The distribution of the sample also needs improvement. The dis- 
tribution of banded samples should correspond to the distribution 
of the continental waterfowl population in August and early September. 
Considering what is known of the distribution of mallards at that 
period, it is apparent in table l that the Prairie Provinces of 
Canada were very poorly represented by the banding in 1959 and 1960. 
In order to attain the objectives of the pre-season banding program, 
it will be necessary to band more mallards in Canada and more black 
ducks in eastern Canada. 
The accuracy of the age and sex determinations made at the time 
of banding also affects the reliability of pre-season banding infor- 
mation. It is known that immatures are more likely to be shot than 
adults. Since immatures had the higher direct recovery rate, it seems 
likely that most banders used reliable aging techniques. By late 
September some immature mallards and black ducks have molted their 
juvenile tail feathers. Therefore, all ducks with "adult" tail 
feathers must be aged by cloacal examination. 
Band recovery rates will not give reliable information about the 
population if the banded birds are subjected to an atypically high or 
low shooting pressure. Most pre-season banding in the United States 
takes place on National Wildlife Refuges. There are little data 
available to evaluate the representativeness of refuge banding sta- 
tions. In Minnesota, recovery rates from bandings at the Agassiz 
National Wildlife Refuge (formerly Mud Lake) and Thief Lake Refuge 
and Public Shooting Area offer an interesting comparison (table 1). 
The Agassiz Refuge consists of 60,000 acres and is entirely closed 
to hunting, while Thief Lake, approximately 10 miles away, contains 
18,000 acres of which a large portion is used for public hunting. 
The similarity in recovery rates suggests that the protection from 
hunting provided by the Agassiz Refuge does not exert a pronounced 
effect on the recovery rates of ducks banded there in August and 
September. 
The proportion of recovered waterfowl bands reported also is a 
variable that affects recovery rates. Organized band collecting 
near a pre-season banding station will distort recovery rates upward 
if banded birds remain near the trapping site until the hunting season 
opens. Since immatures usually are more vulnerable than adults near 
the banding site, relative recovery rates also will be affected by 
organized band collecting. In Oregon, the direct recovery rates of 
Sauvie Island and Summer Lake bandings (table 1) are unusually high 
in comparison to other locations. These high rates probably reflect 
heavy hunting pressure near the banding sites, but they also are 
influenced by band collecting to a degree that is difficult to measure. 
