natural selection has had to counteract this increase , so far as length is concerned, and to effect 8 per cent. of shortening besides. If use-inheritance thickens bones without proportionally lengthening them, it would hinder rather than help the evolution of such structures as the long light wings of birds, or the long legs and neck of the giraffe or crane.

VARIED EFFECTS OF USE AND DISUSE.

The changes which we somewhat roughly and empirically group together as the effects of use and disuse are of widely diverse character. Thus bone, as the physiological fact, thickens under alternations of pressure (and the consequent increased flow of nourishment), but atrophies under a steadily continued pressure; so that if the use of a bone involved continuous pressure, the effect of such use would be a partial or total absorption of that bone. Darwin shows that

bone lengthens as well as thickens from carrying a greater weight, while tension (as seen in sailors' arms, which are used in pulling) appears to have an equally marked effect in shortening bones (Descent of Man, p. 32). Thus different kinds of use may produce opposite results. The cumulative inheritance of such effects would often be mischievous. The limbs of the sloth and the prehensile tail of the spider monkey would continually grow shorter, while the legs of the evolving elephant or rhinoceros might lengthen to an undesirable extent. Such cumulative tendencies of use-inheritance, if they exist, are obviously well kept under by natural selection.

Although the ultimate effect of use is generally growth or enlargement through increased flow of blood, the first effect usually is a loss of substance, and a consequent diminution of size and strength. When the loss exceeds the growth, use will diminish or deteriorate the part used, while disuse

would enlarge or perfect it. Teeth, claws, nails, skin, hair, hoofs, feathers, &c., may thus be worn away faster than they can renew themselves. But this wearing away usually stimulates the repairing process, and so increases the rate of growth; that is, it will increase the size produced, if not the size retained. Which effect of use does use-inheritance transmit in such cases-the increased rate of growth, or the dilapidation of the worn-out parts? We can hardly suppose that both these effects of use will be inherited. Would shaving destroy the beard in time or strengthen it ? Will the continued shearing of sheep increase or lessen the growth of wool? What will be the ultimate effect of plucking geese's quills, and of the eider duck's abstraction of the down from her breast? If the mutilated parts grow stronger or more abundantly, why were the motmot's feathers alleged to be narrowed by the inherited effects of ancestral nibbling?

The " use or "work" or "function" of muscles, nerves, bones, teeth, skin, tendon, glands, ducts,

eyes, blood corpuscles, cilia, and the other constituents of the organism, is as widely different as the various parts are from each other, and the effects of their use or disuse are equally varied and complicated.

USE-INHERITANCE IMPLIES PANGENESIS.

How could the transmission of these varied effects to offspring be accounted for? Is it possible to believe, with Mr. Spencer, that the effects of use and disuse on the parts of the personal structure are simultaneously registered in'

cor

responding impressions on the seminal germs ? Must we not feel, with Darwin apparently,1 that the only intelligible explanation of use-inheritance

1 Variation of Animals and Plants under Domestication, ii. 388, 398, 367; Life and Letters, iii. 44.

is the hypothesis of Pangenesis, according to which each modified cell, or physiological unit, throws off similarly-modified gemmules or parts of itself, which ultimately reproduce the change in offspring? If we reject pangenesis, it becomes difficult to see how use-inheritance can be

possible.

PANGENESIS IMPROBABLE.

The more important and best-known phenomena of heredity do not require any such hypothesis, and leading facts (such as atavism, transmission of lost parts, and the general nontransmission of acquired characters) are so adverse to it that Darwin has to concede that many of the reproductive gemmules are atavistic, and that by continuous self-multiplication they may preserve a practical " continuity of germ-substance," as Weismann would term it. The idea that the