in 1931. The data for this and the following year represent the 
smallest samples in the series here considered (156 and 12 respec- 
tively), and the precise effect of this drastic change in regula- 
tions must await study of a larger sample. The next important drop 
in mortality rate occurred when a staggered 30-day season was changed 
to a consecutive 30-day season. It would appear that the staggered 
30eday season of 193 was as lethal to these waterfowl as the 60- 
day season in 1933. 
Western Montana birds.e--As a check on the high mortality 
rates calculated from the British Columbia and Oregon bandings, an 
analysis was made of male mallards banded in the years 1926-29 by 
F. H. Rose in western Montana. This banding station has been 
included by Lincoln (1939, pp. 17-182) in his Pacific flyway. The 
mortality-rate comparison here contemplated should represent a new 
and rather critical test of the flyway as a population concept. 
Seven hundred and sixteen adult males banded in some year previous 
to the one in which they were shot had mortality rates of 63, 6h, 
and 69 per cent for the years 1928, 1929, and 1930 respectively. 
Mortality data for 1655 males, which included juveniles and birds 
shot the same season as banded, yielded mortality rates of 62 per 
cent in 1927, 61 per cent in 1928, 60 per cent in 1929, and 69 per 
cent in 1930. My time did not permit a further expansion of these 
samples to include females; but as table 59 demonstrated earlier in 
this chapter, this is not a serious omission. The mean mortality 
rate for all males examined (62 per cent) agreed well with the 60 
per cent calculated in table 57 for Pacific coast males. The 
Pacific flyway as a population concept appears to have withstood 
this test well. An even more important test involves the compari- 
son of mortality rates between flyway populations. 
Mortality Rates in the Mississippi Valley 
Illinois and Missouri birds in the 1920's,.—-Adult mallard 
samples for both sexes were analyzed for birds banded by F. C. Lincoln 
in Illinois and John Broeker and L. V. Walton in Missouri from 1922 
to 1926 inclusive. If band loss was frequently occurring in that 
early period, the calculated mortality rates should exceed those 
derived from more recent bandings. No such loss was evident. The 
rates computed for both sexes (table 63) were consistently lower 
than those for the males on the Pacific coast. The differences 
may or may not be attributed to my preoccupation with male birds in 
the more western samples. As a check on the fluctuations, I com 
pared these Mississippi Valley mallard mortality rates with some 
first-season recovery rates reported by Lincoln (1930) for all 
species of waterfowl on the continent (figure 19). The agreement 
was sufficiently close to dispell any possibility that the rates 
are grossly inaccurate. I believe we must conclude that important 
fluctuations in the annual mortality rates of these banded adult 
birds took place during this period when hunting regulations were 
relatively stabilized and some other environmental factor was 
operating on these samples of the population. It seems to. me 
116 
