Davis. J 
482 
[May iS, 
deposits would be for tbe most part local in time and place. 
The drainage from the ice would flow now here, now there, 
frequently changing by outburst to a new channel and leaving 
its former course nearly dry. The passages among the half 
separated ice masses at the margin of the decaying sheet would 
become more or less clogged with sand and gravels ; with fine 
sands where the currents were gentle ; with coarser materials 
along the paths of active streams ; and if we may interpret the 
accounts of the Arctic voyagers, such streams as rise from be- 
neath the ice and enter a marginal pond or lake, might be con- 
strained to build their deposits on an ascending slope. The 
impetuous character of some of the streams described in the 
Greenland ice would warrant our supposing that stones of a con- 
siderable size might thus be carried from lower to higher situa- 
tions. Streams flowing on the surface of the ice might in many 
cases melt and wear out channels, in which gravels could accu- 
mulate ; and these would later be dropped on the ground as the 
enclosing ice walls disappeared ; but in such case it might be ex- 
pected that the deposit would be somewhat scattered on one 
side and the other of the original channel. 
4. METHOD OF INVESTIGATION HERE ADOPTED. 
Thus far this sketch follows essentially deductive lines. This 
plan of presentation was chosen advisably, for a double reason. 
It is much easier to appreciate the force of what follows, if the 
general knowledge of the subject treated is summarized before 
the special subject under investigation is entered upon. It is 
furthermore essential that the investigator should not go unin- 
ormed into the field, but should equip himself as fully as pos- 
sible with a knowledge of what has been learned bearing on tbe 
subject of his studies. It must take new form in his understand- 
ing of it ; it must be carefully discussed, the facts of observation 
and their interpretation being sharply distinguished ; but if the 
field study is to be of the highest profit, as little as possible of 
what has already been done well by others should be uncon- 
sciously done over again. The sum of knowledge on the question 
in hand should be reviewed and revised ; but time should not be 
lost in its rediscovery. 
While the general sketch presented above is, on these pages, 
