- bility of a theory is not a scientific proof of its worth, and the best evidence, viz. that from experiment, that we have at pres- ent, does not show that acquired somatic characters are inherited through the germ-cells. Moreover, the advocates of the theoiy' overlook a consideration of prime importance. If the effects of mutilations and of diseases are also inherited, the results would be highly injurious to animals. Considering how often animals are injured, we should expect to find the animal kingdom in a most dilapidated condition if the accidental injuries of all their ancestors were transmitted to subsequent generations. It is unfortunate that many of the best-authenticated cases of the inheritance of acquired character are those relating to disease. In recent years two elaborations of the principle of inheri- tance of acquired characters have appeared. Semon has worked out Hering’s original suggestion that heredity is racial mem- The Inheritance of Acquired Characters 59 ory. Under the title of “Die Mneme,” Semon has analyzed and classified the different forms of inheritance of former ex- periences of the individual which are assumed to become in time the herditary capital of the race. Rignano has developed the idea of the influence of the germ-cells on the soma, and mce versa, from a different standpoint, viz. that the germ- cells influence the soma during development and in turn are at times influenced by the soma. These speculations are based on the assumption that acquired characters are inherited. Since we are concerned here only with the experimental evidence in favor of or opposed to this assumption, it would carry us too far to attempt to deal critically with these elaborations, that assume at the starting point that such characters are inherited.^ T elegony We may next examine the evidence that has been supposed by certain writers, in the main “practical” breeders, to prove that maternal impressions of various kinds can be transmitted to the young in utero. The crudest examples are those in which it is related how the pregnant mother, being impressed by some unusual or revolting sight, has transmitted to her infant a corresponding structural deformity. Somewhat less credulous perhaps are those breeders and “fanciers” who are firmly convinced that if a purely bred animal — the horse and the dog are the stock examples — has first been paired with a mongrel animal, the subsequent offspring to a purely bred father will show evidences of the first birth, i.e. be impure as to their breed. The credulity of men who have not been trained as to the value of evidence is a matter of everyday observation, and it is not going too far to say that most opinions or statements of the “practical” breeder must be put to a rigorous scientific test before they can be trusted. This has proven to be the case with telegony. * For a criticism of Semon’s argument, see Weismann’s recent (1906) review. 6o Experhnental Zoology The best- known case, and one that Darwin himself believed to have been “perfectly well ascertained,” is that of Lord Mor- ton’s mare, a full account of which is given in Darwin’s “Ani- mals and Plants.” A pure Arabian mare was bred to a quagga, and the first offspring was, of course, a hybrid. The second time, the mare was bred to a stallion of pure stock. The colt showed cross markings on the legs, that were said to he much more developed than on colts of pure pedigree. Dar- win thought the evidence sufficient to show that in some unknown way the mother had been affected, either by the sperm of the quagga or by the hybrid embryo in utero, and this effect was again in some unknown way transmitted to the sec- ond offspring. Darwin advanced this case as one of several in support of his hypothesis of pangenesis. The statements remained practically undisputed until within recent years. A few years ago Ewart undertook to repeat the experiment, which he was enabled to do by having a small experimental farm placed at his disposal. Careful series of cross-breeding experiments, followed by pure breeding, were carried out. Ewart discusses his results in his book entitled “The Penycuik Experiments,” and reaches the conclusion that no clear evidence of infection, if I may so call it, can be pro- duced in this way. He also points out that it is not uncommon for colts, purely bred, to show stripes as distinct as those in Lord Morton’s case. Fortunately a picture of this colt is still in ex- istence, and its examination shows that the markings are not more distinct than those that sometimes occur in the case of purely bred animals. There was, then, merely a coincidence, and not a causal connection. Bell has given an excellent summary of the evidence in regard to the possible influence of a previous sire, and has carried out a number of experiments on horses and dogs.^ In none of these cases was any influence of the previous progeny or of a ‘ Two other cases were reported to Bell, one for pigeons and the other for a cross between a negro and a white woman. In neither case were the later off- spring, by a father of their own kind, affected by the first union. The Inheritance of Acquired Characters 6i previous sire to be seen in the later offspring. The experiments were made with pedigree stock, and the results are convincing, and indicate that the belief in an influence of this kind is another breeders’ myth. Minot has also obtained negative evidence with guinea pigs, and I have obtained similar negative evidence with mice. Xenia It has been claimed that in plants the influence of the pollen is sometimes shown in those parts of the fruit or of the seed that are derived from the mother plant.^ The flower of the orange, fertilized accidentally by pollen from the lemon, is said some- times to produce fruit that may show a stripe in the peel like that of the lemon, although the peel is a product of the tissues of the mother plant. Similar cases have been recorded for the seed-color, and even for the pods of peas. The most familiar ' case is that of the color of the grains of maize. In recent years it has been shown that some, at least, of these ' cases can be explained as the result of a process of double fer- tilization. It has been found that there enters the embryo sac : not only the sperm-nucleus, that unites with the ovum, but 1 another nucleus that fuses with the “polar” nucleus or nuclei. The latter combination gives rise to the endosperm, which is I therefore hybrid in origin, and may show the influence of the i sperm nucleus if this contains the dominant character. In regard to the other cases, where the color is not in, the endo- : sperm, the results cannot be ascribed to the sperm nucleus. Giltay, who has made some experiments to test this point in peas, found no instances where the color of the pods could be assigned to the influence of pollen plant. ^ Theories of Transmission of Somatic Influences The only theory of any prominence that pretends to indicate : how changes in the somatic cells may affect the germ-cells is * From the nature of the case the process could not be expected to occur in animals. ^ See Davis, B. M., American Naturalist, XXXIX, August, 1905. 62 Experimental Zoology Darwin’s provisional hypothesis of pangenesis. Many of the assumptions of this theory are scarcely in accord with our pres- ent cytological knowledge. For instance, few cytologists would be likely to admit that the germ-cells are built up of living par- ticles representing the different tissues and organs of the body that are collected by the reproductive organs. In a modified form, however, Darwin’s hypothesis could no doubt be brought up to date, if it v/ere desirable to do so ; but is it worth while to speculate further in this direction until we have a better basis of fact on which to rest the speculation ; for, as has been pointed out, the experimental evidence in favor of the inheritance of acquired characters is unsatisfactory? The idea that the cell is made up of smaller morphological units that represent the various potentialities of the cell has been a favorite assumption of modern writers. Thus we have the physiological units of Spencer, the gemmules of Darwin, the pangens of de Vries, the plasomes of Wiesner, the micellae of Nageli, the plastidules of Haeckel, the biophores of Weismann, the biogens of Verworn, the idioblasts of Hertwig, etc. It is perhaps needless to point out that the kind of reasoning on which this method of treating the problem of heredity rests is of the sort that gives only the appearance of a real explanation, for the responsibility is only shifted to invisible and imaginary units that can be worked like puppets, at the will of the philoso- pher. Grossly ignorant as we are concerning the chemical and physical basis of cell activity, it is not probable that such guesses can be much more than fictions or at most symbolic. A single citation from Darwin will serve to bring the main points of his theory of pangenesis before us. “It is universally admitted that the cells or units of the body increase by self -divi- sion or proliferation. . . . But besides this means of increase I assume that the units throw off minute granules which are dispensed throughout the whole system, that these, when sup- plied with proper nutriment, multiply by self -division, and are ultimately developed into units like those from which they were originally derived. These granules may be called gemmules. The Inheritance of Acquired Characters 63 They are collected from all parts of the system to constitute the sexual elements, and their development in the next genera- tion forms a new being. . . A few experiments have been made to test this view. Galton ^ transfused the blood of one variety of rabbit into the veins of both sexes of another species, and then bred together the latter. If there are gemmules in the blood, the germ-cells of the rabbits containing the transfused blood might possibly show the influ- ence of the other variety. No evidence of such an influence was found. Darwin did not admit that this experiment was decisive, and Galton himself admitted that the results are not convincing. Darwin thought that the few gemmules present in the blood at any one time might not succeed in supplanting the similar kinds of gemmules supposed to be already present in the germ- cells. Another experiment was carried out by Romanes. Wild rabbits supplied the blood, and Himalayan rabbits received it. Several transfusions were made. In one case the blood of three wild rabbits passed through the veins of the domesticated individual. No evidence of any “foreign” influence was found in the offspring. Romanes said later that he had discovered that this experiment could not have been expected to give any positive results, because rabbits when crossed do not produce young having intermediate character. The force of the admis- sion is not very convincing, however, for the offspring might still have been expected to show the effects — if such influences are transmitted in this way — of the dominant breed, if this had been used to supply the transfused blood. Moreover, Castle has shown that in some breeds of rabbits certain characters at least are intermediate in the hybrid — the length of the ears, for example. Other zoologists who have refused to accept the doctrine of the inheritance of acquired characters — Weismaim, for instance have nevertheless used another idea contained in Darwin’s ‘ Proc. Roy. Soc., 1871. 64 Experimental Zoology hypothesis. The elementary characters of the cell are assumed to be contained in minute, living elements, gemmules, pan- genes, etc., that multiply and increase independently of each other; but their migrations are now limited to the individual cell. The nucleus is believed to be the storehouse of these units that issue forth at times to perform any function whatever that may be assigned, ■pro tem., to the cell. De Vries’s theory of intracellular pangenesis rests on this assumption. How such bodies eat up or replace the rest of the cell contents to dominate its function we are not told, and details of a chemical nature are deplorably lacking. Since this army of immigrants is confined within the boundaries of the cell, they have no importance to us in this connection. The historical origin of the ideas in regard to these bodies is, however, not without interest here. The superficial analogy between the theory and that of the atomic theory of the chemists has sufficed to lure many writers into this fascinating and facile mode of speculation. LITERATURE, CHAPTER V Barthelet. Experiences sur la teiegonie. Compt. Rend. Paris. Tom 131. 1900. Bell, A. L. The Influence of a Previous Sire. Journ. Anat. and Physiol. XXX. 1896. Brown-Sequard. NouvellesRecherchessurl’epilepsie. Archiv de Physi- ologie Normal et Pathol. 1869. Remarques sur I’epilepsie causde par la section du nerf sciatique. Ibid. 1870. Quelques faits nouvelles relatifs h I’epilepsie. Ibid. 1871. Faites nouveaux dtablissait I’extreme frequence de la transmission par I’heredite d’etats organiques morbides produits accidentelle- ment chez les descendents. Compt. Rend. 1882. Charrin, a., Delamare, G., et Moussu, G. Transmission experi- mentale aux descendants des lesions developpees chez les ascendents. Compt. Rend. Acad. Sc. Paris. CXXXV. 1902. Davis, B. M. Studies on the Plant Cell. Am. Natural. XXXIX. 1905. Ewart, J. C. The Penycuik Experiments. London. 1899. Fischer, E. .Experimentelle Untersuchungen ueber die^ Vererbung er- worbener Eigenschaflten. Allgem. z. f. Entomologie, VI. 1901. Galloway, T. W. The Experimental Evidence for the Inheritance of Acquired Character in Organisms. Cumberland Presbyt. Quar- terly, I. 1902. . Hering, E. Ueber das Gedachnis als eine allgemeine Function der organischen Materie. 1870. The Inheritance of Acquired Characters 65 Meyer, J. de. L’heridit^ de caract^re acquis est-elle expdrimentalement. verifiable? Archiv de Biologic, XXI. 1905. Minot, C. S. An Experiment with Telegony. Report 74. Meet. Brit- ish Assoc. Cambridge. 1904. Nussbaum, M. Die Vererbung erworbener Eigenschaften, Sitzungsber, der Niederrheim. Gesellsch. f. Natur- u. Heilkunde zu Bonn. - 1903- Rabl, K. Zuchtende Wirkung Functionelle Reize. 1904. Rignano, E. Sur la transmissibilitd des caracteres acquis. Hypoth^se d’une centro-epigen&se. Paris. 1906. Semon, R. Die Mneme als erhaltendes Princip im Wechsel des organ- ischen Geschehens. 1904. Ueber die Erblichkeit des Tagesperiode. Biol. Centralb. XXV. 1905. SiTowsKi, L. Biologische Beobachtungen ueber Motten. Bull. Acad. Science de Cracovie. 1905. Spencer, H. Principles of Biology. 1864. Revised ed. 1898. The Inadequacy of “Natural Selection.” Contemp. Rev. 1893. Prof. Weismann’s Theories. Contemp. Rev. 1893. Weismann, a. Das Keimplasma, eine Theorie der Vererbung. Jena. 1892. Die Allmacht der Naturzuchtung, eine Erwiderungen H. Spencer. Jena. 1893. Ueber Germinal-selection. C. R. z. Congres internat. Zool. Leyde und Jena. 1896. Richard Semon’s “Mneme” und die “Vererbung erworbener Ei- genschaften.” Arch. f. T.assen- und Gesell. Biologie,TII. 1906. K CHAPTER VI EXPERIMENTAL HYBRIDIZING Within the limits of the Linnaean species it has been found that varieties, or races, or breeds, are generally perfectly fertile when crossed, and in recent years these fertile crosses have been much studied. In some cases it has been found that a character that is different in two parents blends in the offspring, and may or may not separate again; in other cases, however, a character that differs in the parents does not blend, and all the offspring of the first generation are like one of the parents. The inheritance is alternate. If hybrids of this kind are bred to each other, the original character of one of the grandparents may reappear in some of the offspring, the contrasting char- acter of the other grandparent in others. These cases follow what is known as Mendel’s law. It is sometimes stated that Mendel’s law applies only to crosses between varieties, and this is true for many cases ; but characters that are entirely new to the race may also follow Mendel’s law ; and if the appearance of one new character suf- fices to characterize a new form as an elementary species, we must conclude that the characters of some elementary species also follow Mendel’s law of alternate inheritance. MendeVs Law In 1865 Mendel published the results of an elaborate scries of experiments that he had made with varieties of peas. It is strange that so important a discovery should have been entirely neglected for thirty-five years, especially since the question of 66 Experimental Hybridizing 67 heredity and evolution were being actively discussed during that time. It was not until 1900 that de Vries, and simul- taneously Correns and Tschermak, independently obtained re- sults that brought to light again the long-forgotten discoveries of Mendel. Mendel found that when the flowers of one race of peas are fertilized artificially with pollen from another race, the hybrid offspring {F\) of the first generation are like one of the parents in each particular character, and not intermediate in character. If, however, these hybrids were self-fertilized or inbred, both grandparental types reappeared in their offspring and in definite proportions. The character of one of the parents that appears in the first hybrid generation {Fp is called the dominant, and the contrasted character of the other parent that disappears in the first hybrid generation is called the recessive- When these first hybrids are inbred as stated above, there ap- pears in the second generation of hybrids {F^} three individuals showing the dominant character to one individual showing the recessive. This, however, is by no means the whole discovery ; for Mendel found that the recessives of this second generation, if inbred, give always recessives and nothing else. Those that show the dominant character, on the other hand, do not all breed true. A third only are pure and give rise only to domi- nants, while two- thirds of them produce both dominants and recessives. The matter can be graphically expressed as fol- lows ; — If we call the dominant character A, the recessive B, then the first generation {Fp of hybrids will be A{B). This means that while the hybrids show outwardly only the dominant character A, the recessive character {B) is also present in an undeveloped condition. When these hybrids (Fj) are inbred, the A -char- acter dominates in one fourth of the offspring, the 5-character in one fourth, and the A{B) character in two fourths, i.e. in the proportion of 1:2:1. Mendel showed by a simple In practice A{B) can only be distinguished from A by the kind of progeny that each produces. 68 Experimental Zoology assumption how this numerical relation could be explained. When the male, and the female, 5, germ-cells unite, every cell of the hybrid will contain both A and B\ in which case one dominates, namely. A, giving A{B). If we assume that in the germ-cells of the hybrid the characters A and B separate again, and go to different cells, half of the germ-cells will con- tain the one character only, and the other half the other char- acter. This is supposed to take place both in the male and in the female individual. The male germ-cells containing A may meet egg-cells containing A or B, and conversely the male-cells B may meet egg-cells containing A or B. The possible com- binations that result are shown in the following diagram : — I /B '' A/^B V The chances are that, on the average, A will meet B twice as often as A meets A, or that B meets B. Hence the combination A{B) will occur twice as often as or BB. The outcome will be I AAj 2 A{B), I BB. Thus according to the assumption of two kinds of germ-cells in the hybrid the numerical results agree with the actual results of the experiments. For this reason Mendel’s assumption of two kinds of gametes has been generally accepted. Furthermore the theory can be tested in several ways, as will be shown later, and it has so far, on the whole, stood the test. When in addition to this it was found that in the germ-cells a mechanism exists that seemed capable of carry- ing out the postulated process of purification, it appeared to a number of modern zoologists that Mendel’s assumption of two kinds of germ-cells in hybrids of this sort is a real and not a fictitious explanation of the results. An actual example may make clearer Mendel’s principle and its interpretation. If a gray house mouse, A, is crossed with a white albino mouse, B, the offspring, (Fj), will be all gray like the house mouse. If these gray hybrids* (Fj) are bred to each * Following Mendel the cross between two races or varieties is called a hybrid, although this term has been usually employed for crosses between species. Experimental Hybridizing 69 other, their offspring {Fp will be gray or white in the proportion of 3:1. If the white individuals are inbred, they will give only white, and this is true for all of their descendants. They are said therefore to be “pure.” The gray individuals, how- ever, show themselves to be of two kinds; one third of them, if inbred, produce only gray, and all of their descendants will be gray. They, too, are said to be “pure.” The other two thirds, if inbred, produce both white and gray mice. If these offspring are examined by further crossing, it is found that the whites are pure and give only whites; that some of the grays are “pure” gray, but the others are gray-dominant- white- recessives, A{B), and again in these we find the proportion i A : 2 AB : I B. The following scheme will show at a glance H the succession of generations : — A A 1, A 2, A (B) 1, B B B B 'A. practical consideration of some importance is obvious from ; these results. Pure races can be obtained from the hybrid, 1 A{B), by selecting the offspring with “pure” germ-cells, A or B. On the other hand, the A{B) hybrids always produce some ' A{B), so that all their offspring do not return entirely to the Ivwo parental types, but in every succeeding generation they will continue to split off some “pure” yl’s and B’s. As has been stated, Mendel’s assumption in regard to the two u'nds of germ-cells has been tested in other ways and found to 70 Experimental Zoology conform to expectation. One way has been to breed back the first hybrids A{B) to the parent form, either A or B-, the other way has been to apply the rule to more than one character. These two methods may now be illustrated. If a hybrid, A {B), is bred back to the parent type, B, half of the offspring should be A{B) and half BB. This must occur be- cause, on the assumption, the germ-cells of A {B) are A and B, while those of BB are B and B ; thus — A B B ^ A{B) BB A{B) BB If, on the other hand, A{B) is bred back to the other parent type, A, all the offspring will be like A, although only half are pure ^’s, the others being A{B)\ thus — A B A A A A A(B) A A A{B) The most interesting test that Mendel made of his theory involves the heredity of two dominant characters and two re- cessive characters. Thus, if two varieties, AB and ab, are crossed, the first hybrids {F-l) will be AaBb. Since A and B dominate, these individuals will all resemble AB externally. The germ-cells of the hybrid individual, AaBb, will be of as many kinds as there are possible combinations of A, a, B, b, provided that each combination contains some A (or a) and some B (or b), i.e. one or the other kind of the two characters, Thus the only possible combinations arc AB, Ab, aB, abb ‘ It may seem that these four combinations do not exhaust all the possible combinations of the letters, because A A, Aa, aa, BB, Bh, bh, might be sup- posed to appear, but this is not the case, because on the assumption of paired characters A (or a) must always be accompanied in the germ-cell with B (or b) characters. Similarly, B (or b) must always be accompanied by A (or a) characters. Hence the six combinations just given are excluded. Experimental Hybridizing 7 1 Each of the four kinds of egg-cells may be met by any one of these same four kinds of male cells, giving in all nine combina- tions; viz. AABB, AABh, AaBB, AaBb, A Abb, Aabb, aaBb, aaBB, aabb. But since these are combinations of both dominant and recessive characters, the offspring will appear to be of only four kinds. Thus the first four terms will belong to the type AB] the next two kinds will be Ab] the next two aB] and the last ab. It will be found by making all the possible combina- tions that the AB type occurs nine times, Ab three times, aB three times, and ab only once. Mendel carried out an experiment of this sort in which two contrasted characters were involved. The results fulfilled the expectations of the hypothesis. He used two races of peas, in one of which the form of the seed was round. A, in the other angular, a. In the round seeds the albumen was yellow, B; in the angular seeds the albumen was green, b, thus — AB, seed parent ab, pollen parent A, form round a, form angular B, albumen yellow b, albumen green The outcome of the experiment conformed to the scheme given above. It should be noted that in the first generation, A Bab, the hybrid will contain all of the possibilities, although A domi- nates a and B dominates b. In the germ-cells the characters separate on the assumption of pure gametes, but in such a way that A will always be associated with B or b, i.e. the round form will always be yellow or green. Similarly, a will always be associated with B or b. Hence, as stated above, the only combinations possible will be — AB, round and yellow, Ab, round and green, aB, angular and yellow, ab, angular and green. From the foregoing account it will be obvious that the prob- lem will become more complex when three contrasted charac- 72 Experimeiital Zoology ters are involved. Mendel found that the results with three characters agree also with the expectations. As the number of characters increases further, the results will be very complex and difficult to detect except by an exhaustive series of experi- ments, although each single character can easily be traced and found to follow the Mendelian law. Under these circumstances we might anticipate that types differing in many characters would give results too complicated for analysis, especially if some of the characters follow Mendel’s law and others follow other laws of inheritance. The generally accepted statement that species hybrids are intermediate in character between the parental types does not appear to hold in all cases critically examined for all the characters. It is evident that in the future the heredity of each character must be studied by itself. MendeVs Law and the Germ-cells On the assumption that the characters of the animal or plant are represented by primordia or elements or unit-characters in the chromosomes, the following attempt to account for the purity of the germ-cells, assumed on Mendel’s hypothesis, has been suggested by Sutton. In the early germ-cells, the spermatogonia and oogonia, the number of chromosomes is the same as the number in the body-cells, i.e. the somatic number; but just before the two maturation divisions there is a synapsis stage, in which the chromosomes come into closer connection with each other, and, as Montgomery has shown, it is probable that at this time the chromosomes pair with each other in such a way that each paternal chromosome unites with its homologous * maternal chromosome; and for the working out of Sutton’s scheme it is essential that each paternal unites with its homologous maternal, i.e. that the paternal do not unite with any other maternal or with each other. ‘ Homologous chromosomes arc those that have the same form, or, according to some writers, similar characters. 73 Experimental Hybridizing At one of the two maturation divisions the united pairs of chromosomes separate again and move into opposite cells (Figs. I and I A), so that one cell gets one and the other cell the other of each of the homologous chromosomes. Thus each cell will contain some paternal and some maternal chromosomes, but the number of the maternal may be different from the number I a B c D E Fig. 6. Scheme to illustrate the two maturation divisions as seen in the spindle and chromosomes. The clear circles represent paternal chromosomes and the black dots the maternal chromosomes. In the first division, represented by I and I A, some of the maternal and some of the paternal chromosomes move toward each pole of the spindle. In the second division each paternal and each maternal chromosome divides into equal parts. In II the chromosomes of the cell derived from the upper end of I and I A is represented; in II A those from the lower end. of the paternal. At the other maturation division^ each chro- mosome divides equallyFigs.il and II so that the daughter cells are exactly alike.