were thus formed.^[36] These latter cases appear at first sight quite distinct from the congenital production of additional digits in the higher animals; but theoretically, as we shall see in a future chapter, they probably present no real difference. The larvæ or tadpoles of the tailless Batrachians, but not the adults,^[37] are capable of reproducing lost members.^[38] Lastly, as I have been informed by Mr. J. J. Briggs and Mr. F. Buckland, when portions of the pectoral and tail fins of various

fresh-water fish are cut off, they are perfectly reproduced in about six weeks' time.

From these several facts we may infer that supernumerary digits in man retain to a certain extent an embryonic condition, and that they resemble in this respect the normal digits and limbs in the lower vertebrate classes. They also resemble the digits of some of the lower animals in the number exceeding five; for no mammal, bird, existing reptile, or amphibian (unless the tubercle on the hind feet of the toad and other tailless Batrachians be viewed as a digit) has more than five; whilst fishes sometimes have in their pectoral fins as many as twenty metacarpal and phalangeal bones, which, together with the bony filaments, apparently represent our digits with their nails. So, again, in certain extinct reptiles, namely, the Ichthyopterygia, "the digits may be seven, eight, or nine in number, a significant mark," says Professor Owen, "of piscine affinity."^[39]

We encounter much difficulty in attempting to reduce these various facts to any rule or law. The inconstant number of the additional digits—their irregular attachment to either the inner or outer margin of the hand—the gradation which can be traced from a mere loose rudiment of a single digit to a completely double hand—the occasional appearance of additional digits in the salamander after a limb has been amputated—these various facts appear to indicate mere fluctuating monstrosity; and this perhaps is all that can be safely said. Nevertheless, as supernumerary digits in the higher animals, from their power of regrowth and from the number thus acquired exceeding five, partake of the nature of the digits in the lower vertebrate animals;—as they occur by no means rarely, and are transmitted with remarkable strength, though perhaps not more strongly than some other anomalies;—and as with animals which have fewer than five digits, when an additional one appears it is generally due to the development of a visible rudiment;—we are led in all cases to suspect, that, although no actual rudiment can be detected, yet that a latent tendency to the formation of an additional digit exists in all mammals, including man. On this view, as we shall more plainly see in the

next chapter when discussing latent tendencies, we should have to look at the whole case as one of reversion to an enormously remote, lowly-organised, and multidigitate progenitor.

I may here allude to a class of facts closely allied to, but somewhat different from, ordinary cases of inheritance. Sir H. Holland^[40] states that brothers and sisters of the same family are frequently affected, often at about the same age, by the same peculiar disease, not known to have previously occurred in the family. He specifies the occurrence of diabetes in three brothers under ten years old; he also remarks that children of the same family often exhibit in common infantile diseases the same peculiar symptoms. My father mentioned to me the case of four brothers who died between the ages of sixty and seventy, in the same highly peculiar comatose state. An instance has been already given of supernumerary digits appearing in four children out of six in a previously unaffected family. Dr. Devay states^[41] that two brothers married two sisters, their first-cousins, none of the four nor any relation being an albino; but the seven children produced from this double marriage were all perfect albinoes. Some of these cases, as Mr. Sedgwick^[42] has shown, are probably the result of reversion to a remote ancestor, of whom no record had been preserved; and all these cases are so far directly connected with inheritance that no doubt the children inherited a similar constitution from their parents, and, from being exposed to nearly similar conditions of life, it is not surprising that they should be affected in the same manner and at the same period of life.

Most of the facts hitherto given have served to illustrate the force of inheritance, but we must now consider cases, grouped as well as the subject allows into classes, showing how feeble, capricious, or deficient the power of inheritance sometimes is. When a new peculiarity first appears, we can never predict whether it will be inherited. If both parents from their birth present

the same peculiarity, the probability is strong that it will be transmitted to at least some of their offspring. We have seen that variegation is transmitted much more feebly by seed from a branch which had become variegated through bud-variation, than from plants which were variegated as seedlings. With most plants the power of transmission notoriously depends on some innate capacity in the individual: thus Vilmorin^[43] raised from a peculiarly coloured balsam some seedlings, which all resembled their parent; but of these seedlings some failed to transmit the new character, whilst others transmitted it to all their descendants during several successive generations. So again with a variety of the rose, two plants alone out of six were found by Vilmorin to be capable of transmitting the desired character.

The weeping or pendulous growth of trees is strongly inherited in some cases, and, without any assignable reason, feebly in other cases. I have selected this character as an instance of capricious inheritance, because it is certainly not proper to the parent-species, and because, both sexes being borne on the same tree, both tend to transmit the same character. Even supposing that there may have been in some instances crossing with adjoining trees of the same species, it is not probable that all the seedlings would have been thus affected. At Moccas Court there is a famous weeping oak; many of its branches "are 30 feet long, and no thicker in any part of this length than a common rope:" this tree transmits its weeping character, in a greater or less degree, to all its seedlings; some of the young oaks being so flexible that they have to be supported by props; others not showing the weeping tendency till about twenty years old.^[44] Mr. Rivers fertilized, as he informs me, the flowers of a new Belgian weeping thorn (Cratægus oxyacantha) with pollen from a crimson not-weeping variety, and three young trees, "now six or seven years old, show a decided tendency to be pendulous, but as yet are not so much so as the mother-plant." According to Mr. MacNab,^[45] seedlings from a magnificent weeping birch (Betula alba), in the Botanic Garden at Edinburgh, grew for the first ten or fifteen years upright, but then all became weepers like their parent. A peach with pendulous branches, like those of the weeping willow, has been found capable of propagation by seed.^[46] Lastly, a weeping and almost prostrate yew (Taxus baccata) was found in a hedge in Shropshire; it was a male, but one branch bore female flowers, and produced berries; these,

being sown, produced seventeen trees, all of which had exactly the same peculiar habit with the parent-tree.^[47]

