USGS Home
Contact USGS
Search USGS

The situation is somewhat different for nests on islands. First, small nesting areas and high density and visibility of nests often permit many destroyed clutches to be found. Second, nest success is often high on islands. Third, nesting on islands is often fairly synchronous, so nests can be found before many of them are destroyed. Fourth, mortality to clutches in island nests is more often catastrophic; the Mayfield method assumes that a steady mortality rate applies.

This paper examines the question: Which method--the apparent or the Mayfield--is better for island-nesting waterfowl? Further, how can the biologist tell which method to use?

We simulated several situations and evaluated competing estimators under each. Nest populations had different levels of 1) nesting synchrony, 2) constancy of mortality rate (steady mortality rate versus catastrophic mortality), and 3) survival rate of clutches. Simulated searches were made with 4) different frequency, and 5) various detectability of clutches.

Preliminary results follow. With steady mortality, the Mayfield estimator generally performed well regardless of nesting synchrony or mortality rate, although two or more searches were often necessary for the highly synchronized population and three or more for the less synchronized one. The apparent estimator was adequate if detectability was fairly high and three or four searches were done on a less-synchronized population or two searches on a synchronized one. If only one or two searches were done on a low-synchronized population, or one on a synchronized population, very high detectability was necessary to obtain an accurate estimate with the apparent rate.

When mortality was mostly catastrophic, the Mayfield method performed poorly, unless nesting was in low synchrony and several searches were made. For the apparent method, one search made before the catastrophe provided reasonably close values, but a single search afterward did not unless detectability was high. With two or more searches, results were mostly satisfactory.

General recommendations: 1) A minimum of three searches of a population nesting asynchronously, and two of a synchronous one, is required for accurate estimates of hatch rate, unless detectability of nests is known to be high. 2) If mortality occurs steadily, the Mayfield method generally gives better estimates than the apparent, unless detectability is high. 3) If mortality occurs catastrophically, the timing of the searches critical, and accurate results may be only fortuitous.