^ Thus there will be two cells of one kind in regard to the single character (or group of united characters) contained in each chromosome, and two cells of the other kind. The same process occurs both in the egg and in the sperm- cells. In the egg three of the cells, the three polar bodies, are * In some species the first, in others, the second, is the equation division. ^ The meaning of this equation division has been much discussed, but nothing is known about it. 74 Experimental Zoology incapable of developing, and only the egg continues the history of the race. Half of the eggs, however, will on the average con- tain a particular kind of chromosome, and the other half the homologous kind, as shown in Figs II and II A. It is of special importance to notice in this connection that the pairs of chromosomes are assumed to lie haphazard on the spindle, so that while in one pair the maternal chromosome may be turned toward a given pole, in another the paternal chromosome may be turned toward the same pole. In other words, there are no grounds for assuming that all the paternal chromosomes turn toward one pole, and all the maternal toward the other, but “accident” alone determines which way they come to lie on the spindle. Hence the possibility of various combinations of chromosomes in the different cells is given. The evidence in the favor of the assumption of the accidental position of the chromosomes is indirect, and is deduced from the way in which the characters appear in the offspring of Mendelian hybrids when more than a single character is taken into account. Without this assumption the chromosomal hypothesis given above will not apply for more than one character. \Wiether this assumption is entirely satisfactory, will be considered later. A special case may make this discussion clearer. Let us assume that the character albinism of a white mouse is contained in one chromosome, and the gray character of the gray mouse in the homologous chromosome of the gray mouse. \\Ten these individuals are bred together the white chromosome, so to speak, and the gray chromosome are both present in the fertilized egg, which gives rise to the gray hybrid of the first generation, because the gray dominates the white. In the germ- cells of these gray hybrids the changes described above take place. At the synapsis stage the white chromosome pairs with the gray chromosome. Later, at one of the maturation divi- sions, the two separate and go to separate cells. Hence each germ-cell becomes “pure” and carries only one kind of color. If these hybrids (Fj) with white and with gray germ-cells are paired, there will be formed by chance unions of the geim- Experimental Hybridizing 75 cells the three kinds of individuals of the second generation. A white spermatozoon meeting a white egg, as one may say for brevity, produces a “pure” recessive individual; a gray sperma- tozoon meeting a gray egg gives a “pure” dominant individual; but when a white spermatozoon meets a gray egg, the off- spring will be mixed, a heterozygote, and like the first hybrid {F^. If this hypothesis is true, we ought to be able to determine which groups of characters — where several are involved — are contained in a given chromosome, for the number of chro- mosomes is often small, and therefore the actual number of combinations is limited. All characters contained in the same chromosome should “Mendelize” together. If we consider two characters, the principles stated above will apply. Let us take the case of the round and yellow, angular and green peas. The round form. A, may become combined with the yellow, B, or with the green color, h, giving AB or Ah\ the angular form a with B or h, giving aB or ab. To work this out we must suppose the color to be contained in a different chromosome from the form, for otherwise they could not shift over. When the chromosomes unite in synapsis the T-bearing chromosome can only unite with a, and B can only unite with h. Hence since | ^ and | ^ may be turned either way on the spindle, A may pass into a cell with B or h, and also a into a cell with B or h. Simple and logical as Mendel’s assumption appears to be on the hypothesis of each character being contained in only one of the chromosomes, yet I do not think it can be accepted in this form because the primary assumption that each character is contained in a single chromosome and not in others is highly arbitrary, and also because there are certain actual results that are difficult to explain on the assumption of “pure” germ-cells. In fact, it has not been shown beyond question that the chro- mosomes are the bearers of the hereditary qualities. The evi- dence that is generally supposed to establish the assumption of 76 Experimental Zoology the hereditary nature of the chromosomes is that the spermato- zoon brings into the egg only the chromosomes of the male germ-cells. While it is undoubtedly true that the largest part of the sperm-head is made up of chromatin material derived from the nucleus, it is also true that the protoplasm of the origi- nal germ-cell is not lost. It is partly used up in the formation of the tail of the spermatozoon, but also forms a condensed layer around the sperm-head. There is every reason to assume that the latter may become incorporated as a part of the cyto- plasm of the egg. There is also another serious objection to the explanation of the purity of the germ-cells given above, for, even if the chromosomes be the bearers of the hereditary qualities of the egg and sperm, it does not follow that each unit character would be contained in only one chromosome. If it be assumed that each chromosome carries all the hereditary qualities, it is im- possible to account for the purity of the germ-cells on this as- sumption. For instance, if we assume that each of the chro- mosomes contains all of the hereditary characters, the germ-cells of the hybrid, A{B), will contain, before synapsis, half of their chromosomes bearing the character A, and half bearing the character B. If these unite at the synapsis in pairs, and then come to lie, haphazard, on the spindle, some turned one way, some the other, the resulting germ-cells will contain all mixtures of A and B, and hence be impure. If we give up the idea of “purity” and assume that the relative number of ^ or of B chromosomes determine the character of the result- ing cells, the three types of the Mendelian ratio might be ac- counted for, provided the reduced number of the chromosomes is an odd number. If it were an even number, it must often happen that equal numbers of a character, of A or B, would be contained in the same germ-cell, and hence there would be an exact balance, which on the theory should give neither result. This, however, is not in harmony with the facts. In the light of these theoretical difficulties it seems to me that the chromosomal theory must be applied to Mendel’s law with caution, and that while at first sight it appears to offer an Experimental Hybridizing 7 7 explanation of the assumed purity of the germ-cells in the Men- delian cases, yet more careful consideration shows that in order to do so certain assumptions are necessary that are not above suspicion. It may be seriously questioned, I think, whether the germ-cells of Mendelian hybrids are pure. It is true that the Mendelian proportion, 1:2:1, in the second generation can be most easily accounted for by assuming two kinds of male and two kinds of female germ-cells, each kind existing in equal numbers ; but the assumption that the two kinds must be pure germ-cells meets with serious difficulties when certain results are considered. It will suffice to point out here that the main difficulty lies in the behavior of the so-called extracted reces- sives which ought to be a pure strain on the assumption of “pure” germ-cells, but which have been shown on the con- trary to contain in a latent condition the dominant character. I have tried to show that the results may still be accounted for even if the germ-cells of the hybrids of generation {FI) are not pure in regard to any pair of contrasted characters, such as gray and white, but that both characters are present in all the germ-cells. The two kinds of germ-cells that the hypothesis calls for may be referred to the alternating dominance in the germ-cells of each of the two contrasting characters. The Mendelian proportion can be accounted for on this assumption as well as on the accepted interpretation of pure germ-cells, and the latency of the dominant char- acter in the extracted recessive can also be explained on my view, but not on the other. An example may make my meaning clearer. Suppose a white and a gray mouse are paired. The germ- cells of the white mouse are white-producing, or briefly white, those of the gray mouse are gray-producing, or briefly gray. The fertilized egg will contain both characters, and since the gray dominates the white, the symbol G{W) will represent the con- dition in the mouse itself. In its germ-cells both the gray, G, and the white, W, exist, presumably combined in some way. 78 Experimental Zoology At some time in the history of these germ-cells one or the other of these two characters comes to dominate in each cell, so that half of the cells will be G{W) and half W{G). This will be true both for eggs and sperm. Chance meetings of the two kinds of sperm and the two kinds of eggs will give — G(W) W(G) G{W) W(G) I G(W) + 2 G{W) W(G) + I W(G) “ This is the characteristic Mendelian proportion. The first term, G(W), is a gray mouse, the so-called extracted dominant, i.e. it is a mouse gray in color, which, if bred to other gray extracted dominants, will produce only gray mice. This means that the latent white remains in the latent condition in its germ-cells, all of which are G(W). The second term, G(W) W (G), of the proportion represents also a gray mouse, since the gray, G, dominates the white, W, when both occur in the “free” condition in the same body-cell. It will be noticed that the presence of the two “free” colors, G and W, in the cells of this type indicates that the type is the same as the first hybrid formed by crossing G with W ; and it is im- portant to find that when inbred this type gives exactly the same results (i.e. the Mendelian proportion again) as do the first hybrids, GW. By the “free” condition I mean to imply that the two characters, G and W, have not been brought into the intimate relation to each other that is assumed to occur in the germ-cells at the time when the alternating dominance and latency occurs. The third term, W(G), of the proportion is the extracted reces- sive. It represents a white mouse containing gray in a latent condition. If inbred, these white mice produce only white mice, but if crossed in certain ways the latent gray color can be brought out again. The preceding example will suffice to show how the Mcn- delian proportion can be accounted for on the assumption of alternation of the contrasted characters in the germ-cells. 79 Experimental Hybridizing The question may be asked whether this method of account- ing for the results can be referred to the changes that take place in the chromosomes of the germ-cells. The hypothesis de- mands that the contrasted characters come into relation with each other, and the union of the chromosomes might suggest such a possible combination. After uniting, the fused char- acters must be halved again (quantitatively not qualitatively), and the reduction division might suggest a possible method of accomplishing this result. On the other hand, there is no apparent need to assume such a complicated mechanism to bring about the union of the characters in the same cell, nor for their subsequent separation. Moreover, by referring the process to the chromosomes, we introduce the further assump- tion that the characters of the cell are contained only in those bodies — an assumption that is not itself established. For the present, therefore, it seems premature to connect these results I definitely with any known change in the germ-cells, and the : same statement holds also, as we have seen, for the alternative assumption of pure germ-cells. Until we know more of the way in which characters are represented in the germ-cells, we can only offer purely specula- ; tive views of what we suppose might take place in order to give 1 the Mendelian results. The formulae that we use are merely symbols for handling these results. The fact that the char- acters that “Mendelize” are different types or permutations of j! the same characters suggests that they may represent stereo- i i metric relations of the material basis of the characters, i.e. of it) the molecules representing them. Thus we might represent jt the characters gray and white in the hybrids as right- and left- jl handed forms of the same molecule, and indicate this by GW jjand WG. Such germ-cells meeting each other would give — ! ^GW , i GW , jWG I 1 GfF I IFG t WG' i ■and these might be taken as representing the three Mendelian i groups. Interesting as such a speculation might be, could we 8o Experimental Zoology find evidence in its support, it is unprofitable as an interpreta- tion until or unless we can show that it is at least probable. The preceding discussion is based on the idea that there must be two kinds of germ-cells in order to give the Mendelian re- sults. This is perhaps the simplest way to formulate the prob- lem, but it should not be overlooked that the results can also be explained by assuming that all the germ-cells of the hybrid are alike, containing both dominant and recessive characters, but that after fertilization internal or external factors determine whether the one or the other character dominates. The diffi- culty of this view is to account for the middle term of the Men- delian proportion, in which, although the dominant develops, the germ-cells seem to return to the two original kinds. Pos- sibly this difficulty could be met by assuming that the two con- trasted characters are so nearly balanced in half of the progeny, that while the dominating character dominates when differen- tiation begins, it is not able to do so in the germ-cells. It is evident, however, that this interpretation is more complicated than that of two kinds of germ-cells, and fails to account for the three sharply marked groups. LITERATURE, CHAPTER VI Bateson, W. Mendel’s Principles of Heredity. 1902. Note on the Resolution of Compound Characters by Cross-Breeding. Proc. of the Cambridge Philos. Soc. XII. 1902. On Mendelian Heredity of Three Characters Allelomorphic to Each Other. Proc. Cambridge Philos. Soc. XII. 1903. Castle, W. E. The Laws of Heredity of Gabon and Mendel, and Some Laws governing Race Improvement by Selection. Proc. of the Amer. Acad. Arts and Sci. XXXIX. 1903. Mendel’s Law of Heredity. Proc. Amer. Acad. Arts and Sci. XXXVIII. No. 18. 1903. Mendel’s Law of Heredity. Science, n. s. XVIII. 1903. CoRRENS, C. G. Mendel’s Regeln ueber das Verhalten der Nachkommen- schaft der Rassenbastarde, Ber. deutsch. bot. Gesellsch. XX. 1900. Darbishire, a. D. On the bearing of Mendelian Principles of Heredity on Current Theories of the Origin of Species. Mem. and Proc. of the Manchester Liter, and Philos. Soc. XLVIII. 1904. Focke, W. Die Pflanzen-Mischlinge. Berlin, Borntraeger. 1881. Mendel, G. Versuche ueber Pflanzenhybriden. Verb, naturf. Vereins in Briinn. IV. 1866. Experimejital Hybridizmg 8i Ueber einige aus kiinstlicher Befruchtung gewonnenen Hieradum- Bastarde. Verb, naturf. Vereins in Briinn, VIII. 1870. Moll, J. W. Die Mutationstheorie. Biol. Centralbl. XXIV. 1904. Morgan, T. H. The Assumed Purity of the Germ-Cells in Mendelian Results. Science, XXII. 1905. Are the Germ-Cells of Mendelian Hybrids “Pure”? Biol. Cen- tralbl. XXVI. 1906. Prentiss, C. W. Polydactylism in Man and the Domestic Animals, with Especial Reference to Digital Variations in Swine. Bui. of the Mus. of Comp. Zool. at Harvard Col. XL. 1903. Tschermak, E. Weitere Beitrage ueber Verschiedenwerthigkeit der Merk- male bei Kreuzung von Erbsen und Bohnen. (Vorlaufige Mitthei- lung.) Ber. deutsch. bot. Gesellsch. XIX. 1901. Die Theorie der Kryptomere und des Kryptohybridismus. I. Mit- theilung. ueber die Existenz kryptomerer Pflanzenformen. Bot. Centralbl. Beihefte, XVI. 1903. Ueber die gesetzmassige Gestaltungsweise der Mischlinge. (Fort- gesetzte Studie an Erbsen und Bohnen.) Zeitschr. f. landwirtsch. Versuchswesen in Oester. V. 1902. Wilson, E. B. Mendelian Inheritance and the Purity of the Gametes. Science, XXHI. 1906. De Vries, H. Sur la loi de disjonction des hybrides. Compt. Rend. CXXX. Paris. 1900. Das Spaltungsgesetz der Bastarden. Ber. deutsch. bot. Gesellsch. XVHI. 1900. Anwendung der Mutationslehre auf die Bastardierungsgesetze. Be- nch. der Deutsch. Bot. Gesellsch. XXI. 1903. CHAPTER VII EXPERIMENTAL HYBRIDIZING {Continued) Experiments with Mice The ease with which mice can be kept in captivity, their rapid rate of multiplication, and the occurrence not only of an albino race, but of other fancy breeds as well, have made these animals a favorite subject for experimental work. The house mouse (Mus musculrs) is of a gray, sometimes called cinnamon or agouti, color. The white mouse is an albino race derived without much doubt directly from the wild form. Albino mice have been recorded as occurring in situa- tions where there was no reason to suppose that their origin could be due to the escape of domesticated albinos. It is also known that in many groups of the animal kingdom albino indi- viduals often suddenly appear. The color of the hair of the gray mouse is due, according to Bateson and Durham, to three kinds of pigment often asso- ciated in each hair: (i) opaque black, (2) less opaque brown, (3) transparent yellow. In albinos all three colors have dis- appeared as well as the pigment in the eyes. As many as a dozen or more races of fancy mice are known. It appears that the color of the race is determined according to which of the three pigments found in the gray house mouse predominates or exists alone. Thus the golden agouti, of a tawny color, con- tains yellow and brown pigments, but not black; chocolate mice contain only the brown pigment ; yellow contain only the yellow; black contain black and brown. Variegated mice are those in which irregular small spots of black or chocolate occur on a 82 Experimental Hybridizing 83 white background ; spotted or pied or piebald mice are those in which these various colors appear in splotches or marks. In addition to these kinds, a white mouse with dark eyes is known, which is probably — to judge from other white animals with black eyes — not derived from an albino, but from a spotted animal in which the spots of dark pigment have disap- peared except in the eyes. The so-called dancing mice that whirl around at times are said to be of Japanese origin, and may have originated from a different wild variety. Experiments on mice have been carried out by a number of investigators, the principal results being those of von Guaita, Cuenot, Parsons, Darbishire, Castle, Allen, Davenport, Schuster, Haacke, and others. Von Guaita’s results were not consid- ered by him in the light of the Mendelian ratio, but Bateson and Davenport have more recently examined them and have pointed out that many of them appear to follow Mendel’s law. Cuenot’s results conform closely to the Mendelian law. He found that when a gray mouse is paired with an albino, the off- spring in the first generation (F^) are always gray mice — the gray dominating over albinism. The next generation (Fp from the inbred dominant gray mice gave 198 gray and 72 albino mice, i.e. in the ratio of 2.75 : i, a near approach to the expecta- tion of 3 : 1. Later Cuenot reported that his “pure ” gray ^ mice of the third filial generation (F^) when crossed with albinos gave several black Tnice. These black mice when bred with certain albinos gave black mice, which appeared to be B{W), i.e. black dominant, white recessive, for, when bred inter se, the offspring were three blacks ,to one albino. Some of these blacks were shown to be “pure” and produced a race of pure blacks. Wlien individuals of this black strain were bred to ordinary gray mice, the black was recessive, giving in the second generation three grays to one black. Cuenot made a further discovery of great interest. He found when he bred the black mice to albinos that the results were These were the offspring of the second generation that had bred true and shown themselves to be pure G’s. 84 Experimental Zoology different according to the kind of albinos that he used. For example, albinos that had been derived through a gray ancestor, i.e. the so-called extracted albinos, when bred with the black strain produced gray offspring. If, however, albinos that had been derived through a previous cross with a black mouse were bred with a black, the offspring were black. Again, albinos derived from a previous cross with a yellow gave either mixed yellow and gray, or mixed yellow and black. These results show that although albinos may all appear alike and breed per- fectly true with each other, they belong in reality to different classes, whose latent characters are dependent on previous influences. Results of this sort are difficult to account for on the supposition that the germ-cells are pure. This property of latency is not something peculiar to albinos as certain pubhshed statements might have lead one to infer, but holds for the differ- ent colors also. A black race that breeds true may carr}^ another latent color that can be brought out by crossing. Un- less this fact is also taken into account the crosses with albinos may be misinterpreted. Experiments with waltzing mice have been made by Haacke, von Guaita, and Darbishire. These mice are black, or white, or yellow and white; the mixed colors introducing a complication into the results, so far as color inheritance is concerned. As has been said, von Guaita’s results with these mice were not considered from the point of view of Mendel’s law, but Bateson, who has later analyzed the data, finds in some cases what seems to be an approximation to the expected proportions; in other cases this is not evident. A few of von Guaita’s facts and their possible interpreta- tions may be given. When the black-and-white waltzers were bred to ordinary albinos, the first offspring {F{) were gray, like the house mouse, and of the same size, which is larger than that of the waltzing race. They also show the wild disposition. These mice (Fj) when inbred produced albinos and four colored types, — black, gray, black-and-white, and gray-and- white. There were 14 albinos and 30 colored individuals. The relation 85 Experimental Hybridization of the colored types to each other as regards their inheri- tance is too obscure to make it profitable to discuss the result here. Darbishire has carried out experiments with pink-eyed, spotted, waltzing mice crossed with albinos. In respect to their coat color, he recognizes six groups forming a continu- ous series, depending on the extent to which the pigment spots cover the surface of the mice. The colors of the spots were yellow, gray, black, lilac, or chocolate. When these mice were . crossed with albinos, supposed to be pure, spotted mice were produced with dark eyes. None of the mice of this generation t exhibited the waltzing habit. These hybrid mice when bred i inter se gave the following kinds of mice {Fp : — Albino 137 Colored or piebald with dark eyes . . . 287 Colored or piebald with pink eyes . . . 13 1 Of these mice 97 showed the waltzing habit and 458 did not. The Mendelian expectation for waltzers is 138.75. The actual -results fall considerably below the expectation, nevertheless it may be that some of the mice that did not waltz were poten- tially waltzers and might have transmitted this habit as do “pure” recessives. It is interesting to note that, whether the 'Mendelian proportion is or is not given, the waltzing habit dis- lappears in the first generation of hybrids {F-^ and reappears in the second generation {F^ as do other Mendelian characters. The expectation for albinos is approximately realized as well as the expectation for the other two types. Since neither grand- parent had dark eyes, this character must have been latent in one of them, because it appears in all of the offspring Dar- bishire points out that his results do not conform in all respects 'to the Mendelian rule ; but some, at least, of these difficulties are pot insuperable, I think, if it be granted that the so-called “pure” recessives and “pure” dominants are really impure with latent jbharacters that come out on crossing. Darbishire contends, and I think justly, that the behavior in inheritance of extracted 86 Experhnental Zoology recessives and of extracted dominants cannot be accounted for, if they are supposed to have been formed by the union of pure, germ-cells. On the contrary, he thinks that the results become intelligible only when the ancestry of these forms is taken into consideration. In other words, if “pure” recessives and “pure” dominants are really pure (as modern Mendelians have as- sumed), the ancestry of such forms could be ignored ; but since the results are inexplicable on this assumption, the most rea- sonable conclusion is that the germ-cells are not pure. So far I am in agreement with Darbishire; but if this conclusion is meant to imply that the Galtonian assumption in regard to inheritance in these cases is the only alternative, I should dis- sent. I have tried to show how the Mendelian results may still apply without assuming that the gametes are pure, but by assuming that they show alternate dominance and latency. Darbishire also crossed some of his first hybrids (F-P) with albinos and obtained 368 albinos and 378 dark-eyed piebald (or sometimes uniformly colored) mice. This gives a close ap- proximation to the Mendelian expectation. He found, more- over, in this generation that the gap between the albinos and the least- colored individuals was greater than among the offspring of hybrids (F^) when inbred. It is also interesting to note that with this cross the pigmented eyes appeared in all the piebald offspring. In other words, the latent color, eye-pigment, is maintained in the cross, since it dominates the pink-eyed type. Darbishire notes especially that these offspring (obtained from the hybrid and the albino) have more white in their immediate ancestry than have the offspring of the hybrids when inbred ; yet the offspring show actually less white, and show more often the wild color and black (as compared with yellow and fawn color). The meaning of this he finds obscure, but possibly the results may be accounted for on the assumption of the latency of pig- ment in one or both parent types, which is brought out by crossing. The subsequent history of the three classes of individuals ob- tained by inbreeding the first hybrids (Fj) is as follows : (i) The Experimental Hybridizutg 87 extracted albinos give always albinos. Only 2 out of 94 waltzed. (2) The dark- eyed mice with colored coats belonging to the middle term of the Mendelian series should also give, if inbred, the three types again in the proportion of i : 2 : i. In this connection Darbishire points out that on Gabon’s law of inheritance the farther the individuals of this middle class are removed from the first class, the fewer the albinos that should appear, since they are farther removed from the original ances- tor that was white. On the Mendelian law the members of this middle term should always continue to give the same pro- portion, 1:2:1. Experiments that Darbishire made to test this point seemed to show that the results follow more nearly the expectation of Gabon’s law; but the purity of the types used may have seriously affected his results. Especially interesting, it seems to me, are Darbishire’s experi- ments with the extracted mice of the second generation having pink eyes and colored coats. If these are really pure, they should, if paired with pure albinos, produce animals similar to those of the parent cross between colored mice with pink eyes and albinos, i.e. there should be produced only spotted mice with dark eyes. This, however, was not the result obtained. Of 98 young, 12 were albinos, while one pink-eyed (colored) indi- vidual also appeared. These results are complicated by the fact that the albinos used were also the extracted offspring of hybrids paired with albinos. Nevertheless, even granting this, it offers no explanation of why albinos should appear, and the only explanation that seems reasonable is that the albino, latent in the pink-eyed mice, has affected the result, pre- sumably being brought out again by crossing. Sixty-three of the 98 young were obtained from such contaminated albinos (extracted recessives). Only seven of the unions were between such pink-eyed mice and albinos which did not contain pink- eyed, spotted waltzers in their immediate ancestry. From these pairs 35 young were obtained, of which 10 were albinos — a relatively higher number of albinos, and approximately one fourth of the whole. The results seem to show that the extracted 88 Experimental Zoology I “pure ” race was not pure, since it may produce some albinos when paired with albinos. Allen has also carried out a large number of experiments with mice. Crossing the gray house mouse with the albino gave gray offspring, as most other experimenters have found. The second generation gave approximately the Mendelian ratio of three grays to one white, the former being partly pure grays, partly mixed. Crossing the dominant recessives (heterozygote) with the extracted albinos, where equality of grays and whites would be expected, gave 84 pigmented young and 64 whites. While there is only an approximation to equality here (74 of each expected), the deficiency in white may be due to insuf- ficiency of numbers, but possibly to some other factor. Allen found that when spotted mice were bred to albinos the off- spring were spotted, — often with less white than the original spotted parent, — and in some cases the white almost, or even completely, disappeared. Thus, although we may look upon the spotted condition as a unit-character that is dominant, its extent appears