These facts, it might have been thought, would have been sufficient to render it probable that a pendulous habit would in all cases be strictly inherited. But let us look to the other side. Mr. MacNab^[48] sowed seeds of the weeping beech (Fagus sylvanica), but succeeded in raising only common beeches. Mr. Rivers, at my request, raised a number of seedlings from three distinct varieties of weeping elm; and at least one of the parent-trees was so situated that it could not have been crossed by any other elm; but none of the young trees, now about a foot or two in height, show the least signs of weeping. Mr. Rivers formerly sowed above twenty thousand seeds of the weeping ash (Fraxinus excelsior), and not a single seedling was in the least degree pendulous: in Germany, M. Borchmeyer raised a thousand seedlings, with the same result. Nevertheless, Mr. Anderson, of the Chelsea Botanic Garden, by sowing seed from a weeping ash, which was found before the year 1780, in Cambridgeshire, raised several pendulous trees.^[49] Professor Henslow also informs me that some seedlings from a female weeping ash in the Botanic Garden at Cambridge were at first a little pendulous, but afterwards became quite upright: it is probable that this latter tree, which transmits to a certain extent its pendulous habit, was derived by a bud from the same original Cambridgeshire stock; whilst other weeping ashes may have had a distinct origin. But the crowning case, communicated to me by Mr. Rivers, which shows how capricious is the inheritance of a pendulous habit, is that a variety of another species of ash (F. lentiscifolia) which was formerly pendulous, "now about twenty years old has long lost this habit, every shoot being remarkably erect; but seedlings formerly raised from it were perfectly prostrate, the stems not rising more than two inches above the ground." Thus the weeping variety of the common ash, which has been extensively propagated by buds during a long period, did not, with Mr. Rivers, transmit its character to one seedling out of above twenty thousand; whereas the weeping variety of a second species of ash, which could not, whilst grown in the same garden, retain its own weeping character, transmitted to its seedlings the pendulous habit in excess!

Many analogous facts could be given, showing how apparently capricious is the principle of inheritance. All the seedlings from a variety of the Barberry (B. vulgaris) with red leaves inherited the same character; only about one-third of the seedlings of the copper Beech (Fagus sylvatica) had purple leaves. Not one out of a hundred seedlings of a variety of the Cerasus padus, with yellow fruit, bore yellow fruit: one-twelfth of the seedlings of the variety of Cornus mascula, with yellow fruit, came true:^[50] and lastly, all the trees raised by my father from a yellow-berried holly (Ilex aquifolium),

found wild, produced yellow berries. Vilmorin^[51] observed in a bed of Saponaria calabrica an extremely dwarf variety, and raised from it a large number of seedlings; some of these partially resembled their parent, and he selected their seed; but the grandchildren were not in the least dwarfed: on the other hand, he observed a stunted and bushy variety of Tagetes signata growing in the midst of the common varieties by which it was probably crossed; for most of the seedlings raised from this plant were intermediate in character, only two perfectly resembling their parent; but seed saved from these two plants reproduced the new variety so truly, that hardly any selection has since been necessary.

Flowers transmit their colour truly, or most capriciously. Many annuals come true: thus I purchased German seeds of thirty-four named sub-varieties of one race of ten-week stocks (Matthiola annua), and raised a hundred and forty plants, all of which, with the exception of a single plant, came true. In saying this, however, it must be understood that I could distinguish only twenty kinds out of the thirty-four named sub-varieties; nor did the colour of the flower always correspond with the name affixed to the packet; but I say that they came true, because in each of the thirty-six short rows every plant was absolutely alike, with the one single exception. Again, I procured packets of German seed of twenty-five named varieties of common and quilled asters, and raised a hundred and twenty-four plants; of these, all except ten were true in the above limited sense; and I considered even a wrong shade of colour as false.

It is a singular circumstance that white varieties generally transmit their colour much more truly than any other variety. This fact probably stands in close relation with one observed by Verlot,^[52] namely, that flowers which are normally white rarely vary into any other colour. I have found that the white varieties of Delphinium consolida and of the Stock are the truest. It is, indeed, sufficient to look through a nurseryman's seed-list, to see the large number of white varieties which can be propagated by seed. The several coloured varieties of the sweet-pea (Lathyrus odoratus) are very true; but I hear from Mr. Masters, of Canterbury, who has particularly attended to this plant, that the white variety is the truest. The hyacinth, when