Normally, to obtain the most accurate population 

 estimates, multiple aerial surveys are recommended for each 

 sampling period. The number of replicates required depends 

 upon the count as a proportion of the total population and the 

 number of marked deer as a proportion of the total population 

 (Rice and Harder 1977). As with most studies, our funding was 

 limited. Because of this, our earlier experience, and our 

 objective of concurrently estimating deer numbers and 

 collecting other types of data, we directed our efforts toward 

 3-4 population surveys at different times of the year rather 

 than 3-4 replicates during one period. 



Although this approach initially reduced our confidence 

 in any one estimate, over time it had the advantage of 

 providing estimates at shorter time intervals that could be 

 reconciled with sex and age composition and mortality data 

 from other sources. Also, because the survey data were 

 collected over many years, data sets were developed that could 

 be analyzed similar to replicate counts made near the same 

 time in one year. For example, the surveys flown with the 

 same pilot and observers during late December or early January 

 each year from 1980 through 1986 could be used as replicates 

 to determine the average percentage of deer observed at that 

 time of the year. Although the percentage of the population 

 marked (generally 4-9% during winter and spring) was lower 

 than desirable (Robson and Regier 1964), the total count as a 

 percentage of the total population (60-83%) was much higher 

 than usual . 



The 3-4 population estimates made each year during 

 1976-87 generally supported one another and, when correlated 

 with total counts, provided support or basis for improving 

 estimates for earlier years when population size was estimated 

 from a single complete coverage survey in winter or spring. 

 Any estimate, before being finalized, had to include 

 reasonable reconciliation of numbers in the various sex and 

 age categories. These accounted for proportion observed in 

 concurrent classifications as well as known or estimated 

 mortality and natality based on hunting statistics, survival 

 of marked deer, dispersal of marked deer, and population 

 numbers and composition from prior and subsequent estimates. 

 Our estimates based on the Lincoln index for years when 

 adequate numbers and distribution of marked deer were present 

 were seldom far from numbers estimated from known or estimated 

 natality and mortality, especially during early winter and 

 spring. 



The proportion of marked deer observed during aerial 

 surveys (observability) was generally consistent within a 

 season from year to year when the same pilot and observer were 

 used. Observability was most consistent and precise and 

 population estimates were most accurate during early winter 



31 



