INCREASING POINT-COUNT DURATION INCREASES STANDARD ERROR

We examined data from point counts of varying duration in bottomland f orests of west Tennessee and the Mississippi Alluvial Valley to determ ine if counting interval influenced sampling precision. Estimates of s tandard error increased as point-count duration increased for cumulati ve number of both individuals and species in both locations. Although point counts appear to yield data with standard errors proportional to means, a square root transformation of the data may stabilize the var iance. Using long (>10 min) point counts may reduce sample size and in crease sampling error, both of which diminish statistical power and th ereby the ability to detect meaningful changes in avian populations.