It is clearly a preformation hypothesis. It is true that the actual organism is not contained in the germ, but all the parts of the latter, even the colours of the eyes or hair, are present in it in the form of the determinants. Obviously it involves a mechanism of almost incredible complexity. But if we regard it as a working hypothesis of development this complexity of detail does not matter; its truth would be indicated by the fact that all analysis of the processes involved would tend to simplify it and to smooth out the complexity. But this is exactly what has not happened, for all subsequent investigation has necessitated subsidiary hypothesis after hypothesis. As a theory of development it has failed entirely. If, after one of the blastomeres in the frog’s egg at the 2-cell stage be killed, the egg is then turned upside down, the results of the experiment become totally different; the uninjured blastomere develops into a _whole_ embryo, differing from the normal one chiefly in that it is smaller. If the uninjured egg in the 2-cell stage be turned upside down _two whole embryos_, connected together in various ways, develop. In the frog’s egg the two first blastomeres cannot be separated from each other without rupturing them, but in the egg of the salamander they can be separated. After this separation two perfect, but small, embryos develop. In the egg of the newt a fine thread can be tied round the furrow formed by the first division. If this ligature be tied loosely it does not affect development, and then it can be seen that the median longitudinal plane of the embryo does not correspond, except by chance, with the first division plane. If the ligature be tied tightly, then each of the blastomeres gives rise to an entire embryo. If it is tied in various places monsters of various types are produced. Therefore there is no segregation of the determinants in the first two blastomeres. These results, moreover, are not exceptional, for similar ones have been obtained with other animal embryos, in fishes, _Amphioxus_, ascidians, medusæ, and hydrozoa, and in some cases even each of the first four blastomeres develops into an entire embryo when it is separated from the rest. In the sea-urchin embryo the blastomeres can be shaken apart; or by removing the calcium which is contained in sea water the blastomeres can easily be separated from each other. It was then found by Driesch that each of the blastomeres in the 16-cell stage could develop into an entire embryo. It is plain, then, that up to this stage at least there has been no segregation of the determinants.