260 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. 
over the continental shelf, in summer, because, as the profiles show _ 
(p. 233), the tendency must have been just the reverse. And the very 
considerable difference in density between surface and deep water — 
off the coast must be a bar to upwelling, even though it may not 
absolutely prevent it, as it does in stratified waters where the layers 
of different densities are sharply defined (Sandstrom, 1908; Wedder- 
burn, 1908). 
It is not to be expected that our work could conclusively settle 
such a complex problem. But considering that the evidence of 
temperature, salinity, and density agree, and that it is hardly con- 
ceivable that one or other of them would not have revealed upwelling, 
it is safe to say that no widespread vertical movement of this sort 
was taking place over the continental shelf in July, 1913. And the 
fact that the cold water over the shelf south of Marthas Vineyard is 
usually separated from the cold water of the abyss by a zone of warmer 
bottom water in summer, suggests that this conclusion holds good 
for that season normally. It is true that during one summer, 1882, 
the whole shelf is known to have been bathed by cold water; but it 
is as likely that this resulted from an unusual accession of northern 
water or from abnormal winter cooling, as from upwelling. 
Upwelling may be more important in winter, for, as Kriimmel 
(1911) and Schott (1912) point out, the prevailing north and north- 
west winds, which often rise to storm strength, would have more 
tendency to produce this type of circulation, than the southwest, 
long-shore winds of summer. Furthermore, density is not so effee- 
tive a barrier to upwelling in winter as in summer, because its vertical 
range is much smaller then. Nevertheless, it is probable that upwell- 
ing caused by off shore winds would be from a comparatively shallow 
depth, say 100 to 200 fathoms, both because the direction of the wind 
is not constant but often reversed, and because the abyss water must 
be considerably heavier than coast water even in winter. And 
gravity would similarly resist any upwelling which the effect of the 
rotation of the earth might tend to produce along the inner edge of a 
current moving parallel to the coast. Upwelling of this latter type 
may play a very important part in the movements of ocean waters, 
as pointed out by Ekman (1905a) and recently by McEwen (1912); 
but until the movements of the bottom water of the North Atlantic 
are better understood, discussion of this theoretic aspect of the case 
may well be postponed. 
The real explanation of the low temperature of the coast waters 
is to be found neither in upwelling, nor in a northern current, but 
7. 
