plasm to one-half their original number. According as chance brings together, in amphimixis, two cells containing a majority or a minority of the ids A, will the embryonic cells, and consequently the cells in the growing organism, have a tendency to a plus or to a minus variation. Any germ-cell which contains a majority of plus determinants may unite with another cell containing likewise a majority of plus determinants, and the character represented by these determinants will be the more strongly accentuated in the offspring.

When a given characteristic becomes of pronounced utility to the life of the species, the determinants of that characteristic will be more likely to vary in an ascendant than in a regressive direction. For instance, the peculiar luminous apparatus developed by fishes inhabiting the dark depths of the ocean is obviously adapted to their conditions of life. Some of them possess rows of miniature organs with luminous secretion on their sides or along their ventral surface; others possess similar organs on the head; and it is extremely probable that these luminous organs serve to attract the small animals which form the prey of these fishes, just as the electric light above ground attracts insects, which are lured by it to their destruction. When such luminous organs became directly useful to the species, they were bound to increase, for those individuals whose determinants tended to a plus variation of these organs would obviously be favoured in the struggle for existence; amphimixis would tend to disseminate the determinants favourable to this new adaptation, and ultimately the possessing of luminous organs would become a characteristic of the species.

Natural selection operates only when the variations effected in the germ-plasm by perturbations in the balance of intragerminal nutrition attain selective value-that is to say, when they become of vital importance for the species in the struggle for life. If advantageous, natural selection will further the development of such variations by eliminating gradually and

GERMINAL AND PERSONAL SELECTION

53

progressively those individuals who do not possess them; and thus does natural selection tend to produce uniformity in the germ-plasm within a given species under given conditions. The more individuals of a species there are who possess a majority of determinants A1, corresponding to a variation A, the less chance is there of these determinants being eliminated in the reduction of the ids at maturation or in amphimixis.

But can germinal selection bring about lasting changes in the character of a species without the co-operation of natural selection, or, as Weismann calls it, personal selection? The answer must be in the negative. Germinal selection is not by itself capable of transforming a species, although it is the origin of those variations which personal selection operates on in bringing about such a transformation. By itself, germinal selection can effect changes which possess a purely morphological value, and which do not influence in any way the life of the individual or the life of the species; but once a variation has attained selective value, such variation must either be maintained or eradicated by the action of personal selection.

For instance, the possession of six fingers on the hand, or of a tiny scar behind the ear, does not in any way influence the life of the individual in the struggle for existence; consequently, there is no reason for natural selection to intervene, either by the maintenance or by the eradication of such abnormalities, which have a purely morphological, as distinct from a biological, value. Such variations are exclusively the work of germinal selection, and their disappearance is effected solely by perturbations of intragerminal nutrition in the germ-plasm. On the other hand, the peculiar colouring of certain insects is the result of adaptation to peculiar conditions of life, and is the work of natural selection supplementing the action of germinal selection. There is a certain species of butterfly, Phyllodes ornata, a native of Assam, which shows the result of adaptation in a very highly developed degree, although numerous equally significant cases

could also be cited. The hind-wing of this butterfly is for about three-quarters of its size deep black, and for about one-quarter yellow. When the animal is at rest, this wing is covered by the fore-wing. The latter is of reddish-brown colour sprinkled with black, and with a black pattern on it resembling with extraordinary precision the venation of a leaf. A perpendicular line commences at the end of the wing, and breaks off a little more than half-way along, and three black lines flank the perpendicular one on either side (Fig. 4). After breaking off in the centre of the wing, the perpendicular line begins again, but is shadowy

and indistinct, as are also its flanking lines. The insect, when at rest, exactly resembles a withered leaf, and it must be impossible to distinguish it from the leaves among which it lives. This is a case of adaptation to the conditions of the environment. The colouring of the fore-wing of Phyllodes ornata possesses biological value for that species, enabling it to escape detection by its enemies. Natural selection has here been at work, for those insects whose fore-wing was less perfectly adapted to the colour of the surrounding leaves were more likely to be detected and destroyed, and thus it has come about that, as only those individuals which were best adapted survived,

EXCESSIVE VARIATION

55

the adaptation has become ever more perfect- that is to say, in ever-increasing harmony with the environmental conditions.

There being only two possible kinds of variation among the determinants of the germ-plasm-namely, plus or minus variations-it is obvious that the chances of a variation in one or other of these directions are far greater than the chances that the plus or minus determinants of a given organism simply balance one another as a result of reduction and amphimixis. This dominant tendency to variation possessed by the determinants may seem to contradict the fact that many species are undoubtedly constant in their characters, and have existed without noteworthy change for a very long time. Among the Cephalopods, the Nautilus has certainly existed since the Silurian Age. The fact that in past ages the Nautili were extremely numerous, whereas at present the few remaining species are confined to the Indian and South Pacific Oceans, does not alter the fact that their constancy of form has been preserved throughout an immense period. How can we reconcile this fact with the tendency to variation possessed by the determinants?

Professor Emery of Bologna was, we think, the first to maintain that the tendency of certain variations to be carried to excess was an important factor in the extinction of species. But there seems no reason for supposing any variation to have an irrepressible or irresistible power, incapable of being checked by natural selection except by the extirpation of the whole species. There are several remarkable cases of adaptation known in which the particular variation stops just when the adaptation is completed-that is to say, just when such variation attains the limits of biological value, and when any further development would be harmful. These cases show that there exists some relation of proportion between variation and

adaptation. The fact is that at any point natural selection may limit the further progression of a plus variation. To minus variations, in the case of organs or characteristics harmful to the species, there is no limitation other than of the complete disappearance of the organ in question. But plus variations are arrested as soon as they attain their object-that of adapting the organism to its environment. Excessive variation is not a cause of the extinction of the species, but results only in the suppression of individuals who may exhibit it.

Personal selection is a very important—perhaps the most important factor in maintaining the constancy of a species. Every excessive variation which breaks the harmony of the species is at once eliminated by personal selection. The rôle of germinal selection is a less important one. Theoretically, there is no reason for supposing that the germ-plasm may not have itself the power of checking the plus variations of any sort which may arise within it. Support is lent to this theory by the fact that numerous morphological variations which are below selective value disappear after a while. And the fact that many species are undoubtedly constant throughout very long periods shows that not every movement among the determinants towards a plus variation continues sufficiently far to attain biological value.

But the very fact that only those variations which are of purely morphological value are capable of being controlled by germinal selection demonstrates that the chief factor in regulating the constancy of species is personal selection. Weismann has distinctly admitted that as soon as any plus variation has gone beyond a certain point germinal selection alone is powerless to hinder its further progress. In the case of ancient and consequently constant species, the germ-plasm is fixed, the distinctive determinants of the species stand little or no risk of being weakened or eliminated, and variations which may arise are immediately checked before they attain biological value—that