in degree blocks; the median number was 0.8 and the mode 0.0. 
Because of the high frequency of degree blocks containing no 
observed black ducks and the occasional degree block with dis¬ 
proportionately high numbers, the median was used to divide the 
degree blocks into high and low density categories for examining 
distribution patterns. 
Since there are two alternatives in each degree block (above 
or below median), it is possible to state the amount of agreement 
between years among the matched degree blocks, to compare the result 
with the 50 percent agreement expected from chance, and to learn if 
there is a consistent distribution pattern upon which to base a 
stratified sampling plan. 
Comparison of the distribution pattern of 1955 with that of 
1956 showed an amount of agreement (67 percent) which exceeded 
the 99-percent confidence limits (34 percent to 66 percent). 
Comparisons of other possible combinations of years did not reveal 
significant agreement, as shown in the following table. Comparison 
of the 1963 and 1964 patterns involved identical sample segments, 
as did comparisons between the two earlier years, but the amount of 
agreement did not differ much from that expected by chance. 
Years compared Percent agreement 
1955 and 1956 
1955 and 1963 
1955 and 1964 
1956 and 1963 
1956 and 1964 
1963 and 1964 
67 
51 
47 
44 
44 
51 
Ten of the degree blocks were consistently above or below 
median for all 4 years. Nine such occurrences (one-eighth of the 
sample) are to be expected from chance. 
These comparisons indicate that degree blocks do not seem to 
provide a basis for stratification of the black duck survey, since 
the counts divided in this manner show inconsistencies between 
years. Although there was some agreement on a similar pattern in 
the surveys of 1955 and 1956, that pattern did not agree with the 
pattern of the two recent years of survey. Also, patterns in the 
two recent years of survey show little agreement with each other. 
Our inability to detect consistent differences in the density of 
black ducks may be due to (1) a relatively even dispersal of black 
ducks in the area, (2) a sample too small, in numbers of birds seen 
and intensity of coverage, to detect differences in density that 
4 
