Читать книгу Darwin's On the Origin of Species - Daniel Duzdevich - Страница 11

IN CONSIDERING THE INDIVIDUALS OF A DOMESTICATED plant or animal variety, it is striking that they are generally more diverse than those belonging to varieties or species in the wild. The vast diversity of domesticated organisms, which have varied under many different climates and treatments, suggests that greater variability results from the conditions under which domestication occurs – conditions unlike those encountered by the parent species in the wild. This variability may partly be connected with excess food, as proposed by Andrew Knight. It seems clear that organisms must be exposed to a new environment over several generations for it to cause appreciable variation, and once organization begins to vary, it usually continues to do so for many generations. There is no case of a variable organism ceasing to be variable under domestication. Established domesticated plants such as wheat still often yield new varieties, and animals domesticated long ago are still capable of rapid improvement or modification.

It is disputed whether the causes of variation – whatever they may be – act during the early or late stage of embryonic development or at the instant of conception. Isidore Geoffroy St. Hilaire’s experiments show that unnatural treatment of the embryo causes monstrosities, which cannot be clearly differentiated from mere variations. I strongly suspect that variability is most frequently caused by effects on the egg or sperm before conception, mainly because of the remarkable influence of cultivation or confinement on the functions of the reproductive system, which appear far more susceptible to environmental changes than any other component of organization. Nothing is easier than taming an animal and nothing more difficult than getting it to reproduce in confinement, even when the male and female mate. This is generally attributed to impaired instincts, but many cultivated plants are vigorous yet do not seed. In some cases, minor changes, like a little more or less water at a particular period of growth, determine whether or not a plant will produce seeds. I will not go into the copious details I have collected on this curious subject, but to illustrate the strangeness of the rules that govern the reproduction of captive animals, consider that, with the exception of bears, carnivorous mammals, even from the tropics, breed freely in Britain under confinement, whereas carnivorous birds rarely lay fertile eggs. Many exotic plants have pollen as useless as that of the most sterile hybrids. Some domesticated plants and animals that are otherwise weak and sickly breed freely under confinement; but tame, long-lived, and healthy individuals taken young from the wild may have reproductive systems so seriously affected by unknown causes that they are nonfunctional. Unsurprisingly, then, when the reproductive system actually works under confinement, it does so irregularly, producing offspring that are different from the parents. Finally, some organisms breed under very unnatural conditions – like rabbits and ferrets kept in hutches – demonstrating that their reproductive systems have not been affected. So some organisms withstand domestication and vary only slightly, perhaps hardly more than in the wild.

Sterility is a horticultural nuisance, but variability, the source of all the choicest productions of the garden, shares a cause with sterility. There are many plants (called “sporting” plants by gardeners) that produce single buds or offshoots with novel characteristics, sometimes very different from the rest of the plant. Such buds can be propagated by grafting or other techniques, and sometimes by seed. These “sports” are rare in the wild but common under cultivation. In this case, manipulation of the parent affects a bud or offshoot but not the ovules or pollen. According to most physiologists, however, there is no essential difference between a bud and an ovule in the earliest stages of formation. Therefore, sports show that variability may be largely attributed to the effect on the ovules, pollen, or both by treatment of the parent prior to conception. In any case, these examples demonstrate that variation is not necessarily connected with the act of generation, as some authors have suggested.

Seedlings from the same fruit and young from the same litter sometimes differ considerably from each other even though both parent and offspring have apparently been exposed to the same conditions, as Müller has remarked. This shows how unimportant direct environmental effects are in comparison to the laws governing reproduction, growth, and inheritance. If the influence of environment were direct, then variation would be the same among offspring. Judging the extent to which heat, moisture, light, food, and other factors have an impact on variation is difficult. My impression is that such agents produce very little direct effect on animals, but apparently more on plants. (Mr. Buckman’s recent experiments on plants are valuable here.) When all or nearly all individuals exposed to certain conditions are identically affected, the resultant changes appear to flow directly from the conditions. But in some cases opposite conditions generate similar structural changes. Nevertheless, some slight amount of change may be attributed to direct environmental action, as in certain cases of increased size from greater food intake, altered coloration from particular kinds of food and light, and perhaps the thickness of fur from climate.

Habit also has a deciding influence, as with the flowering period of plants transported from one climate to another. The effect is greater in animals. For example, I find that the wing bones of a domestic duck weigh less and the leg bones weigh more in proportion to the whole skeleton than do those of a wild duck. I presume this results from the domestic duck flying much less and walking more than its wild parent. The large inherited udders of cows and goats in countries where they are habitually milked is another example of the effect of use. There is no domestic animal that in some region does not have drooping ears. As suggested by some authors, this probably results from the disuse of ear muscles, the animals being rarely alarmed.

Many rules regulate variation; some of them can be dimly seen and will be briefly mentioned. Here I will only allude to “correlated growth.” For example, a change in the embryo or larva often entails changes in the mature animal. With monstrosities, correlations between distinct parts are very curious.¹ Breeders maintain that long limbs are often accompanied by an elongated head. Some correlations are whimsical; for example, cats with blue eyes are invariably deaf. There are many remarkable cases among plants and animals of coloration and constitutional peculiarities going together. Observations collected by Heusinger suggest that white sheep and pigs are affected differently by poisonous vegetables than individuals with coloration. Hairless dogs have imperfect teeth; long-haired and coarse-haired animals tend to have long or many horns; pigeons with feathered feet have skin between their outer toes; pigeons with short beaks have small feet, and those with long beaks have large feet. If humans select and augment a peculiarity, they will probably unintentionally modify other parts as a consequence of the mysterious rules of correlated growth.

