Regional Uniformity 
Some preliminary explorations of the banding data for the 
Mississippi Valley involved comparisons of mallard mortality rates 
for specific years. An example given in table 60 shows how fairly 
close in agreement many of these sample mortality rates were. 
Figure 16 demonstrates how 5 out of the 6 mortality series in this 
table displayed a marked relative increase in hunting reports for 
the year 194-45. For the birds cited in table 60 this in the number 
shot represents an increase in mortality "rate" of the order of 20 
per cent over the previous year (that is, from 39 for 19h3-hh to 17 
per cent for 194-45). Although the comparison involves incomplete 
mortality data (in that additional mortality reports could be expected 
subsequent to the time of compilation), the mortality "rates" cal- 
culated for a given cohort are equally biased and general conclusions 
remain reliable. 
Mallard mortality rates were also compared for two banding 
operations carried out by the Fish and Wildlife Service on different 
refuges in the same state, North Dakota. In 1939 C. J. Henry and 
Merrill Hammond banded 918 mallards of all sexes and ages at the 
Lower Souris National Wildlife Refuge. The mean annual mortality 
rate for 233 of these subsequently reported was 56 per cent. In the 
same year S. H. Low banded 000 mallards at the Des Lacs National 
Wildlife Refuge; for 741 of these reported the mean annual mortality 
rate was 54 per cent. These means include both juvenile and adult 
birds shot the same season they were banded. The seasonal recovery 
rates for the two bandings differed; 15.3 per cent for Lower Souris, 
and 10.1 per cent for Des Lacs. Mortality rates calculated for the 
year 1939-40 were 56 per cent for the former, 55 per cent for the 
latter. 
As Lincoln (1939, p. 164) and Hawkins (199) have shown, 
the mallards in the Mississippi Valley are not an easily defined 
population unit. I finally assembled hunters! reports of this species 
for birds banded in the 3 prairie provinces and in the states actually 
bordering on the Mississippi River. The hunting covered from 1939-0 
to 1945-46 inclusive. In this aggregation, 05 Illinois adults banded 
up to August 31, 1939, had a mean annual mortality rate of 6 per cent, 
identical to that of 457 adults banded elsewhere in this region. For 
the year 1939-0, the Illinois sample gave an annual mortality rate of 
46 per cent, the non-Illinois sample 48 per cent. These were dynamic 
life-table calculations with 3 birds estimated as the mortality in 
1947-8 and 1 estimated for 198-9. 
On the basis of this evidence, it would seem that aggrega- 
tions of banding data for a species like the mallard display a 
geographic uniformity within a region and that they should be useful 
in analyzing changing mortality rates for specific years. 
139 
