50 



constitution, as several different kinds of rock will not generally 

 be found assembled in one small area. This is also partly the 

 case with this sample No. 29, as will be seen later, for, out of 

 29 rock-fragments over 4™™, this has 23 consisting of granite 

 or gneiss, the remaining ones being sedimentary. We then 

 arrive at the result that the greater part of the granitic material 

 has its origin in the sea-floor, while the rest together with 

 the sediments originates from icebergs or moraine-formations. 

 Even supposing this last-named material be covered over by 

 later deposits, wherever it is situated in a somewhat flat place, 

 yet, wherever it was originally deposited in a somewhat more 

 undulating tract, it may still project in some places and con- 

 tribute its share to the constitution of the samples. 



Sample No. 30 deviates in its constitution greatly from 

 the surrounding ones, as it is almost devoid of coarser ingre- 

 dients over 0-5и1°1, and has very few ingredients between 0'05 

 and 0'5™™. This peculiarity is directly connected with its 

 occurrence at a greater depth viz. 241 Danish fathoms. Owing 

 to this unusually large quantities of finer material are depo- 

 sited, the conditions for such deposition being less good in 

 places at a higher level, where the deep-seated currents some- 

 times may be very strong. Hence what may origiuaUy have 

 existed of morainic deposits is covered, while those derived 

 from icebergs are so slight compared with the amount of finer 

 material that they cannot be noticed in the sample. For the 

 sample to be so fine as is the case here, it is also necessary 

 that no solid rocks should exist in the neighbourhood, as the 

 presence of such will always bring about the deposition of 

 coarse ingredients in large quantities. The very regular curve 

 shows that the deposit or at any rate the greater part of it, 

 must have originated from one single source; the greater part 

 of it was evidently derived directly from land. 



The samples No. 31 and 32 are of a quite different appear- 

 ance. Compared with the last-named sample they are of much 



