ANALYSIS OF BAND DATA 
The use of band data in analyses of bird populations was pioneered by Nice (1937) in 
her study of the song sparrow. Leopold et al. (1943) combined band and census data to deter- 
mine mortality of ring-necked pheasants on the University of Wisconsin Arboretum refuge. On 
the other side of the Atlantic, Lack (1943a, b, c) published studies, based entirely on band re- 
covery figures, on the survival of several English birds. 
Recently Farner (1945), Buss (1946), Marshall (1947), and Hann (1948) have published ~ 
on the survival and mortality of, respectively, the robin, ring-necked pheasant, herring gull, 
and oven-bird. Deevey (1947) presented a searching analysis and review of life tables as used 
to determine survival of animals in nature. 
The present study deals with the survival and mortality of the mallard, black duck, 
and blue-winged teal in the Mississippi flyway as determined for the most part by means of 
10,718 recoveries (bands recovered from ducks shot and bagged or ducks found dead) from 
51,297 individuals banded by the Natural History Survey at the Chautauqua National Wildlife 
Refuge, near Havana, Illinois, fig. 1, or at McGinnis Slough, in and with the co-operation of the 
Forest Preserve District of Cook County,near Orland Park, Illinois. The actual banding was 
done during the fall months, 1939-1944, at Lake Chautauqua and, 1940-1945, at McGinnis Slough; 
recovery data extend through 1946 for Lake Chautauqua and 1947 for McGinnis Slough. Inclu- 
sion of the McGinnis Slough data in this report was made possible through the co-operation of 
Roberts Mann, Dr. David H. Thompson, and John Jedlicka. 
VALIDITY OF SHOT RECOVERIES.--Previous investigations indicate that satisfac- 
tory appraisals of longevity can be made from band recoveries of shot birds. Paynter (1947) 
found that there was no significant difference in mean length of life between herring gulls dy- 
ing as a result of shooting, trapping, or other human interference and those dying as a result of 
natural or unknown causes. Farner (1945) showed that, in robins, life-expectancy figures ob- 
tained from the number of birds shot and from the number killed by cats were similar to those 
calculated from band recoveries representing the total population. 
Band recoveries from shot ducks are many times as numerous as records of live re- 
captures or other returns, and they provide the largest sampling on which to base life tables. 
The number of band recoveries from ducks in a banding-class* shot in a particular 
year following banding is dependent on two principal factors: (1) the number of ducks in the 
banding-class left alive to be bagged and (2) the shooting pressure. If shooting pressure in 
two specified years is equal, then the difference in the number of band recoveries in the two 
years would be a measure of total population loss from the first year to that following. How- 
ever, there has been a year-to-year variation in shooting pressures. In Chautauqua-banded 
mallards the season-of-banding bag ranged from 6.7 to 14.9 per cent of birds banded, 1939- 
1944, table 9. This year-to-year variation results in a sampling error when the population 
loss is measured for any specified year considered separately. But because ducks of most 
* Banding-class refers to the ducks banded in a specified year. 
