highly unlikely that the difference in mortality rates between the 
Pacific coast and Mississippi Valley birds is solely the result of 
sexual differences in the samples. The evidence points to different 
rates for the Pacific and Mississippi Flyway mallard samples examined 
ae a proof on this point was not gathered in the present , 
Study. 
; Illinois birds in the 1930's.--An interesting comparison 
with the 1920's Is afforded by the banding of mliaedocey Ce J. Goetz 
in Fulton County, Illinois, from 1933 to 1939 inclusive (table 6h). 
A high overall mortality rate for the year 193-35, when a split 
season governed hunting, may be due in part to chance. No marked 
fluctuations in mortality rates are evident from 1935 to 190 inclu- 
sive. The mean annual mortality rate for 835 birds reported shot up 
to mid-19l,7 was 4.7 per cent in contrast to 6.5 per cent for 1185 
banded in Illinois and Missouri during the 1920's, 
Fluctuations in mortality reports in the 19)0's.--By com 
bining hunters’ reports of birds panded tn-saskatchswan, \ianitoba, 
Minnesota, Wisconsin, Missouri, Tllinois, Arkansas, and Louisiana 
(table 65), a preliminary picture can be obtained of fluctuations 
in adult mallard mortality rates for years (beginning on September 1) 
from 1939 to 1946. During this study the events were so recent that 
dynamic lifetable compilations were impossible, but a time-specific 
analysis should at least indicate the approximate nature of changes 
taking place in the samples. The basic data assembled are given 
in table 65, where it is evident that the survival-—index values 
obtained with cumulatively totaled cohorts (part B of the table) are 
slightly higher than those obtained with single cohorts (part A of 
the table). This is another manifestation of a phenomenon noticed 
earlier in this chapter: increasing age in adult mallards makes 
for a decreasing mortality rate. 
In figure 20 it is further evident that the survival-index 
values are inversely correlated with first-season recovery rates 
.for mallards banded in Canada and northeastern Illinois. This does 
not mean that the index is an accurate substitute for survival rate, 
but it gives some assurance that the two statistics are measuring 
the same phenomenon. The time-specific mortality rates in table 65 
average l7~8 per cent per year, in contrast to 3-7 per cent for 
ynaged birds banded by the Illinois Natural History Survey during 
this same period (see table 58). 
On the whole, it seems to me that seasonal recovery rates 
emerge from this discussion as being more reliable and more sensitive 
indicators of the effect of hunting regulations than survival indices 
computed by the time-specific method. There can be no doubt that 
important fluctuations in adult mallard mortality rates exist from 
one year to another, that they are measurable, and that they often 
reflect differences in hunting. 
149 
