the present time, it seems likely that the survival rates calculated 
here will refer-—-for the most part-~to nonmigrants. 
Abridged Life Table 
An abridged life table for barn owls is given in table 9, 
where a mortality series is computed as the number reported dead per 
1000 banded birds available for each age interval. The number of 
banded birds was ascertained by taking the recovery cards in the 
Fish and Wildlife Service files and tabulating the numbers that the 
successful cooperators actually banded. The number banded by unsuc- 
cessful banders could not be ascertained without a prolonged and 
exhausting survey. As a check on possible discrepancies introduced 
by this method, a second life table was constructed using Lack's 
(1943a) method on birds banded up to 1935. For 55 birds found dead 
or reported shot, the mean mortality rate was 2 per cent per year 
(in contrast to 0 per cent in table 49). On the theory that cap- 
tured barn owls are equivalent to dead ones, another table involving 
80 recoveries of all kinds was constructed using Lack's method on 
birds banded to 1935. This again yielded a mean mortality rate of 
42 per cent for all age intervals. Both these tables (which are not 
shomm here) had the same first-year mortality rate (51 per cent) and 
nearly identical mean adult mortality rates (35 and 37 per cent per 
year respectively). 
The adult mortality rate in table 9 was next checked by 
computing the mean mortality rate for 33 birds banded as adults up 
to 1925. Using recoveries of all types but excluding birds recovered 
within one week after banding, this new calculation yielded a mean 
rate of 3h per cent per year. Excluding 1 adults that were recovered 
within a year after banding dropped the sample to 19 birds and the 
mean mortality rate to 30 per cent per year. 
Age Ratios and Productivity 
It seems certain that Barn Owls breed when one year old. 
The 172 adults in table 49 (colum 1',) could keep the population 
stable if each pair raised 1.33 young per year. This is so far 
below the .2 young observed in successful nests by Gallup (199) 
in California that the discrepancy calls for some explanation. 
Several possibilities require comment: 
(1) That sampling errors occur in the life table. It is certain that 
these exist but I fail to see how they could exaggerate and dis- 
tort the age ratios and required productivity to the extent 
described here. 
(2) That the construction of a life table (mo. 49), starting as of 
the date of banding, is false. A review of table-l7 will con- 
vince one that there is at least no hiatus of banding recoveries 
in the weeks or months following the banding date. There are 
110 