The dimly seen or unknown rules of variation yield infinitely complex and diverse results. Treatises on some established domesticated plants, such as the hyacinth, potato, and even the dahlia, reveal a surprising number of slight structural and constitutional differences between varieties and subvarieties. The whole organization seems to have become plastic, with a tendency to depart somewhat from the parental type.

Variations that cannot be inherited are unimportant to this argument. What nevertheless remains is an endless number of diverse heritable structural deviations of both slight and considerable physiological importance. (Dr. Prosper Lucas’s two-volume work is the best treatment of this subject.) The strong propensity for inheritance is known by breeders, whose fundamental belief is that “like produces like.” (Only theoretical writers have thrown doubt on this principle.) When a commonly occurring deviation is observed in both parent and offspring, it may result from the same cause acting on both. But when a very rare deviation due to some extraordinary combination of circumstances appears, say, once in several million individuals all apparently exposed to the same conditions, and it reappears in an offspring, the mere doctrine of chance almost compels us to attribute its reappearance to inheritance. Everyone has heard of albinism, prickly skin, hairy bodies, and other such peculiar characteristics reappearing in several members of the same family. If strange and rare deviations really are heritable, then surely commonplace deviations are also heritable. Perhaps the correct view is to take inheritance of every characteristic as the rule and non-inheritance as the anomaly.

The laws of inheritance are unknown. It is unknown why some given peculiarity of individuals within the same species, or of individuals among different species, is sometimes inherited and sometimes not, why a child reverts to characteristics found in a grandparent or more remote ancestor, or why some peculiarities are inherited in a gender-dependent manner.² Peculiarities appearing in the males of domestic breeds are often transmitted exclusively, or more strongly, to male progeny. A more important rule is that the age at which a peculiarity first appears tends to be the same in the parent and in its offspring (although sometimes earlier in the offspring). In many cases this cannot be otherwise. For example, the inherited peculiarities of cattle horns can appear only as the offspring mature, and peculiarities in the silkworm are known to appear at the corresponding caterpillar or cocoon stage. Hereditary diseases and some other examples suggest that this rule is generally applicable: even when there is no apparent reason for a peculiarity to appear at a particular stage, it tends to appear in the offspring at the same period of development as in the parent. This is very important to illuminating the rules of embryology. These remarks are, of course, confined to the first appearance of a peculiarity and not to its primary cause, which may have acted on the egg or sperm. If the offspring of a short-horned cow and a long-horned bull develops long horns, then it’s clearly due to the sperm.

Naturalists often argue that when domestic varieties run wild, their characteristics gradually but surely revert to those found in the original stocks, and that, consequently, deductions drawn from domestic varieties cannot be applied to species in nature. I have tried without success to find the decisive facts on which this statement is so often and so boldly made; it would be very difficult to prove, because many established domestic varieties could not possibly survive in the wild. In many cases we do not know what the original stock was and could not tell whether or not reversion had ensued. It would also be necessary to turn loose only one variety to avoid the effects of intercrossing. Nevertheless, varieties sometimes do partially revert to the parental form. For example, if various strains of cabbage were cultivated in very poor soil for many generations, they would probably revert wholly or largely to the wild stock. (However, some effect would have to be attributed to the direct action of the poor soil.) Whether or not the experiment would succeed is not particularly important to the argument, because the experiment necessarily alters the environment. If a strong tendency for reversion – that is, a loss of acquired characteristics under constant conditions in a large population so that free crossing, by blending, checks slight deviations of structure – were demonstrated in domesticated varieties, I would grant that nothing deduced from domestic varieties would apply to species. But there is not a shadow of evidence in favor of this view. To assert that we could not breed cart and racehorses, long- and short-haired cattle, and poultry of various breeds, and cultivate edible vegetables for an almost infinite number of generations is contrary to all experience. When the environment changes in nature, variations and reversions probably do occur, but natural selection, as will be explained, determines how far such new characteristics are preserved.

As already mentioned, there is less uniformity of character among individuals of a domestic variety than among individuals of a true species. Also, domestic varieties of the same species often have a monstrous character, by which I mean that although they differ in some minor respects from one another and members of the same genus, they often differ extremely in some one part. With these exceptions and that of the perfect fertility of crossed varieties (discussed later), domestic varieties of the same species differ from one another in a manner similar to the way closely related species of the same genus differ in the wild. There are very few domestic varieties of plant or animal that have not been classified by some competent judges as just varieties and by others as descendants of distinct parent species; if there were any significant distinction between domestic varieties and species, this source of doubt would be less common. Contrary to frequently made assertions, I think domestic varieties differ from one another in generic characteristics,³ which naturalists disagree in defining because all such valuations are currently empirical. Given the following examination of the origin of genera, there is no reason to often expect generic differences in domesticated organisms.

