PHYSICAL OCEANOGBAPHY OP THE GULF OF MAINE 
867 
In general, the dominant set has been found most rapid in the region of Cape 
Cod and Nantucket Shoals, averaging about 8 miles daily. The average velocity 
(about 7 miles per 24 hours) is nearly as great for the stations along the west coast 
of Nova Scotia and in the Bay of Fundy; but the resultant set into this side of the 
gulf is not so rapid, because most of the stations show components either to the west 
or to the east. Perhaps 5 miles per day approximates the rate at which a bottle 
might be expected to drift northward along Nova Scotia by this evidence. 
EXPERIMENTS WITH DRIFT BOTTLES 
Measurements with the current meter, such as have just been discussed, give 
both the direction and the rate of the dominant set, as well as of the tidal currents, 
at that particular place and time, assuming always that the observations are taken 
at frequent enough intervals and extended over a sufficient period of time. 
The setting free and recovery of a drift bottle can never yield information so 
definite, because only the two end points of its journey are known, the route it has 
traveled from the one to the other always remaining a matter for deduction. Our 
drift bottles, furthermore, reflect the dominant movement of the uppermost stratum of 
water only; a fathom or two deep, at most, for the bottles with the longest drags. 
Neither does the drift of a bottle necessarily reproduce the drift that would have been 
followed by a particle of water, because the bottle floats on the surface, while the 
water may sink to lower levels by vertical currents, while new water may well up to 
the surface from below to take its place. 
Because only the end points of the drifts are known and the intervening tracks 
can only be assumed, their value depends on a number of factors, especially on their 
consistency, one with another; the length of time they are adrift; the extent of the 
oceanic area covered; and on general information from other sources as to the local 
currents. In all these respects the Gulf of Maine has proved an especially favorable 
region for the study of the dominant circulation by the drift-bottle method. Since 
all the drifts from all the lines set out have, without exception, proved reducible to 
one scheme, entirely consistent with the current measurements (p. 866) and with gen- 
eral report as to the dominant set along various parts of the coast, with temperature 
and salinity, with the distribution of the plankton, and with the internal hydrostatic 
forces (p. 936), I believe they may be taken as representing the main features usually 
prevailing in spring, late summer, and early autumn. 
The greater the time interval between release and recovery, the greater does 
the uncertainty become, because the longer the bottle is afloat, the greater distance 
it may have covered in its journey — i. e., the farther its track is apt to have diverged 
from the direct point to point line. By this same reasoning, when bottles are 
released in numbers the time interval becomes an important factor in deducing their 
probable tracks. If, for example, bottles released near Cape Elizabeth were to drift 
repeatedly to a point in Nova Scotia in as short a period as bottles released at 
Mount Desert, it is a fair assumption that the latter have diverged enough from the 
direct route to make their journey approximately as long as that of the former, 
assuming, of course, an approximately equal rate of drift for both. I should also 
