566 APPENDIX. 



would find their way into such subglacial streams, and thus stones, and 

 even boulders of some size, might come to be carried often for long dis- 

 tances out of the route they had followed when imprisoned in the moraine 

 profonde. Indeed, it is quite possible that now and then portions of the 

 ice itself, charged with rock-fragments, might fall into subglacial waters 

 and travel many miles in the most contrary directions. We have now 

 only further to conceive some change taking place in the course of the 

 subglacial channels — caused probably by modifications in the ice-flow 

 overhead — when the bed of the subglacial river would be invaded and its 

 detritus become commingled with the unmodified drift or boulder-clay of 

 the moraine profonde. In this manner erratics from widely-separated dis- 

 tricts might occasionally become mixed up in one and the same subglacial 

 accumulation. 



But the most potent cause of all remains to be mentioned. It has 

 been shown that the direction pursued by the ice in the basin of the 

 Baltic was different at different periods. When the mer de glace reached 

 down to Saxony and Silesia, boulders and smaller stones, derived from 

 Sweden, were dragged over the bed of the Baltic, and carried south to the 

 farthest limits reached by the ice-flow. But when the mer de glace was on 

 the wane, and had melted away over a large part of Germany, the ice in 

 the Baltic basin followed the direction of that trough towards the south- 

 west. Now, as we have seen, the great mer de glace invaded the low 

 grounds of Germany three times at least. And these epochs of glaciation 

 were separated by long intervals of milder conditions, during which the 

 ice disappeared, and left the land to be reclothed and repeopled by plants 

 and animals. With each successive advance and retreat of the ice, there- 

 fore, fresh accumulations of boulder-clay would be formed, but we cannot 

 doubt that the moraine profonde would in many places consist to some 

 extent of the rearranged morainic materials which had been left behind 

 during the disappearance of each preceding mer de glace. The modifica- 

 tions which must have been brought about by this means are more than 

 sufficient to account for all the abnormal cases of " erratic distribution " 

 which have been referred to. Indeed, the wonder is that these are not 

 more numerous than they appear to be. It is highly probable that, when 

 the work of correlating the various boulder-clays of Germany and the 

 adjacent regions has been worked out in detail, it will be possible to map 

 out the area covered by the ice-sheet during each separate glacial epoch, 

 and not only so, but to determine approximately the prevailing directions 

 followed by the separate ice-flows. Already, indeed, it has been observed 

 that the upper and lower boulder-clays of one and the same place often 

 contain very different percentages of stones and boulders. Thus, accord- 

 ing to Harting, the upper boulder-clay in the island of Urk contains only 

 2*2 per cent of flints, while the lower dark-gray till shows 38'5 per cent. 



I have made special reference to these cases of abnormal distribution 1 

 of which I have been speaking, because they have been held by some as 



1 Mr. Helland has given an interesting and useful summary of a number of the 

 more remarkable examples in his very able paper descriptive of the glacial deposits 

 of the low grounds of Northern Europe (Zeitschr. der deutsch. geol. Ges., 1879, p. 

 63), to which, and to Dr. Penck's paper (Op. cit. p. 117), I would refer the reader 

 for a much fuller discussion of the subject than I can enter upon in this place. 



