Because fluctuations in the relative number of recoveries between 
the Mississippi and Central Flyways were small as compared with the 
average proportion of total recoveries, and because variations resulted 
from unpredictable conditions that occurred during the hunting season, 
it appears that the average distribution of recoveries from a given 
reference area provides the most useful information for predicting 
the location of the kill. 
Sex and age variation 
Sex and age variation in the distribution of recoveries was 
examined using the large sample of recoveries from bandings in 
Saskatchewan. The distribution of direct recoveries from south¬ 
western Saskatchewan is illustrated in figure 10, and the distri¬ 
butions of indirect recoveries of males and females are presented 
in figures 18 and 19. The data illustrated in these figures are 
summarized in table 41. 
Important features of kill pattern that should be noted in 
figures 18 and 19 are (1) direct and indirect recoveries of both 
sexes cover essentially identical geographical areas, (2) except 
for the high incidence of direct recoveries near the banding station, 
the direct and indirect recoveries from bandings in Saskatchewan have 
similar distribution patterns, and (3) a highly developed homing 
tendency is indicated by indirect recoveries that are no more 
scattered from the area of banding than the direct recoveries of 
birds banded as locals. The tendency to return to the general 
region of origin is apparent among males as well as females and 
does not appear to have been changed by habitat deterioration during 
drought years. However, hunting pressure in regions north of the 
prairies is so low that analysis of band recoveries may have failed 
to show a northward displacement of mallards, if such did occur, 
during the drought in the prairie region. 
Although the distributions of recoveries of males and females 
were essentially similar, there were several small but statistically 
significant differences (table 41). For direct recoveries, there 
was a significant difference (p<.05) between the proportion of males 
(26.8 percent) and the proportion of females (31.1 percent) recovered 
in the reference area of banding. Because there was not a significant 
difference between the proportions of immature males and females 
taken in Canada (virtually all of which were in the general region of 
banding), and because the proportions of males and females taken in 
other flyways or latitudinal areas were similar, it might appear that 
the statistical difference is not realo However, data from south¬ 
eastern Alberta and southeastern Saskatchewan (tables 42 and 43) 
indicated a similar tendency for this difference to occur and it is 
necessary to conclude that the difference is real. 
19 