to be variable. In fact, a latent character may also come to light here that is not seen in either parent, but must be potential in one or in both of them. While the white parent might have been expected to add more white to the offspring, — on Gabon’s hypothesis, — the result is exactly the opposite. When pigmented, heterozygote individuals (F^) were inbred, they produced 159 pigmented young and 55 albinos (53.5 being the Mendelian expectation). In another experiment, pig- mented, heterozygote individuals (F^) were bred to pure albi- nos, giving 69 pigmented and 69 albinos, exactly the anticipated ratio. Allen carried out a number of experiments made to test the inheritance of partial albinism, as he calls the condition when white areas are present along with colored areas. When “par- tial albinos” are bred to pure albinos, the young (Fj) were more nearly totally pigmented (as stated above), some showing no trace of white, others had white toes, or a white tip to the tail or even a few scattered white hairs. One only had a white spot Experimental Hybridizing 89 on the belly. Haacke and von Guaita found similar results. “The influence of the albinos in these cases seems to be to upset the condition of localization of the pigment, so that the pigment patches become more extensive, tending to cover the entire body surface, as in totally pigmented animals. In expla- nation of this observation it is suggested that the character total pigmentation may be “transmitted by albinos, and when so transmitted dominates over the spotted condition.” Cuenot has offered a suggestion to account for this possibility. The pigment is assumed to be due to the action of a ferment upon a chromogene substance. The albino may transmit the ferment but not the substance. The germ-cell of the black-and-white individual would then be assumed to convey the pigment, and, when to it more ferment is added by the white gamete, more pigment is produced in the offspring. Cuenot has given a very clear and important analysis of his results with mice in his third contribution to “The Heredity of the Pigmentation in the Mouse.” J He uses the letter C to denote any colored character, and A for the albino character; G for the color gray; B for the black; and Y for the yellow. Thus the wild gray mouse will be represented by CG, and the ex- tracted albinos, having potentially the gray color, by AG. The black mouse will be represented by the formula CB ; extracted albinos derived through black ancestors AB; the yellow mouse by CY ; the extracted albino through yellow ances- tors by A Y. I When a germ-cell containing the character CG unites with I one containing CB, the gray, G, dominates. When gray, CG, I meets yellow, CY, the latter dominates. When CB meets ' CY, the yellow again dominates. When a colored germ-cell, G, I meets an albino. A, the individual that develops has black { eyes, but the color of the hair depends on which color accom- > panics C or A. As an example Cuenot gives this case : a black I mouse, CB, crossed with an albino, AY (yellow latent), gives a i dihybrid, CBA Y. This hybrid has black eyes, because for ‘ Archiv. Zool. Exper. et Gen. 1904. Ser. 4, T. 2. Notes et Revue, p. xlv. I! 90 E>xperimental Zoology the eyes the color, C, always dominates; but the color of the hair is yellow, because yellow dominates over black. The four types with which Cuenot has experimented give eighteen different combinations, which he shows by means of the following table : — . COLOR HOMOZYGOTES HETEROZYGOTES OR PURE RACES Monohybrids Dih\"brids CGCB CGAB Gray CG (wild) CGAG Black CB CBAB Yellow CY CYCG CYAG CYCB CYAY CYAB Albinos AG AGAB AB • AGAY AY ABAY Of these eighteen types there are six that are pure races,^ i.e. they produce germ-cells that are all alike (homozygotes). In- bred, they give always the original types in all successive gen- erations, and this holds also for the three albino types, having the latent characters gray, black, and yellow, TG, AB, and A 1 . The other twelve types are heterozygotes, resulting from the crossing of pure races. Of these, nine are monohybrids, hav- ing only one pair of antagonistic determinants; and three are dihybrids, having two pairs of antagonistic determinants. It will be seen that Cuenot thinks that albinos are not nec- essarily all alike, although they may breed true to the albino type, but that they are different according to the latent char- acter that each contains. The latent character may appear in * According to Cudnot’s nomenclature. The three albinos, homozygotes, be- long in my opinion to a different category, for although they breed true, }et the) contain a latent color that may come out in crossing. Experimental Hybridizing 9 1 some of the descendants, if these albinos are crossed with colored types. Thus if an albino be crossed with an albino AB, the albino offspring will be AGAB. Its germ-ceUs will separate into AG and AB, but these are albinos. If, on the other hand, a black mouse, CB, be mated with an albino A F, containing yellow as a latent character, the offspring will be CBA Y (yel- low), whose germ-cells will be of four kinds, CB, AY, CY, and AB. Crossing this yellow mouse (with its four kinds of germ- cells) with a white mouse, AGAB, obtained in the way just described, eight possible combinations may follow. The whole process is indicated in the following table ; — Parents ist generation 2d generation AG (albino) AB (albino) \ / AGAB (albino) CB (black) \ CBAV AY (yellow) / (yellow) r AGCB gray (one) ABCB black (one) (four) AGAV] AGAB „ . albinos ABAB . CYAG\ „ , CYAB ) Cuenot performed this experiment and obtained in the sec- ond generation 15 1 young, distributed as follows according to color : — 81 albinos, 34 yellow, 20 black, 16 gray. The probability according to the formula 4W + 2 w + w + w is: 76 albinos, 38 yellow, 19 black, 19 gray. The agreement is so close that there can be little doubt that the hypothesis is substantially correct. Heredity of Piebald or Spotted V arieties. — The piebald condi- tion is regarded by Cuenot as a special mutation, and not one due to crossing colored and white forms. The piebald charac- ter appears in crossing to be dominated by the uniform colora- tion, whatever may be its tint. For example, a spotted gray- and -white mouse crossed by a uniformly black mouse gives a uniformly colored gray mouse, showing that the coupled 92 Experimental Zoology characters, spotted versus color uniform, are independent of the coupled characters gray-black. Cuenot adds, therefore, a new character to the formula representing the uniform color, U, viz. its antagonist the spotted color, S. From the point of view of heredity the spotted condition is peculiarly interesting, since it appears to be continually vary- ing, and by suitable selection this spotted character may be carried through a regular progression until the dark color may almost disappear. The depigmentation begins at the tail, toes, and ventral surface of the body; more rarely there is a small spot on the top of the head. This is the condition found not infrequently in the wild gray mouse. Through selection the caudal and ventral white areas enlarge and the latter in- vades the flanks, right and left, finally meeting dorsally, pro- ducing a white girdle. The white invades the muzzle, then the head, where it may unite below with the ventral spot. Fi- nally there remain two pigmented regions, both dorsal, one anterior and one posterior. The eyes always remain black. Cuenot is not sure that the development of the white may be carried so far that the black totally disappears from the hair, his experiments on these points not being sufficiently complete. Cuenot studied the problems connected with the heredity of the spotted condition by making the following combinations : — 1. Cross between spotted and uniform coat. 2. Cross between a form much spotted (with white) and a form bearing the least possible amount. 3. Cross between two forms much spotted. These experiments may now be considered in turn. I. Cross between Spotted and Unijorm Coat. — The couple uniform, U, and spotted-pigmented, S, follow rigorously the Mendelian rule of dominance with disjunction of the gametes. The spotted character is dominated in the first generation by uniform color. If a much-spotted mouse is crossed by one uniformly colored, the offspring that result show the dominant color without trace of spots. This result is all the more para- doxical, because if we cross a much-spotted mouse with an Ex'perimental Hybridizing 93 albino, descended (extracted) from an animal with a uniform coat, one might be led to suppose that the white of the albino might go to augment the white of the spotted parent, but on the contrary these hybrids are uniformly colored. In the next gen- eration the two characters, V and 5, separate, i.e. disjunct, and the offspring give the Mendelian proportions V^-zJJS -h 5, i.e. one spotted to three uniform. II. Cross between Two Forms unequally spotted withWhite. — If a much-spotted individual is bred to one very little marked with white, having for example only a little white at the end of its tail, the offspring shows that the maximal dark mark- ing is the dominating character. All the young have the tail par- tially white, but no other white marks on the body. Yet these young are not all exactly like the darker parent, since the degree of tail marking, for example, may be quite variable. In the I next generation, when the young are inbred, the phenomenon j of disjunction appears. Two groups of offspring arise, one ! oscillating around the least amount of white (one grandparent j type), the other around the most white-spotted type (the other i grandparent type). III. Cross between Two Much- spotted Forms. — Without ex- •I ception the young are spotted, but in variable degrees. The j cross may even produce albinos if the two parents are hybrids, 1 including the character A. These albinos, in turn, possess la- ( tent the spotted character. Progression of the Spotted Condition by Selection. — Beginning ; with mice little marked with white and excluding, in each gen- : eration the less-marked individuals, Cuenot found that the white : areas could be increased slowly but regularly, so that in two and a half years mice were obtained that contained much white, and differed to a large extent from the first forms used. The :t details of the spotted type seem not to be represented in the ■j germ-plasm, because the young and the parents are not identi- |i cal. Local factors appear to determine the limits of variation. ! In regard to the characters of albinos, it has been pointed out ji that they carry in a latent form the characters of the race from 94 Experimental Zoology which they have sprung/ If, for example, the characters spotted and uniform coloration be considered, the number of possible latent characters contained in albinos is given by Cuenot in the following table : — HOMOZYGOTES OR PURE RACES HETEROZYGOTES Monohybrids Dihybrids AGU AGUABU AGUABS ABU AGU AYU AGU AYS AYU ABU AYU ABU AYS AGS AGS ABS ABS AGS A YS AYS ABS AYS All these forms “prove to exist” and may lead to diverse re- sults when different albinos are bred. Only by hybridizing can the latent characters of the albino races be brought to light. Cuenot thinks that the results of a number of authors find their • correct interpretation in the latent character in the albinos employed. In a later communication Cuenot gives the results of some further experiments with gray, G ; black, B ; brown, R ; and yellow, F; and with the corresponding albinos, AG, AB, AR, and AY. A most remarkable result was found in the case of the behavior of the yellow race. It dominates all the other colors, yet when a yellow mouse is crossed, for instance, with a gray, half of the offspring only are yellow, the other half being gray, or black, or brown (according to the recessive colors pres- ent). In Mendelian terms this means that the yellow mouse never produces pure yellow gametes alone, but some yellow and some of another color (gray, black, or brown). The same result follows when white mice, having recessive yellow, arc crossed with gray, black, or brown. There result not only yel- low offspring, but the other colors as well. If the yellow is a ‘ Whether an albino mutant, differs in this respect from an extracted albino cannot be stated. Cuenot appears to deal with both types, regardless of their origin. 95 Experimental Hybridiziiig heterozygote, it should give, when crossed with a pure race, half yellow and half the other color (as stated above), according to the formula: CYCG X CGCG = CYCG (yellow) + CGCG (gray). The gray offspring should be pure and never produce any yellow. Such was found to be the case. Thus of 355 young there were 177 yellow and 178 gray or black. The gray or black do not include the recessive yellow — at least yellow mice never appear in their progeny.^ The remarkable fact, referred to above, is that it is impossible to obtain a pure yellow race, GY. Theoretically one would expect to obtain, Cuenot thinks, such a race by crossing two similar, heterozygote, yellow mice. Thus — CYCG X CYCG = GFGF+ 2 CGCY 3 yellow CGCG I gray One third of the yellows should be pure yellow with gametes all CF’s. They should breed true to their color. Of 81 yellow mice obtained by such combinations, not one proved to be pure CF’s. In connection with this result, Cuenot finds that there are always fewer yellows than expected on the last formula, which gives 75 per cent yellow to 25 per cent grays. If we : assume that this is due to the absence of CYCY, then we should expect 66.6 per cent yellows to 33.3 per cent grays; but neither ' does this occur,^ and the proportions are more nearly 75 per cent to 25 per cent. Therefore Cuenot offers the following ex- planation: Since his yellow mice were never homozygotes, it follows that the combination of GF with CY can never occur in fertilization, although Cuenot thinks that other results ■ show that a disjunction of GF from the other colored charac- ij ters takes place. This means that these gametes, GF, can 11 never meet to give the zygotes having the formulas G YC Y, i| j' [ * Nevertheless yellow must have been latent, as in the case of albinos that ^ contain a latent color and still breed true. 1 i, ^ Professor E. B. Wilson has pointed out that even if the CY spenn never | Ijl fertilizes the CY eggs, the expectation would still be 75 :: 25, because the CY 1 I eggs would be fertilized by other sperm. I f 96 Experimental Zoology AY AY , and CYAY. In other words, selective fertilization occurs. It seems to me that this is improbable and that a simpler assumption may account for the results. Cuenot’s yellow mice were obtained through an albino of unknown ancestry. He crossed them with gray (or with black) mice, and obtained dominant yellow mice, to which he assigns the formula CYCG. These when inbred should give, according to Cuenot, on the theory of disjunction of the gametes, CY- and CG-gametes. The offspring would then give the Mendelian proportion — I CYCY + 2 CYCG + I CGCG. But no mice represented by CYCY were obtained. It seems to me more probable from the results that the yellow does not separate from the other colors, and if so all the germ-cells would be on my view CY(CG) or (CY)CG. Such forms inbred would not give CYCY, as Cuenot assumes, but the dominant heterozygote CY(CG). This point of view assumes that the yellow is so slightly prepotent in the extracted dominant, CY(CG), that the gray may dominate in half the germ-cells, giving CY(CG) and CG(GF). If this is true, gray mice would appear in one- fourth of the offspring of these dominants. The yellows differ on this point of view from all other extracted dominants in the failure of the yellow to remain dominant in the germ-cells. Schuster has made a number of pairings between gray and white mice. Seventy of such families (Fi) were gray ; ^ two families contained yellow mice and gray mice ; one family con- tained four chinchilla mice only; and one contained two chin- chillas and one gray. The appearance of these yellows in the first hybrids is ascribed by the author to the presence of yellow in the white parents, — the yellow dominating the gray of the first hybrids. Whether the same explanation will account for the chinchillas is not known, because the dominance of the chin- chilla has not been tested. * Containing 342 mice. I Experime^ital Hybridizing 97 The gray hybrids were paired with albinos and 537 1 young produced, of which 261 were albinos and 276 colored — a close approximation to the expected equality. The colors . were of various kinds, but gray greatly predominated. When ; gray hybrids were inbred, they produced 119 albinos and c 308 colored — a rough approximation to the expectation of I I to 3. Haacke has published the results of a large number of crosses ■ between white mice, colored mice, and Japanese waltzing mice, i. Although his experiments were not carried out to test the Men- ! ' delian law, yet, as the author points out, they closely conform I to this law. Haacke states his general conclusion in a some- f 'what involved way, namely, that the possible races of species f I equal the number of possible kinds of germ-cells, i.e. equal ? a sum that consists of as many factors as a particular species -has independently variable qualities or germinal portions. In r.the sum each factor equals the number of all possible modifica- K tions of its properties. LITERATURE, CHAPTER VII -Casile W^E and Allen, G. The Heredity of Albinism. Proc Amer . , Arad, of Arts and Sd. XXXVIII. igi, nroc.Amer. “°5,uris ’"Amh‘‘7„d“E'‘"''‘ o'*' Pig^'^tation ches les Muris. Arch. Tool. Exper. et Gen. Sdr. 3, X. Notes et Revue. tmStT'Lr Zod^Expdn'et a Oarbishire, a. D Note on the\p<;i U ^905- ing Mice with Eurok:" AYbl't^e^' BTomefrift““ TuroperAlbrno Mcr'^ometrika® 98 Experhnental Zoology Third Report on Hybrids between Waltzing Mice and Albino Races : On the Result of crossing Japanese Waltzing Mice with “Extracted” Recessive Albinos. Biometrika, II. 1903. On the Result of crossing Japanese Waltzing Mice with Albino Mice. Biometrika, III. 1904. Davenport, C. Review of von Guaita’s experiments in breeding mice. Biol. Bull. II. 1900. Mendel’s Law of Dichotomy in Hybrids. Biol. Bull. II. 1901. Color Inheritance in Mice, Wonder Horses and Mendelism. Sci- ence, n. s. XIX. 1904. Guaita, G. von. Versuche mit Kreuzungen von verschiedenen Rassen der Hausmaus. Ber. Naturf. Ges. zu Freiburg, X. 1898. Zweite Mittheilung ueber Versuche mit Kreuzungen von verschiede- nen Hausmausrassen. Ber. Naturf. Ges. zu Freiburg, XI. 1900. Haacke, W. Ueber Wesen, Ursachen und Vererbung von Albinsmus und Scheckung, und ueber deren Bedeutung fur vererbungs theo- retische und entwicklungs-mechanische Fragen. Biol. Centralbl. XV. 1875. Die Geseze der Rassenmischung und die Konstitution des Keim- plasmas zuchtanalytisch ermittelt. Arch. f. Entw. mech. der Or- ganismen, XXI. 1906. CHAPTER VIII EXPERIMENTS WITH OTHER MAMMALS AND WITH BIRDS Experiments with Guinea Pigs The hair of wild guinea pigs or cavies shows the same three pigments present in mice, viz. black, chocolate, and yellow. Domesticated cavies may show the following colors : the agouti cavy with the hair containing the black, yellow, and chocolate pigment. The yellow cavy with only yellow pigment. The red cavy is a dark, yellowish red. The chocolate cavy has chocolate pigment predominating. The black cavy has black pigment predominating over the other two. The albino cavy is principally white, although an entirely white individual seems never to occur ; for pigment is generally found in the extremities of the body, on the feet, or nose, or ears, and sometimes in hairs I ■ on the body also. In addition to these there occur spotted or pied races. Any I one of the colors may be combined with white. A great variety ■ of markings result in this way, since the pigment areas may be of different colors. I Brindled cavies have black and red hairs interposed in the • same patches. Roan animals have white hairs interspersed with red ones. Silvered cavies have black and white hairs. 1 Castle has carried out through several years experiments with differently colored and marked races of these animals. > Crossing Uniformly Colored Types. — The self-colored races, as well as the albinos, breed true : thus Castle found albinos bred i inter se gave 1 56 young, all albinos ; pure pigmented individ- ■J uals gave 261 young, all pigmented. On the other hand, al- 1 binos mated to pigmented animals gave in the first generation lOO Experimental Zoology 314 young, all pigmented ; and these last give in later generations some albinos. The results agree fairly closely with Mendclian expectations. Castle has found, as had Allen and Cuenot for mice, that albinos, although breeding true, may carry latent or suppressed Fig. 7. Upper figure, short-haired, smooth coat red guinea pig. Lower figure, long-haired, rough coat albino. (After Castle.) colored characters. Thus one albino individual, a male, when mated with red females produced offspring marked with black. Another albino crossed with the same and with other red females gave young marked with black and red in about half the cases, the other half showing only red or yellow, but no black. A third albino mated with red females produces only red or yellow offspring, never black ones. In the light of these results, Castle lOI Experiments with other Mammals makes a distinction between recessive and latent characters. Recessive characters are those that disappear in the Mendelian sense when brought into contact with a dominant character. This recessive character is transmitted distinct from the domi- nant character in half the gametes of the hybrid. A latent character is a “condition of inactivity in which a normally domi- nant character may exist in a recessive individual or gamete.” This latter is illustrated by the latent pigment carried by ex- tracted albinos. Cross-breeding is necessary to bring the latent character to activity again. When elementary pigmented types, or so-called pure races, of different colors are crossed, neither color dominates perfectly in the offspring. For example, when a black individual is mated with a red one, the offspring are blackish, but not so black as the black parent, red being also present, although masked somewhat by the black. The agouti type carries the three pig- ments black, chocolate, and red-yellow. When crossed with pure black it gave in one case one agouti and two black indi- viduals. Agouti crossed with red gave four young, three of them agouti spotted with red, one red spotted with agouti and white. Black animals seem to contain always some red pigment, which may appear when crossed with albinos. Red animals may, however, be free from black. Red crossed with white gives results that depend on the latent pigment characters borne by the albino, so that black offspring may appear among the others. A case of special interest is found in white animals with black eyes, which are therefore not albinos.^ They do not seem to he al- binos but may contain recessive albinism. They arose from spotted ancestors, and Castle regards them as spotted animals themselves with the pigment spots obliterated except in the eyes. These animals bred inter se or with albinos produce offspring with colored patches of greater or less extent. Whether by selec- tion the white animals with black eyes could be made into a fixed race remains to be shown, and judging from what breeders of ^ Two such animals appeared in Castle’s experiments. 102 Experimental Zoology cattle have been able to do it seems not improbable that this might be accomplished. Castle draws attention to the curious point that red and yel- low cavies, having no black pigment in their coats, do not trans- mit black coat-pigment to their offspring, although they do transmit black eye-pigment. It would be erroneous, he thinks, to conclude from this that the eye-pigment is something alto- gether different in its inheritance from coat color, because when mice with coat patches but devoid of eye-pigment are mated with albinos, the offspring have pigmented eyes — a character that neither parent possessed Heredity of the Rough Coat. — Some races of domesticated guinea pigs show the hair arranged in whirls or rosettes. When best developed the rosettes are found around the following paired centers: (i) the eye, (2) a point immediately behind the ear, (3) the shoulder, (4) a point dorso-lateral on the body, (5) the hip, (6) the groin, (7) each of the single pair of mammae; and from two unpaired centers, viz. (8) the middle of the fore- head, and (9) the navel. The direction of the hair is also re- versed on the toes. These rough-coated individuals breed true. Wlien crossed with smooth-haired individuals the rough character dominates. The rough character of the offspring is usually as fully developed as in the rough parent. However, certain smooth individuals when crossed bring about a weakened condition of the rough character, some of the rosettes being less developed or even absent. These partially rough individuals may transmit to their descendants the fully rough condition. The result is important in that it shows that what we must regard as a new character in the species, viz. a rough coat, dominates when a back cross is made.^ On the other hand, Castle has also found that repeated crossing of rough individuals with prepotent smooth ones results in further weakening of the rough character until it is almost eliminated — one after another of the rosettes disap- pearing. The weakening does not follow a definite decline, but ' See Castle’s analysis, pp. 47-50. Experiments zvith other Mammals 103 in some cases may be slow, in others sudden, so that the inter- mediate steps are passed over at once. Heredity of the Long Coat. — A race of long-haired cavies is known, either smooth (Angoras) or rough (Peruvians). The two sets of characters, long versus short, and smooth versus Fig. 8. Upper figure, long tiaired, smooth coat albino guinea pig. Lower figure, short-haired, rough coat, black-red pigmented guinea pig. (After Castle.) rough, are independent of each other in their behavior in heredity. The long-haired is recessive in relation to the short-haired or ordinary types. The long hair is not present at birth. The coat of the short-haired guinea pigs reaches its maximum length (about 4 cm.) not far from the age of one month, and is then gradually shed and renewed. On the other hand, the hair of the long-coated animals is apparently not shed at this period, but 104 Experimental Zoology keeps on growing. At three months it is 6 to 9 cm. long. If shedding now takes place the animal never acquires longer hair; but animals that reach the age of four months without shedding their longest hairs may ultimately acquire hair 14 to 16 cm. long. It has been stated that long hair is recessive to short hair. When two long-haired individuals are mated, having hair of different lengths, the offspring have hair like that of the shorter- haired individual. In the next generation both long- and short-haired individuals occur, but the number of long- coated individuals exceeds the Mendelian expectation. A pure albino, long-haired, and rough-coated individual, was crossed with nine different pure, pigmented, short-haired, and smooth females. Twenty-nine young were produced, all pig- mented, short-haired, and rough-coated. Thus while two of the dominant characters come from the smooth-haired father, the third comes from the rough mother — the rough coat dis- sociating from the long hair to influence the arrangement of the short hair of the progeny. Castle and Forbes have recently extended Castle’s earlier experiments with short- and long-haired guinea pigs, ^\dlen crossed the short hair dominates, but the hair of the hybrid is nearer the upper limit of variation of the pure, short-haired type, showing perhaps the influence of the other character. Wlien the hybrids were inbred not only the two grandparental types reappeared, but a new intermediate type was present. This type contained individuals of all intermediate grades between the long- and short-haired types. It contained, first, individuals whose hair grew continuously from the age of twenty days on, but much more slowly than does the long hair, and shows a ten- dency to break off at lengths much less than that of the long hair. The long hairs were less numerous, as if some of the hairs only were continuous in growth, while the other hairs ceased to grow after a time. The alternative characters behaved somewhat like a mosaic. In other individuals the hair ceased to grow, i.e. was determinate in growth, but not until it had reached a length of 60 to 80 mm. These cases seemed to be a blend, but no Experiments with other Mammals 105 sharp line existed between the blend and the mosaic condition. In all there were 29 short-haired, 12 intermediate, and 10 long- haired individuals. The Mendelian expectation for these 51 individuals would be 38 short-haired and 13 long-haired. Thus there are fewer long-haired and fewer short-haired individuals, and the presumption is that each kind has contributed to the intermediate group. When a long-haired individual was mated with a short- haired dominant-recessive (which gives in other Mendelian cases equal numbers of dominant- recessives and recessives) there were produced 14 short-haired, 17 intermediate, and 19 long-haired offspring. The results seemed to show that many of the germ- cells of the dominant-recessive had intermediate characters, and were not segregated into the two groups of pure Mendelian gametes. The authors conclude that the intermediate group is probably a new hybrid combination, and that its germ-cells do not split up into long- and short-haired types. A few other experiments were also carried out that seemed to show that cross-breeding produces a “contamination of the gametes.” Instead of pure types separating in the germ-cells of the hybrids, some mixing occurs. Another cross increases the amount of contamination, or at least it produces a larger number of intermediate forms. The authors appear to look upon the “contamination” as a partial or incomplete separation of the characters in the germ-cells of the hybrid. The evidence may be equally well interpreted, I think, to mean that the results are due to incomplete dominance of the dominant character. In other words, it seems to me that these results may be more easily “explained” on my interpretation of the behavior of the gametes in Mendelian cases than in the •“ modified ” Men- delian explanation of the authors. Castle and Forbes found in a family of short-haired guinea pigs a few individuals with hair about twice as long as that of their parents. Mated together they produced all long-haired olT- spring of the same kind. By selecting the best long-haired individuals for two generations a race of imperfectly long-haired io6 Experimental Zoology guinea pigs was produced, which compared favorably with the intermediate groups described above. Crossed with long-haired individuals two kinds of offspring were produced, with long and intermediate hair, but with no definite line of separation. Experiments with Rabbits Hurst has carried out a series of experiments with rabbits that have given results of unusual interest, especially in connection with the inheritance of color and of length of hair. Two races that were known to breed true were used, namely, white An- goras and Belgian hares. The former is an albino breed with pink eyes and silky hair. These animals have a peculiar habit • of swaying the head when at rest. The Belgian hare has a pig- mented skin, dark eyes, and short yellow-gray fur. Wlien crossed the hybrids were pigmented like the Belgian hares, but the hair had lost the yellow color and was gray, like that of the common wild rabbit. When these first hybrids were inbred they produced 14 distinct types in the second generation, viz. : — 1. Hair short, pigmented, gray, uniformly colored. 