
While the movement of the ducks was similar in both years, there 
are notable differences in when and where the ducks were harvested 
along the way. A greater portion of the birds moved farther and 
faster in 195h than in 1955, This apparently accounted for the dif- 
ference between the years in the proportion taken in Canada am the 
proportion taken in the Central as compared with the Yississippi 
Flyway, It appears, therefore, that distribution to the four flyways 
is not entirely predictable, However, it is evident that weather 
during the fall migration period is the controlling factor. Freezing 
weather arrived unusually early in the prairie portions of Canada and 
the United States during the fall of 195, while the fall of 1955 was 
extremely mild. Whereas the bulk of the waterfowl population leave 
the-Canadian prairies before mid-November in an average fall, large 
mumbers were still present during mid-December of 1955 and a fair 
munber actually wintered there. It would seem that the years 195) 
and 1955 have presented the extremes of weather conditions that are 
likely to be encountered and that during most years the differences 
in distribution may be less, 
. The fact that differences occur points to the necessity of ob- 
taining samples of banded birds within each of the important breeding 
areas for a series of years in order to arrive at an average expect= 
ancy of distribution. For example, if our banding had been done 
any in 1954, our measure of distribution would have been con- 
siderably in error during 1955 and somewhat in error during a year 
with average fall weather. An average of data from 195 and 1955 
would be less in error with regard to either an early or a late 
year phenologically, and might be quite close during a normal years 
Although it is difficult to assess at this point, it seems likely 
that banding data from or 5 years should be obtained from each of 
the important breeding areas before the data can be considered eerire 
factorye . 
