nor political subdivisions have been used to describe the breeding 
ground origin of the birds; rather, the distribution patterns from 
various banding locations were examined and those with similar 
characteristics were combined into single units termed "areas of 
reference." This was done to examine the effect on distribution 
of such factors as weather, shooting regulations, and yearly vari¬ 
ations in migration. 
DATA INCLUDED 
Band recoveries used in this report include only those from 
wild ducklings captured in traps, by hand, or with the use of dogs, 
and banded in a normal manner. Recoveries were further limited to 
those of ducks that were shot or found dead between September 1 and 
February 29. Except for a few tables where yearly variation is 
examined, data from all years of banding are combined. Thus they 
reflect a more or less average distribution of the harvest of immature 
ducks in periods of both high and low populations and liberal and 
restrictive seasons and bag limits. It should be recognized, however, 
that the number of ducks banded was larger for many species during the 
period of high populations in the mid 1950*s. In some cases this pro¬ 
duces results more representative of a period of high populations and 
liberal regulations. Further, the distribution of banding effort did 
not agree particularly well with the distribution of the breeding popu¬ 
lation. Thus, many ducklings have been banded in the Prairie pothole 
region, but few or none have been banded in the important production 
areas of northern Canada (fig. 1). 
The number of recoveries of each species summarized in this 
report are as follows: 
Mallard 
9,272 
Gadwall 
981 
American widgeon 
1,282 
Green-winged teal 
74 
Blue-winged teal 
1,884 
Shoveler 
951 
Pintail 
1,859 
Redhead 
2,081 
Canvasback 
609 
Lesser scaup 
1,159 
It should be noted that these 
are the maximum numbers of recoveries 
used in any of the tables. Numbers used in individual tables may be 
less for various reasons. For example, both direct and indirect recov¬ 
eries were used in some of the mallard tables whereas only direct 
2 