2. Hair short, pigmented, gray, marked. 3. Hair short, pigmented, gray, Dutch marked. 4. Hair short, pigmented, black, uniform. 5. Hair short, pigmented, black, marked. 6. Hair short, pigmented, black, Dutch marked. 7. Hair short, albino, white. 8. Angora, pigmented, gray, uniform. 9. Angora, pigmented, gray, marked. 10. Angora, pigmented, gray, Dutch marked. 11. Angora, pigmented, black, uniform. 12. Angora, pigmented, black, marked. 13. Angora, pigmented, black, Dutch marked. 14. Angora, albino, white. This epidemic of variation ” in the second generation of hybrids has been the common experience of experimenters both in ani- mals and plants, and before the Mendelian principles became Experiments with other Mammals 107 known remained practically unexplained. By the aid of the Mendelian principles we are able to see at once that there are at least four pairs of distinct characters concerned in the offspring of the second generation, each pair being inherited independently of the other; namely, short hair versus longhair, pigmented coat versus albino, gray versus black coat, uniform versus marked coat.^ Hurst takes up these four pairs of contrasting characters and deals with them separately. Short Hair versus Long Hair. — The short hair of the Belgian breed rarely exceeds one inch, while the long hair of the Angora breed may exceed six inches. The first hybrids had short hair, the influence of the Angora not being apparent. A careful ex- amination, however, revealed what appeared to be faint traces of the Angora influence in both length and texture. The hairs of the hybrid coat were slightly longer, seemed softer to the touch, and were apparently more densely distributed than in the pure short coat. These traces of Angora influence are slight and might easily be overlooked. \\flien these short-coated hybrids {F-P were inbred, there were produced 171 young, of which 70 reached the age of two months or more, when the character of the hair becomes manifest. Of these 53 were short- and 17 long-haired. This is a close ap- proximation to the Mendelian expectation of 51 to 17. The short hair was like that of the grandparent, the long hair like that of the other grandparent. When the long-haired Angoras were bred together they produced in the next generation only Angoras. The short-coated individuals mated together gave both short-haired and Angoras in Mendelian proportions. Pigmented Coat versus Albino. — The cross gave in one case 26 totally pigmented individuals; in another case the fore-feet showed some white markings, which Hurst thinks is not due to the albino influence, but to Dutch marking latent in the albinos. The hybrids, when inbred, produced 132 pigmented and 39 albinos, the Mendelian expectation being 129 : 43. In the marked coat there is white on the ends of the feet and the tip of the nose. The Dutch is an extreme form of this marking. io8 Experimental Zoology The albinos subsequently bred true ; the pigmented types were of two kinds, — pure and hybrid. Gray versus Black. — When the yellow-gray Belgians were mated to the white Angoras, wild gray hybrids were produced. These, as stated above, when inbred, gave both colored and white offspring in Mendelian proportion, but some of the colored in- dividuals were black instead of gray. There were 85 grays to 25 blacks. The grays, as stated, were like the wild gray instead of the yellow gray of one grandparent, although a few appeared to contain somewhat more yellow than the wild type. The blacks had no gray, but it is interesting to note that after the first molt a few white hairs appeared, which increased in number with each molt until some of the individuals resembled the silver-gray breeds (chinchilla). The blacks when inbred produced only blacks, the grays were of two kinds, — pure and hybrid. “The sudden appearance of the black character in the sec- ond generation was quite unexpected as there had been no black individuals in the ancestry of either of the original parents . . . for at least eight generations, and probably many more. The fact that these black individuals appeared in about the proportion of one quarter, and bred true at once, was very significant from the Mendelian point of view. It suggested that the hybrid grays of the first generation were giving off gametes, one half of which contained the factor for black coat color. That it was not intro- duced by both is clear from the absence of black in the first gen- eration.” It could not have been introduced with the Belgians because these mated to black gave only grays. The black must, therefore, have been introduced with the albinos. Hurst carried out some further experiments that seemed to substantiate this view. One male, albino Angora mated with four black docs produced 16 black young; but another albino female mated with black produced 5 black and 6 gray young. Hurst inter- prets these results to mean that the first albino gave off gametes that all carried black, while the second albino gave off gametes some of which carried black, others gray. When the white male and the white female used in these c.xperiments were mated, only Experiments with other Mammcils 109 white offspring were produced ; but when one of these offspring was mated with a pure black, 5 blacks and i gray were produced. The albino must have carried, therefore, both gray and black. Unijorm versus Marked Coat. — It was found that the uniform coat of the Belgians breeds true, but mated with albinos gave, in two cases, somewhat different results: in one case the off- spring showed no trace of markings (white) ; in the other case 1 5 of the young were marked with more or less white on the fore- feet, shoulders, breast, nose, and forehead. The subsequent history of these two sets of offspring was as follows: (a) Three of the uniform or self-colored individuals were bred together and with a pure self-colored individual, pro- ducing 35 self-colored and 2 slightly marked individuals having a few white hairs on the tip of the right paw. The third gen- eration of the uniform individuals gave only uniform offspring. (h) Four of the marked individuals produced 67 young, of which 16 were uniform, 34 were slightly marked, and 17 had the maxi- mum of white (Dutch markings). In the third generation three of the uniform individuals produced 14 uniform and i slightly marked. Also in the third generation the Dutch marked produced 10 young, all Dutch; and three of the marked indi- viduals produced 3 uniform, 12 marked, and 2 Dutch. Hurst interprets these results to mean that one of the Angoras, al- though of pure stock, contained Dutch markings, latent, and when crossed these appeared. When these gametes united with the pure uniform gamete the slightly marked individuals of the first generation were produced. Afterward segregation of the gametes occurred, so that subsequently gametes for uni- form and for Dutch markings appeared in about equal numbers with the result that in the second generation there were one quarter pure uniform, one half hybrid marked, and one quarter pure Dutch. It is interesting to note that this analysis leads to the conclusion hat “ the coat-pattem characters — unlike the previous characters dealt with — are neither dominant nor recessive toward one another, but when crossed give intermediate hybrids in the first generation. In the second and third gen- I lO Experimental Zoology erations, however, these characters appear to follow the ordinary Mendelian rules of segregation and gametic purity.” Hurst calls attention to the similarity of these results to those of Cuenot with spotted mice. The results are in agreement as far as the marked coat is a unit-character following Mendelian lines, and in so far as it may be carried by albinos in a latent state. The two results differ in that the marked coat in mice is completely re- cessive to uniform coat, while the Dutch markings in rabbits are neither dominant nor recessive at first, but give variable hybrids. Castle has made a few experiments with rabbits, but the re- sults gave little that was new in principle. “A cross between two different types of albino rabbits, Himalayan and pure white, shows imperfect dominance of the Himalayan character in the offspring, but complete segregation among their gametes.” Long-haired rabbits bred to short-haired individuals give off- spring with short hair. When rabbits with ears of different length are mated, the offspring have ears intermediate in length. In this character, blending appears to take place, and neither dominance nor segregation. Experiments with Rats Rats have been used much less than mice, and the results seem to be more comphcated. Crampe has published the results of a large number of experiments, extending over ten years ; but as the experiments were made before the importance of Mendel’s theory was appreciated, it is difficult to interpret from this point of view the data obtained. No more striking instance could be given of the insight into cross-breeding experiments furnished by Mendel’s law than a comparison of the work before and after this period. Confused and irregular as the earlier results appear, they arrange themselves into orderly groups in the light of this law. It is, of course, difficult now to show in all cases that Men- del’s law unravels Crampe’s results, since the records arc often incomplete on important points, where further tests arc requi- site to interpret the result. Nevertheless Bateson’s analysis of Crampe’s data indicates that the outcome shows in many Ill Experiments with other Mammals cases the Mendelian expectation. Crampe found that when wld gray rats, Mus decumanus, were bred to albinos, the off- spring (i^i) were of two kinds, viz. either (I) gray like the wild rat, or (II) gray with white marks. If the former (I) were bred inter se, the following types appeared : — 1. Self -gray. 2. Gray with white marks. 3. White and gray. 4. White (albino). 5. Black-and-white. 6. Black with white marks. 7. Black without marks. If the other group of offspring (II) was used, i.e. inbred, all of the preceding types except 3 (white and gray) and 5 (black-and- white) were produced. Bateson states that the great variety of types that appear here is difficult to interpret, but that such occurrences are by no means uncommon. He suggests that two classes of germ-cells may be present either in the albinos or in the wild gray rat. The albino is recessive to all the other six types, as shown by crossing these with albinos. The extracted albinos bred inter se, whatever their origin, gave only albinos. In this connection Crampe makes another statement of interest. Albinos that had been bred true for several generations behaved differently from extracted albinos. The former albinos were simply re- cessive on being crossed with colored rats, while the extracted albinos gave a mixture of ancestral types when crossed with colored types. The result appears to be similar to Cuenot’s with mice where the ancestry of the albino appears as a factor in the product. On breeding inter se each of the seven types given above (i^i), Crampe found that the offspring {Fp belonged to the following types : — Type I might give types 12 4 67 Type 2 might give types 1234567 1 1 2 Experimental Zoology Type 3 might give types 3 4 5 Type 4 might give types ■ 4 Type 5 might give types 4 5 Type 6 might give types 4 567 Type 7 might give types 4 6 7 Thus the wild [ type I is dominant to all the others, i.e. its offspring may belong to any one of the other types which must have been recessive in its germ-cells. The gray forms i, 2, 3, are also' dominant, in the same sense, to the black forms. The albinos give albinos only. “ It appears that types 3 and 5 could be ultimately bred true. As to 6 and 7, the evidence is not very clear; but as I understand the account, neither was completely freed from throAving the other. The breeding in these types was the least successful and extensive. Possibly they are illus- trations of the Mittel-rassen of de Vries. It is especially noteworthy that the gray-and-white type 3 and the black-and- white type 5 do not give rise to self-gray gametes or to self- black gametes, a fact found again in mice. We see, therefore, that there are gametes for black-and-white and for gray-and- white, each of which may behave as a single character and domi- nate over albino.” ^ When pure black-and-white rats were crossed with the vdld gray rats, all the colored types might appear in generation {Fp except albinos. In other words, the black-and-white do not separate, they are not resolved in the germ-cells, as other experi- ments also indicate. Crampe found further that black-and- white individuals that gave albinos in the first generation when bred inter se also gave albinos when bred to albinos. In this case the black-and-white individuals had probably arisen from a cross between black-and-white and albino, so that the albino (and not the white of the black-and-white) gave the white mice just mentioned. On the other hand, Crampe found that when the black-and-white rats did not themselves throw albinos, they did not do so in the first generation when bred to albinos. * Bateson, Proc. Zool. Soc. Experiments with other Mammals 1 1 3 In the light of this analysis that Bateson has made of Crampe’s data, there can be little doubt that Mendel’s law applies to many at least of the phenomena of heredity in rats. This is made highly probable by the recent results of Doncaster. He finds that there are only two color types, brown (or gray) and black, and only two color ■ patterns in the colored -and-white individuals. In addition there are albinos, but these may carry in a latent condition either the uniform black or brown color, or the pie- bald markings. Crosses between the black or the brown wild rat with the albino may bring out the latent characters of the al- binos. Doncaster points out that Crampe’s work shows that brown (gray) dominates black, and both brown and black dominate white. The self or uniformly colored races i, 2, 6, 7 (and those having small white areas below) dominate the pie- bald condition 3 and 5. Doncaster states that Crampe’s brown forms i, 2, 3, correspond exactly with the similar black forms 7, 6, 5, “but are less simple to work with since they may contain recessive black.” He finds two varieties of type 6, — one with much white and one mth very little. The latter belongs, in his opinion, to the uni- form type. Crampe failed to make this deduction, so that one of his forms was probably heterozygous. Evidence that the inheritance in rats is Mendelian was found by Doncaster in a number of the crosses made to test this question. Experiments with Cats Doncaster has brought together a number of records, ob- tained from owners of pedigree cats, that show the color in- heritance of certain breeds. He examined more especially the question as to why tortoiseshell cats are nearly always females. His conclusions, as will be seen, have an important bearing on the problem of dominance in relation to sex. Tortoiseshell kit- tens may be obtained in any of the following matings : — (a) Tortoiseshell 9 by tortoiseshell $ (b) Tortoiseshell 9 by black or blue S 1 14 Experune7ital Zoology (c) Tortoiseshell 9 by orange ^ id) Orange 9 by orange $ {e) Orange 9 by black or blue $ (/) Black or blue 9 by orange $ In all of these matings, in addition to tortoiseshell, kittens of other colors may appear, viz. : — {a) Tortoiseshell 9 by tortoiseshell $ gives tort., orange, black. {h) Tortoiseshell 9 by black or blue $ gives tort. 9, orange $ , black ^ , 9 . (c) Tortoiseshell 9 by orange $ gives tort., orange, black. (d) Orange 9 by orange $ gives either tort., orange (or blue) or only orange. {e) Orange 9 by black $ gives tort., 9, orange $. (/) Black 9 by orange $ gives tort., black (and probably orange). {g) Black 9 by black $ gives only black (or blue). From these results it appears that tortoiseshell is a heterozy- gous color produced by the meeting of orange and black gametes. The explanation that tortoiseshell cats are nearly always females and rarely males is owing to orange nearly always dominating in the male over black, while in the female the dominance of the orange is incomplete, so that tortoiseshell results. In other words, in the female sex the orange and the black both exist to- gether, while in the male sex the yellow usually dominates. A few examples will make the conclusion clearer. For instance, in mating (e) when an orange female is crossed with a black male, only tortoiseshell and orange kittens arc produced ; if both the orange and the black breeds are “pure,” the female offspring are tortoiseshell and the males yellow. In the reverse mating (/), where a black female is crossed with a male orange, the male may be heterozygous {i.e. having both black and orange germ-cells), hence black kittens may also be produced. The kittens vill be black males or females, tortoiseshell females, and orange males. When a tortoiseshell female is mated with a black male, the male offspring will be orange, because the tortoiseshell is hetero- Experiments with other Mammals 1 1 5 zv^^ous. It is also evident why orange females are very rare, although orange males are common, since in all matings m which one of the parents is black, orange can appear only in the male offspring. “If, therefore, the great majority of orange males contain recessive black, when they are paired with tor- toiseshells, only a quarter of the kittens will be pure orange, and only half of these females.” The preceding statements show the relation of the colors orange and black. The inheritance of two other colors was also ex- amined ; namely, cream and blue. Cream appears to be a dilute form of orange, and blue of black. The blues breed true (when derived from yellow ancestors) and are therefore recessives or homozygous. A cream female and a blue male give blue tor- toiseshell (blue and cream), cream males, but no blues, since the cream dominates incompletely in the female, completely in the males. On the other hand, a blue female and a cream male give blue tortoiseshell females, blues of both sexes, and possibly cream males. These and other results show that the dilute forms behave in the same way as do the stronger colors. Thus cream is dominant over blue in the male, but when blue and cream meet in the female a tortoiseshell results. It has been stated that male tortoiseshell cats are known, although they are rare. It must be assumed that in such cases the dominance of the yellow is incomplete as in the female. This means that while complete dominance is usually associated with the male character, it is not necessarily always associated with this sex. It is interesting to find that when a male tor- toiseshell is mated with a female of the same color, the kittens are tortoiseshell, orange, and black. This is what is expected on the assumption that the germ-cells of the tortoiseshell are black and orange (with the alternate character latent on my view). The prepotency of different tortoiseshell individuals (males) seems, however, to vary. It should also be pointed out that the colors described above may be associated with a certain amount of white which reap- pears in the offspring without, however, affecting the inheritance ii6 Experimental Zoology of the other colors. The piebald character stands as a unit con- trasted with uniform coat, but is independent of any particular color. Data for Other Mammals and Man A few other cases in mammals, that seem to show discontinu- ous inheritance, are known. Castle and Davenport have both called attention to cases of so-called wonder-horses, i.e. horses with remarkably long mane and tail. In the case of “Linus I” the mane was i8 feet long and the tail 21 feet. The parents and grandparents of these horses also had unusually long hair, which increased in successive generations. The data are insuffi- cient to show the relation of dominance and recessiveness in this case, but the persistence of the long hair seems to indicate its dominance.^ Harper and Hurst have recently examined certain data in regard to the inheritance of coat color in horses. Harper deals with the problem from the standpoint of prepotency of certain colors in regard to ancestry, selection, age, and sex. Hurst shows that bay and brown colors dominate completely chestnut, and there are definite indications that these two colors follow Mendel’s law. Some statistics recently published (1904) by A. G. Bell have furnished Davenport with material to study the relation of black color to white color in sheep. The data show that when three white individuals having as far as known white ancestors were crossed with black sheep, the 13 lambs resulting {F-l) were white, showing the dominance of white. Of 20 offspring from black parents all were black. ^ When a black (recessive) individual was mated with a dominant white (one of whose parents was white and one black), 26 lambs were white and 25 black, which is the Mendelian expectation. When a dominant-recessive white was mated to a dominant-recessive white, 40 were white and 7 were black. The expectation is 25 per cent black. The ’ In guinea pigs the long hair is recessive. * One uncertain case of while is given that is not above suspicion. Experiments with other Mammals 117 number of black lambs is too small on the assumption that chance meeting of equipotent “pure” germ-cells brings about the results. Poulton ^ has given some records of polydactyl cats that appear to be explicable, so far as they go, along Mendelian lines. Three young were produced from a polydactyl female by an unknown father. They were all polydactyl. If polydactylism dominates over the normal condition, this result is simple dominance. One of these individuals (Tj) produced three litters (by unknown fathers), in which four normal and six abnormal kittens ap- peared. If the father was normal, five normal and five poly- dactyl young would be expected. Thus : — P + N N -f N 2 NP + 2 NN Only two kinds of discontinuous inheritance that may possibly follow Mendel’s law have been shown for man. Albinism, ac- cording to certain data collated by Castle, may perhaps follow this rule. The cases referred to were albino negroes. Albinism is, of course, different in this respect from white. In the latter case, blending of the black and white occurs to produce mulattoes. The other case is that of polydactylism. Fachenheim has given some statistics,^ that Davenport has examined from the point of view of Mendelism. The accompanying table gives the inheritance through. three generations: — Gen. NxP I. P N N(xN) P(xN) N(xN) P(xN) N(xN) PxN J L ! I I I II. 6N3P4N3P 7N 3N 2P 8N 2P 2N P (xN) J 2 N 3 P N(xN) P(xN) N(xN) N(.xN) I l_ 1 I 12N3N2P 3N N ^ Nature, 1883. ^ Jena Zeitsclirift, XXII, 1888. ii8 Experimental Zoology If polydactylism is dominant and the normal condition is reces- sive, the chances are that any polydactyl person has had one nor- mal parent and the germ-cells are therefore P and Nd Paired with a normal individual (N + N), half the children should be polydactyl and half normal. In the above case there were in fact in the first generation four normal and four polydactyl chil- dren. In the second generation when normal offspring paired with normal, 5 polydactyl children and 21 normal were pro- duced ; and when the polydactyl descendants paired with nor- mal, 7 polydactyl children and 12 normal were produced. For the small number recorded, the latter result is not veiy different from 1:1, the Mendeiian ratio. Again in the third generation when P was mated to N, 5 normal and 5 polydactyl children were born. Struthers gives the following case of polydactyl inheritance in man : — I. Px ? I P(xN) 10 P II. 3P. iP(xN) III. 4N 4P The result can only be explained on the Mendeiian view by assum- ing that both parents of the first generation were polydactyls, i.e. produced germ-cells bearing polydactyhsm. It is necessary to make this assumption in order to account for the second genera- tion that descended from one of the first filial generation. Here a polydactyl parent married a normal individual and produced only polydactyl children, showing that no normal germ-cells were present in one parent. Had there been some normal germ- cells, some normal children would be expected, provided the numbers are really large enough to give this result a chance to appear. In the third generation an equal number of the two kinds of offspring are expected, and such are found. A third case is ^ Or P(N) and N(P) on my view. Experiments zvith other Mammals 119 given by Struthers/ A normal man married a woman who had six fingers on the left hand. There were 18 children, only one of whom was abnormal. In this case the polydactylism was not dominant except in one case ; but among the normal chil- dren in the third generation one polydactyl individual is re- corded, indicating that polydactyhsm was in the strain. The failure of the polydactyl condition to dominate in this case, except in one instance, shows how unsafe it is to argue from a few cases to all others. The same character may be a dominant one in certain strains and not in others.^ The preceding records and observations are made much clearer by Castle’s recent experiments with polydactylous guinea pigs. There was born of normal parents a male guinea pig with an extra toe on the left hind foot. The toe bore a claw, which was not connected to the foot by appropriate muscular and tendinous connections.^ From the polydactylous male were obtained 15 individuals with extra toes out of a total of 77 offspring. In sub- sequent generations, partly inbred, the number of extra- toed offspring varied. When the male was paired with females from families in which polydactylism was not known, there were pro- duced about 6.25 per cent extra-toed young. Females, de- scended from the original father, that gave the first polydactylous male, gave 25 per cent extra-toed offspring. Females, them- selves polydactylous, gave 44 per cent polydactylous young. Many of the young of the first male had extra toes on both hind feet, and in several cases they were better developed than in the original male. The extra toes were supplied with all the muscles characteristic of functional toes. Castle has traced the descent of this race through five generations, and has obtained some important data regarding the inheritance of the anomaly. He finds that the “potency” of certain individuals is a more im- portant factor in the transmission of their characters than is their ' Edinburgh New Philosophical Journal, 1863. ^ Several other cases of inherited polydactylism are given by Gregg Wilson. * The same father that produced this “sport” subsequently produced also five others out of 147 offspring. 