Vo 1. 17, No. 12 
Page 2 
These data indicate that differences in pheasant population levels between 
the two areas are probably largely the result of higher productivity from road¬ 
sides on the FCMU. 
Table 1. Estimated average number of successful pheasant nests per square mile, 
bibley study area (SSA) and Ford County Management Unit (iCMU), 1973 and 1974. 
Hay 
Pasture 
Sma11 grains 
Strip/Nonagricul tural 
Roadsides 
1973 
1974 
FCMU 
6 
2 
2 
SSA 
2 
0 + 
1 
FCMU 
2 
0 + 
I 
6 
10 
2 
2 
3 
9 
26 
7+ 
15 + 
Ecology and Management of Squirre1s q ^ Nixon 
S. P. Havera 
In last month's report (MWRL 17(10):1-2), we pointed out the probable error 
m the est.mated squirrel harvest for Illinois. There are two possible sources 
of error involved in the present mail survey--response and nonresponse. 
We recognize that response bias will be more difficult to cope with than 
nonresponse bias. The 5 . 5 -month (August-January) delay between the end of 
squirrel hunting and the receipt of questionnaires undoubtedly contributes to 
errors in reporting game kills. Two possibilities for reducing this bias are; 
(l; to compile a mailing list of potential squirrel hunters from individuals 
who purchased licenses the previous year, contact them before the season, tell 
them that they have been selected to help in estimating the kill of squirrels, 
and urge them to keep records; and (2) to mail a separate squirrel hunter 
questionnaire to these hunters in mid-to-late November. This procedure would 
reduce by about 2 months the time interval between the end of squirrel huntinq 
and the arrival of questionnaires. 
Nonresponse bias may best be overcome through repeated mailings. In 
Michigan, the Game Division has achieved a better than 90 percent rate of 
return using repeated mailings of up to five questionnaires for estimating 
harvests of small game. They send new questionnaires with each mailing, repeat 