Attempts to estimate the amount of structural difference between domestic varieties of the same species are hampered by our ignorance of whether they have descended from one parent species or several; it would be interesting to clear up this problem. For example, if it were shown that the greyhound, bloodhound, terrier, spaniel, and bulldog, which propagate their kind truly, are derived from a single species, the supposed immutability of the many closely related natural species (such as the foxes) would be brought under considerable doubt. I do not believe that all dog breeds have descended from one wild species (see below),⁴ but there is tentative or even strong evidence that some other domestic varieties have.

Humans are often assumed to have chosen for domestication those plants and animals that possess an extraordinary inherent tendency to vary and to withstand diverse climates. Although such capacities have added significantly to the value of many domesticated productions, how could primitive humans have possibly known when first taming an animal that it would vary in succeeding generations and endure other climates? The limited variability of the ass and the guinea fowl, and the low tolerance for warmth by the reindeer and for cold by the common camel did not prevent their domestication. If plants and animals equal in number and belonging to equally diverse classes and regions to existing domesticated organisms were taken from the wild and bred for an equal number of generations under domestication, they would vary on average as much as the parent species of already domesticated organisms have varied.

I think it is impossible to ascertain with complete certainty whether established domesticated plants and animals have descended from one or multiple species. Those who believe in the multiple origin of domestic animals argue mainly that ancient records, especially on the monuments of Egypt, reveal a great diversity of breeds, some of which resemble or are identical to existing ones. Even if this were found to be more strictly and generally true than I believe is the case, it suggests only that some of our breeds originated there four or five thousand years ago. Based on Mr. Horner’s research, civilization advanced enough to manufacture pottery probably existed in the Nile valley thirteen or fourteen thousand years ago; it is not known how long before these ancient periods peoples like those of Tierra del Fuego or Australia, who possess a semidomesticated dog, may have existed in Egypt.

I think the whole subject must remain vague. Nevertheless, without going into details – but based on geographic and other considerations – I think it is likely that domestic dogs have descended from several wild species. I cannot form an opinion with respect to goats and sheep. Information about the habits, voice, constitution, and other features of humped Indian cattle, communicated to me by Mr. Blyth, indicate that it descended from a different stock than European cattle, which, in turn, have more than one parent, according to several judges. And for reasons I cannot cover here, I am doubtfully inclined to believe, in opposition to several authors, that all the varieties of horse have descended from one wild stock. Mr. Blyth – whose opinion I value highly, drawn as it is from his large and varied stores of knowledge – thinks that all poultry breeds have proceeded from the common wild Indian fowl. Duck and rabbit breeds, which differ considerably from one another in structure, have all descended from the common wild duck and rabbit.

Some authors carry the doctrine of plural descent to an absurd extreme, believing that every variety that breeds true, even those possessing only very slight distinctive characteristics, has a distinct wild prototype. At this rate there must have existed twenty species of wild cattle and sheep – and several goats – in Europe alone, and even several within Great Britain. One author believes that there once existed eleven unique wild sheep species in Great Britain! Britain barely has any unique mammals, France only a few distinct from Germany, and vice versa, and the same is true of Hungary, Spain, and other localities. Yet each of these kingdoms possesses several peculiar breeds of cattle, sheep, and other animals, meaning that many domestic breeds originated in Europe because it lacks sufficient unique species as parent stocks. This is also true for India. Even in the case of the domestic dogs of the whole world, which I admit probably descended from multiple wild species, there has been immense heritable variation. Who can believe that animals closely resembling the Italian greyhound, bloodhound, bulldog, or Blenheim spaniel – so unlike all wild dogs – ever existed in the wild? Crossing a few original dog species would result in only intermediate forms, and accounting for domestic dogs by this process requires the previous existence of extreme forms, similar to the breeds mentioned, in a wild state. In addition, the possibility of generating distinct varieties by crossing is exaggerated. Obviously a variety can be modified by occasional crossing if mongrels with desired characteristics are carefully selected, but a variety intermediate between two extremely different varieties or species could not be obtained. (Sir J. Sebright experimentally attempted precisely this and failed.) The offspring from the first cross between two pure breeds is fairly uniform – and as I have found with pigeons, sometimes very uniform – and everything seems simple enough. But when these mongrels are crossed with one another for several generations, few will be alike and the utter hopelessness of the task becomes apparent. An intermediate between two very distinct breeds could be obtained only with extreme care and continuous long-term selection, but I cannot find a single recorded case of a variety formed this way.

Believing that it is best to study one particular group, after deliberation I took up domestic pigeons. Pigeons have been watched, tended with the utmost care, and loved by many people. They have been domesticated for thousands of years in several parts of the world. Professor Lepsius has pointed out to me that the earliest known record of pigeons is in the fifth Egyptian dynasty (ca. 3000 BC), but Mr. Birch informs me that pigeons are given a bill of fare in the previous dynasty. In Roman times pigeons fetched a high price. Pliny writes, “Nay, they are come to this pass, that they can reckon up their pedigree and race.” The court of India’s Akber Khan (ca. 1600) always traveled with at least twenty thousand pigeons. “The monarchs of Iran and Turan sent him some very rare birds,” and, continues the court historian, “His Majesty, by crossing the breeds, which method was never practiced before, has improved them astonishingly.” At about this same period, the Dutch were as enthusiastic about pigeons as the Romans had been.