1 20 Experhnental Zoology ancestry. If the various sires are arranged “in the order of the respective amounts of polydactylous ancestry which they possess, we see at once that this is not the order of their potencies, for those having the same amount of polydactylous ancestry often differ much in the potency with which they transmit the poly- dactyl character.” Wlren polydactylous individuals were mated with normal ones, the results were far from being uniform. Some of the offspring have the extra toes greatly weakened ; in other cases there is no toe at all, while in still other cases the extra toe may be fairly well developed. “The inheritance is neither sharply alterna- tive (Mendelian) nor completely blending.” It is clear that in its inheritance the extra toe of these guinea pigs does not follow Mendel’s law. Castle concludes that the extra toe is inherited in a manner intermediate between blending and alternative in- heritance. The gametes, he thinks, only partially blend in the zygote, producing a variable result. “If the inheritance were sharply alternative, we should expect to get, not a series of gradu- ated forms, but two or at most three sharply distinct groups, but this is not the result observed. If, on the other hand, the inheritance were fully blending, all the offspring of two pure par- ents, or of two cross-bred parents should be alike, but this is not the result observed. We are forced to conclude, therefore, that there occurs a partial blending of gametes [characters] in the zygote, and a partial segregation as the zygote gives off gametes.” Castle points out, further, that partial blending is the more common result of hybridizing, since both sharply alternative inheritance and complete blending are rare. By selection the breeder is able to produce an almost pure race by picking out the more potent individuals in each generation. It is interesting to note that the potency of the male is a germinal variation, tend- ing toward determinate inheritance, and not simply an extreme fluctuating variation due to external conditions. Hence, in selecting prepotent individuals the process involves the choice of certain individuals that transmit certain qualities in a high degree. Experiments with Poultry 121 rather than involving the selection of extreme somatic fluctua- tions. These two kinds of selection may be superficially similar, but involve in reality an important difference in principle. Experiments with Poultry The different breeds of poultry have furnished Bateson and his co-workers with excellent material for experimental study. The domesticated breeds differ not only in color, but also in the char- acter of the combs, in the feathered or unfeathered condition of the shanks, in the number of toes, in crested and uncrested heads, and the habit to sit or not to sit on the eggs. These characters are inherited discontinuously, showing dominance and recessive- ness, and also often giving the Mendelian ratio. In his earlier work (begun in 1898 and published in 1902), Bateson used principally Indian Game and white Leghorn, [ but subsequently brown, and white Dorkings and one Wyan- I dotte were used. He found that as a rule pea comb, rose comb, I and extra toe are dominant characters, while single comb and j normal foot are recessive. Nevertheless, the first generation j sometimes shows blending in various degrees, and in consequence i the dominance may be considerably reduced. When the j are inbred, some of their offspring show one character, and I others the other “in proportions following Mendel’s law with some consistency,” but here again the results do not always conform to the expectation. Other conflicting results are also recorded that are difficult to explain, j In a recent communication (published in 1905) by Bateson and I Punnett further details are given ; and in a supplementary paper I by C. C. Hurst, some experiments with Leghorns, Houdans, i black Hamburgs, and buff Cochins are described. As the re- sults of Hurst are more easily presented in a less technical form, I have relied on them mainly in the following account. The leaf comb of the Houdan is dominant over the single comb of the Leghorn and Cochin. In a few cases the dominance is complete, but in the majority of cases it is incomplete — interme- diate combs being produced. 122 123 Experiments with Poultry The Hamburg rose comb is domi- nant over the single comb of the Leg- horn and Cochin, the dominance be- ing complete. The Hamburg rose comb is domi- nant over the Hou- dan leaf comb, al- though the latter, as stated above, is itself dominant in other combinations. In the second gen- eration (i^2) Mendelian expecta- tion is largely rela- ized. The white plum- age of the Leg- horn dominates over the black of the Houdan and Hamburg, and also over the buff of the Cochin. In only a few cases, however, is this dominance of white complete. In the majority of Fig. io, A. Buff Cochins. {^Reliable Poultry . . . Journal.') cases it IS incom- plete, the white feathers being ticked with black, or there are patches of buff or brown. A few exceptional cases were noted where white did not appear to dominate. Experimental Zoology 1 24 The black plumage of the Houdan and the Hamburg domi- nates over the buff of the Cochin, but incompletely, the black being marked and shaded with brown. When the hybrid dominant whites (jPj) were mated, the off- spring were dominant whites and recessive blacks in the pro- portion of 3.1 : 1. When the hybrid dominant whites (F{) .vere mated with pure recessive blacks, there were produced dominant whites and recessive blacks in the proportion of i : i . When the hybrid dominant whites (F{) were mated with a pure buff, they gave whites and blacks in the proportion of i : i. The experiments in which animals with the normal number of toes are crossed with races having an extra toe give results of unusual interest. In general, the extra toe (of the Houdan) is dominant over the normal foot (Leghorn, Hamburg, Cochin). In some cases the dominance is complete, i.e. the extra toe is full size ; in other cases all gradations in the size of the extra toe were found “down to the mere duplication of the nail.” The extra toe was found in some cases only on one foot, the other appearing as in the normal. There were some cases in which the normal foot appeared to dominate, but whether such cases are real dominance of the normal, or the failure of the extra toe to appear in an individual that has it potentially present can only be deter- mined by subsequent breeding. When the dominant extra-toed hybrids {F-P) were bred together, they gave dominant extra toes and recessive (apparently normal) individuals in the proportion of 3.8 ; i ; when the Fps were bred to pure recessives without extra toes, they gave dominant extra toes and recessives, apparently without extra toes, in the propor- tion of 1 : 1.5. There were two exceptional cases of F^, in which the normal foot seemed to dominate. These were a male and a female. When mated they gave 22 chicks, of which 14 had an extra toe and 8 had normal feet. The result shows that the parent birds are really RD’s, since chicks with extra toes appeared when the birds were bred together. This conclusion was confirmed by breeding the cockerel to another pure individual with (reces- Experiments with Poultry 125 sive) normal feet, which gave ii individuals with and 13 without extra toes. A similar experiment with the hen {p j) gave analo- gous results. Evidently in this case a character usually domi- nant has become recessive. It is clear that unless great pre- cautions are taken, such cases might easily be put down, in other experiments, amongst the recessives. Fig. 10, B. Houdans. {Reliable Poultry Joiir^tall) The shank feathering of the Cochin dominates over the clear shank of the Leghorn, Houdan, and Hamburg, but the dominance is always incomplete. When the E^’s were bred together, they produced a large number of E2 with feathered shanks, and a few recessive clear shanks in the proportion of 28.7 : i. The Mende- lian expectation is 3 : i. In other combinations the expectation is much more nearly realized. Hurst concludes that “ the Men- delian principles are at work in these aberrant phenomena, but are masked by something not yet perceived.” Hurst’s general conclusions are as follows: Dominant char- acters are rose comb, white plumage, extra toes, feathered shanks. 126 Experimental Zo'dlogy white and blue shanks, crested head, brown egg color, and broodi- ness ; while leaf comb, single comb, black plumage, buff plum- normal foot, clear shanks, uncrested head, white egg color, and non-broodiness are all recessive to the dominants given above. Some o] these recessives may, however, be dominant over others. Xhus leaf comb and black plumage are dominant ov’er single comb and buff plumage that remain recessive. Fig. io, C. Silver-spangled Hamburgs. (^Reliable Poultry Jour7iaV) Hurst points out further that dominance may be complete, when it is indistinguishable from pure dominance, or incomplete, showing the influence of the recessive character in different de- grees. For some characters the dominance is always complete; in some it is always incomplete; and in others it is sometimes complete, but more often incomplete. The incomplete domi- nants appear to be about twice as numerous as the complete. In the second generation F.^. dominants are again complete and incomplete. It is to be remembered that all the preceding characters behave Experiments with Poultry 127 independently of each other, there being no correlation between the characters usually associated with a particular breed, so far as their inheritance is concerned. One of the most curious results in crossing black and white fowls is the occasional appearance of a blue (or Andalusian) color, which consists of a minute patchwork of black and white. It has been known for some time that this color does not breed true, but a pair of such blue fowls gives rise to 25 per cent black, 50 per cent blue, and 25 per cent white. The explanation of this is that the germ-cells of the blue individuals are black and white, hence the result ; but the special condition that leads to the occasional formation of blue by the combination of black and white is not known. Possibly the presence of latent colors determines the result. The most recent experiments with poultry are those of Daven- port. His work confirms many of the results already obtained by Bateson and Hurst, and also establishes the relation of domi- nance and recessiveness for some new characters. It is these latter points that will be especially considered here. A cross was made between the single comb, black Minorca and white- crested, black Polish. These races and their hybrids are shown in Fig. ii ; 1-6. In the hybrid (5 and 6) the comb is single anteriorly and bifurcated behind. There is much variety in the extent to which the comb is split. In fact, it was single in one case. Neither parent type can be said to dominate, the Minorca having a large single comb, and the Polish a much- reduced bifid comb. Davenport suggests that these two types of comb may both be dominant types that combine to form the Y-shaped comb. When the hybrids were inbred, there resulted, in a total of loi offspring, 29.7 per cent showing single comb, 46.5 per cent Y-shaped comb, and 23.8 no comb (oronly papillos). Two interpretations are possible. On the assumption of two dominant and two recessive types, viz. (i a) median comb and (i b) no median, and (2 a) no splitting and (2 b) splitting, the re- sults agree more or less with the expectation. But on another assumption the results conform even more closely, viz. if we Fig. II. Male and female black Polish fowls, Figs, i and 2; male and feniale black Minorca, Figs. 5 and 6; and hybrids, male and female, higs. 3 an 4. (After Davenport.) form constantly reproducing itself. Further work must decide between these alternative views. 128 Experimental Zoology assume that single comb and V-shaped comb (that of the Polish) are contrasted characters, and the Y-shaped comb is a combined 129 Experiments with Poultry This idea of combined dominance was first suggested by Bate- son and Punnett to explain the walnut comb of certain hybrids. Thus when rose comb and pea comb are crossed, walnut comb results. The gametes produced by these hybrids are of four types and in equal numbers ; namely, single, rose, pea, and wal- nut, giving when inbred 9 walnut (rose-pea), 3 rose, 3 pea, i sin- gle comb. The interpretation offered by Bateson and Punnett for these facts is that the characters of the original parents with rose and pea combs are rose and no pea, and pea and no rose. The contrasted characters are rose and absence of rose, pea and absence of pea. When rose and pea bearing gametes meet, the walnut comb is produced. The results follow the rule for two characters. In other words, rose and pea combs are not them- selves contrasted characters, but the allelomorph of each is its absence. The authors point out, however, the danger involved of making general assumptions of this sort when two characters meet. The nostrils of the Minorca are slitlike, and this dominates the wide-open nostril of the Polish, but the dominance is imper- fect in the first hybrids. In the next generation, the split nostril is present in 21 per cent, but even in this generation the high or dominant nostril is frequently imperfect. The Polish fowls have a large cerebral hernia on the top of the head, covered by a hardened layer of outer brain coat or dura mater, and by the skin. The Minorca breed has a normal head. In the hybrids, P’2, not a single case of hernia occurred, but most of them showed evidence of their mixed ancestry in the pres- ence of a frontal eminence. In the second hybrid generation, P’2J the hernia reappears again in 23.5 per cent of cases — a close agreement with Mendelian expectation. A crest is present in all the hybrids, but always reduced in size. “The crest is dominant, but dominance is imperfect.” The crest is larger in the female hybrids, as it is in the female Polish breed. In the second hybrid generation the crest was absent in about 30.7 per cent. Davenport concludes that in this cross the dominance is incom- K 1 30 Experimental Zoology plete in all cases; the nearest approach to typical Mendelian inheritance is exliibited in the crest, but in the first generation it is always reduced. The white color of the crest is recessive in the male hybrids, but is not entirely absent in the females. The high nostril is recessive, yet it shows its influence in the first hy- brids. The comb in the first hybrids is different from that of either parent, yet in the second generation there is a partial return to the two parent types. An interesting cross was made between the Japanese long- tailed fowl or Tosa fowl (Fig. 12 ; 2) and the white Cochin Ban- tam (Fig. 12 ; i). In the Tosa fowl the feathers of the tail show continuous growth, reaching in extreme cases 18 feet, and generally 7 to 8 feet. There is a marked sexual difference in the Tosa breed, but not in the white Cochin. The male hybrids had the coloration of the Tosa cock except that every feather was barred with white (Fig. 12 ; 3). The female hybrids were like the Tosa hen, excepting that the shafting was much broadened, and the saddle feathers and the secondaries were black and buff barred. In the second generation the two original types reappeared. There were 28.1 per cent white (Fig. 12 : 4) and 7 1.9 per cent pigmented individuals. However, of the whites, only five were without reddish pigment, showing that they were contaminated by the cross. “The 41 pig- mented individuals showed a curiously mixed lot of coloration. Of 41 mature females 6 are like the female Tosa fowl, without barring, but sometimes with wider shafting than the male Tosa fowl. The remainder have feathers of the back and wing coverts barred with lighter, even with white — a condition not found in the female first hybrids. One of these shows a mixture of female Tosa and female Partridge Cochin coloration. As no Partridge Cochin is involved in the immediate ancestry, this looks like a ‘rever- sion ’ ; the characteristic has probably lain latent in the White Cochin. Of 10 males, 2 showed no trace of white, and may consequently be considered as homozygous. The remainder are more or less barred with white. One bird shows a remarkable mixture of Tosa and male Partridge Cochin coloration.” The Experiments with Poultry 13^ statement shows besides the contamination of the second gen- eration the complexity of the result due to latent characters other than those patent in the breeds used. In regard to tail length, it was found that the first hybrid Fig. 12. White Cochin Bantam, Fig. i; male Tosa fowl, Fig. 2; hybrid, Fi, male, between last. Fig. 3; second generation hybrid extracted recessive. Fig. 4. (After Davenport.) males developed abnormally long middle-tail feathers, but the tail was not so long as in the Tosa fowl. It will be observed that in this cross white is apparently not dominant, as it is as a rule in other combinations ; but evidence of the white is seen in the barred feathers of the first hybrids. 132 Experimental Zoology “The white Cochin has no sexual dimorphism in plumage color, while the Tosa fowl is strongly dimorphic. Every one of the first hybrids is dimorphic in plumage coloration, the two sexes resembling, except for the white, respectively the female and the male Tosa fowl. It is striking to see how from a germ-cell of the male Tosa fowl either a bird colored like the male Tosa or a bird colored like the female Tosa may arise. The male germ- cells contain the Anlagen not only of the male characteristic but also of the female characteristic.” In another case two races, both having sexual differences, were crossed. These were the Tosa and the dark Brahmas. The hybrids were also sexually different, .showing the dominant color of their respective sex. Thus the red wing- bar and white wing- bar are found in the males, and the shafting and penciling of females in the female hybrids. There are two races of fowls that have aberrant feathers. In the Frizzled jowl (Fig. 13 ; 2) the contour feathers have a shaft con- vex inward so that the feather is lifted up and even turned forward. The primaries of the wings show groups of the barbs that are twisted in corkscrew fashion. In ih.e. Silky jowl (Fig. 13;!) the contour feathers are like down feathers, with a weak shaft, and the barbs are subdivided, producing a fluffy effect. The quill feathers of the wing and tail are less modified. The Silky fowls used in the experiments were white, the Frizzle were dark (black, red, and buff). Some of the hybrids were white (Fig. 1314) and some were dark (77.4 per cent). Here the white is neither dominant nor reces- sive, but the explanation of the result is not clear unless one or both races are impure in their color inheritance. None of the hybrids showed any silkiness, which is, therefore, seen to be reces- sive to non-silkiness. Only 6 of the 10 hybrids were frizzled, the other four having flat feathers. Davenport explains this ap- proach to equal numbers on the ground that his frizzled animals produced both frizzled and plain bearing germ-cells. In another cross the breed of black-breasted red Game (Fig. 14; 2), in which the tail and rump are entirely absent, was Experiments with Poultry 133 crossed with the white Leghorn breed (Fig. 14 ; i). Of 24 hybrids, 12 were white (Fig. 14; 3), or prevailingly so; black was usually present, and more rarely some buff. The other 12 hybrids were 4 I Fig. 13. Silky fowl, Fig. i ; Frizzle fowl, 2; extra toe, Fig. 3; [ between Silky and Frizzle, Fig. 4. (After Davenport.) hybrid, Fi, black and white barred, or black with reddish. The tail and rump were normal. Rumplessness is recessive in the strain used by Davenport. 134 Experimental Zoology Aside from the details of Davenport’s work, his chief results consist in showing that in nearly all characters examined the influence of the recessive character is to be seen in the first hy- brids, and even in the second hybrids, so far as obtained, the 2 Fig. 14. Single-comb White Leghorn, male, Fig i ; rumpless Game male, Fig. 2 ; hybrid, Fi, between last two races. Fig. 3. impurity of the offspring is often apparent. As yet in only a few cases have the experiments been carried to the second generation. The results are evidently complicated by the im- purity of some of the strains that were used and by the presence 135 Experiments with Pigeons of latent characters in these strains. Despite the fact that some of the birds were bought as “pure” stock, i.e. stock that breeds true as long as inbred, yet the presence of latent qualities in them is admitted. Under these circumstances it does not seem to me “contrary to experience” to admit that “pure” strains carry latent characters. Davenport contrasts the dominance versus the recessiveness of new characters with the original characters of fowls as indi- cated by their presence in the wild parent . species of Indian game and of Aseel. He finds that the new characters dominate as often as they are recessive. Hence there' is no reason to suppose that new characters are at a disadvantage in respect to dominance as contrasted with old characters. Experiments with Pigeons Darwin has given the results of several experiments in crossing pigeons. There are also several other recorded results of hy- bridizing races of pigeons. The hybrids appear to be variable in color, but some of the markings peculiar to the wild Rock Pigeon are apt to appear. For example: A male “Nun” that is white, with head, tail, and primary wing feathers black, was crossed with a red “Tumbler.” Neither parent had any blue in the plumage, nor bars on the tail or wings. Of the sev- eral young, one was red over the back, but the tail was as blue as that of the Rock Pigeon; two others were quite similar; a fourth was brownish and the wings showed a trace of a double bar; a fifth was pale blue over the back, breast, and tail, but the neck and primary wing feathers were reddish ; the wing had two distinct red bars. Wdien a black “Barb” was crossed with a red “Spot,” the young were black, or dark, or pale brown, sometimes slightly piebald with white. Six of these birds had double wing bars, etc. When a black Barb was crossed with a snow-white Fan- tail, some of the hybrids were black with a few white feathers, others were dark, reddish brown, and others snow-white. None of them had wing bars. 1 36 Experimental Zoology These and other experiments with pigeons show that the results are more complicated than in the case of fowls. It is not possi- ble to bring the results under a single point of view at present. Ewart’s experiments^ with pigeons indicate that while certain kinds of crosses may give rise to offspring resembling the Rock Pigeon, yet in other cases a more immediate ancestral color may come out. A dark blue Fantail, having all the characteristic bars of the Rock Pigeon, was bred to a less pure blue Fantail. On two occasions an absolutely pure white Fantail was produced. This result, Ewart thinks, is due to a reversion to a white grand- parent. Ewart also crossed a white Fantail with a white Pouter. The offspring was white in color, but in form resembled the Pouter. A hybrid between an “archangel” and an “owl” was bred to a white Fantail. The two offspring were blue, one of them being almost identical with the wild Rock Pigeon, more es- pecially with the Indian variety. Not only was there reversion in color, but in form as well. Ewart seems to think that reversion amongst closely inbred races of dogs, horses, and pigeons leads to a sort of rejuvenes- cence of the stock. Those individuals showing the ancestral characters prove to be stronger and more active. The reversion to the type of the Rock Pigeon, that seems to play often so conspicuous a role in these experiments with pigeons, recalls the return to the gray color in mice when fancy breeds are crossed, but in the second generation of mice there is a re- turn in some of the forms to the parent types. How far this occurs in pigeons is not clear from the evidence at hand. A brief but important note in alternate inheritance in pigeons is given by Staples-Browne. Webfoot sometimes suddenly appears in pigeons. A pigeon of this sort crossed with another having normal feet produced six normal-footed offspring. These individuals (Fj), inbred, produced in one case nine with normal feet and three with webbed feet. Another pair (F^), however, produced seventeen normal birds. Extracted web-footed individuals produced six web-footed. It appears that the latter condition is recessive to normal feet. ' The Penycuik Experiments. Experiments ivith other Mammals 137 LITERATURE, CHAPTER VIII Ahlfeld. Missbildung des Menschen. Leipzig. 1880. Ballowitz, E. Ueber hyperdactyle Familien und die Vererbung der Vielfingerigkeit des Menschen. Arch. f. Rassen und Gesellschafts Biologic, I. 1904. Barrington, A., and Pearson, K. On the Inheritance of Coat Colour in Cattle, Part I. Biometrika, IV. 1906. Bateson, W., and Saunders, E. Experimental Studies in the Physiology of Heredity. Reports to the Evolution Committee of the Roy. Soc. Report I. London. 1902. Report II. 1904. Bateson, W., and Punnett, R. C. A Suggestion as to the Nature of the “Walnut” Comb in Fowls. Proc. Cambridge Phil. Soc. XIII. 1905. Bardeleben, K. V. Hand und Fuss. Verhandl. Anat. Gesell. VIII. 1894. Castle, W. E. The Heredity of “Angora” Coat in Mammals. Science. 1903. Note on Mr. Farabee’s Observations (on Negro Albinism). Science, n. s. XVII. 1903. Heredity of Coat Characters in Guinea Pigs and Rabbits. Carnegie Institution of Washington. 1905. Recent Discoveries in Heredity and their Bearing on Animal Breeding. The Pop. Sci. Monthly. 1905. The Origin of a Polydactylous Race of Guinea Pigs. Carnegie Insti- tution, XLIX. 1906. Castle, W. E., and Forbes, A. Heredity of Hair-Length in Guinea Pigs and its Bearing on the Theory of Pure Gametes. Carnegie Insti- tution, XLIX. 1906. CoRRENS, C. Ueber Bastardirungsversuche mit Mirabilis-Sippen. Erste Mittheilung. Ber. deutsch. bot. Gesellsch. XX. 1902. Crampe, H. Kreuzungen zwischen Wanderratten verschiedener Farbe. Landwirtsch. Jahrbucher, VI. 1877. Die Gesetze der Vererbung der Farbe. Landwirtsch. Jahrbucher, XIV. 1885. Doncaster, L. On the Inheritance of Tortoiseshell and Related Colours in Cats. Proc. Cambridge Phil. Soc. XIII. 1905. On the Inheritance of Coat Color in Rats. Proc. Cambridge Phil. Soc. XIII. 1906. Ewart, J. C. Variation: General and Environmental. Sci. Trans. Roy. Dublin Soc. VII. 1901. ‘ Farabee, W. C. Notes on Negro Albinism, Science, XVII. No. 419. 1903. Haacke, W. ‘Ueber Wesen, Ursachen und Vererbung von Albinismus und Scheckung und ueber deren Bedeutung furvererbungstheoritische und Entwickelungsmechanische Fragen.’ Biol. Centrbl. XV. 1895. Die Gesetze der Rassenmischung und die Konstitution des Keim- plasmas. Arch. Ent. mech. XXI. 1906. Harper, E. Studies in the Inheritance of Color in Percheron Horses. Biol. Bull. IX. 1905. Hurst, C. On the Inheritance of Coat Color in Horses. Proc. Roy. Soc. B. LXXVII. 1906. N(^es on the “Proceedings of the International Conference on Plant Breeding and Hybridization, 1902.” Tour, of the Roy. Horticul. Soc. XXIX. 1906. 138 Experimental Zoology Experimental Studies on Heredity in Rabbits. Jour, of Linnean Society, XXIX. 1905. Kallmann. Handskelet und Hyperdactylie Verb. Anat. Gesell. 1888. Raynor, G. H., and Doncaster, L. Experiments on Heredity and Sex- determination in Abraxas grassulariata. Report 74, Meet. British Assoc. Cambridge. 1904. Rijkebusch. Bijdrage tot de Kennis der Polydactylie. Archiv Neer- landaises, XXII. Schuster, M. A. Hereditary Deafness. Biometrika, IV. 1906. Starles-Browne, R. Experiments on Heredity in Web-footed Pigeon. Report 74, Meet. British Assoc. Adv. Sc. 1905. Strutbcers. On Variation in the Number of Fingers and Toes. Edin. New Phil. Journ. 1863. Wilson, G. Hereditary Polydactylism. Joum. of Anat. and Phynol. XXX. 1896. Woods, F. A. Mendel’s Law and Some Records in Rabbit Breeding. Biometrika, II. 1903. Zander. 1st die Polydactylie als Theromorphe Varietat oder als MissbU- dung anzusehen. Virchows Archiv, CXXV. 1891. 