higher reproductive success than the Utah and Montana studies, the 
average size brood being 7.17 for 1310 observa:ions. No one f amiliar 
with the breeding grounds of waterfowl] will doubt that the reproduc- 
tive success of this group is subject to fluctuations. These hatching 
and brood data are not, I think, contradictory, but may be actually 
representative of typical variations encountered in the population 
dynamics of the mallard and other waterfowl, 
Table 25.—-Reproductive Success of North American Mallards-—~II 

Per eee ee ee eee eee 

er ean 
Where When of Number, Number References 
Studied Studied Broods Young!/ per Brood 
British Columbia - 8h 509 6.0 Munro (193) 
British Columbia 19)8 4,7 310 6.6 - Munro (1948) 
Alberta 1948 70 472 6.7 Smith (19)8b) 
Sask.-Alberta 1934-35 32 227 Tel Kalmbach (1937) 
Saskatchewan 193-35 15 87 5.8 Furniss (1938) 
Saskatchewan 1937 7 39 5.6 Furniss (1938) 
Saskatchewan 19447 77 552 702 Lynch (19,8) 
Saskatchewan 1948 53 389 7.3 Soper (19,8) 
Manitoba 19,7 65 03 6.2 Hawkins (1918) 
Manitoba 19,8 17 118 6.9 Hawkins & Cooch (1948) 
North Dakota 1947 66 Lhd 6.8 Hammond (1948) 
North Dakota = 1948 19 152 8.0 Stoudt & Davis (1948) 
Minnesota 1937-8 758 5685 7.5 Stoudt (19,8) 
Total 1310 9392 Tel 
i Calculated for the most part from other statistics in the original 
paper cited. 
Recent attempts to gather data on the shrinkage in mallard 
brood size have not been very successful (table 26), although Stoudt's 
(1946, 1948) work in the Chippewa National Forest, Minn., seems to be 
a satisfactory appraisal of the early survival of young. This work 
covered a period when the continental waterfowl population was thought 
to be increasing. Much of it is a by-product of census work. While 
the data may be biased by field difficulties that prevent counts from 
being complete for the younger ducklings, it seems possible that the 
narrow time~span of the censuses also created a time-specific life- 
table effects In other words, the size of the mallard clutch was 
gradually shrinking throughout the season. This means that the size 
of broods at hatching also decreased as the nesting season progressed. 
On a transect of the breeding grounds, the census taker would thus 
‘tend to record the early-hatched young in Classes II and III (table 26) 
and the recently hatched young from renestings in Class I. Excellent 
evidence for the progressive shrinkage in clutch size in other species 
has been given by Leopold (1933, p. 363), Srrington and Hamerstrom 
(1937), Hamerstrom (1939), and J. B. Low (1945); and an important 
summary of the known data has been published by Lack (1947). Stoudt's 
63 
