Date of Award


Degree Type


Degree Name

Master of Science (MS)


Environmental Science and Biology


Declining populations of grassland breeding birds have generated considerable concern from biologists and managers, and an increased understanding of the habitat preferences of grassland breeding birds and improved management techniques are being used to guide planning and conservation efforts. The success of these efforts is often determined by the collection of data on bird response, primarily changes in occupancy or abundance as determined by point counts or similar techniques. However, anthropogenic grasslands in the Northeast may serve as ecological traps when mature birds choose fields that appear to be suitable breeding habitat, but intensive agricultural practices or other detrimental management occurs prior to the successful fledging of young. Therefore, merely quantifying the presence or abundance of adult birds does not indicate the quality of a field as breeding habitat, and impacts to productivity should be quantified as the true measure of grassland bird response.

The traditional metric for productivity has been nest searching and monitoring. However, due to the challenge of locating well camouflaged nests, along with the potential to unnecessarily disturb nesting attempts while searching for nests, interest is mounting in developing methods that involve indirect estimates of productivity. A potentially valuable method was introduced by Vickery et al. (1992); this combines territory mapping (from which density can be calculated) with observations of behaviors associated with stages in the breeding cycle to create an index of productivity for each territory. Estimates of rates of breeding success (productivity) can then be calculated, although this method does not provide any information on the number of young fledged. In addition, the estimated rates can be artificially inflated by misclassifications of breeding failure as successes due to nest parasitism by Brown-headed Cowbirds (Molothrus ater). Before this method should be widely implemented, its accuracy and suitability as a substitute for nest searching and monitoring should be assessed. The research presented in this thesis expands upon a preliminary effort by Rivers et at. (2003) that compared results from the reproductive index and nest searching and monitoring for Dickcissels (Spiza americana) in Kansas. The authors found that the reproductive index may be unsuitable substitute when studying Dickcissels, and concluded that the method should not be implemented for other species without further assessment.

This thesis presents the results of a rigorous assessment of the reproductive index using Savannah Sparrows (Passerculus sandwichensis) in western New York during the 2002 and 2003 breeding seasons. I found a weak correlation in classification of the breeding stages of the monitored territories among multiple observers (r=0.398), as well as large differences between plot-level success rates estimated by the reproductive index and modified nest success rates using data from standard nest searching and monitoring (9.8% and 41 .7%, respectively) . Most importantly, I made territory-level comparisons of reproductive index estimates of success with actual fate as observed through detailed nest monitoring, and found the reproductive index correctly predicted actual nest fates for only 43% of the monitored nests. When treated as a continuous predictor of nest success, the reproductive index rank did have a positive slope (odds ratio 1.55, P = 0.09), but treated as a categorical predictor, predicted nest success did not increase smoothly with increasing index rank. In short, the reproductive index exhibited neither internal consistency, nor the ability to predict nest fate at the plot or territory level, and functioned poorly as a substitute for traditional nest searching and monitoring for Savannah Sparrows in my study.