4^ ooooo OCXXX) ooooo X ooooo ooooo ooooo ooooo ooooo 1 2345 12345 ooooo ooooo ooooo ooooo ooooo 12345;' 12345 12345 r ooooo i/ 12345 Fig. 15. Helix hortenses. First (top) line, a bandless and a banded shell. When indm these are inbred they give bandless and banded (third line) in the proportion of 6 to 2 (i.e. 5 sives, as shown by inbreeding (fourth line). (After Lang.) =345 345 ^ 12345^ 12345 ooooo d 12345 12345 ^ ooooo / ■ 12345 ooooo IJM5 ^ 12345 12345 12345' I CXXX30 ooooo d 12345 r 12345 ooooo 12345^ ooooo d 12345 12345 12345 of these two kinds are paired the first generation (second line) are all bandless. When Of these two are extracted dominants, four are dominant-recessives, and two extracted reces- CHAPTER IX EXPERIMENTS WITH SNAILS, MOTHS, AND BEETLES The European snails, Helix hortensis and Helix nemoralis, have been studied by Lang, whose experiments in breeding them have extended over several years and, though still in progress, have already yielded decisive results on a number of important points. The young snails require at least two years, generally three, and often four years to reach maturity. The animals, although hermaphroditic, do not self-fertihze, as isolation experi- ments have shown. The sperm, received during copulation, may remain alive for several years in the receptaculum seminis, therefore wild individuals cannot be used safely in breeding ex- periments, but the young snails must be first reared and isolated and paired in order that the parentage of their offspring can be certainly known. The bands of pigment on the shell are the chief characters that Lang has studied in his experiments. Colonies of snails are sometimes found in which only two kinds of individuals are met with, — those with five bands and those without bands, no intermediate types existing in such colonies. These colonies furnish the best materials for breeding experiments. In such colonies the banded individuals are “pure,” that is, they breed true to their type. It is more difficult to obtain “true” indi- viduals without bands. Those are most likely to breed true that are found in colonies in which only bandless individuals exist. If virgin banded and bandless individuals are allowed to pair and are then separated, each will make a nest in the ground and deposit from 40 to 60 eggs, or more. The banded indi- 139 140 Experimental Zoology viduals may be designated by — and the bandless by 12345 00000 The young of the first generation are all without bands. In other words, the banded condition is the recessive. The formula used IT r 1 00000 d by Tang for these individuals is • 12345 When the individuals of the first generation are paired, the offspring of the second generation are of two sorts (like the grand- parents). Those having the dominating character (bandless) are to those with bands as 3 to i . In other words, the Mendehan ratio appears. That this is really the case is shown by further 12345 12345’ and produce only banded offspring. The bandless are of two kinds, one third true or “pure” dominants, and two 00000 There are no external char- experiments. The banded individuals are the recessives. thirds dominant recessives. 00000 12345' acters that distinguish the pure dominants from the dominant recessives, since both are bandless ; but if they are separated in pairs and their offspring obtained, it will be found that in some cases all of the offspring will be bandless. They have arisen , , . - , , . 00000 00000 , by the union of two extracted dominants, x , or by 00000 00000 the union of one extracted dominant and of one dominant reces. 00000 00000 sive, X . In other cases two thirds of the offspring 00000 12345 are bandless and one third banded, showing that two dominant 00000 00000 recessives have paired, x 12345 12345 Lang has also carried out some experiments with snails in which two characters are involved. The individuals of Helix nemoralis or of H. hortensis differ not only in their banding but also in the color of the shell, which may be yellow or red. These colors also form a pair of antagonists that follow the hlendelian law. The double combination of banded and bandless, yellow and red, gives the Mendelian expectation for two contrasted characters. Experiments with Snciils^ Afot/is, and Beetles 14 1 Individuals of these two species also show other antagonistic characters. There are size differences, and these also Mendelize. The form of the navel also differs and gives a fourth character. The preceding breeding experiments relate to differences within the species, but Lang has also carried out experiments between the two species, H. hortensis and H. nemoralis. The results will be described in the next chapter. One point of interest must, however, be mentioned here. As stated above, each species contains individuals that are banded or bandless, red or yellow, etc. These are varietal differences. The point of interest is that when the two species are crossed the offspring — species hybrids — show that one of the antagonistic characters dominates in the same way as when the varieties within the spe- cies are crossed. For example, if a bandless individual of Helix hortensis is paired with a five-banded individual of Helix nemo- ralis, the hybrids are in some cases entirely bandless, in other cases partly bandless and partly five-banded. Lang accounts for these results by assuming in the first case that a “ pure ” bandless form was used, all of whose germ-cells were “ pure ” and hence dominated ; while in the other case a dominant recessive was used which would produce both kinds of germ-cells. On the other hand, . one should not lose sight of the fact that the hybrid-crossing may itself set “free” latent characters, as in mice, so that the results may have arisen in this way. Lang concludes that those varietal characters that Mendelize within the species behave in the same way when different species are crossed. Experiments with Silkworms Elaborate series of experiments with silkworms have been carried out by Coutagne.^ The results of ten years’ work were published in 1903. Although ample evidence is furnished of alternative inheritance, the results are not treated by the author from the point of view of Mendel’s law, although there are indica- tions in many places that some at least of the results might profit- ‘ Bulletin Scientifique de la France el de la Belgique, XXXVII, 1903. 142 Experimental Zoology ably have been thus considered. On the other hand, there are also numerous instances where it seems probable that the inheri- tance is of a different kind. To what the results are due is not clear in all cases, but it seems not improbable that some of the domesticated races of silkworms have originated from different wild species, or even genera; while others have arisen under domestication as sudden variations. Several instances of this sort are given and invite special attention. Some of the races may be crosses, i.e. hybrids, and even although breeding true inter se the individuals may carry in a latent state the quahties of other strains. In the light of these possible complications we can do little more than examine Coutagne’s results as they stand. It is to be hoped that this most promising field of inquiry may be further investigated. Coutagne distinguishes between (i) an “alliage homogene,” or fusion, in which the hybrid character is something new and intermediate between the parental characters (which are united or fused “fondus”); (2) a “mdange heterogene,” or mixture, in which some of the hybrids are like one parent and some hke the other in respect to a particular character, and still others inter- mediate; (3) a “liquation” or separation in which there is no fusion of characters in any of the individuals, but they are strictly like one or the other parent type. The inheritance of the following characters was examined by Coutagne : — The caterpillars (“ worms ”) ; White (with or without a masque) a, I mode albus Black a, 2 “ nigcr Zebra a, 3 “ virgatus The cocoons: Yellow b,i “ flavus White b,2 “ nivetis The moths: White, with or without “ cendre ” markings c,i “ canus Black C,2 “ casteneus Experiments with Snails^ Moths, and Beetles 143 Coutagne states that any one of the three kinds of larvae, white, black, or zebra, may be associated with either of the two kinds of cocoons, yellow or white. This gives six combinations. Any of these six may be associated with either of the two kinds of moths, white or black, giving a total of twelve possible combi- nations. The following results were obtained : — (1) Crossing individuals with two different characters often gives offspring in the first generation that are intermediate in character. Thus the race Chang-hai has a white cocoon that is small and spherical, while the race Jaune Var has a rose-yellow cocoon that is large and ellipsoidal with a constriction. The cocoons of the hybrid have a pale yellow tint, about intermediate in color, ^ the form also is intermediate — ellipsoidal, but less elongated, and the constriction absent or scarcely marked. (2) Crossing individuals of different races may give in the first generation a jusion of two characters, but subsequent generations descended from the first show a mixture of the two characters in question. The statement is also illustrated by reference to cocoon-char- acters, but the line of separation of the colors does not appear ver}^ sharp. (3) When the crossing has given a fusion in the first generation, and in the second generation a mixture (“melange heterogene”), it is possible to produce a homogenous race, by means of selec- tion, that shows the characters that fused. In each successive generation the individuals presenting the united characters, i.e. those “fondus,” must be selected. This statement Coutagne puts in the form of a question, because he has not, he says, indisputable facts in support of it. (4) Crossing individuals with different characters often gives a separation in equal parts of the two characters. For example, an individual of a race having white worms, white cocoons, and white moths was crossed with an individual of a race having ' Two to three per cent, however, were pure white, small, and spherical (type Chang-hai) ; and 4 to 5 per cent were yellow, nearly of the yellow type but less yellow. These Coutagne suggests were due to accidental mixing. 144 Experimental Zoology striped worms, yellow cocoons, and black moths. The results are shown in the table, in which the horizontal lines represent the stages of each individual. No. OF Individuals Worms Cocoons Moths I16 striped yellow black 124 white yellow ‘ black III white white black 108 striped white black ■ Thus there are nearly equal numbers of white and striped worms (235 and 224) ; the same holds nearly for the two characters of the cocoons (240 and 219). No worms intermediate in color were found nor were there intermediate conditions between the cocoons.^ It will be seen that all the moths were black (mela- nitic), yet Coutagne thinks some influence of the white was present. (5) Crossing individuals with different characters gives at times offspring all like one parent, without the other character show- ing any influence in the first generation. For example, an individual of a race with white worms, cocoons, and moths was crossed with an individual having white worms, but yellow co- coons and black moths. All of the cocoons were white. An individual of a race having black worms, white cocoons, and moths was crossed with an individual having white worms, yellow cocoons, and black moths. All of the cocoons were white. An individual of a race having black worms, white cocoons, and white moths was crossed with an individual having white worms, yellow cocoons, and black moths. All the worms were black, about half the cocoons were yellow and half white (262 and 248), without any intermediates. All the moths were black or blackish, but rarely one was almost white. In a third case an individual of a race having black worms, white cocoons, and white moths was crossed with an individual * Certain double cocoons that were rejected did not show intermediate colors. Experiments with Snails, Moths, and Beetles 145 having white worms, yellow cocoons, and white moths. All of the worms were black, half of the cocoons were yellow, and half were white (180 and 188), without any intermediate ones. In a fourth case an individual (female) with white worms, cocoons, and moths was crossed with another (male) having black worms and moths, but yellow cocoons. All the cocoons were white. The worms were 265 black like the father and 253 white like the mother; some of the moths were white, others black, others intermediate. In the group of 265 black worms there were no more black moths than in the group of 253 white worms. Thus the characters black or white worms and the characters black or white moths have no mutual relation. In a fifth case, the female of a variety “ Jaune Var” was used that has the black character of the worm and of the moth almost completely “fixed,” but fixed only recently, while the yellow of the cocoon had been fixed for a long succession of generations. The male was of the race “Blanc des Alpes” and had white characters throughout. The result of the cross gave white co- coons, the paternal type — the reverse of what occurred in other cases where the white was maternal. The worms were 258 white like the father and 182 striped like the mother (3 to i). This predominance Coutagne attributes “to the ancestors of the white worms of the mother; possibly if the striped char- acter had been fixed for a greater number of generations there would have been an equality between the white and striped worms.”' The questionable character of this explanation is at once apparent in the next experiment that Coutange gives, in which the same types were crossed as before and all the worms were striped. (6) It has been found that white cocoons dominate over yellow, but in another combination the reverse was found to be the case, showing that there is no absolute rule for all races in regard to the inheritance of white versus yellow. (7) After crossing individuals of two races which give offspring showing the character of one parent, it often happens that the concealed character reappears in the following generations. 146 Experimental Zoology When Blanc des Alpes are crossed with Jaune Var, all the off- spring have yellow cocoons. The next generation gave yellow and white cocoons with none intermediate; the proportion of yellow averaged in fifteen lots 75.2. Here there seems to be an approach to one fourth that strongly suggests the Mendelian ratio. In other cases, however, an average of 49.3 per cent was obtained. The evidence here is opposed to this interpretation, unless one of the races itself had a latent character that did not appear until the second generation. (8) When two crossed races give a separation in equal parts of the two characters, the following generations give equally again the separation of the two characters without its being pos- sible to realize their fusion in a single individual. For example, two types were crossed, each having striped worms and yellow cocoons. The offspring gave four classes : — Worms Cocoons I 236 Striped yellow 2 80 striped white 3 89 white yellow 4 34 white white In another case the female belonged to a race with white worms and cocoons, the male striped and yellow cocoons. The crosses were : — Worms Cocoons I 95 Striped yellow 2 84 striped white 3 103 white yellow 4 109 white white Seven other combinations are given, but until some principle running through these cases can be formulated it is needless to recount all the results here. It is true that Coutagne attempts to show that the rule in subsequent generations is such that the inheritance of the contrasting characters follow the sequence of Experiments with Snnits^ Moths^ cind Beetles 147 _}. 1 ^ 4- 1 ^ + Jg j which is the same as Galton’s law, but so far as I can interpret his data this is not strikingly apparent. Failing to carry the results discussed above through subsequent genera- tions leaves the matter in an unsatisfactory condition. In only one experiment is the result of the next generation given, which shows, if I interpret it correctly, that each of the four types gives, when inbred (?), individuals amongst which most of the same types reappear, but in very different proportions ; the most striking result being that each type gives a much higher percentage of its own kind. (9) In the course of separation of two characters which takes place during a series of generations it happens at times that when two similar individuals are paired, i.e. both having the same character, the other contrasted character never appears again in their descendants. It has not simply become latent, but has gone entirely. It is obvious that this would happen, according to Mendel’s formula, whenever the individuals are pure dominants or pure recessives. (10) Crossing individuals with different characters often gives in the first generation a mixture, “melange heterogene,” of two characters with marked predominance of one of them in the com- bination, and in this case it is easy, by means of selection in suc- cessive generations of the individuals having the most marked character, to fix rapidly this character and even to exaggerate its relation to the other. In cases of this sort it is not clear that Mendel’s law holds at all, and some other principle must be involved, especially if the author means that the results are obtained by simply discarding the individuals having the disappearing character and allowing the rest to breed together. (11) Crossing two individuals with a different character often gives in the first generation a separation of these characters in a part of the offspring and in the other part a union constituting a new character. Subsequent generations show a separation or liquidation of these characters. 148 Experimental Zoology For example, a female of a race with black worms and white cocoons was mated to a male with striped worms and yellow co- coons. The offspring were : — No. OF Individuals Worms Cocoons 89 black and striped yellow 86 black and striped white 77 black yellow 77 black white The mother had been a white moth, the father a black moth. The offspring (of this table) consisted of a small number of whites, and the majority of a “gris-marron ou moues fonce,” in other words, a melange hdt6rogbne of the white and the black character, but without predominance of either. “The fact of greatest interest in this cross is the fusion, or at least the close juxtaposition, of the black and the striped charac- ters in half of the caterpillars of this lot. These worms at the same time black and striped are most curious and constitute a new character without any intermediate.” The further ' evolution of these new types was as follows: The group with black-and-striped worms and yellow cocoons gave: — No. OF Individuals Worms Cocoons 127 - 90 individuals black and striped yellow .37 black and striped white 47 ( 29 black yellow (18 black white 20 - white yellow 7 white white 129 - 66 striped yellow 63 striped white The group with black-and-striped worms and white moths gave : — Experiments with Snails^ Moths, and Beetles 149 No. OF Individuals Worms Cocoons , f 70 black and striped yellow black and striped white 00 f 27 black yellow «M6i black white white yellow (23 white white I16 striped yellow l77 striped white Coutagne calls attention to the reappearance of the white char- acter that is atavistic on his interpretation. It will be noticed in each group the numbers are nearly exactly halves, thus 70 and 189, 27 and 61, ii and 23, 39 and 77. Although it is not apparent that an application of the Mende- han law is competent to explain all the results of these experi- ments, it is probable that some such rule lies behind several of the obsen^ed cases. Other cases clearly show blended inheritance, and still others show in some characters one kind and in others other kinds of inheritance. It is difficult in many cases to under- stand just what really occurs, but the results show plainly how complicated the problem of inheritance in a single group of forms may be. In a note pubhshed later, Coutagne compares his results with the Mendehan formula and points out that certain classes of his results conform to this law. It seems to me not improbable that if the latent quahties of some of the races be taken into consid- eration, the conformity may be greater than Coutagne admits. On the other hand, it appears probable that some characters do not dissociate according to the Mendehan expectation. A short paper by Toyama on Mendel’s law as applied to silk- worm crosses has very recently appeared, in which it is shown that many of the same characters studied by Coutagne give the Mendehan expectation. Unfortunately no reference is made to Coutagne, although a comparison would have been valuable. When moths of the Siamese breed, having either yellow or 150 Experimental Zoology white cocoons, are crossed, the offspring {Fp produce only yel- low cocoons. When these hybrids are inbred they give two kind of individuals ; namely, those producing white and those producing yellow cocoons in the proportion of 25,03 per cent whites to 74.96 per cent yellows. The whites are extracted re- cessives, and are found to breed true. The yellows are of two kinds, one kind giving only yellow-producing offspring (the ex- tracted dominants) and the other kind producing yellow as well as white in the proportion of 3 to i (the dominant reces- sives). A more complex result was obtained when Japanese “whites” and Siamese “yellows” or when Japanese “whites” and European “yellows” are paired. The first generation give offspring that produce yellow cocoons. When these indi- viduals are paired,, they give in the second generation four kinds of cocoons: (i) pure yellow, 70 cases; (2) pale pinkish yellow, 2 1 cases ; (3) greenish white, 24 cases ; and (4) pure white, 12 cases. The results show apparently that we have to do here with three characters, one of the races used being a monohybrid and the other a dihybrid. In subsequent generations the pure white type breeds true. Some of the yellows give only yellows; others give white (25 per cent) and yellow (75 per cent); others give yellow (75 per cent), flesh-colored (25 per cent), and still others white and greenish white (25 per cent), yellow (56 per cent), and flesh-colored (19 per cent). The flesh-colored forms of the second generation gave white (25 per cent) and flesh-colored (75 per cent). The greenish white type of the same generation gave white and greenish white offspring. Toyama used two breeds of caterpillars, pale whites character- ized by the absence of markings, and striped whites having dark stripes. The former breeds true, while the latter produces some pale whites and is therefore a cross breed form. The hybrids produced by uniting these breeds were white and striped worms in equal numbers. In subsequent generations the whites re- mained true while the striped kind gave both whites and striped. Experiments with Siiails^ Afoths, and Beetles 151 Other crosses were made between a brood showing no markings and yellow cocoons, and striped with white cocoons, and gave the usual Mendelian results of sphtting and recombination. Experiments with Beetles In the Californian beetle, Lina lapponica, two types exist, a spotted type shown in Fig. 4 and a black type shown in Fig. 6. The heredity of these two types has been studied by Miss Mc- Cracken. When the beetle emerges from its pupal case the wings are nearly pure white. The large median black spot and the two lateral spots on the prothorax are present (Fig. i). In the course of ten minutes faint indications of spots appear on the wings (Fig. 2), and after fifteen to twenty minutes these become more distinct, as seen in Fig. 3. In the course of forty-five min- utes the spotted type has reached its final condition (Fig. 4). The black type passes through a stage like that of Fig. 3 to that of Fig. 5, and finally after forty-five minutes to Fig. 6. In a sense the black type passes through the spotted type, the back- ground becoming as dark as the spots themselves, so that the spots can no longer be seen. If these two kinds of individuals are collected at random and isolated in pairs, it is found that the spotted pairs sometimes (in half the cases) produce their own kind only, and in other cases pro- duce both spotted and black types. There are no intermediates. When the blacks are paired they produce in some cases broods that are all black, and in other cases broods in which some of the individuals are black and some are spotted, in the proportion of I spotted to 1.7 black. If a black and a spotted individual are paired, mixed broods are produced with a preponderance of the spotted type. The results for the second generation are especially interesting. If those broods from black parents are picked out in which all of the individuals are black and these are paired, all of their off- spring, without a single exception (in 4985 cases), are black. The third generation is also black. Hence there is a pure race of blacks. The black is evidently the recessive form. 152 Experimental Zoology The results of the second generation are different when the spotted forms from the pure spotted brood produced by spotted parents are paired. In this case both spotted (1021) and black FIG. 16. Development of the two-color patterns of Lina Japonica. (After McCracken.) Experiments with Snails, Moths, and Beetles 153 (345) are produced in the proportion of 3 : i. In two cases, how- ever, all the offspring (265 in one case, 182 in the other) were spotted. If these are paired, they produce only spotted forms. Clearly we have to deal here with extracted dominants. The results may be briefly summed up as follows : the two types of beetle show alternate dominance and recession, the spotted character being dominant, the black being recessive. By isolation both types may be obtained “pure.” In nature both are continually crossing and recrossing, so that the chances are that most individuals are impure, but by. selection pure breeds can be quickly obtained from them. Experiments with the Currant Moth There is a curious case, reported by Raynor and Doncaster for the currant moth (Abraxas grassulariata) , in which there is a rare variety, A. lacticolor, that had previously been found only in the female sex. The variety is recessive in the first generation when crossed with the parent form. The offspring, however {F-l), produce males, all of which are the ordinary variety grassu- lariata, and females, hah of which are like the males and the other hah are var. lacticolor. When, however, a lacticolor fe- male is paired with a {F-l) male hybrid (L $ x G (L) ^), some of the male offspring are lacticolor (and others female). The ex- planation of the transference to the male of the female character is not apparent. Experiments with Tephrosia A series of hybridizing experiments between the moths Tephrosia bistortata and Tephrosia crepuscularia have been described by Tutt (based on records by Riding and Bacot). These' two species are sufficiently similar to have been put to- gether as one by some entomologists because occasionally indi- viduals have been found that could not be referred with cer- tainty to either species, but Tutt describes a number of constant differences between the two forms and regards them as distinct. Both species have a melanitic variation, that of T. bistortata 154 Experunental Zoology being exceedingly rare and almost confined to South Wales, but that of T. crepuscularia is widely distributed and known as aberratio delamerensis. The first generation of hybrids (F-^) is described as containing individuals that are like the one or the other parent form, although each kind may show traces of the other species. Tutt states that he has avoided the use of the term “intermediate” when describing these hybrids because the term has been made to cover so many different things ; but that nevertheless all of the hybrids with few exceptions are inter- mediate to varying degrees, i.e. “almost every specimen appeals, in some part of its facies, to a specialist, as resembhng T. bis- tortata, whilst the same specimen, in other particulars, strikes one as resembling T. crepuscularia.” When inbred the hybrids produce a large percentage of individuals differing much from either parent-form. The crossing of the hybrids, obtained from original reciprocal crosses, tends to produce a mixed prog- eny, some referable to known forms of the crossed species, others quite unlike anything ever obtained in nature. . . . These experiments support Timer’s view that sexual combination can lead to the production of new forms. I doubt, however, very much whether they could be perpetrated without selection,” for if crossed with one of the present species they would “in my opinion” revert to the wild form. Some interesting results were also obtained in regard to the sex of the hybrids. “The greater vigor of the male results largely in the production of female offspring. When the male is of the dominant species, females are developed in fair propor- tion ; when the female is of the dominant species, males are largely in excess.” LITERATURE, CHAPTER IX CouTAGNE, G. Recherches experimen tales sur I’h^r^dit^ chez les vers ^ soie. Bull. Sci. France et Belg. XXXVII. 1902. Sur les facteurs dl^mentires de I’heredite. Comp. Rend. Acad. Sci. Paris, CXXXVII. 1903. Sur les croisements entre taxies diflerentes. Comp. Rend. Acad. Sci. Paris, CXXXVII. 1903. Experiments with Snails^ Moths ^ and Beetles 155 Lang, A. Kleine biologische Beobachtungen ueber die Weinbergschnecke s (Helix pomatia, L.). Viertelzahrschrift d. naturf. Gesell. Zurich. XLI. 1896. Ueber Vorversuche zu Untersuchungen ueber die Varietatenbildung von Helix hortensis Muller und Helix nemorales, L. Fest. von Haeckel. 1904. Ueber die Mendelschen Gesetze, Art und Varietatenbildung, Muta- tion und Variation inbesordere bei unsem Hain- und Garten- schnecken. Verb. d. Schweiz. Naturf. Gesell. 1905-1906. McCracken, I. A Study of the Inheritance of Dichromatism in Lina lapponica. Jour. Exp. Zool. II. 1905. Toyama, K. Mendel’s Laws of Heredity as applied to the Silkworm Crosses. Biologisches Centralblatt, XXVI. 1906. Tutt, J. W. Some Results of Recent Experiments in Hybridizing Te- phrosia bistortata and Tephrosia crepuscularia. Trans. Entomol. Soc. London. 