have made changes on the face of the land, although the first-time 
visitor still is impressed by the vast areas in which no signs of 
man's activity are apparent. In spite of the improvements, there 
are still several locations where it is necessary to stockpile fuel 
and oil for next year's survey. 
Stratifying Samples According to Black Duck Densities 
/ 
Identification of discrete areas having either high or low 
densities of black ducks would allow stratification of aerial 
sampling. Such treatment would strengthen the basis for population 
estimates. Previously, those who examined aerial survey data from 
eastern Canada referred to five strata based on forest types 
(fig. 1). Our recent analysis of the data indicated that it was 
not possible to demonstrate statistically significant differences 
in black duck density among these strata. 
Since we believed that forest-type strata might be too large 
to show population differences, we divided the surveyed area into 
15 units based upon both topography and forest types. Units not 
only showed great differences in black duck densities for any given 
year, but also varied among years to such an extent that they 
appeared to be of no value as stratified samples. Further, using 
the 15 units, we found no correlation between the quantitative 
results of the production survey and those of the breeding population 
survey. In other words, the two surveys did not appear to be 
sampling the same population of black ducks when the data were 
examined in this manner. 
Degree blocks of latitude and longitude were used as areal 
units for examining the distribution of observed black ducks. The 
data for degree blocks were summarized as ducks per 18-mile segment 
of transect, since the number of transect segments per block was 
not uniform. This plan permitted comparing data for all 4 years 
even though transects in 1955 and 1956 were at different locations 
from those in 1963 and 1964. 
A degree block in eastern Canada contains approximately 3,000 
square miles. Transects were spaced about 1 degree apart and 
averaged about 2.5 segments per degree block. For this reason, an 
analysis of degree blocks as reference units departs little from 
an analysis by 18-mile segments. 
A frequency graph (fig. 2) of black ducks per segment in the 
312 degree blocks sampled over a 4-year period shows a distribution 
strongly skewed toward zero. Black ducks per segment averaged 1.3 
3 