I have kept every pigeon breed that I could purchase or obtain and have been kindly favored with skins from several parts of the world, especially by the Hon. W. Elliot from India and the Hon. C. Murray from Persia. Many treatises in several languages have been written about pigeons, some of them very important because of their age. I have consulted several eminent breeders and joined two of the London Pigeon Clubs. The diversity of breeds is astonishing. Compare the English carrier to the short-faced tumbler and see the wonderful differences in their beaks, with corresponding differences in their skulls. The carrier, especially the male, is remarkable for the carunculated skin about its head and the accompanying greatly elongated eyelids, large nostrils, and widely gaping mouth. The outline of the short-faced tumbler’s beak is like that of the finch’s beak, while the common tumbler has the strictly inherited singular habit of flying at a great height in a compact flock and tumbling in the air head-over-heels. The runt is large with a long, massive beak and big feet. Some runt sub-breeds have long necks; others, long wings and tails; still others, short tails. The barb is related to the carrier but has a short and broad beak instead of a long one. The pouter’s body, wings, and legs are elongated; it glories in inflating its enormously developed crop, which may elicit astonishment and even laughter. The turbit has a short and conical beak, with a line of reversed feathers down its breast and a habit of slightly expanding the upper part of its esophagus. The Jacobin has feathers along the back of the neck that are so reversed that they form a hood; for its size, it has relatively elongated wing and tail feathers. The trumpeter and laugher, as their names suggest, coo very differently from the other breeds. The fantail has thirty or even forty tail feathers – even though the normal number in all members of the pigeon family is twelve or fourteen – and they are kept expanded and carried so erect that in good specimens the head and tail touch; the oil gland is aborted. Several other less distinct breeds could be mentioned.

Skeletal structure also differs among breeds. The length, breadth, and curvature of face bones differ enormously. The size and shape of the ramus of the lower jaw, the size and shape of the apertures in the sternum, and the degree of divergence and relative sizes of the two arms of the furcula vary remarkably. The number of caudal and sacral vertebrae; the number of ribs, along with their relative breadth and presence of processes; the proportional width of the mouth; the proportional length of the eyelids, nostrils, tongue – not always in strict correlation with the length of the beak – the size of the crop and the upper esophagus; the development and abortion of the oil gland; the number of primary wing and caudal feathers; the relative lengths of the wings and the tail to each other and to the body; the relative lengths of the legs and feet; the number of scutellae on the toes; and the development of skin between the toes are all structural features that vary. The period at which complete plumage is acquired, the state of the down with which nestlings are clothed when hatched, and the size and shape of the eggs also vary. Flying style and, in some breeds, voice and disposition differ remarkably. Lastly, in certain breeds the males and females have come to differ slightly from each other.

There are at least twenty pigeon breeds that an ornithologist would classify as well-defined species if he were told they were wild birds. I doubt any ornithologist would place the English carrier, short-faced tumbler, runt, barb, pouter, and fantail in the same genus, especially because for each there are several true-breeding sub-breeds, or, as he would call them, species.

Despite the great differences among pigeon breeds, I agree with the common opinion of naturalists that they have all descended from the rock pigeon (Columba livia), a category that includes several geographical varieties or sub-species differing from one another in minor respects. Some of the justifications for this conclusion are somewhat applicable in other cases, so I will briefly give them here. If pigeon breeds are not varieties and have not descended from the rock pigeon, then they must have descended from at least seven or eight original stocks, because it would be impossible to generate the present domestic breeds by the crossing of any fewer. For example, how could a pouter be produced by crossing unless one of the parental stocks possessed the characteristically enormous crop? The supposed original stocks must all have been rock pigeons – that is, not breeding or perching in trees. But besides C. livia only two or three other species of rock pigeon are known, and these lack the characteristics of domestic breeds. Therefore, the supposed original stocks would have to either (1) still exist unknown to ornithologists in the regions where they were first domesticated, or (2) have become extinct. Because of the rock pigeon’s size, habits, and remarkable characteristics, (1) is unlikely. And birds that are good fliers and breed on precipices are unlikely to become extinct. The common rock pigeon has the same habits as the domestic breeds and hasn’t been exterminated on several British islets or the shores of the Mediterranean, making (2) a rash assumption. Furthermore, the above-mentioned breeds have been introduced to all parts of the world, so some of them must have wound up in their supposed native regions, yet not one has ever became wild or feral (although the dovecot pigeon, which is the rock pigeon in a slightly altered state, has become feral in several places). All recent experience demonstrates the difficulty of breeding wild animals in confinement, but the hypothesis of plural origin for domestic pigeons implies that at least seven or eight species were so thoroughly domesticated in ancient times by half-civilized humans that they were quite prolific under confinement.

An important argument, also relevant to other cases, is that although these breeds are generally the same in constitution, habits, voice, coloring, and most structural parts as the wild rock pigeon, they are extraordinarily abnormal in other parts of structure. We could search the entire pigeon family in vain for a beak like the English carrier’s, short-faced tumbler’s, or barb’s; for reversed feathers like the Jacobin’s; for a crop like the pouter’s; or for tail feathers like the fantail’s. So it would have to be assumed not only that half-civilized humans thoroughly domesticated several species but also that they intentionally or unintentionally picked abnormal species, and that these species are now all unknown or extinct. So many strange contingencies seem very improbable.

