
Comparison of Mid-winter and Christmas Season Counts 
We can make a further test by plotting the figures obtained through 
the mid-tinter count, together with their estimated error, to the same 
scale as corresponding figures from the Christmas season counts. 
rizures 1 and 2 show the bands within which the actual ponulation index 
would be expected to lic according to the two separate methods of esti- 
mation. The median line in Figure 1 connects the total snipe per narty- 
hour for each year from 1952 through 1956. The vertical limits of the 
band represent 20 nercent deviations (see Table l:) from this line. 
Similarly, the median line in Figure 2 represents number of snipe per 
10C party-hours of Christmas Count coverage, increased by a constant 
percentage (7.2) to make the average figure by this method coincide 
with the average by the other method. The width of the band indicates 
a deviation of 15% from the center line (sec Table l:). Since the width 
of both bands is based on a percentage, the width is greater for high 
values of the center line than for low values. 
In Figure 3 the bands based on the two types of count are super- 
imposed. If the Counts by the.two methods had been sinilarly distrib- 
uted an? if there had been no change in populations between late 
December and late January, We would have expected some overlapping 
area in each of the five years. The actual index figure of snipe 
abundance for each year would lie somewhere within this overlapping 
arca. It will be noted that there is fairly close agreement between 
the Mid-winter Count band and the Christmas Season Count band and 
that a downward trend in a viven year by one method is balanced by an 
even or an uwpyard trend by the other method. In other words, the 
snipe population has actually been more stable from year to year than 
has been indicated by any one mthod of estimating vopulation trends. 
It is clear from Figure 3 that the continental snipe ponulation has 
changed little if any during the five-year veriod. 
Figure lh shows the average (mean) abundance index for each year, 
as obtained from the two series of Counts. Although the trend aopears 
to be downward, it is still within the margin of srror demonstrated 
in Tigure 3. There is no indication that re-opening the shooting 
season has had any measureable effect on the snipe population. 
Suggestions for Future Coverage 
Between 30 percent and hO percent of the winter snipe counts can- 
not be used for year-to-yzar statistical comparison because they either: 
a) are not taken in the same areas for two or more successive years; or 
b) are taken entirely by car or by boat instead of on foot, and so can- 
not be used in statistical comparisons which are based on number of 
party-hours of foot coverage. fs shown in Table 4, a total of 9h Mid- 
winter counts are required to enable us to detect a 20 percent change 
in snipe nopulation. 
62 
