in assessing the importance of waters in the study area as nesting, resting, or 

 feeding habitat, size of breeding populations, nesting habitat requirements, 

 dates of nesting, clutch and brood sizes, productivity, dates of earliest arrival, 

 peak of migration, and end of migration for waterfowl found in the study areas. 

 A more detailed discussion of the categories of data recorded is given in 

 Appendix C. The locations of waterfowl observed were not mapped, as the lo- 

 cations were coded into the data sheets. 



Computer programs were developed to summarize and analyze raw data recor- 

 ded on these two types of data sheets. For both types, programs were prepared 

 to: code data from data sheets; print the data in expanded form with column 

 titles; edit the data; and list the data chronologically or alphabetically by 

 species. Additional programs were developed for the general wildlife observa- 

 tion data sheets to: summarize sex and age classifications by species; analyze 

 the data in terms of vegetation, topography, slope, aspect, activity, and time 

 of day for each species; and sort these analyses by month or season, study area 

 level, observer, vehicle, and by completeness of classification (that is, whether 

 or not all individuals in a group were classified by sex and age). This allows 

 speedy retrieval and analysis of data; for example, one can quickly obtain a 

 printout which shows the sex and age classification of pronghorn antelope groups 

 (fully classified groups only) observed by Blanchard during November and December 

 aerial flights over the study area. This same printout will show the numbers 

 and percent of individuals in each sex and age class which were seen in each 

 habitat category. 



This method of ground and aerial surveys has the advantage that a relativ- 

 ely large number of observations can be recorded over a large area with a min- 

 imum of effort and special equipment, and characteristics of the animals obser- 

 ved (e.g., percentage use of habitat categories, sex ratios) can be statistic- 

 ally determined for species with a fairly large sample size. However, neither 

 aerial nor ground surveys provide a random sampling technique, as both are 

 biased toward large, conspicuous animals or groups and toward open habitats 

 where the animals are most visible. Also, both aerial and ground surveys are 

 restricted to daylight hours, with few exceptions, and as such sample only 

 diurnal habitat use, animal activity, distribution, and group characteristics, 

 which may be quite different from the nocturnal situation. While aerial sur- 

 veys provide an essentially random sample of habitats, ground surveys are heav- 

 ily biased toward roadside habitats, especially those which happen to occur 

 along major roads. Aerial surveys give the most reliable picture of general 

 distribution patterns, and also the minimum total numbers of large vertebrates, 

 as each individual is observed and recorded only once. Ground surveys, however 

 sample the same individuals many times, and observations cluster heavilv along 

 main-traveled routes. Ground surveys, in essence, sample the population, return 

 the sample to the population, and sample again. Thus, while 500 ground observations 

 of pronghorn antelope may be made in one week, the actual number of pronghorn 

 antelope present may by only one hundred, even though every effort was made to 

 avoid recording observations for the same group twice on the same day. The 

 results obtained by thus pooling samples are by no means a census, but rather 

 provide a description of the group sampled. If we assume that the character- 

 istics of the sample apply also to the entire population, we can theoretically 

 extrapolate descriptive statistics to areas and individuals not sampled. For 



27 