1898. CHAPTER X OTHER KINDS OF HYBRIDIZING Blended Inheritance We have seen in the cases that come under MendePs law that the contrasted characters do not both develop at the same time, the offspring in the first generation being often like one or the other parent. Yet in some of these cases there is evidence that the dominant character may be weakened by the recessive one. We may now consider cases in which the contrasted characters of the two parents fuse or blend completely in the offspring. Cases of the sort are found not only between races, varieties, and elemen- tary species, but this method of union has long been supposed to be a characteristic feature of hybridization when Linnaean species are crossed. The most famihar and striking case of fusion or blending of two characters is found in the mulatto — the result of union of a white and a black individual. The mulatto breeds true in all successive generations, neither the white nor the negro ever appearing again in the pure form. If the mulatto again crosses with the white stock, the dark color is again lessened, but even after several generations of crossing with the white stock traces of the dark pigment remain. Conversely crosses between the mulatto and the black race produce ever increasing shades of darkness in successive generations of offspring. Not only the color, but the character of the hair also shows a tendency to blend in the hybrid. Flourens made crosses between the domestic dog and jackal, the latter being, however, “prepotent.” The horse and the ass give the mule, that is intermediate in many respects, but the 156 Other Kinds of Hybridizing 157 characters of the ass are more prepotent. The lion has been crossed with the tiger and an intermediate hybrid produced. The brown bear, crossed with the polar bear, gave a mixing of colors with the head and neck white. Darwin states that the pheasant crossed with domestic fowls gives a hybrid, showing the pheasant characters “prepotent.” Darwin describes a cross between the penguin variety of the common duck and the Egyp- tian goose that is intermediate in character. Mosaic Inheritance A mosaic character sometimes appears when differently colored individuals are mated. Each character appears in its pure form over certain regions. Thus when a gray and a white rat are crossed, individuals sometimes appear that are mosaics, but it is questionable in this case, and perhaps in all such cases, whether the results may not be due to a latent mosaic character coming to light. As has been pointed out, offspring of the same litter may be different, some being of a single color, others mosaic. Whether spotted orpiebald, domesticated races — horses, cattle, dogs, cats, etc. — owe their origin to a cross between two uniformly but differently colored parents, or are themselves sports that breed true, or have been back-crossed, is an open question. In pigeons, as we have seen, the mosaic character of the offspring is apparent. The results are complicated, how- ever, by the ancestral (latent) blue color appearing in parts of the body. The inheritance of the mosaic pattern in mice and guinea pigs has been already discussed. Hybridization between Linncean Species Most wild species of animals and plants differ from each other in more than in a single character. In the great majority of cases it is perhaps not going too far to state that species differ in all their characters — in some parts more, in others less. As already pointed out this is due, on the mutation theory, to the 158 Experimental Zoology saltations having been in most cases more than a single one, and often in different directions. Now it is almost a universal rule — with exceptions, however — that wild species are infertile when crossed with other wild species, the degree of infertility differing enormously in different species — from complete fecundity to complete sterility. It is also a general rule — again with exceptions — that the more widely different two species are the greater the difficulty in crossing them. When the species are so different that they are put into different genera, the chance of their crossing is small. If the species belong to different families, the chance of crossing is much smaller still ; and if to different orders, there is scarcely any chance at all of their crossing. Crosses between the do- mestic horse and zebras of different species produce infertile hybrids. A jack-ass crossed with a mare (horse) gives a mule, which is sterile. Conversely, a she-ass crossed with a pony gives a hinny, also sterile. On the other hand, the American bison has been crossed with the wild ox of Europe and has produced a fertile hybrid. Similarly, the humped cattle of India crossed with the domesticated ox produces fertile offspring. Crosses between the common goose and the Chinese goose, which are very differ- ent species, give fertile hybrids. Simliarly, for the common duck and the pintail duck, and for different species of pheasants.^ A cross between a fowl (a langshan cock) and the common guinea hen has been brought about, but the hybrid is sterile.^ In regard to the characters of these hybrids no general state- ment can be made. Sometimes the hybrids appear to be inter- mediate in one or more characters ; sometimes the character of one or of the other parent predominates, and in still other cases the hybrid may have characters peculiar to itself. Latent char- acters may also be brought to the front by hybridizing, and these dormant characters seem sometimes to be characters that the ancestors of the species may be supposed to have possessed. ! The preceding cases are quoted from Ewart’s “Penycuik. Experiments,” 1899. ^ Guyer, M. F., Science, XXI, June, 1905. Other Kinds of Hybridizing 159 De Vries has found that some of the elementary species of the evening primrose also show a certain degree of infertility when crossed ; and there can be little doubt that infertility may begin with the appearance of elementary species, and increase in pro- portion to every new change in the germ that takes place. Whether infertility is a general rule for elementary species, may be questioned. Many cases of crossing between wild species of animals and plants have been recorded as occurring in nature, and many more cases have been experimentally brought about, especially in plants, with wild forms kept under domestication. The re- sults are different in different cases, but it is a generally accepted opinion that the species-cross is generally intermediate be- tween the parents. This conclusion needs, perhaps, careful revision in the light of the results of recent years on crossing types that differ in only one character, where in many cases discontinu- ous inheritance is the rule. Darwin was so impressed with the difference in the results of crossing Linnaean species and sports that he concluded that wild species could not have arisen, as sports, since the latter when crossed show discontinuous inheri- tance, while wild species give intermediate forms. Since Dar- win’s time our knowledge of the results of hybridizing has greatly increased, and his argument seems less conclusive, because, in the first place, a single mutation may show incipient infertility, as in de Vries’s oenotheras; in the second place, because the results of crossing elementary varieties and elementary species with the parent forms or with each other do not always show discontinuous inheritance ; thirdly, because wild species have undergone so many changes of different kinds that the results are too complicated for an analysis of single characters; and, fourthly, because discontinuous inheritance may sometimes oc- cur between wild species, if unit characters rather than the ensemble of characters is considered. It is the failure to recog- nize this last point that has probably led to an exaggerated idea of the difference between the inheritance of single variations and of complex variations that characterize Linnaean species. i6o Experime7ital Zoology Lang has studied the hybrids between the closely related species of Helix hortensis and Helix nemoralis. These species are very similar, and it has often been disputed whether they are separate “species.” They differ principally in size, in the form of the peristome of the shell and in the color of the lip, in the form of the “dart” and in the finger-formed gland. The hybrids are infertile with each other. Within the limits of each of these species there are the same kind of varietal differences, and these, as pointed out in the previous chapter, dominate and recede in the first species-cross. It is not possible to test their further behavior, since the species-hybrid is infertile. In re- gard to other characters, Lang states that these also dominate and recede. The hybrids are not intermediate, but have the form of the peristome of H. hortensis and the pigmentation of the lip and of the throat of the shell of H. nemorahs. The dart and the finger-shaped gland are exactly those of H. hor- tensis. Here it is evident that the hybrids are mixed, but that in some characters they are true to one species and in others to the other species. If the numbers of such characters were larger, the hybrid might appear to be a blend of the different characters, while in reality it might be only a mixture of one and the other parental characters. It is evident that the charac- ters must be studied separately in such cases before we can con- clude whether species-hybrids show blending of the parental characters or whether they give mixtures (mosaics) in their char- acters. Some characters may blend, others alternate in their inheritance.^ Dimorphism The word “dimorphism” is sometimes used for cases in which the male and female differ markedly in form, but it is also used for those cases in which two forms of the same sex exist. I shall use the term here in the latter sense. Few cases of this sort exist amongst animals, and no experiments have been made to test the inheritance. The male hercules beetles occur under two ‘ Correns has described similar results in plants. Othej' Kinds of Hybridizing i6i forms, one with the other without horns. There is a large and a small form of earwig. Papilio glaucus has two forms of females, one yellow, the other black. The latter is found over the south- ern range of the species. “In this region both yellow and black forms have been reared from eggs produced by a single female.” Fig. 17. A number of cases are known in the higher plants, and their inheritance has been examined in a few instances. These are all hermaphroditic forms in which two kinds of flowers — both pro- ducing ova and pollen — are present on the same or on different plants. The European cowslip. Primula veris, occurs in two forms, the long-styled and the short-styled (Fig. 17). Each plant bears flowers that belong to one type only ; the two types 1 62 Experimental Zoology never appearing on the same plant. The long-styled, as its name implies, has a long style reaching to the top of the corolla ; the corolla is larger, the stigma globular, the papillae longer, and the pollen grains larger than those in the short-styled form. This long-styled form flowers first, on an average, but the short-style averages more seeds. The stamens are in the middle of the tube in the long-style form, and at the top of the tube in the short-style form. Darwin carried out a series of important experiments on these plants. He calls illegitimate unions those in which long-styled flowers are fertilized by pollen from the same flower or from a similar flower of the same or of a different plant. Similarly for the short-styled form : legitimate unions are those between long-styled and short-styled flowers. The behavior of the offspring from seeds of legitimate and ille- gitimate unions is most surprising. In one case ^ an illegitimate union between short-styled forms produced seeds that germinated so badly that only 14 plants were obtained, of which 9 were short- styled and 5 long- styled. In another experiment the stigma of a long-styled flower was fertilized by the pollen of a long-styled flower. Three long- styled plants resulted. From these, in turn, self-fertilized, 53 long- styled offspring were obtained ; from their seed 4 long-styled plants ; from their seed 20 long-styled ; and from their seed 8 long- styled and 2 short- styled. In another plant, Lythrum salicaria, 3 forms occur: the long-styled, the mid-style, and the short-styled type. The sta- mens also occur under the same three lengths. Figure 17 shows the conditions of the three kinds of flowers. There are 6 possible legitimate unions and 12 illegitimate ones. To test these two kinds of unions, legitimate and illegitimate, 18 distinct kinds of crosses must be made. The results of these experiments are shown in the following table : — * “DifTcrent Forms of Flowers,” p. 217. Other Kinds of Hybridizing 163 Nature of Union No. OF Flowers . Fertilized No. OF Capsules Average No. OF Seeds PER Capsule Average No. of Seeds per Flower Fertilized The legitimate unions 75 56 9639 71.89 The illegitimate unions 146 36 44.72 11.03 The fertility of the legitimate to the illegitimate was found to be as 100 to 33, judged by the flowers that produced capsules, and as 100 to 46, as judged by the average number of seeds. Darwin concluded that only the pollen from the longest stamens can jully jertilize the longest pistil; only pollen from the mid- length stamens can fully fertilize the mid-length pistil, and only the shortest stamens can fully fertilize the shortest pistils. The meaning of this difference is entirely obscure. It is of much interest to find a condition of this sort between individuals of the same species. It suggests a comparison with the infertihty that exists between different species; but in point of fact the results are just the reverse, for, in the present case, it is the same kind of flowers that imperfectly fertilize each other, while the flowers having a different form are more fertile. In the case of dimorphic or trimorphic plants Darwin makes a determined effort to show that selection of fluctuating variations has brought about the two kinds of flowers. This argument is so instructive that I shall give it in full. Since heterostyled plants occur in fourteen different famihes of plants, it is probable, Darwin thinks, that this condition has been acquired independently in each family and “that it can be acquired without any great difficulty.” The first step in the pro- cess he imagines to have been due to great variability in the length of the pistil and stamens. “As most plants are occasion- ally cross-fertilized by the aid of insects, we may assume that this was the case with our supposed varying plant but that it would have 'been beneficial to it to have been more regularly cross-fertilized.” “This would have been better accomplished if the stigma and the stamens stood at the same line; but as the 164 Experimental Zoology stamens and pistils are supposed to have varied much in length, and to be still varying, it might well happen that they could be reduced much more easily through natural selection into two sets of different lengths in different individuals than all to the same length and level in all individuals.” Darwin points out that the mutual sterility of these plants could not have resulted from natural selection, and although he thinks that the difference in the length of the stamens and pistils has resulted from a process of natural selection, yet he admits that one of the most striking facts in the case is that the individuals have in consequence become partly sterile to half the other indi- viduals in one case and to three fourths in the other. This con- clusion in itself shows, it seems to me, how futile it is to apply the theory of selection of fluctuating variations to the process of evolution of these forms. Bateson and Gregory have examined the inheritance of hetero- stylism in Primula and have found that the Mendehan rule is followed. In the case of P. sinensis, the short-styled is dominant over the long-styled form. When long-styled was crossed with long-styled — pure forms being used — all the offspring were long-styled. When these were inbred again, only long-styled forms {Fp were produced. The short-styled plants that were obtained for experimental work proved to be heterozygous (DR). When these short-styled were crossed with short-styled forms, there were produced 26 short- and 10 long-styled — the Mendelian expectation being 3:1. Of these 26 short-styled forms some were pure dominants (DD) and others dominant- recessives (DR). The latter (DR) inbred gave short (24) and long (4). It was found that other combinations also conformed to the Mendelian expectation.^ Bateson and Gregory also examined the inheritance of a peculiar form of Primula sinensis known as equal-styled. The anthers are at the same level as in long-styled flowers, but the style is short and does not reach above the level of the anthers. The corolla has a central yellow flush extending over half of each ‘ A few departures difficult to explain were also met with. Other Kinds of Hybridizing 165 petal. “The flush is transmitted independently of the length of the style or the size of the pollen grains, for it may be trans- ferred to the true short-styled or ‘thrum’ type. But when the flush is developed in plants which by gametic composition would be long-styled, the style does not pass through the anthers and the equal-styled condition is produced. Why the development of the yellow flush in these flowers should entail the reduction of the style, we cannot in any way suggest.” The discovery, that Primula follows the same law of inheri- tance as do other discontinuous variations, some of which are known to have appeared suddenly, furnishes an argument in favor of the view that the dimorphism in Primula owes its origin to discontinuous variation. Compared with the involved argu- ment by which Darwin attempts to show how natural selection has brought about the result, the mutation theory offers a much simpler and in my opinion a much more plausible interpretation. Reversal 0} Symmetry In some species of snails the spiral of the shell turns to the right, in other species to the left. Occasionally in a right- handed species an individual that is left-handed is found, and there can be little doubt that such a form may suddenly arise. It is a discontinuous variation, but whether it is a mutation is not so clear, since the result may be due to an accidental shift- ing of the blastomeres on each other at the time when the unsymmetrical mesoblast cell is laid down. Nevertheless, the fact that entire species are characterized by the right- or the left- handed condition indicates that the factor that produces the one or the other result may at times be impressed on or arise in the and be inherited. There are also species in which some individuals are right-handed, others left-handed. Here both possibilities seem to exist in the egg; but whether this can be referred to alternating dominance and recession, or to purely local conditions that arise during segmentation, is unknown. Somewhat similar conditions occur in the two kinds of chela; of crabs, prawns, and other decapods ; and perhaps the right- 1 66 Experimental Zoology handedness and left-handedness in man belongs to the same category. It is also known that in man all of the viscera may be in a reversed position, and the aortic arch also. Lang has bred together some left-handed snails of the right- handed species, Helix pomatia. They produced only right- handed individuals. These were also inbred and produced only right-handed young. It is evident that this new character is not inherited in this instance and also that it does not behave as a recessive to the right-handed condition, for, if it did so, it would have reappeared in the grandchildren. Physical difficulties seem to interfere with the union of the sinestral individuals of Helix with dextral forms, so that as a rule the sinestrals are precluded from breeding, unless by chance they meet one of their own kind ; but even if this occurs all the offspring will be dextral again, and no opportunity exists to per- petuate the new race. If, however, in other species the left- handed condition should be heritable, such individuals might, if once started, establish a variety alongside of the parent forms, although the two forms might be prevented by their difference in structure from pairing. That the left-handed condition must sometimes be inherited is shown by the existence of species of this sort. Conclusions regarding Mendelian Inheritance The examples of discontinuous inheritance given in the pre- ceding chapters make it clear that Mendel’s law accounts in many cases for the results, and is therefore an invaluable acqui- sition to our method of interpretation ; yet in some other cases it is evident that the inheritance is not strictly Mendelian.^ Used with discretion the law may still unlock many problems, but if attempts are made to force it to interpret cases that do not belong to its proper field of action, especially in regard to dissocia- tion in the germ-cells, harm rather than good may temporarily result. ‘ The most striking exceptions are those recorded by Standfuss and by de Vries (oenothera). Other cases involve blending, etc. Other Kinds of Hybridizing 167 The Mendelian inheritance of coat color in mice and in some other animals may give an exaggerated idea of the inheritance of color in general. That the inheritance is not alAvays of this sort seems to be shown by a number of cases, some of which have been given. Pearson’s examination of the inheritance of color in the coat of dogs, horses, and of the eye color in man, has led him to conclude that in these cases there is no evidence of Mendelian inheritance. He points out that when the whole range of the ancestry is examined we get more nearly an idea of what the color of the offspring will be. For example, in one case, where the color of the eye of the mother and of the father was blue, only two of ’the children had blue eyes ; but in another case, where the father and the mother and all four of the grandparents and five recorded grandparents had blue eyes, four children had blue eyes. Single cases of this kind in themselves do not show much, but these are only samples of what is generally found in many cases of the sort. Pearson thinks that the prediction of what the color of the eyes is likely to be will be closer when we use the ancestry and not the parents alone. If any of these eye colors follow the Mendelian rule of dominance and recession, some evi- dence of this would appear in the statistics, but nothing of the sort has been found. ^ It is evident, nevertheless, that these cases require careful reexamination, since there has probably been great intermingling of different colors in the past. De Vries and some other students of mutation have laid much stress upon the immutability of unit characters. De Vries as- sumes that transitions between unit characters exist as little as between the molecules of chemistry. It cannot be maintained, I think, from the evidence that we possess, that unit characters are immutable, for there are some cases in which it appears that the unit character may be halved by every crossing. It is true that some of these cases may be explained by antagonistic char- acters both developing and mutually influencing the result ; but if they do not subsequently separate, it is impossible to tell whether or not a new unit character has been formed by combination. ‘ “ The Law of Ancestral Heredity,” Biometrica, II, 1903. 1 68 Experimental Zoology Cases of blended inheritance especially seem to come under this heading. But so long as we do not know definitely what occurs in these cases, it seems to me arbitrary to speak of unit characters as immutable and quite unnecessary to make this idea a cardinal point of the mutation theory. The behavior of cer- tain characters in heredity shows that they do act as units, and it is a great convenience to deal with them as such, but unnecessary to push the matter so far as to hold that they are immutable. If unit characters can be halved, altered, added to, or changed in any way, their immutability does not seem to be an essential point of their characterization, and, as has been said, there is some evidence to indicate that such changes may take place. The idea of immutability is not likely to suggest itself when results of this kind are considered, but in cases where complete domi- nance of one form over the other occurs, the idea of unit char- acters is more likely to arise. If, however, as I think probable, we are dealing here with alternate or contrasted characters that cannot both develop .at the same time, the isolation of the char- acters in question is due to their mutual exclusion rather than to their immutability. Even if it be true that mutations take place by definite steps without the presence of intermediate forms, it does not necessarily follow that these steps may not subsequently become subdivided by each crossing with the parent form. The central idea of mutation remains, even if the unit- character that marks the steps is capable of being changed. The results of hybridizing of forms differing by a single unit- character seem to show that when in the first generation the hybrids {F^ are strictly like one of the two parents, the hybrids of the new generation {Fp also show complete development ^ of one of the two characters in the extracted dominants and of the other in the extracted recessives. This may be called the law of incompatibility. If, on the other hand, the dominance is incom- plete, i.e. if both dominate in the first generation, there is in- complete dominance in the second generation. This may be called the law of compatibility. If further facts establish these ‘ In the sense of contrast, not necessarily of actual separation. Other Kinds of Hybridizing 169 laws as general, the results suggest certain questions of great theoretical importance. For instance, the results seem to me to indicate that we are dealing not with a question of partial purity of the germ-cells, but with the question of the relation of dominance and recessive- ness of contrasted characters. In the first generation there can be no doubt that both characters are present. In the first class of cases the activity of one character completely suppresses the activity of the other — the characters are mutually exclusive, i.e. the development of one suppresses the development of the other. In the second class of cases both characters may become active either at the same time in the same cell, producing blend- ing ; or in different parts of the soma, — different cells or groups of cells, — producing a mosaic or a piebald condition. If, then, the same condition holds in the second generation, the most probable conclusion is, I think, that there has really been no separation of the contrasted characters in the germ-ceUs, but only a condition of relative dominance and latency estab- lished that is akin to but not identical with the dominance and recession in the first generation. Such a conclusion seems to me more in conformity with the results than that which tries to explain the facts of the second class as due to imperfect separation in the germ-cells of the two contrasted characters ; for on my view the results in the first and second generation are accounted for on the same assumption ; while in the current interpretation it is not apparent why imperfect separation in the germ- cell of should not occur as often in the first class of cases as in the second. In the preceding pages the heredity of a large number of char- acters of domesticated animals has been described. Relatively few facts regarding wild species have been given. The objec- tion has sometimes been raised that our domesticated animals are contaminated to such an extent by crossing that they offer questionable material for studies in heredity, and that the study of wild forms is more profitable. This objection is misleading, since it directs attention away from the point at issue and rests on several false or questionable assumptions. 