The chromosomes are now completely split, each into its two daughter-segments, which glide apart (Fig. 7, G, _ep_), and pass each to its own pole of the spindle, stopping just short of the centrosome (I). Thus, on the inner side of either centrosome is found an aggregation of daughter-segments, each of which is sister to one at the opposite pole, while the number at either pole is identical with that of the segments into which the old nucleus had resolved itself at the outset. The daughter-segments shorten and thicken greatly as they diverge to the poles, and on their arrival crowd close together. A distinct wall now forms around the aggregated {28}daughter-chromosomes (J), so as to combine them into a nucleus for the daughter-cell. The reorganisation of the young nucleus certainly varies in different cases, and has been ill-studied, probably because of the rapidity of the changes that take place. The cytoplasm now divides, either tapering into a "waist" which finally ruptures, or constricting by the deepening of a narrow annular groove so as to complete the formation and isolation of the daughter-cells. We might well compare the cell-division to the halving of a pumpkin or melon, of which the flesh as a whole is simply divided into two by a transverse cut, while the seeds and the cords that suspend them are each singly split to be divided evenly between the two halves of the fruit; the flesh would represent the cytoplasm, the cords the linin threads of the nucleus, and the seeds the chromatin granules. In this way the halving of the nucleus is much more complete and intimate than that of the cytoplasm; and this is the reason why many biologists have been led to regard the nuclear segments, and especially their chromatic granules, as the seat of the hereditary properties of the cell, properties which have to be equally transmitted on its fission to each daughter-cell.[40] But we must remember that the linin is also in great part used up in the formation of these segments, like the cords of our supposed melon; and it is open to us to regard the halving in this intimate way of the "linin" as the essence of the process, and that of the chromatin as accessory, or even as only part of the necessary machinery of the process. The halving or direct splitting lengthwise of a viscid thread is a most difficult problem from a physical point of view; and it may well be that the chromatin granules have at least for a part of their function the facilitation of this process. If such be the case, we can easily understand the increase in number, and size and staining power of these granules as cell-division approaches, and their atrophy or partial disappearance during their long intervening periods of active cell life. Hence we hesitate to accept the views so commonly maintained that the chromatin represents a {29}"germ-plasm" or "idioplasm" of relatively great persistence, which gives the cell its own racial qualities.[41]