Form and shape of pottery are often quite or very persistent. We cannot understand why the base of so many jars was left rounded, or in some old lake-dwellings pointed, when it might easily have been flattened, apparently to good advantage. But even the form, and still more the ornament, changes according to time, place, and fashion; hence these are very useful in tracing periods and cultures and their relations. Where different types meet there is usually more or less change or modification, often difficult to interpret. Our knowledge of European pottery is still small and unsatisfactory, but it has already been of much use in tracing migrations of culture and relations between provinces often widely separated. CHAPTER VIII NEOLITHIC CHRONOLOGY "We must imagine Europe in upper Paleolithic times again as a terminal region, a great peninsula toward which the human emigrants from the east and from the south came to mingle and to superpose their cultures. These races took the grand migration routes which had been followed by other waves of animal life before them; they were pressed upon from behind by the increasing populations from the east; they were attracted to western Europe as a fresh and wonderful game country, where food in the forests, in the meadows, and in the streams abounded in unparalleled profusion.... Between the retreating Alpine and Scandinavian glaciers Europe was freely open toward the eastern plains of the Danube, extending to central and southern Asia; on the north, however, along the Baltic, the climate was still too inclement for a wave of human migration, and there is no trace of man along these northern shores until the close of the Upper Paleolithic, nor of any residence of man in the Scandinavian peninsula until the great wave of Neolithic migration established itself in that region."[118] We must now attempt to determine the succession of these great changes in the climate and face of Europe, and then see if we can fix any dates for some of the changes and for the introduction of new cultures. In the oscillations of the ice-front marking the final retreat of the Alpine glaciers there were three epochs of advance. Two of these, the Bühl and Gschnitz advances, with the interval of retreat between them, were occupied by the Magdalenian or last epoch of Upper Paleolithic time. The third advance, the Daun Epoch, or perhaps the latter part of the Gschnitz and the first part of the Daun, is represented by the Azilian-Tardenoisian Epoch, a period of transition from Paleolithic to Neolithic time. These changes have been clearly traced by Osborn.[119]