Singing rate and detection probability: an example from the least Bell's Vireo (Vireo belli pusillus)
This article is part of a larger document. View the larger document here.Abstract
We used 4-hr tape recordings to assess singing activity of a Least Bell's Vireo (Vireo belli pusillus) on 7 d distributed throughout the breeding season of 2001. We logged 9873 songs, for a mean singing rate of 5.84 min-1, but despite this level of activity, the bird was silent during 33 percent of the 1691 minutes of the study. Availability, the proportion of minutes in which singing activity occurred (and hence the probability of at least one song occurring in any randomly selected minute) varied from 0.36 to 0.96. As availability is a major component of detection probability, we tested several ways of estimating availability from data that can be obtained during a brief point count. Singing rate (songs/min), estimated in 10 or more 1-min samples randomly distributed through the morning provided a good (r² = 0.8) estimate of availability for the entire morning. Measures based on a comparison of singing activity (presence/absence) in contiguous 1-min samples yielded poor estimates of availability, regardless of sample size, presumably because of the highly clumped bout structure (i.e., serial auto-correlation of singing and silence).

