were lost as well as those that were 
retrieved. 
2. Number of people, adults and juniors, 
who planned to hunt (potential hunters). 
3. Number of people who did hunt (active 
hunters). 
4. Average number of days active hunters 
were in the field hunting. 
5. Species composition of ducks 
geese in the waterfowl bag. 
Additional information on the proportional 
periodic distribution of the duck bag andthe 
percentage of hunters participating during 
the different parts of the season has been 
recorded by Atwood and Wells, Jr. (1961). 
The sampling system for the survey is 
based upon post offices where Migratory 
Bird Hunting Stamps are sold, because there 
exists no listing of waterfowl hunters that 
might be sampled. Within each State, a 
series of sales outlets at post offices is 
selected at random. For sampling efficiency, 
the post offices are grouped in two cate- 
gories: those with single sales outlets 
(usually in smaller towns) and those with 
multiple outlets (usually in larger cities), 
and each category is sampled separately. 
The postmaster at an outlet selected is 
asked to give an address card to each 
person who buys a Migratory Bird Hunting 
Stamp; this postage-paid card is tobe filled 
out and mailed. The card requests the stamp 
buyer's name and address, number of 
stamps bought, purpose of purchase (hunting, 
stamp collecting, or resale), and names of 
junior hunters (under 16) living in the same 
household who are expected to hunt during 
the current season. Hunters who mail these 
cards as requested, and junior hunters 
listed on the cards, receive the question- 
naire. The number of sampling outlets used 
and the number of questionnaires mailed and 
returned are shown in table A-1 (p. 43). 
Estimates of total kill are made on the 
basis of hunter responses and the sales 
volume of stamps, by State. For kill esti- 
mates to be available by early August, the 
July through March summaries of stamp 
sales supplied by the Post Office Depart- 
ment must be used, but they are substantially 
the same as the final sale figures that are 
available later. 
The information requested from each 
and 
Icee Literature Cited, page 40. 
hunter includes, inpart, the total bag (by 
ducks, geese, and coots), dates hunted and 
the bag for each date, and the species com- 
position of the kill. 
Certain minor parts of the 1960-61 survey 
are known to be unreliable and are so 
designated in the tables. Through aclerical 
error undetected until too late for correc- 
tion, distribution of the eastern and western 
forms of the questionnaire was somewhat 
mixed. Judging from a sample of the 
responses, the eastern form went to more 
than 99 percent of the hunters questioned 
in the Pacific Flyway and to about 77 
percent in the Central Flyway, while the 
western form went to about 61 percent in 
the Mississippi Flyway and to about 15 
percent in the Atlantic Flyway. Correct 
forms went to the others. The two question- 
naires are identical inmost characteristics, 
therefore for most purposes this mistake 
probably had no effect. However, the eastern 
form (received by most western hunters) 
did not list the cinnamon teal and the 
western form (received by so many hunters 
in the Mississippi Flyway) did not list the 
black duck. Very likely the major decreases 
recorded for the two species reflect these 
omissions. Other differences are that the 
western form lists the cackling goose and 
the black brant, while the eastern form 
lists, instead, the Hutchins's goose and 
the American brant. 
Accuracy of the data in this report is 
affected both by reporting bias (misre- 
porting and differential nonreporting) and 
by sampling error. Reporting biases tend to 
exaggerate the estimates of hunting activity 
and kill, and corrections are made by 
methods that have been employed for several 
years.? 
The species composition of the kill is 
recorded as reported by hunters and com- 
pared with the species composition esti- 
mated independently from the wing col- 
lection survey (table A-2, p.43). Agreement 
between the results of these two surveys is 
encouraging. 
In general, the statistical reliability of the 
report is greater for areas with the larger 
numbers of outlets sampled and higher 
2These errors and methods for making the corrections are 
partially described in two papers by Atwood (1956, 1958). A 
mimeographed description of the methods used also is available. 
