Limitations of the Data 
In attempting to compute mortality rates for specific years 
(like the year 1939-0), one mst realize that we are no longer study- 
ing the "theoretical" populations of Part II; all of these were banded 
over a period of years during which annually fluctuating mortality 
factors would appear as average environmental influences. The graphs 
showing recovery rates in Chapter 12 demonstrate that hunting mortality 
among mallards varies importantly from one year to another. When this 
information from hunters is either the only or the principal source of 
original data in a dynamic life table, how reliable are the annual 
mortality rates that are calculated for specific years? 
This pertinent question is explored in table 61 by calcula- 
ting what would happen to a cohort of 1000 hypothetical waterfowl 
when hunters kill anywhere from 15 to 35 per cent in different years, 
Table 61.~-Effects of Wild Hunting-pressure Fluctuations on 
Mortality Rate Calculations of a Single Hypothetical 
Cohort of Waterfowl 
(These birds were assumed to have been banded in the summer of 1939; 
and a dynamic life-table analysis based on the shot sample is used 
below to calculate mortality rates for successive years each begin- 
ning September 1. Only a part of the original calculations are 
shown here.) 
ears = = = = = 
Alive at Start 1000 600 282 180 108 52 
Per Cent Shot 20 35 15 20 35 25 
No. Shot 200 210 2 36 38 13 
Nonhunting Deaths (25%) 200 98 60 36 18 10 
Total Deaths 00 308 102 72 55 23 
Annual Mortality Rates (%) 
Actual 0 cl 36 Ke) Sl Lh 
Life Table 36 59 31 33 5h 2 

Under the rigorous conditions thus laid down, it is clear that life- 
table mortality rates for specific years are of extremely limited 
reliability. For the most part, they do seem to reflect actual 
changes upward or downward, but these changes may be importantly 
exagcerated. It is only when recovery rates are relatively constant 
or when a series of cohorts is grouped together in a life-table that 
we can expect approximations of mortality rates actually occurring 
in wild populations. 
General Trend: 1929 to 1938 
It is evident in table 57 at the start of this chapter that 
the annual mortality rate of 64 per cent for age interval x + 1 to_x + 2 
1y2 
