Estimating Nest Success: The Mayfield Method and an Alternative

摘要: -Mayfield's method for calculating the success of a group nests is examined in detail. The standard error his estimator developed. Mayfield's assumption that destroyed are at risk until midpoint interval between visits leads to bias if visited infrequently. A remedy suggested, Mayfield-40% method. I also present competing model, which recognizes actual destruction date failed nest unknown. Estimated daily mortality rates and errors developed under this model. comparison original Mayfield method, new incorporates an unknown destruction, shows or modified performs nearly as well more appropriate requires far easier calculations. technique statistically comparing offered; one proposed by Dow (1978) claimed be misleading. Finally, give detecting heterogeneity among improved estimator, it found. Received 5 March 1979, accepted 28 July 1979. THE well-being avian population lies delicate balance natality mortality. Biologists attempt infer status species estimating births deaths and, through their comparison, determining former sufficient offset latter. For most populations wild birds, none crucial characteristics dynamics easy measure. One component seems gauge percentage hatch, often used indirect measure reproduction. (1961) has demonstrated, however, serious ordinary rate: dividing number observation into those ultimately hatch. To overcome difficulties he recognized, (1961, 1975) alternative hatch rate. In accounts fact normally not all from day of. initiation but discovered various stages development. Nests found late stage likely than early one, because they have already survived part requisite time. Combining nests, regardless development, apparent rate will result severely biased estimator. places on comparable basis using only information period during was observation. length termed exposure, although may term. Thus, 10 May still active 18 had 8 days exposure. Had been May, would credit with 4 half period. 651 Auk 96: 651-661. October 1979 652 DOUGLAS H. JOHNSON [Auk, Vol. 96 From calculates total exposure nest-days. This divided were while resultant value, expressed losses per nest-day, estimated nests. example, (1961: 258) analysis Kirtland's Warbler (Dendroica kirtlandii), 154 seen incubation represented 882.5 (Mayfield's data reanalyzed here; some results differ slightly presentation.) Thirty-five lost (destroyed deserted), yielding 35/882.5 = 0.04 nest-day. determine probability survives entire incubation, must know period; 14 days. survival 0.96 (=1 0.04), so surviving throughout 14-day times itself times, 0.9614 0.56. Although major advance treating nesting data, criticized (Green 1977) its homogeneous, i.e. subject same addition, provided neither variance estimates nor tests underlying assumptions. paper intended augment (1961 1975). derive heuristically, formal context. can calculated derivation. implications midway considered propose realistic does require assumption. Estimators obtained model compared Mayfield. discuss importance variation rates, both identifiable nonidentifiable causes. Methods such variability it, exists, presented.