In attempting to determine the date of the beginning of the Neolithic period we may begin with a remote point of departure for comparison and select the Bühl stage and the beginning of the Magdalenian Epoch. Nuesch made a careful estimate from the deposits at Schweizersbild near Schaffhausen, Switzerland. His method of estimating is described fully by Obermaier.[125] He places the beginning of the Neolithic deposits here at 6000 B. C., and considers 20,000 years as a fair estimate for the time elapsed since the first occupation of this locality by Magdalenian hunters at some time during the Bühl Epoch. Obermaier, summing up the evidence, concludes that the beginning of the Magdalenian Epoch could not have been later than 16,000-18,000 B. C., and that it ended not far from 12,000 B. C. Osborn says: "Bühl moraines in Lake Lucerne are estimated as having been deposited between 16,000 and 24,000 years B. C." He also appears to place the Maglemose culture at about 7000 B. C.[126] We may now turn to the great Scandinavian ice-sheet, whose retreat may have begun somewhat later and proceeded more slowly on account of its more northerly position. Here De Geer has made a report based on a very careful study of the annual layers of deposition formed during the glacial retreat. We have already seen that the material brought down by the spring freshets differs in color and texture from that of late summer and autumn. Hence these annual layers are almost as distinct and as easily counted as the rings in the trunk of a tree. This method promises great accuracy of results, and the thickness and character of the layers and their included organic remains throw much light on the climatic and other conditions under which they were laid down. But even here the length of certain periods of halt in the glacial retreat can be only very roughly approximated. The number of annual layers of deposit in the Swedish Lake Ragunda lately drained shows the number of years since the lake was uncovered almost at the end of the retreat of the Scandinavian ice. Says Sollas: "The Ancylus Lake was in existence at a time when the ice had very nearly, though not quite, accomplished its full retreat, _i. e._, a little more than 7,000 years ago (the length of post-glacial time); and Baron de Geer, although he has not yet been able to bring the beach of the lake into connection with his system of measurements, thinks, as he has kindly informed me, that its probable date may be 7,500 years counting from the present."[127] Menzel, in a chart embodying the results of his study of De Geer's work, places the beginning of the retreat of the ice in Germany at 21,000 B. C., the maximum of the Littorina depression and epoch of kitchen-middens at 6000 B. C., full Neolithic at 4500 B. C., beginning of Bronze period 1700 B. C.[128]