Banding Data 
Effect of capture methods 
As mentioned in the introduction, this report includes ducklings 
caught either in traps (by bait or drive methods) or by dogs and 
banded in a normal manner. When the data from all sources were 
summarized, ducklings caught by dogs and assumed to be uninjured 
had direct recovery rates much lower than those of birds captured 
in traps. Hie low recovery rate associated with dog-caught birds 
suggests that serious injuries were frequently undetected and that 
these injuries resulted in a mortality of approximately 20 percent. 
Such mortality probably occurred soon after banding. 
Because the location of bandings may affect recovery rates, it 
was necessary to determine whether this cause, rather than method 
of capture, had produced the differences described above. Recovery 
rates of birds caught by dogs are compared with those of trapped 
birds from the same Province in table 20. Even though recovery 
rates varied geographically, those for dog-caught birds varied in 
the same direction and magnitude as trapped birds, but in each case 
with lower recovery rates. This was true in all cases except for 
the State of Washington where recovery rates for dog-caught birds 
were higher. However, this is likely explained by small sample size. 
Although it was definite, that use of dogs resulted in higher immediate 
mortality than use of trapping methods, it was concluded that deter¬ 
minations of geographical distribution would not be affected by 
method of capture. Thus both dog-caught and trap-caught birds were 
used collectively when determining recovery distributions. 
Geographical distribution 
The geographical distribution of banding effort reflects both 
the relative distribution and magnitude of breeding populations and 
the accessibility of these populations to banders (fig. 4). Thus, 
north to about 54° of latitude, bandings may roughly indicate the 
distribution of the breeding population. Beyond this, relatively 
few bandings are available for large and important populations. The 
relation of the size of the breeding population to banding intensity 
as indicated by recoveries is summarized in table 21. It may be 
noted that areas including Alaska, Yukon, Northwest Territories, and 
all other regions north of 54° of latitude support about 35 percent 
of the breeding population, but are represented by only 1.3 percent 
of the total recoveries. Although this difference is inflated by 
use of population data gathered in 1961 when prairie populations 
were low, there can be no doubt that northern populations are inade¬ 
quately represented in our sample. 
12 
