5i8 T. C. CHAM BERLIN 



When this datum point is chosen there arise at once special 

 difficulties in determining what were the landward and bottom 

 boundaries at the start, for even the landward edge at the surface 

 is largely concealed by later deposits, or has been cut away, while 

 its lower parts and the bottom are generally inaccessible except 

 as diastrophism has forced them into view. In few cases — prob- 

 ably in no case — -has a terrace shelf throughout its history been built 

 oceanward solely. In most cases the sea has advanced upon the 

 land and the topset deposits have crept out upon the continent as 

 well as oceanward. The landward advance has in many cases been 

 interrupted by diastrophic movements and the shore line forced 

 seaward, sometimes even beyond the initial line. Later a new 

 advance has taken place from this line of retreat and so the oscilla- 

 tion has gone on. As a result of this the innermost edge of sedi- 

 mentary deposits is liable to lie farther landward than the line from 

 which growth began at the opening of the epoch chosen. The 

 border sediments that thus lap upon the continent — -defined as it 

 stood at the outset — must of course be taken into account but not 

 as extensions of the continent; they are rather superficial replace- 

 ments or over-placements and, so far as determinable, are usually 

 limited in depth and relatively small in quantitative value. 



When pre-Paleozoic formations come out to the present coast- 

 line the location of the inner border of the terrace formed 'after the 

 initial epoch is little better than conjectural. Some suggestions of 

 value may spring from the trend of the border line in adjacent regions 

 and from the consistency of value of any assigned configuration. 



Probably the best approximation to the initial line of growth 

 may be reached by setting aside for compensation purposes all 

 sediments that are known to lap upon the continent as it stood at 

 the outset, while such border sediments as are known to be deep 

 and to show no signs of overlap are regarded as belonging to the 

 continental outgrowth. In this it is assumed that any over- 

 estimates in the volume of the basal parts will be made good by the 

 sediments of the overlap set aside for this purpose. It will be 

 shown presently why the basal parts are not as likely to lead to 

 overestimates as their apparent uncertainty may lead one to think. 

 They are not wholly uncertain. 



