How well do predation rates on artificial nests estimate predation on natural passerine nests?

Daily survival rates (DSR) of three nest types were compared among three passerine species in woodland habitats of the Czech Republic in 1998-99. The species were: Yellowhammer (YH) Emberiza citrinella, Blackcap (BC) Sylvia atricapilla, Song Thrush (ST) Turdus philomelos. The nest types were: AA, artificial nests with artificial plastic eggs, both mimicking the real models (total n = 432); NA, natural nests left in their original position and baited with artificial plastic eggs (n = 706); and NN, active natural nests observed during the egg stage (n = 596). Rodents were dominant predators of the near-ground YH nests while predation by corvids slightly prevailed in the shrub nests of BC and ST There was no consistent relationship between nest type (AA vs. NA) and type of predator. Effects of nest type (AA vs. NA or AA vs. NN, respectively), species, date and study plot on DSR were examined by fitting logistic regression models. Marginally significant (< P < 0.05) differences were found between AA and NA nests in both years. Due to the interaction with date in 1998, the model predicted lower/higher DSR for AA nests before/after about 25 May respectively. In 1999 the DSR of AA nests was consistently higher (difference: 0.011; 95% Cl: 0.001 to 0.021). Marginally significant differences between AA and NN nests were found only in 1998: the DSR of AA nests was consistently lower (-0.014, 95% CI: -0.027 to -0.001). Comparisons of mean DSR within each species-year sample showed that the differences between nest types were of both directions and varied from almost zero to 0.02 (ST, BC; ns) or 0.04 (YH; P < 0.05). Although the experimental nests accurately reflected the mean values as well as seasonal trends and interspecific differences in DSR of the real nests in some subsets of data, the pattern was not consistent across species and between years. This study suggests that artificial nests could not be used to assess survival of natural nests without careful validation of the experimental results.