But we must think of all these struggles and mutual influencings as taking place on the smallest possible scale, so that it is only by long summation that they can produce any visible effect, and we must never forget the essential significance of the plurality of ids, for these 'spontaneous' variations may take place in a different and quite independent manner in each individual id. If this were not so no intervention of personal selection would be possible, natural selection would not exist, and the adaptation of the organism from the single cell up to the whole would remain wholly unexplained. The whole crop of spontaneous germ-variations, whenever it ceases to be 'indifferent,' and becomes either 'good' or 'bad,' comes under the shears of personal selection and under its almost sovereign sway.

On the other hand, the sudden first appearance of a saltatory variation takes place quite independently of personal selection, depending on similar variations in a number of ids, which remain latent until they have by the process of reducing division which precedes amphimixis, chanced to attain a majority. In sudden bud-variations we may perhaps suppose that reducing division occurring in some still unverified abnormal manner is the reason why the germinal variation suddenly makes itself visible—a supposition previously suggested as the explanation of the reversion of these sports.

The rarity of bud-variation is thus explained, while the greater frequency of saltatory variations in plants propagated by seed may be accounted for by the regular occurrence of reducing division in sexual reproduction. But that the same or similar variations may occur in several, it may be in many, ids at the same time must depend upon similar general influences which affect the plant as a whole, as happens through cultivation, manuring, and so on. I shall return to this when discussing the influence of the environment.

In some quarters this whole conception of germinal selection has been characterized as the merest figment of imagination, condemned on this ground alone, that it is based on the differences in nutrition between such extremely minute quantities of substance as the chromosomes of nuclear substance within the germ-cell. The quantity of substance is certainly minute, but it needs nutriment none the less, and can we believe that the stream of nourishment for all the invisibly minute vital elements is exactly alike? It may be admitted that the nourishment outside the ids is usually abundant, although undoubtedly fluctuations occur in it also, but it certainly does not follow from this that every vital unit within the id is similarly disposed in relation to the nutritive supply, or has food in equal quantities at its command, or even that each has as much as it can ever need. To make an assertion like this seems to me much the same as if an inhabitant of the moon, looking at this earth through an excellent telescope and clearly descrying the city of Berlin with its thronging crowds and its railways bringing in the necessaries of life from every side, should conclude from this abundant provision that the greatest superfluity prevailed within the town, and that every one of its inhabitants had as much to live upon as he could possibly require.

We certainly ought not to conclude from the fact that we cannot see into the structure and requirements and methods of nutrition of a very minute mass of substance that its nutrition cannot be unequal, and that it cannot, by its inequalities, give rise to very material differences, especially when we are dealing with a substance to which we must attribute an extraordinarily complex organization built up of enormous numbers of extremely minute particles. That this complexity is undeniable is now admitted by many who formerly thought it possible to believe in the simple structure of the germ-substance. How complex not only the germ-substance but every cell of a higher organism is in its structure, and how far below the limits of visibility its differentiations and arrangements reach, is pressed upon our attention by the most recent histological researches, such as those we owe to Heidenhain, Boveri, and many others. The whole scientific world was amazed when it came to know the mysterious nuclear spindle in the seventies, and since then this has been quite thrown into the shade by the discovery of the centrosphere, the centrosome, and more recently even the centriole, and now we believe that these marvellous centres of force may, or must, possess their own dividing apparatus! In the face of discoveries like these no one is likely to be able to persist in recognizing as existing only what is disclosed or even hinted at by the most powerful lenses; no one can any longer doubt that far below the limit of visibility organization is still at the basis of life, and that it is dominated by orderly forces. To me, at least, it seems more cogent to argue from the phenomena of heredity and variation to an enormous mass of minute vital units crowded together in the narrow space of the id, than to argue from the calculated size of atoms and molecules to the number which we are justified in assuming to be present in an id. In my book on the germ-plasm I made a calculation of this kind, and I arrived at figures which seemed rather too small for the requirements of the germ-plasm theory. This has been regarded as a proof that I disregard the facts for the sake of my theory, but it should rather be asked whether the size of the atoms and molecules is a fact, and not rather the very questionable result of an uncertain method of calculation. Undoubtedly modern chemistry has established the relative weight-proportions of the atoms and molecules with admirable precision, but it can make only very uncertain statements in regard to the absolute size of the ultimate particles. It is therefore admissible to assume that these have a still greater degree of minuteness when the facts in another domain of science require this.

We must assume determinants, and consequently the germ-plasm must have room for these; the variations of species can only be explained through variations of the germ-plasm, for these alone give rise to hereditary variation. It is upon this foundation that my germinal selection is built up; whether I have in the main reached the truth the future will show: but that I have not exhausted this new domain, but only opened it up, I am very well aware.