SUMMARY. Palaeontological evidence suggests that the beginning of each phylogenetic cycle, irrespective of the systematic category concerned, is marked by an intensification of evolution. In the later phases, the rate of evolution is much smaller and the ability to change decreases. An instructive example is the evolution of the placental mammals in the early Tertiary. In the upper Cretaceous the Insectivora appear, initiating the evolutionary cycle of the Eutheria. At the boundary of Cretaceous and Tertiary, and in the early Paleocene all other known placental orders become differentiated from the original group. Twenty-five orders, representing the entire morphological range of the subclass, appear in the relatively short period of 10–15 million years, whilst during the subsequent, post-Paleocene, period of 60 million years not a single new order is added. Evolution, of course, is not arrested in this second period. Specialization and adaptation continue within the framework of the orders, which split up into lower systematic categories. The chief features of the organization of orders, like the whales, or the bats, or the armadillos appear already at the beginning of the cycle. The upper Devonian Cephalopod family Manticoceratidae acquired a median saddle in its suture-line. Once more, intense splitting occurred at the very beginning, and within the very short period of a sub-division of a Goniatite horizon, of the duration of perhaps half a million years, at least 6 genera were formed. The most highly specialized types, however, soon became extinct, but the simpler genera survived through a period of about 3 million years without change except in insignificant specific characters. A phylogenetic cycle therefore, comprises three phases, the first and rapid one of typogenesis, the drawn-out second one of typostasis, and a final one (not here discussed) of typolysis or disintegration of the structural characters which precedes extinction. The typogenetic phase is Simpson's ‘quantum evolution’ and Zeuner's ‘explosive phase’. How can it be explained ? Stratigraphical sequences of fossiliferous deposits are so complete that, in the cases considered, gaps in the palaeontological record have to be ruled out. It has been suggested that early forms were often of small size, and thus would have had a more rapid succession of generations increasing the effects of selection. But since Zeuner demonstrated as long ago as in 1931 and as Simpson rediscovered independently in 1944, that the number of generations appears to have little effect on phylogenetic evolution, this explanation has to be abandoned. Another interpretation assumes very small populations with high selection pressure at the beginning of an evolutionary cycle. This is a possibility, but in many investigated cases it does not apply. Similarly, occupation of ecological niches or empty spaces, with increase in the effects of selection, is conceivable, but the condition that niches are available is often not fulfilled. An increase in the rate of mutation does not afford a satisfactory explanation either, since in such process disadvantageous mutations would preponderate, shift the balance of mutability and selection to the negative side and undoubtedly lead to an overall loss in adaptive efficiency. Thus, there remains only the assumption of an unusually large size (or effect) of the mutational steps. Single mutations or short series of mutations would interfere with the essential elements in the organization, and therefore change it in a short time and present the processes of natural selection with an altogether new type of organization. They have been called ‘Grossmutationen’. Starting from palaeontological evidence one thus arrives at conclusions not unlike those of the geneticists.