Coloration is worth considering. The rock pigeon is slate blue with a white rump (the Indian sub-species, Strickland’s C. intermedia, has a bluish rump). The tail has a terminal dark bar with the bases of the outer feathers edged in white; the wings have two black bars; some semidomestic and apparently wild breeds also have black checkers on the wings. These marks do not occur together in any other species of the whole family. All of these marks, down to the white edging of the tail feathers, sometimes develop perfectly in thoroughly well-bred specimens from every one of the domestic breeds. And when two birds from distinct breeds lacking these marks and blue coloring are crossed, the mongrel offspring are very likely to acquire them. I crossed some white fantails with black barbs and got mottled brown and black offspring. Then I crossed these together and one of the offspring had as beautiful a blue color, white rump, black double wing bar, and barred white-edged tail feathers as any wild rock pigeon! These observations can be understood through the well-known principle of reversion to ancestral characteristics, if all domestic breeds have descended from the rock pigeon. The alternative would require making one of two very unlikely assumptions: (1) all of the supposed multiple original stocks were colored and marked like the rock pigeon, even though no such species exists today, so that each breed reverts to the same characteristics, or (2) even the purest breed has been crossed by the rock pigeon within twelve or twenty generations.⁵

Hybrid offspring from between all domestic pigeons are fertile. I can state this from my own crosses, intentionally made between the most distinct breeds. It is difficult, if not impossible, to suggest a single case of fertile hybrid offspring from two unambiguously distinct species. Some authors believe that long-term domestication eliminates this strong tendency for sterility. This hypothesis may be true if applied to closely related species, based on the history of the dog, but is unsupported by a single experiment. However, it would be rash to extend the hypothesis and claim that supposedly original “species” as distinct as the carrier, tumbler, pouter, and fantail could have produced fertile offspring when crossed.

I feel no doubt that all domestic pigeon breeds have descended from Columba livia and its geographical sub-species. To reiterate, the reasons are: (1) it is unlikely that primitive humans got seven or eight supposed pigeon species to breed under domestication, with none of these supposed species existing today and none of the breeds having become feral in their supposed native regions; (2) these species have certain abnormal characteristics with respect to the whole pigeon family but are like rock pigeons in other respects; (3) the blue color and marks of the rock pigeon occasionally appear in all breeds both when kept pure and when crossed; and (4) mongrel offspring are fertile.

There is even further support for my assertion. The rock pigeon has been domesticated recently in Europe and India, agreeing in habit and many structural characteristics with all domestic breeds. Furthermore, it is possible to make an almost perfect incremental series between extremes of structure using sub-breeds within any one breed, especially if we include specimens from distant regions. Also, the main distinctive feature of each breed is highly variable. These considerations will be invoked in discussing selection as explaining the immense amount of variation pigeons have undergone. The reason that the breeds often have such monstrous characteristics will also be explained.

When I first kept pigeons, I felt as much difficulty in believing that they have descended from one parent as any naturalist would about the many species of finches or other large bird groups in the wild. It was striking to me that every breeder of domestic animals and every cultivator of plants with whom I talked or whose treatises I read is convinced that each breed has descended from a distinct original species. A celebrated breeder of Hereford cattle would laugh with scorn at the suggestion that his livestock have descended from long-horns. I have never met a pigeon, poultry, duck, or rabbit breeder who was not fully convinced that each main breed has descended from a distinct species. In his treatise on pears and apples, Van Mons rejects that the several varieties (such as Ribston pippin and Codlin apple) could ever have proceeded from seeds of the same tree. There are innumerable other examples. The explanation, I think, is simple: long-term study impresses on the mind differences between breeds, and although they know that individuals of each breed vary slightly – prizes are won by the selection of such slight differences – they fail to sum up in their minds how, over many generations, slight differences can accumulate into large differences. There are naturalists who know less about inheritance and no more about intermediate links in the lines of descent than breeders but nevertheless admit that many domestic varieties have descended from common parents. Yet they deride the idea of species in nature being lineal descendants of other species. Perhaps they should be more cautious.

What are the steps by which a domestic variety arises from one or several related species? Environmental conditions and habit may play a minor role, but they cannot account for the differences between a dray and a racehorse, a greyhound and a bloodhound, or a carrier and a tumbler pigeon. One of the most remarkable features of domesticated organisms is that we see in them adaptations, but to human use or fancy rather than the animal or plant’s own good. Some of these useful variations probably appeared suddenly. According to many botanists, the fuller’s teazle, with its hooks unrivaled by any mechanical device, is a variety of the wild Dipsacus that arose suddenly in a seedling. The same is probably true of the turnspit dog and known to be true of the ancon sheep. But on comparing the dray horse and racehorse; the dromedary and camel; the various sheep breeds fit for cultivated land or mountainous pasture and each with wool for a different purpose; the various dog breeds each uniquely useful to humans; the gamecock (so pertinacious in battle) with breeds that are not quarrelsome, with “everlasting layers” (which never sit), with the bantam (so small and elegant); and the host of agricultural, culinary, orchard, and flower-garden breeds of plants useful to humans at different seasons and for different purposes, or so beautiful in their eyes, mere variability does not suffice. Every breed could not have been suddenly produced perfectly useful; in some instances this is historically known not to be the case. The key is the human power of cumulative selection: nature provides successive variations and humans add them up in useful directions, thus producing different breeds.

