ginnings and the random effect of predation served to spread out the season. 
Brood populations rose gradually to a peak and then tapered off. The same 
situation was reported by Sowls (1951) for the Delta Marsh in Manitoba. 
Unless some widespread nest destruction such as a storm or flood should occur, 
there would be little chance of more than one peak of hatching. Even then, 
the individual variations between hens in the amount of delay before renesting 
could be expected to stagger the hatch. 
POTHOLE UTILIZATION 
Barly arrivals 
The relatively few spring migrants that used the area were generally 
gathered in groups and tended to favor the larger waters. Dabbling ducks used 
mainly the deep and shallow marshes, while the divers concentrated in the open~ 
water areas. There were conspicuous exceptions to this, and groups of mi- 
grants, even divers, might be seen on temporarily flooded grainfields or hay 
meadows. Apparently, they congregated where there was open water that was 
easy to keep ice-free at night and where feeding conditions were favorable. 
Breeding pairs 
Resident birds, except for the earliest arrivals, tended to scatter out 
over the landscape and not to favor any particular type of water area. The 
earliest arrivals made use of areas not occupied by ducks at any other time. 
Table 9 shows the use of potholes of the various size classes by breeding 
pairsi during the peak of the breeding season in May and early June of the four 
years. 
Pothole use varied inversely with size: the smallest, apparently least 
attractive areas received the heaviest use per acre, even though many were dry 
during much of the season. In the last column of table 9, the acreage of a 
pothole is included in the total for the size class only if it held water, and 
the use by ducks is considered in terms of water alone as it occurs in the 
various size classes. This illustrates what might be expected to occur if 
there were no variation in water levels and potholes of all size classes were 
relatively permanent. On this basis, there were 13 times as many birds per 
acre (8:0.6=13.3:1) on the smallest bodies of water as on the largest. It 
appears that the best distribution of a given amount of water for pairs only 
would be in the form of many small, relatively permanent areas available to 
the birds throughout the breeding season. 
The size of the breeding population,. as well as the availability of 
water areas, appeared to have considerable influence on this behavior. Table 
10 shows the use of areas under 2 acres in size by pairs of blue-winged teal, 
gadwall, mallard, and pintail in 1951 (the only year in which enough breeding- 
season coverages of the study area were made to permit comparison). 

1 Breeding pairs throughout thig section refer to pairs, lone drakes, 
and lone hens. 
39 
