similar trends in all routes, the extremely close parallel between Routes 
2 and 3 (possibly indicating that populations of similar magnitude were 
being sampled), and the fact that early peaks in Routes 1 and 2 were al- 
most identical in height with late ones, all lead to the conclusion that 
these counts are quite sensitive. 
Since the calling level is fairly stable for several weeks during 
the midseason plateau, this period seems to be the best time during 
wnich counts should be made. Table 1 presents statistical summaries 
of the three 1951 Wisconsin routes, comparing measurements on counts 
run throughout May and June, and counts run only between May 15 and 
ie 26. In comparing the Coefficients of Variation for the two 
analyses, it is obvious that by running counts between the early May 
and July peaks less variation is encountered. Tess variation should 
enable detection of smaller population changes. 
Table 1.--Statistical Summary of Wisconsin Call Counts Taken During 
May and June o 


No. of Counts Mean « Deve Coef. of Var. 
We) ‘oO Wo) 
C N A ret) 
gS Fs F > 3 
a ef 
ee eee zw OUS 
Route 1 9 6 32.9 29.8 5.10 3.61 15.56 12.1% 
" 2 6 Tl 26.8 24.3 4.53 2,26 17.3% 6.2% 
t 3 8 —~6 23.8 23.1 4.68 2.10 19.9% h.& 
It must be remembered that these are tentative conclusions based 
only on one year's work. These seasonal calling curves therefore need 
further scrutiny before it can be said with certainty at what periods 
call counts should be run. As yet we know nothing of year-to-year 
phenological changes or effects of different population levels on calling 
pattern. 
Rural Mail-Carrier Surveys | 
In an attempt to get quantitative data on geographical abundance of 
breeding doves in Wisconsin, a rural mail-carrier survey was employed. 
The peak of nesting during the first brood was selected as the time during 
which to run this survey. At this time there are virtually no young of the 
year in the population and the carriers truly would be counting breeding 
adults. It was also hoped that at this period there would: be virtually no 
48 
