common species they are sure to be found; that they are everywhere of considerable amount, often reaching 20 per cent of the size of the part implicated; and that they are to a great extent independent of each other, and thus afford almost any combination of variations that may be needed.

It must be particularly noticed that the whole series of variation-diagrams here given (except the three which illustrate the number of varying individuals) in every case represent the actual amount of the variation, not on any reduced or enlarged scale, but as it were life-size. Whatever number of inches or decimals of an inch the species varies in any of its parts is marked on the diagrams, so that with the help of an ordinary divided rule or a pair of compasses the variation of the different parts can be ascertained and compared just as if the specimens themselves were before the reader, but with much greater ease.

In my lectures on the Darwinian theory in America and in this country I used diagrams constructed on a different plan, equally illustrating the large amount of independent variability, but less simple and less intelligible. The present method is a modification of that used by Mr. Francis Galton in his researches on the theory of variability, the upper line (showing the variability of the body) in Diagrams 4, 5, 6, and 13, being laid down on the method he has used in his experiments with sweet-peas and in pedigree moth-breeding. I believe, after much consideration, and many tedious experiments in diagram-making, that no better method can be adopted for bringing before the eye, both the amount and the peculiar features of individual variability.

Variations of the Habits of Animals.

1

Closely connected with those variations of internal and external structure which have been already described, are the changes of habits which often occur in certain individuals or in whole species, since these must necessarily depend upon some corresponding change in the brain or in other parts of the organism; and as these changes are of great importance in relation to the theory of instinct, a few examples of them will be now adduced.

1 See Trans. Entomological Society of London, 1887, p. 24.

The Kea (Nestor notabilis) is a curious parrot inhabiting the mountain ranges of the Middle Island of New Zealand. It belongs to the family of Brush-tongued parrots, and naturally feeds on the honey of flowers and the insects which frequent them, together with such fruits or berries as are found in the region. Till quite recently this comprised its whole diet, but since the country it inhabits has become occupied by Europeans it has developed a taste for a carnivorous diet, with alarming results. It began by picking the sheepskins hung out to dry or the meat in process of being cured. About 1868 it was first observed to attack living sheep, which had frequently been found with raw and bleeding wounds on their backs. Since then it is stated that the bird actually burrows into the living sheep, eating its way down to the kidneys, which form its special delicacy. As a natural consequence, the bird is being destroyed as rapidly as possible, and one of the rare and curious members of the New Zealand fauna will no doubt shortly cease to exist. The case affords a remarkable instance of how the climbing feet and powerful hooked beak developed for one set of purposes can be applied to another altogether different purpose, and it also shows how little real stability there may be in what appear to us the most fixed habits of life. A somewhat similar change of diet has been recorded by the Duke of Argyll, in which a goose, reared by a golden eagle, was taught by its foster-parent to eat flesh, which it continued to do regularly and apparently with great relish.1

Change of habits appears to be often a result of imitation, of which Mr. Tegetmeier gives some good examples. He states that if pigeons are reared exclusively with small grain, as wheat or barley, they will starve before eating beans. But when they are thus starving, if a bean-eating pigeon is put among them, they follow its example, and thereafter adopt the habit. So fowls sometimes refuse to eat maize, but on seeing others eat it, they do the same and become excessively fond of it. Many persons have found that their yellow crocuses were eaten by sparrows, while the blue, purple, and white coloured varieties were left untouched; but Mr. Tegetmeier, who grows only these latter colours, found that after

1 Nature, vol. xix. p. 554.

two years the sparrows began to attack them, and thereafter destroyed them quite as readily as the yellow ones; and he believes it was merely because some bolder sparrow than the rest set the example. On this subject Mr. Charles C. Abbott well remarks: "In studying the habits of our American birds —and I suppose it is true of birds everywhere—it must at all times be remembered that there is less stability in the habits of birds than is usually supposed; and no account of the habits of any one species will exactly detail the various features of its habits as they really are, in every portion of the territory it inhabits." 1

Mr. Charles Dixon has recorded a remarkable change in the mode of nest-building of some common chaffinches which were taken to New Zealand and turned out there. He says: "The cup of the nest is small, loosely put together, apparently lined with feathers, and the walls of the structure are prolonged for about 18 inches, and hang loosely down the side of the supporting branch. The whole structure bears some resemblance to the nests of the hangnests (Icterida), with the exception that the cavity is at the top. Clearly these New Zealand chaffinches were at a loss for a design when fabricating their nest. They had no standard to work by, no nests of their own kind to copy, no older birds to give them any instruction, and the result is the abnormal structure I have just described." 2

These few examples are sufficient to show that both the habits and instincts of animals are subject to variation; and had we a sufficient number of detailed observations we should probably find that these variations were as numerous, as diverse in character, as large in amount, and as independent of each other as those which we have seen to characterise their bodily structure.

The Variability of Plants.

The variability of plants is notorious, being proved not only by the endless variations which occur whenever a species is largely grown by horticulturists, but also by the great difficulty that is felt by botanists in determining the limits of species in 1 Nature, vol. xvi. p. 163; and vol. xi. p. 227.

2 lbid., vol. xxxi. (1885), p. 533.

many large genera. As examples we may take the roses, the brambles, and the willows as well illustrating this fact. In Mr. Baker's Revision of the British Roses (published by the Linnean Society in 1863), he includes under the single species, Rosa canina the common dog-rose -no less than twenty-eight named varieties distinguished by more or less constant characters and often confined to special localities, and to these are referred about seventy of the species of British and continental botanists. Of the genus Rubus or bramble, five British species are given in Bentham's Handbook of the British Flora, while in the fifth edition of Babington's Manual of British Botany, published about the same time, no less than forty-five species are described. Of willows (Salix) the same two works enumerate fifteen and thirty-one species respectively. The hawkweeds (Hieracium) are equally puzzling, for while Mr. Bentham admits only seven British species, Professor Babington describes no less than thirty-two, besides several named varieties.

A French botanist, Mons. A. Jordan, has collected numerous forms of a common little plant, the spring whitlow-grass (Draba verna); he has cultivated these for several successive years, and declares that they preserve their peculiarities unchanged; he also says that they each come true from seed, and thus possess all the characteristics of true species. He has described no less than fifty-two such species or permanent varieties, all found in the south of France; and he urges. botanists to follow his example in collecting, describing, and cultivating all such varieties as may occur in their respective districts. Now, as the plant is very common almost all over Europe and ranges from North America to the Himalayas, the number of similar forms over this wide area would probably have to be reckoned by hundreds if not by thousands.

The class of facts now adduced must certainly be held to prove that in many large genera and in some single species there is a very large amount of variation, which renders it quite impossible for experts to agree upon the limits of species. We will now adduce a few striking cases of individual variation.

The distinguished botanist, Alp. de Candolle, made a special study of the oaks of the whole world, and has stated some