The great power of selection is not hypothetical. Several eminent breeders have drastically modified some sheep and cattle breeds even within a single lifetime. To fully appreciate what they have done, it is necessary to read some of the many treatises on the subject and actually inspect the animals. Breeders habitually speak of animals’ organization as plastic, something they can mold almost as they please. (If I had space I could quote numerous passages to this effect from expert authorities.) Youatt was a good judge of animals and probably more knowledgeable about the work of agriculturalists than anyone; he describes selection as “that which enables the agriculturalist, not only to modify the character of his flock, but to change it altogether. It is the magician’s wand, by means of which he may summon into life whatever form and mold he pleases.” Speaking of breeders’ feats with sheep, Lord Somerville says, “It would seem as if they had chalked out upon a wall a form perfect in itself, and then had given it existence.” The skillful breeder Sir John Sebright used to say about pigeons that “he would produce any given feather in three years, but it would take him six years to obtain head and beak.” The importance of selection in breeding merino sheep is so fully recognized in Saxony that the use of the principle has become a trade. The sheep are placed on a table and studied like a picture by a connoisseur three times over the course of several months. The sheep are marked and classed so that the very best can ultimately be selected for breeding.

What English breeders have effected is proven by the high prices given for animals with a good pedigree, which have been exported almost everywhere in the world. The improvement is not generally a result of crossing different breeds – a practice opposed by the best breeders, with the exception of occasional crosses among closely related sub-breeds. When a cross is made, careful selection is even more important than usual. If selection involved simply separating some very distinct variety for breeding, then it would be too obvious to even discuss. But its importance lies in the dramatic effect of unidirectionally accumulating minute differences imperceptible to the untrained eye – mine included – over many generations. Very few men are discerning enough and in possession of the proper judgment to become eminent breeders. Even with these gifts, he will only make improvements by studying his subject for years and devoting a lifetime to his task with indomitable perseverance. Few would believe the talent and practice necessary to become even a skillful pigeon breeder.

These concepts also apply to horticulture, although the variations tend to be more abrupt. No one supposes that the choicest productions arose by a single variation in an original stock. This can be proven in some cases, for records have been kept; the steadily increasing size of the gooseberry is one example. The astonishing improvements in florists’ flowers are apparent when present-day varieties are compared with drawings made only twenty or thirty years ago. Seed raisers do not need to pick out the best specimens from a well-established plant variety, but simply pull up the “rogues,” as they call plants that deviate from the defined standard. This kind of selection is also followed with animals, because no one is careless enough to let the worst animals breed.

Another way to observe the accumulated effects of selection is by comparing the diversity of flowers of different varieties of one flower-garden species; the diversity of leaves, pods, tubers, or other valued parts of kitchen-garden species; or the diversity of fruits of a species in the orchard relative to other parts of the same variety. Notice how different the leaves of a cabbage are but how alike the flowers; how different the flowers of a heartsease are but how alike the leaves; how much the gooseberry fruit differs in size, color, shape, and fuzziness but how similar the flowers are. It’s not that varieties differing drastically in one way don’t differ in others. In fact, this is rarely the case. Correlated growth, which should never be underestimated, ensures some differences in other parts. However, as a general rule, selection for slight variations in a specific part will produce varieties differing mostly in that part.

Some may object that the principle of selection has been reduced to methodical practice for less than seventy-five years. It is true that it has been utilized frequently in recent years and accompanied by rapid and important results with many treatises published on the subject. All the same, it is not a modern discovery. I could give several references that acknowledge its importance in works of high antiquity. In rude and barbarous periods of English history, choice animals were often imported and laws were passed to prevent their exportation. The destruction of horses under a certain size was ordered, comparable to the “roguing” of plants mentioned earlier. I found that the principle of selection is clearly given in an ancient Chinese encyclopedia. Some of its explicit rules are written down by Roman classical writers. It is clear from passages in Genesis that the color of domestic animals was attended to at that early period. Indigenous groups sometimes cross their dogs with wild canines to improve the breed, both in the present day and – as attested by Pliny – in the past. The natives of South Africa mate their draft cattle by color, just as some of the Eskimo do for their dog teams. Livingstone reports that good domestic breeds are valued by inhabitants of Africa’s interior who have not associated with Europeans. Some of these examples do not demonstrate actual selection, but they show that breeding of domestic animals was done in ancient times and is now done by indigenous peoples. It would be strange if attention had not been paid to breeding, the inheritance of good and bad qualities being so obvious.

