
USES AND LIMITATIONS OF THE DATA 
Pre-hunting Season Banding 
The banding of an adequate sample of immature and adult wood ducks 
in late summer and early fall, before the opening of the hunting season, 
could provide data for estimating mortality rates, for measuring the 
importance of hunting as a mortality factor, and determining the effects 
of regulations on the kill. Estimates also could be made of the size 
of the pre-hunting season population. Since this is the kind of infor- 
mation mosturgently needed, pre-season banding is considered the most 
important. 
Most of the banding in the present program was pre-season banding, 
but even so, the samples were much smaller and more poorly distributed 
than they should be for reliable results. Recovery rates are shown in 
table 2. The sample sizes for most of the States appearing in the table 
were too small to yield very precise estimates of recovery rates. If 
SO recoveries are obtained, the sampling error of the estimated recoveyy 
rate is about 30 percent at the 95 percent confidence level. Only band- 
ing of immatures in Wisconsin, Illinois,and Vermont, in 1960, produced 
more than 50 direct recoveries to represent each State of banding. When 
band recoveries are distributed among the various harvest areas, the 
reliability of the recovery rate estimate in each area decreases further, 
For reliable data, there should be enough banding in each State to yield 
at least 50 recoveries for each age group. 
_ Perhaps of even greater importance than sample size is the repre- 
sentativeness of the banded sample. Wood ducks were sampled inadequately 
in almost the entire southern United States. Also, portions of the 
northern area are poorly represented. It is particularly important to 
band pre-season samples of wood ducks in the southern States that are 
important harvest areas, such as Texas, Mississippi, Alabama, North 
Carolina, South Carolina, Georgia, and Florida. Louisiana was the only 
southern State where more than a small number of adult and immature wood 
ducks were banded. 
Breeding-qround Banding 
In order to relate production areas to harvest areas it is necessary 
to band birds on the breeding areae The banded birds must be either 
flightless young (locals) or adult nesting birds to be sure that they 
did not come from some other area. During 1959-60, too few locals were 
banded in any State to yield an accurate measure of the direct recovery 
rate or the distribution of the hunting kill (table 2). For example, 
in 1959-60 a total of 676 locals were banded in Ohio and only 16 were 
recovered. Thus, not only were the banded samples small, but the recovery 
