lack of sexual dimorphism in the species. A flock may consist of 
unsegregated pairs or of postbreeding drakes. Both are important 
in indicating the size of the population. 
The ideal aerial count of black ducks would sample the entire 
area at a specific phase of the nesting cycle. Since this is not 
feasible, an alternative may be to select samples with similar 
frequencies of groupings (singles, pairs, flocks, etc.) for com¬ 
parison under the hypothesis that these samples are likely to be 
most similar in nesting phenolog^. 
Stratifying for Phenological Differences 
The area of Canada that was sampled consistently for 4 years 
was divided into southern, central, and northern zones (fig. 4). 
This arrangement is most likely to delineate phenological units, 
since in the south the nesting season may be well advanced while 
nesting has not begun in the north. Although the aerial survey 
begins in the south and progresses northward, it very likely 
occurs relatively late phenologically in the south and becomes 
progressively earlier phenologically as it proceeds northward. 
The total number of black ducks and the frequency per segment 
of the black duck groupings discussed above were ascertained from 
each year's records for each zone. The values obtained are shown 
in table 1. 
Frequencies obtained for each grouping were ranked as high, 
medium, or low, in relation to the mean frequency for the zone 
based on a 4-year average. Values falling within the range of 
the mean plus or minus 2 standard errors were considered medium. 
Values outside that range were considered highs and lows. 
The number of high and low values obtained (31 percent) is 
greater than the 5-percent figure which usually is expected to exceed 
2 standard errors. This indicates that these data depart greatly 
from a normal distribution. The object in using 2 standard errors, 
however, was not to describe variability but to apply a consistent, 
mechanical method for dividing the data into classes. 
A quantitative estimate of agreement or disagreement in 
characteristics between any two zone-year groups was obtained by 
assigning a value of 2 when items agreed (high-high, low-low, 
medium-medium), a value of 1 for partial agreement (high-medium or 
low-medium), and no value for complete disagreement (high-low). 
7 