170 Experimental Zoology In the first place, if we are to study those characters that be- have, as a rule, as units, we can find for this purpose no better material than some of the domesticated races that differ from each other by single unit-characters. We are dealing here with the simplest cases of discontinuous variation, and the simpler the problem the better the opportunity to study it. In so far as the contamination due to previous hybridization is concerned there is introduced, it is true, a complication (but one that can be dealt with, in most cases), for the unit-characters, as such, are not necessarily affected by the latent characters, but can be studied independently of them. Hence it is not a serious difficulty to find that previous contamination has occurred. To ignore this point shows a misconception of the problem of the heredity of discontinuous variation. It is important, in this connection, to bear in mind that the same rule of discontinuous heredity has been found to be true for wild forms also that differ from each other by a single character. Furthermore, how do we know that wild species are not also hybrids ? If evolution has taken place by mutation, then it is possible that many wild species are complex hybrids, even if all of them are not. There will also be recalled in this same connec- tion de Vries’s conclusion that even his elementary species of Oenotheraemust also have arisen by hybridization, since it is im- probable that a mutating germ-cell should meet another of its kind. Therefore, until it is more evident that wild species are not hybrids, this side of the argument carries little weight. It is sometimes said also that wild species that have bred true for hundreds (or thousands) of generations are much purer than our “pure” domesticated races that have been bred for a rela- tively small number of generations. This may be true, provided it can be shown that a hybrid is purer the longer it is inbred. There is little direct evidence that this is the case, but the state- ment is likely to pass unchallenged because it seems so plausible ! Finally, if evolution has taken place by single steps, our first problem is to study the heredity and results of crossing of these single steps. Most wild species must differ from each other by Other Kinds of Hybridizing 171 a large number of steps, and the results of hybridization of such forms may differ from the results of hybridization of single steps. If new elementary species arise by single steps and are subjected to crossing with the parent type, as would almost inevitably take place in^many cases, the problem of paramount importance, from the point of view of evolution, is to study the results of such hybridization. Crosses between already established species are infrequent in nature and seem only in rare cases to give rise to new species, hence their study is of secondary importance from the standpoint of evolution, provided always that evolution has taken place discontinuously. On the other hand, it is un- fortunate in many ways to attempt to estimate the value of the study of unit characters by the supposed bearing of the results on the theory of evolution ; for while those who believe that evolu- tion has taken place by simple discontinuous steps will ascribe the highest value to studies of this kind, those who heli&ve that evolution has taken place in some other way will be likely to underrate the importance of the results ^s a contribution to the theory of heredity. Already the bitter controversies that the pubheation of these results has aroused are less concerned with the results themselves, that are accepted by all, than with the imagined bearing of the: results on the theory of evolution. The intolerance that each side sees in the other is due not to the ac- ceptance or denial of the results of the experimental work, but to the arguments that pretend to show that evolution has or has not taken place by discontinuous variation. Meanwhile there is danger that we forget the importance of the experimental results as a contribution to the study of heredity, whatever their bearing on the theory of organic evolution may be. LITERATURE, CHAPTER X Ackermann. Tierbastarde. Kassel. 1898. Bateson, W. Materials for the Study of Variation. 1894. Bateson, W., and Gregory, R. P. On the Inheritance of Heterostylism in Primula. Proceed. Roy. Soc. of London, LXXVI. 1905. Conklin, E. G. The Cause of Inverse Symmetry. Anat. Anz. XXIII 1903. Experunental Zoology 1 72 Crampton, H. E. Reversal of Cleavage in a Sinistral Gasteropod. Ann. N. Y. Acad. Sci. 1894. Darwin, C. The Effects of Cross- and Self-Fertilization in the Vegetable Kingdom. The Variation of Animals and Plants under Domestication. 2d ed. 1890. Holmes, S. J. The Early Development of Planorbis. Jour. Morph. XVI. 1900. Rabl, C. Homologie und Eigenart. Verb. d. Deutsch. Pathol. Geseli. II. 1900. DE Vries, H. Species and Varieties. 1905. CHAPTER XI BEHAVIOR OF THE GERM-CELLS IN CROSS-FERTILIZATION In the preceding chapters the characters of the hybrids re- sulting from cross-fertilization have been considered. The present chapter will deal with the behavior of the germ-cells themselves when cross-fertilization is attempted. In certain respects this topic covers a wider field than the preceding, since there are many more species in which the eggs may be entered by the spermatozoa of other species (and the early development take place) than of those that produce adult hybrids. Experiments with Amphibia Different species of European frogs have been frequently utihzed in crossing experiments. The most important results are those of Rusconi (1840), Lataste (1878), Pfliiger (1882), Born (1883), and Heron-Royer (1883). It has been found in a number of different forms that the spermatozoa of one species will enter the eggs of other species and start the development. The egg may cleave, generally quite irregularly, but later stages than this may not develop. In some combinations the early, or even the later, gastrula stages may develop, but the embryos perish without going farther. Finally, in a few cases tadpoles, often having a weak constitution, may be formed. Thus the two closely similar species, Rana fusca {$) and Rana arvalis (9), cross readily, and tadpoles have been reared as far as the frog stage. The reverse cross gave no results. Bufo variabilis and Bufo cinereus also cross, and toads may be produced. The different races of Rana fusca intercross as readily with each other, as each race fertilizes its own eggs. 173 174 Experimental Zoology The other extreme is found where the eggs of Rana fusca are fertilized by the eggs of the salamander, Triton, and divide irregularly, but go no farther. Pfliiger concluded from his results that cross-fertilization de- pends less on the similarity of the adults than on the peculiarities of the spermatozoa. Thus the spermatozoa of Rana fusca and Rana arvalis are very different, and while cross-fertilization takes place in one direction it does not in the other. On the other hand the spermatozoa of Bufo cine reus and B. variabilis are much alike, and reciprocal cross-fertilization is successful. In support of his view Pfliiger points out that the spermatozoa are most suc- cessful in crossing that have the thinnest or most pointed heads. Furthermore those eggs are most easily crossed that belong to species whose spermatozoa have the largest heads, because, being as it were so constructed as to admit their own large-headed sper- matozoa, they do not exclude spermatozoa of smaller size. This view of the matter may explain the power of certain kinds of spermatozoa to enter the eggs of other species, but it does not ex- plain why, after entering, certain combinations develop normally and others scarcely at all. The conditions for normal development appear to be most readily fulfilled when the two species are like each other in struc- ture, which usually, though not invariably, means consanguinity. Pfliiger found that the eggs of the frog have the greatest power of being cross-fertilized at the height of the breeding season. Certain experiments that I have made on other forms indicate that this result may be due not so much to the eggs as to the greater mobility of the spermatozoa at this time. Hertwig has questioned Pfliiger’s conclusion, basing his objection on the evi- dence derived from some experiments that he carried out on the eggs of the sea urchin. He found that eggs could be more easily crossed when overripe or stale, as when they have stood for some hours in sea water, or after they have been injured by poisons. It has been shown more recently by Vernon that while in a few cases {i.e. in some species) more eggs can be cross-fertilized if they have stood twelve to twenty-four hours in sea water, yet in 175 Behavior of the Germ-cells most cases this does not hold. Eggs lose, as a rule, rather than gain in their responsiveness to foreign sperm if kept too long. Experiments with Echinoderms A number of investigators have made crosses between different species of sea urchins with varying success. In recent years Vernon, Boveri, Seeliger, Morgan, Driesch, Herbst, Loeb, and Godlewski have carried out experiments with these forms. Vernon found that out of sixty-four possible combinations, forty-nine gave the following results ; twenty-nine developed to the pluteus stage, nine to the segmentation, blastula, or gastrula stages, and in eleven fertilization did not take place. Vernon tried to show that the characters of the hybrid embryo are dependent upon the relative ripeness of the eggs and of the sperm in the two species that are crossed. He carried out his experiments with several species of sea urchins found in the Mediterranean. The breeding period of these animals extends over several months, or even in one species throughout most of the year. The height of the breeding season may be different for. different species. During the time preceding and following that of the full maturity of the eggs and sperm, the eggs may still be fertilized, although fewer of them develop normally. For example: — the eggs of Strongylocentrotus reach their optimum in December or January, their minimum in July or August. Sphaerechinus gives throughout the year mature eggs and sperm, although in summer the percentage of larvae that develop is smaller. If the eggs of Sphaerechinus are fertilized by sperm of Strongylocentrotus during the summer months. May, June, July, the hybrid larvae resemble the Sphaerechinus type (the mother), although some of them or less) show traces of the paternal (Strongylocentrotus) type of larvae. In November the hybrids approach more nearly the type of Strongylocen- trotus (the father), and in December are entirely of this pater- nal type. In other words, as the sperm of Strongylocentrotus becomes more and more mature, it transmits to the larvae its own characters. I y6 Experimental Zoology The reciprocal cross, Strongylocentrotus 9 and Sphserechi- nus $ gave less striking results, because of the greater difficulty in making the cross. In April, May, June, young stages were obtained that died. In July and August 29 per cent reached the pluteus stage. The hybrids showed no indication of their double origin, but were pure Strongylocentrotus (maternal). In November and December no eggs cross-fertilized. Doncaster has carried out experiments in hybridizing sea urchins that lead him to conclude that the different hybrids ob- tained by Vernon at different times of year owe their peculiarities to the temperature of the water in which they develop. Herbst has recently carried out a more elaborate series of experiments that lead him to a similar conclusion, although he thinks that some other condition than temperature is also operative. What the other condition is he did not determine ; but he does not think that it can be due to the relative condition of ripeness of the male sex cells. In fact, his analysis of Vernon’s results, in the light of his own observations, seems to show, for the sperm at least, that Vernon’s evidence is most unsatisfactory. The preceding results apply more especially to the later larvag or pluteus stages. In some respects the results seem to be in- consistent with results that other observers have obtained with the younger stages of these hybrids. Driesch has found, for instance, that the method of cleavage, its tempo, the character of the mesenchyme formation, and of gastrulation are char- acteristic of the egg irrespective of the kind of sperm that is used. In later stages, when the skeleton develops and the pigment ap- pears, the larvae first begin to show their hybrid origin. On the other hand, Boveri thinks that the hybrid characters appear very early, but the principal difference between his view and that of Driesch lies in the age at which each supposes the differences to become apparent. There can be little doubt, however, that, as a rule, in the early stages little or no trace of the paternal ele- ments appear, and only later do they influence the characters of the hybrid. This difference may be interpreted to mean that, at first, the elements introduced by the sperm — the nucleus or 177 Behavior of the Germ-cells the cytoplasm — have not had time to act or to increase suffi- ciently in amount to affect the development. Most embryolo- gists seem inclined to ascribe the effects entirely to the nucleus, which they believe dominates all the changes in the protoplasm. On the contrary, I am inclined to think that it has not been sho^^^l conclusively that this influence is nuclear in origin, but may possibly be due to the protoplasm introduced with the sperm. The slow increase in amount of the introduced proto- plasm might account for the insignificant part it plays during the early development, when it is very small in amount com- pared to that of the egg-protoplasm. If, as others suppose, the chromatin of the nucleus is the all-controlling influence, it is difficult to see why this is not apparent at once, since the nucleus of the hybrid has equal amounts of paternal and ma- ternal material. It may be fairly claimed, however, that the introduced sperm-nucleus requires time to change the protoplasm into its own sort of material. The most striking case of the lack of influence of the sperm nucleus on the egg is that recently given by Godlewski. By following Loeb’s method and making the sea-water alkaline, he has succeeded in fertilizing the egg of the sea urchin (Echinus and others) with the sperm of the crinoid (Antedon). The hy- brids were of the sea-urchin type in all respects observed, includ- ing the pluteus stage. Boveri carried out the ingenious experiment of fertilizing a non-nucleated piece of the egg of one species of sea urchin with a spermatozoon of another species.^ The pluteus obtained was purely paternal. He concluded that the result was due to the introduced nucleus. Both Seeliger and I have taken exception to Boveri’s evidence on the ground that the hybrid pluteus that can be obtained from nucleated pieces or from entire eggs is too variable in its characters to give support to Boveri’s conclusion. In fact, some of these hybrids are so similar to the paternal type that they cannot be distinguished from it. Godlewski’s ‘ O. and R. Hertwig had previously shown that pieces of the sea urchin’s egg without a nucleus may be entered by spermatozoa of the same species. N 1 78 Experimental Zoology experiment, described above, in which he used a non-nucleated piece of the sea urchin’s egg and the spermatozoa of the crinoid produced larvae entirely of the maternal type, which is the con- verse of Boveri’s result. Factors involved in the Entrance of the Spermatozoon The entrance of a foreign spermatozoon into an egg is closely connected with the question of normal fertilization. What brings the sperm and egg together? How does the sperm enter the normal egg, and what delays or prevents its entrance into eggs of another species ? The immense collections of sperm around the egg in normal fertilization has led to the idea that the egg at- tracts the sperm. Certain experiments seemed to support this view. Pfeifer’s experiments with the antherozoids of ferns have often been cited as an instance of such an attraction. He found that when a dilute solution of malic acid was inclosed in open capillary tubes, and these tubes were immersed in a drop of water containing antherozoids, the latter collected around the open ends of the tubes, as though attracted by the malic acid. The evidence in favor of this interpretation has recently been considerably weakened by Jenning’s study of the behavior of protozoa. These also will collect in a drop of acid, not, how- ever, because they are attracted to the drop, but because no re- action takes place when they pass from water into a more acid solution. A reaction does occur, however, in passing from an acid region into water. The reaction involves a backing of the individual into the drop followed by a movement forward again in a new direction. On coming a second time to the edge of the acid area the reaction is again repeated. All individuals that pass by chance into the acid remain there — caught like rats in a trap — so that in time an accumulation occurs that might read- ily suggest that the animals had been attracted to this region. Strasburger claimed that the eggs of the seaweed, Fucus, excrete a substance that attracts the spermatozoon from a dis- tance of two diameters of the egg, but Bordet and Buller have failed to corroborate this observation. 179 Behavior' of the Germ-cells Buller’s experiments with the eggs of sea urchins and starfish have given important results. He pointed out that spermatozoa accumulate around immature eggs, and also around mature eggs that had been killed in osmic acid and then thoroughly washed in sea water. In this case it is highly improbable that any attrac - tion could exist. The results are due to those spermatozoa that accidentally run into the membrane of the egg, remaining stick- ing there as a result of some physical property of the jelly or of some reaction on the part of the spermatozoon. Buller showed by means of the following experiment that the eggs do not se- crete an attracting substance. Eggs were allowed to stand in a little water from two to twelve hours. Capillary tubes were then filled with this water and placed in a drop of sea water containing the sperm. No collecting of spermatozoa around the ends of the tube was observed. He alse tried other substances in the tubes, viz. salts, sugars, ferments, and alcohol, etc., but no evidence of their action in attracting the spermatozoa was observed. Buller thinks that the spermatozoa are sensitive to contact, hence on coming into contact with the membrane bore into it. The spermatozoa swim in spirals in the water, but on entering the jelly they take a straight course, which in most cases will bring them into contact with the egg, although, if they should enter quite obliquely they may bore through the periphery of the membrane and pass out again on the other side. Thus even after entering the membrane, there is no evidence that the egg attracts the spermatozoa toward itself. On the other hand it has been shown by von Dungern that the egg and its membranes may contain and even give off substances that act in some cases injuriously on the spermatozoa of other species. The egg of the starfish, Asterias glacialis, contains a substance that acts as a poison on the spermatozoa of the sea urchins — Echinus or Sphasrechinus. The minimal lethal dose for sperm mixed with two cubic centimeters of sea water varies with different individuals between the limits of to parts for half an hour. The same poison is found also in the skin of the starfish. Cross-fertilization of the eggs of this starfish by i8o Experimental Zoology sea urchin’s sperm is thus prevented by the action of the poison, but it is going beyond the evidence to extend this conclusion, as has been done, to other starfish and other sea urchins without further examination. In the sea urchin, Sphaerechinus, there is a poison in the pedicel- larias that is injurious to the spermatozoa of the starfish, but if this poison exists also in the egg it is not strong enough to prevent the spermatozoon of the starfish from entering, at least in those species of sea urchins in which such a combination has been artificially brought about. The spermatozoa of Sphaerechinus itself are killed by this same poison from its own pedicellariae, but a much stronger dose is required. It has been also found by von Dungern that extracts of the eggs of Echinus, Sphaerechinus, Strongylocentrotus, or Arbacia do not kill the spermatozoa of the starfish even in the strongest solutions. Why, then, do not the spermatozoa of the starfish readily enter the eggs of these sea urchins? Their inability to enter appears to be due to another factor. The egg membrane of these sea urchins agglutinizes the spermatozoa of the starfish, i.e. they stick to it, and this interferes with their penetration. In Sphaerechinus, however, agglutinization does not take place. Why, then, does not cross-fertilization occur here? Von Dun- gern claims to have found still another substance in this sea urchin’s egg that excites even immature sperm to greater activity.^ He believes that these exciting substances may in some cases prevent cross-fertilization because they change the kind of ac- tivity shown by the spermatozoa. He observed that spermato- zoa that do not show rotational movement on meeting a solid, or semi-solid, body may do so when excited in this way and fail in consequence to penetrate. Von Dungern also tries to show that eggs or egg membranes may contain substances that favor fertilization by its own sperm. Eggs of Echinus were rubbed up and mixed with pieces of jelly that had first been carefully washed. Spermatozoa were then ' Immature sperm excited in this way may then even fertilize their own kind of egg, i.e. starfish eggs. BeJidvior of the Germ-cells i8i added and observed to stand vertically to the surface of the jelly as they penetrated. If the piece of jelly was not first mixed with the extract, the spermatozoa simply rotated on its surface. On the other hand, starfish spermatozoa on coming into contact with Arbacia jelly behave as with the simple jelly alone, i.e. they do not stand vertically to it. The vertical position may be due, von Dungern thinks, to the presence of some substance in the extract that lowers the excitability of the spermatozoa to con- tact. Much still remains obscure, but these results show clearly some of the factors involved in fertilization and cross-fertihza- tion. Artificial Helps to Fertilization and to Cross-fertilization A number of embryologists have found that normal fertihza- tion may be assisted by adding certain substances to the water in which the sperm are placed. These substances excite the spermatozoa to greater activity, and in this or in other ways promote fertihzation. It has also been known for a long time that the glands connected with the ducts of the male may secrete substances that make the spermatozoa active. For instance, Kolliker discovered that the secretion of the prostate glands of the male greatly excites the spermatozoa. Ordinarily the sper- matozoa are quiescent as long as they remain in the testes or even in the ducts leading from them, but become active when the secretions are added or when set free in water. Roux states that the fertilization of the frog’s egg is helped by the addition of one fourth per cent sodium chloride to the water, and Wilson has found that the spermatozoa of Patella can fertilize a much larger proportion of the eggs if a little potassium hydroxide is added to the sea water. Torelle and I have found that somewhat immature sperm of the starfish can be made active by ether, am- monia, salt solutions, etc., and will then fertilize the eggs. Von Dungem had previously observed that extracts of the eggs of the sea urchin excite immature spermatozoa to activity. The most striking case of this sort is that of the fertilization of the sea-urchin egg by sperm of the starfish, recently described 1 82 Experimental Zoology by Loeb. If to loo parts of an artificial sea water containing 0.3 cubic centimeter sodium hydroxide the eggs of Strongylo- centrotus and the sperm of the starfish be added, as many as 50 to 80 per cent of the eggs will be fertilized. The eggs segment and it appears that even the early stage of development — more or less abnormal — may be reached. Polyspermy In the great majority of known cases only a single spermatozoon normally enters the egg to fertilize it. The moment one sperma- tozoon penetrates, some reaction takes place in the egg that pre- vents other spermatozoa from entering. The reaction involves, in some cases, contraction of the egg from the surrounding mem- brane ; in other cases the formation of a membrane ; and we may infer that other kinds of changes in the protoplasm may in other cases interfere or prevent the entrance of more than one sperma- tozoon. If the egg is injured or narcotized or is stale, more than one spermatozoon may enter. The result is generally injurious, for, while a regular or more often an irregular cleavage may fol- low, the embryo fails to develop properly. This failure seems to be due, in the main, to the unequal distribution of the chroma- tin or other materials to the different parts of the egg. It has been found in hybridizing that in certain cases more than a single sperm enters the egg, and the irregularities in cleavage that follow seem to be due in part to this condition, but in other cases where one spermatozoon enters, the failure to develop must be ascribed to different causes. In these cases the results may be due sometimes to irregularities in the mechan- ism of division, but at other times to some incompatibility or failure of the two uniting elements to work together to a com- mon end. According to the degree of perfection to which the distribution of the materials of the egg is carried out, the subsequent develop- ment is more or less perfect. Even in cases where the hybrid reaches the adult condition, it has been observed that it may have a weak constitution, and even when it is strong the hybrid is Behavior of the Germ-cells 183 sometimes sterile. This sterility is due apparently, in some cases at least, to irregularities in the division of the germ-cells. When we recall that at one stage in the development of the germ-cells there may be a pairing and subsequent fusion of the ma- ternal and paternal chromosomes, we can readily imagine that any differences in their behavior at this time might lead to disastrous results. LITERATURE, CHAPTER XI Bordet. Contribution a I’Etude de Tlrritabilite des Sperm atozoides ches les Fuccacees. Bull, de I’Acad. Belgique. 3"’“ ser. XXVII. 1894. Boveri, Th. Fin Geschlechtlich erzeugter Organismus ohne miitterliche Figenschaften. Sitz.-ber. d. Ges. f. Morph, u. Phys. Munchen, V. 1889. Ueber die Befruchtungs- und Fntwicklungsfahigkeit kemloser Seeigel- eier. Arch. f. Fntw.-Mech. II. 1895. Ueber den Finfluss der Samenzelle auf die Larvencharaktere der Fchiniden. Arch. f. Fntw.-Mech. XVI. 1903. Noch ein Wort ueber Seeigelbastarde. Arch. f. Fntw.- Mech. XVII. 1903. Bctller, a. H. The Fertilization Process in Fchinoidea. Report 70, Meet. Brit. Assoc. 1901. Is Chemotaxis a Factor in the Fertilization of the Fggs of Animals? Quart. Jour. Micro. Sc. XL VI. 1903. Contributions to our Knowledge of the Physiology of the Spermatozoa of Ferns. Ann. of Botany, XIV. 1900. Calkins, G. Studies on the Life-History of Protozoa, I. The Life-Cycle of Paramoecium Caudatum. Arch. f. Fntw.-Mech. d. Organismen. XV. 1902. Studies on the Life-History of Protozoa, III. The six hundred and twentieth generation of Paramoecium Caudatum. Biol. Bull. III. 1902. Delage, Y. Embryons sans noyau maternel. C. R. Acad, des Sc. V.. , 1898. Etudes sur la Merogonie. Arch, de zool. experim. et g^n. V. 1899.. Sur I’interpretation de la fecundation merogonique et sur une thdorie- de la fecundation normale. Arch, de zool. expdrim. et gen. VII.. , 1899. Etudes experimentales sur la maturation cytoplasmique et sur la. parthenogen&se experimentale. Arch, de zool. experim. et. g^n. IX. 1901. Dewitz, J. Ueber Gesetzmassigkeit in der Ortsveranderung der Sperma- tozoen und in der Vereinigung derselben mit dem Ei. Arch. f. die gesammte Physiol. XXXVIII. 1886. Driesch, H. Ueber rein miitterliche Charaktere an Bastardlarven der Fchiniden. Arch. f. Fntw.-Mech. VII. 1898. Ueber Seeigelbastarde. Arch. f. Fntw.-Mech. XVI. 1903. 184 Experimental Zoology V. Dungern, E. Die Ursachen der specifitat bei der BefruchtunK. Centrbl f. Phys. XIII. 1901. Neue Versuche zur Physiologie der Befruchtung. Zeitschr. f. allgem. Phys. I. 1902. Giard, a. _ Developpement des oeufs d’Echinodermes sous I’influence d’actions kinetiques anormales (solution salines et hybridization). Compt. Rend, des Sc. et Mem. de la Soc. de Biologie. Paris. 1900. Godlewsky, E., Jun. Die Hybridisation der Echinideen- und Crinoideen- familie. Bull. de. I’Acad. des Sc. de Cracovie. 1905. Untersuchungen ueber die Bastardierung der Echiniden- und Crinoi- den familie. Arch. f. Entw.-Mech. XX. 1906. Gurwitsch, a. Zerstorbarkeit und Restitutionsfahigkeit des Protoplas- mas des Amphibieneies. Verb. Anat. Ges. 18 Vers. Jena. 1904. Guyer, M. F. Spermatogenesis of Normal and of Hybrid Pigeons. Univ. of Cincinnati. Bull. LXH. Ser. II. Vol. II. 1900. Hacker, V. Ueber das Schicksal der Elterlichen und Grosselterlichen Kernanteile. Jen. Zeitschr. f. Naturw. XXVII. 1902. Bastardirung und Geschlechtszellenbildung. Zool. Jahrb. Suppl. VII. 1904. Kolliker, a. Physiologische Studien ueber die Samenflussigkeit. Zeitschr. f. wiss. Zool. VII. 1856. Lidforss. Ueber den Chemotropismus der Pollenschlauche. Ber. d. Deutsch. Bot. Gesell. XVII. 1895. Loeb, J. On the Artificial Production of Normal Larvae from the unfer- tilized Eggs of the Sea Urchin. Amer. Jour, of Phys. HI. 1900. Further Experiments of Artificial Parthenogenesis. Amer. Jour, of Phys. V. 1900. On a Method by which the Eggs of a Sea Urchin (Strongylocentrotus purpuratus) can be fertilized with the Sperm of a Starfish (Asterias ochracea). Univ. of California Pub. Phys. I. 1903. The Fertilization of the Egg of Sea Urchin by the Sperm of Starfish Univ. of California Pub. Phys. I. 1983. Ueber die Befruchtung von Seeigeleiern durch Seestemsamen. Pfliigers Arch. CXC. 1903. Further Experiments on the Fertilization of the Egg of the Sea Urchin with Sperm of Various Species of Starfish and a Holothurian. Univ. of California Pub. Phys. I. 1904. Further Experiments on Heterogeneous Hybridization in Echinoderms. Univ. of California Pub. Phys. II. 1904. Weitere Versuche ueber die heterogene Hybridization bei Echinoder- men. Pfliigers Arch. CIV. 1904. On an Improved Method of artificial Parthenogenesis (I, II, HI, Com- munications). Univ. of California Pub. Physiol. II. 1905. Loew, O. Die Chemotaxis der Spermatozoen in weiblichen Genitaltract. Sitzungsber. d. Wienner Akad. ; Math. -naturw. Cl. CXI. 1903. Massart, J. Sur I’Irritabilite des Spermatozoides de la Grenouille. Bull, d. I’Acad. Roy. de Belgique. 3"'“ s^r. XV. 1888. Sur la Penetration des Spermatozoides dans PHluf de la Grenouille. Bull, de I’Acad. Roy. de Belgique. 2“"-’ s