Today, breeders try to make a superior and novel strain or sub-breed by methodical selection with a preconceived object in view, but for this discussion a type of selection I call “unconscious” is more important; it results from everyone trying to possess and breed from the best individual animals. For example, a man who keeps pointers will obviously try to obtain the best dogs he can and then breed from the best of his own. He has no intention or expectation of permanently altering the breed, but this process, extended over centuries, will modify any breed. Using this same process, only more methodically, Bakewell, Collins, and others greatly modified, even during their own lifetimes, the forms and qualities of their cattle. These kinds of gradual and unobservable changes could never be recognized in the absence of measurements or careful drawings made long ago for comparison. In some cases, unchanged or slightly changed individuals of a known breed can be found in less civilized regions where the breed has been less improved. The King Charles spaniel may have been extensively modified unconsciously since the time of its namesake. Some authorities assert that the setter is directly derived from the spaniel, and the English pointer is known to have been greatly changed over the last century, probably by crosses with the foxhound. Importantly, change has been effected unconsciously and gradually but so effectively that although the old Spanish pointer came from Spain, there is no native dog in Spain like the English pointer.⁶

Through a similar process of selection, coupled with careful training, English racehorses have come to surpass their parent Arab stock in speed and size so that by the regulations of the Goodwood Races, they are favored in the weights they carry. Lord Spencer and others have shown that English cattle have increased in weight and early maturity compared to the stock once kept in this country. The stages through which carrier and tumbler pigeons have passed, and how they have come to differ so greatly from the rock pigeon, can be traced by comparing accounts given in old treatises with modern British, Indian, and Persian breeds.

Youatt gives an excellent example of unconscious selection in which the breeders in question could never have expected or even wanted to produce two distinct strains. He remarks that the two flocks of Leicester sheep kept by Mr. Buckley and Mr. Burgess “have been purely bred from the original stock of Mr. Bakewell for upward of fifty years. There is not a suspicion existing in the mind of any one at all acquainted with the subject that the owner of either of them has deviated in any one instance from the pure blood of Mr. Bakewell’s flock, and yet the difference between the sheep possessed by these two gentlemen is so great that they have the appearance of being quite different varieties.”

There may be peoples so barbarous that they never consider the inherited characteristics of their domestic animals, but if an animal is particularly useful to them for some special reason, it will be preserved during famines and other accidents and consequently leave more offspring than its inferior brethren; this is a kind of unconscious selection. The value of animals is demonstrated even among the barbarians of Tierra del Fuego who kill and devour their old women during times of dearth, as of less value than their dogs.

In plants the same gradual process of improvement through occasional preservation of the best individuals can be recognized in the larger size and more intense beauty of modern heartsease, rose, geranium, dahlia, and other varieties when compared to older varieties or parent stocks. (This applies whether or not the individuals can be classified as belonging to a distinct variety on first appearance, and whether or not species or varieties had been blended by crossing.) No one would expect to get a first-rate heartsease, dahlia, or melting pear from the seed of a wild plant. Although the pear was cultivated in classical times, Pliny’s descriptions suggest it was a fruit of very inferior quality. Horticultural essays convey surprise about the gardener’s wonderful skill in generating splendid products from such poor materials, but the art has been simple: the final result proceeds from an almost unconscious process. It has always involved cultivating the best-known variety, sowing its seeds, and selecting slightly better varieties when they happen to appear. The gardeners of classical times cultivated the best pear they could procure with no intention of providing us with such sweet fruit, and yet in part we owe to them our pear, because they naturally chose and preserved the best varieties they found.

The large amount of change that has been slowly accumulated unconsciously in cultivated plants explains why we cannot recognize the parent stocks of many established kitchen- and flower-garden varieties. If it took centuries or millennia of improvement to create useful plants, we can understand why Australia, the Cape of Good Hope, and other regions inhabited by uncivilized peoples have not afforded us a single plant worth culture. It is not that these species-rich countries lack original stocks of useful plants, but that these native plants have not been improved to a standard of perfection by continuous selection as in anciently civilized regions.

Domestic animals kept by uncivilized peoples almost always have to struggle for their own food, at least during certain seasons. Individuals of a single species with slightly different constitutions or structure often succeed better in one environment than another; by a process of “natural selection,” as will be explained, two sub-breeds might form. This may partly explain the observation of some authors that varieties kept by natives are more like well-defined species than the varieties of civilized countries.

Acknowledging the important role played by human selection, it becomes obvious how domestic organisms display adaptations of structure or habit conforming to human want or fancy. It also explains the frequently abnormal traits of domestic varieties and why their external characteristics vary greatly but their internal organs only slightly. Only with great difficulty can man select for structural deviations that cannot be externally observed, and he usually does not care for internal variations anyway. He can select only variations first provided by nature. He could not make a fantail, or think to or even try, until he found a pigeon with an unusually developed tail, or a pouter until he found a pigeon with an unusually large crop. The more abnormal a characteristic upon its first appearance, the more likely it is to catch his attention. But the phrase “to make a fantail” is not correct. The person who first selected a slightly larger tailed pigeon never imagined what its descendants would become through long-term, partly unconscious and partly methodical selection. Maybe the parent of all fantails had only fourteen slightly expanded tail feathers, like the Java fantail, or seventeen tail feathers like individuals of other breeds. Maybe the first pouter did not inflate its crop much more than the modern turbit inflates its upper esophagus, a habit disregarded by breeders because it is not one of the points of the breed.

A major structural deviation is not necessary to catch the breeder’s eye, which perceives extremely small differences. It is human nature to value even a slight novelty in one’s possession. The former value of slight differences cannot be judged by the value that such slight differences might have today after several breeds have been well established. Many slight deviations still arise among pigeons, but they are rejected as faults in the breed’s perfection. The common goose has not given rise to any marked varieties, so it has recently been exhibited at poultry shows as distinct from the Thoulouse goose, from which it differs only in the most fleeting characteristic: color.

