82 SCIENCE PROGRESS 



older than theilower. Here, again, however, there are impor- 

 tant sources of possible error, for the river does not continuously 

 lower its bed, but sometimes raises or " aggrades " it, by 

 depositing upon it masses of gravel, sand, etc., which it can 

 no longer carry away ; and that this process is sometimes very 

 extensive and important may be judged from the fact that 

 the Thames has at some period excavated a channel which 

 extends at least ioo ft. below present sea-level, but is now 

 almost entirely filled with drift. It is evident, therefore, that 

 we must not too readily assume the deposit on a higher terrace 

 to be necessarily older than that on a lower one ; and in fact 

 the same terrace might conceivably contain drift of three dif- 

 ferent periods, belonging to (i) the period of first erosion ; (2) 

 a succeeding aggradation ; and (3) a final re-excavation ; and 

 if the alternation of erosion and aggradation has occurred 

 more than once, as is not improbable, the record may be still 

 further complicated. Seldom, however, will the older deposits 

 remain undisturbed during later movements, and it is therefore 

 possible that extensive fluctuations in level may sometimes 

 have occurred without leaving any definite trace behind. 



So far we have dealt only with true river deposits, but there 

 is another series of beds which play an important part in 

 palaeolithic history, and which for want of a better term we 

 may class together as " subaerial." Foremost among these 

 come the various forms of " hillwash " — muds, sands, and other 

 material washed down the hillsides by the direct action of rain 

 or by the melting of frozen soils ; while materials accumulated 

 under the influence of wind are probably of less frequent occur- 

 rence, though they sometimes attain considerable thickness. 

 It is evident that while true fluviatile beds show us the exact 

 level of the river at the time of their formation, these subaerial 

 deposits may extend to any height above it, spreading from 

 top to bottom of the valley slopes, so that their elevation above 

 the river is no indication of their relative age, and it is only 

 when the deposits of one period differ in character from those 

 of another (which is fortunately often the case) that we are 

 able to determine their succession. This distinction between 

 fluviatile and subaerial deposits is of considerable importance 

 in the study of palseoliths, but it must be admitted that it is 

 one that is often difficult to establish in practice ; and indeed 

 on theoretical grounds we should expect the two forms of deposit 



