

cripples has no effect on the calculation of mortality rate so long as 
this inability is constant; it merely reduces the size of the sample 
available for study. 
(c) The uncooperativeness of hunters who refuse to report the banded 
birds that they shoot likewise has no effect on the computation, of 
mortality rate, but this uncooperativeness is here considered as being 
chronologically constant. 
(d) Badly worn bands lacking the name of the Fish and Wildlife Sarvice 
will obviously lower the number of older birds reported. This distortion 
will affect mortality-rate calculations for all ages in the dynamic method 
of analysis, but in the time-specific calculations it will affect only 
the later stages in the life table. 
(e) In practice, birds are never reported and tabulated as fractions, 
Their deaths occur as discontimuous variables which inevitably produce 
some distortion in analyses. This distortion is minimized in the time- 
specific analysis where it is most noticeable in the later stages of 
the table. In the dynamic analysis, the distortion is carried forward 
into the earlier stages as well; it can be recognized as a pronounced 
tendency for the mortality rate to creep toward 100 per cent as one goes 
from year O-1 to the end of the table. 
Another hypothetical population is illustrated in figure h. 
Here three successive mortality factors are at work: 10 per cent 
initially due to disease and accidents, 60 per cent of the survivors 
taken by hunting, 15 per cent of the remaining population dying from 
predation and other factors throughout the year. These three death 
curves all decrease at a rate of 69.) per cent annually. 
Can mortality and survival reports of banded birds be grouped 
together in constructing a life table? It should be obvious from table 
7 that ot only does each sample of mortality drop 70 per cent per year, 
but the various samples of living birds (those alive at the start of 
each year, those left at the end of the year) all decrease 70 per cent 
per year, too. Thus, if banding recoveries contain reports of birds 
dead as wellas those alive in the same proportion year after year, 
these reports can be safely zrouped in calculations of a constant 
mortality rate. The effects of combining such data are further shown 
in table 8. 
Where mortality factors operate at constant levels, we cannot 
escape the conclusion that a time-specific analysis based on a single 
source of mortality may yield reliable estimates of the overall annual 
mortality rate. Since, under the same condition randomized samples 
of the living population likewise give the same result, they can be 
grouped with mortality data to form a survival series for calculations 
of mortality rate. Thus, the Dutch investigators-—-Kraak, Rinkel and 
Hoogerheide (1910) and L. Tinbergen (19)6)--were justified in grouping 
together all types of recovery reports on banded birds into survival 
series and analyzing these on a time-specific basis. As far as I 
can determine, the recoveries that they used were random samples of 
bird populations reported by the general public to a central agency. 
33 