I think these ideas further explain why we know nothing about the origin or history of domestic breeds. In fact, a breed, like a dialect of a language, cannot be said to have a definite origin. A man preserves and breeds an individual with some slight structural deviation, or takes more care than usual in matching his best animals, thereby improving them, and the improved offspring slowly spread into the immediate neighborhood; they are not yet separately named, and their history is disregarded because they are only slightly valued. With further improvement by the same slow and gradual process, they spread more widely and are recognized as distinct and valuable, deserving of a local name. (In semicivilized regions with limited free communication, a new sub-breed will spread slowly.) As soon as the new sub-breed’s valuable characteristics are fully acknowledged, unconscious selection always enhances the breed. Unconscious selection is perhaps more influential at one period than another, subject to the breed’s popularity, and in one region than another, according to the state of civilization of the inhabitants. The chances of any record describing such slow and immediately unobservable changes are infinitely small.

A high degree of variability is obviously favorable to the human power of selection, because it provides the raw materials for selection to work on; not that mere individual differences are insufficient to allow for the accumulation of extensive modification in almost any desired direction. Manifestly useful or pleasing variations appear only occasionally, but the odds can be increased by keeping a large number of individuals, an important tool for success. Marshall has remarked about this principle with respect to sheep in parts of Yorkshire that “as they generally belong to poor people, and are mostly in small lots, they never can be improved.” Professional plant breeders are generally more successful than amateurs in getting new and valuable varieties, because they raise large stocks of the same plant. When only a few individuals are kept, they are all allowed to breed, effectively preventing selection. Keeping large numbers of individuals naturally requires the creation of conditions favorable to that species to ensure proper breeding. But the most important point is that the organism should be so useful to or valued by humans that very close attention will be paid to even minor deviations in the quality or structure of each individual; without this, nothing can be effected. I have seen it gravely remarked how fortunate we are that the strawberry began to vary exactly when gardeners began to attend to it closely. Surely the strawberry had always varied, but slight variations were neglected; as soon as individual plants with slightly larger, earlier, or better fruit were selected for propagation over several generations (aided by some crosses with distinct species), those many admirable strawberry varieties appeared that have been raised for the past thirty or forty years.

Preventing crosses is an important element of success with animals that have separate sexes – at least in a region already stocked with other varieties. Enclosure of land plays a part in this. Wandering peoples or the inhabitants of open plains rarely have more than one breed per species. It is a huge convenience to the breeder and favorable to the formation of new breeds that pigeons mate for life, so many varieties can mingle in one aviary but keep true. Additionally, pigeons can be quickly propagated in large numbers and inferior birds easily rejected because they can serve as food when killed. Conversely, cats cannot be matched because of their nocturnal rambling habits, and though much loved by women and children, distinct breeds can hardly ever be maintained. The breeds that occasionally are observed are usually imported, often from islands. Although I do not doubt that some domestic animals intrinsically vary less than others, the rarity or absence of cat, donkey, peacock, goose, and other breeds is due mainly to selection not having been employed: in cats because they are difficult to pair, in donkeys because only a few are kept by the poor and little attention is paid to their breeding, in peacocks because they are difficult to rear and keep in large numbers, and in geese because they are valuable only for food and feathers, and because no one has taken pleasure in the display of distinct breeds.

To sum up the origin of domestic plant and animal varieties, environmental influence on the reproductive system is the most important cause of variability. I do not believe, as some authors do, that variability is an inherent and necessary contingency under all circumstances with all organic beings. The effects of variability are modified by inheritance and reversion. Variability is governed by many unknown rules, especially correlated growth. The direct action of the environment, and use and disuse, may have some effect. The final result is thus infinitely complex. In some cases intercrossing of originally distinct species played an important part in the origin of domestic breeds. The occasional crossing of established domestic breeds contributes to the creation of sub-breeds with the aid of selection, but the importance of crossing varieties has been greatly exaggerated both in animals and plants propagated by seed. (The importance of crossing distinct species and varieties is immensely important for plants temporarily propagated by cuttings, buds, etc., because the cultivator disregards the extreme variability of hybrids and mongrels and the sterility of hybrids. Plants not propagated by seed are temporary, so they are of little importance to this discussion.)⁷ Over all these causes of change, the cumulative action of selection, whether applied methodically and relatively quickly or unconsciously and relatively slowly but more efficiently, is by far the most predominant power.

1. St. Hilaire gives many examples in his great work on this subject, Histories des Anomalies.

2. [These phenomena are now well understood through genetics. – D.D.]

3. [“Generic characteristics” refer to those that are relevant at the genus level. – D.D.]

4. [In fact, they have; all dogs belong to the same species, descended from the wolf. – D.D.]

5. There is no support for the idea that offspring revert to ancestral characteristics introduced more than twenty generations ago. In a breed that is crossed with another distinct breed only once, the tendency for reversion obviously diminishes with each generation as the foreign contribution thins. But if parents revert to a characteristic lost by some former generation even in the absence of crosses with a distinct breed, then the tendency to revert can be transmitted undiminished and indefinitely. These two cases are often confused in treatises on inheritance.

6. I am told by Mr. Barrow, who has not observed any.

7. [Grafting and certain other forms of plant propagation do not involve reproductive cells or any other means for the exchange of (what we now know to be) genetic material, so these methods do not actually involve “crossing.” – D.D.]