Observations on the principal Manitoulin Island study area in- 
dicated that 87 percent of the snipe present were recorded during 
the average one-minute stop at the peak of morning activity, as 
compared with 45 percent for a one-minute period during the evening 
peak.- Whether these percentages are sufficiently typical for obtain- 
ing an estimate of the actual number of birds present in a given area 
remains to be determined. Preliminary investigations in areas where 
only one pair of snipe is present suggest the likelihood that winnow- 
ing counts are less reliable in this case since the presence of two 
or more males appears to act as a stimulus for winnowinge 
Summary 
le Breeding ground transects, or area studies must be conducted 
according to rigid specifications if they are to yield significant 
resultse 
2. Winnowing counts should be made between the termination of 
the spring migration period and the time the young birds are about 
two weeks old. In the northern states and southern Canada this would 
be roughly from May 20 to Jume 20, with the optimum time the first 
ten days of June. These dates are subject to further investigation, 
and probably vary a week or more due to local or widespread weather 
conditions or water levels in any particular year. 
3e The daily peak of winnowing occurred during a 10-minute 
period extending from 75 to 35 minutes before local sunrise time; if 
5 minutes are dropped from each end of this period to allow for in- 
dividual variation, the remaining 30 minutes were characterized by 
almost continuous winnowing by all three birds in the area which was 
most intensively covered. This appears to be the period most suited 
for breeding ground census work, 
4. A lesser and shorter peak occurred during the period from hO 
to 70 minutes after sunset. Seldom were more than two out of three 
males in the air at once. The great variability of activity from 
minute to minute and from one evening to another indicates the neces- 
sity of making four or more times as many evening counts to obtain 
the same accuracy as during the early morning peak. Winnowing at 
other times of the day gives little indication of the number of pairs 
pre sent. 
Se The presence of a bright moon encourages winnowing at other 
hours of the night and reduces the amount at the usual times. 
6. The exact manner of coverage can be adapted to specific local 
conditions, The minimum number of stops at one listening station 
necessary to provide significant data for listening periods of l, 2, 
3, 4, and 5 minutes during the morning and evening peaks have been 
computed from the Manitoulin observations and will serve as an approxi- 
mate indication of the amount of coverage required until further data 
50 
