Lincoln's data shows that, although his mean rate was 11.9) per cent, 
the extremes for specific years varied from 10.68 to 15.16 per cent. 
Lincoln has quite properly emphasized that the proximity of a banding 
station to an area of intensive hunting must be an important variable 
in the recovery rates for waterfowl banded in different regions, and 
that these rates undoubtedly vary according to that part of the hunt- 
ing season in which the birds were banded. Recognition of the addi- 
tional possibility that variations in first-season waterfowl recovery 
rates arise from annual differences in hunting pressure has had to 
await the accumulation of more banding data. 
Banding index of hunting ssure.--Bellrose (19h, p. 369= 
370) has used first~-season recovery rates as an index of the cumula- 
tive effect of federal regulations on the continental take of ducks. 
For black ducks in Michigan these rates dropped from 20.5 per cent 
in 1928-3 to 9.2 per cent for 1939-2. He also contrasts a rate of 
about 13 per cent for all waterfowl quoted by Phillips and Lincoln 
(1930) to a rate of from 5.6 to 7.5 per cent for Canadian waterfowl 
in 190-192 reported by B. W. Cartwright of Ducks Unlimited (Canada). 
These changes in seasonal recovery rate led Bellrose to conclude that 
regulations in the past ten years cut about in half the rate of kill 
mate by hunters in the previous decade. 
There are at least two possible criticisms of this method, 
neither of which is necessarily valid but both of which encourage 
skepticism: (1) We know nothing of chronological differences in the 
percentage of hunter cooperation in reporting bands. When hunters 
complain about strict regulations, are they less apt to report bands 
to the federal agency responsible for the regulations? (2) Are 
recovery rates for different banding stations so similar that these 
comparisons are valid? In 1939, C. J. Henry and Merrill Hammond 
banded 918 mallards at the Lower Souris National Wildlife Refuge, 
obtaining a first-season recovery rate of 15.3 per cent. During the 
same year S. H. Low banded },000 mallards at the Des Lacs National 
Wildlife Refuge in the same state and got a seasonal recovery rate 
of 10.1 per cent. This difference could of course represent varia- 
tion arising from differences in the date of banding, a possibility 
that I did not attempt to explore. Although recovery rates have yet 
to be correlated with mortality rates, Bellrose's (1945) use of a 
recovery-rate index immediately impresses one as an original 
intellectual contribution that has great potentialities in the study 
of waterfowl populations. Some of these potentialities are explored 
in the two sections that follow. 
Regional Agreement in Seasonal Recovery Rates 
_ Bellrose and Chase (1950) have made the interesting point 
that (adjusted) seasonal recovery rates have annually varied in a 
somewhat similar manner for two banding stations about 175 miles 
apart in Illinois. In table 55, I have summarized some first-season 
recovery rates for banding stations somewhat farther apart. No 
126 
