Reliability of egg morphology to detect conspecific brood parasitism in goldeneyes Bucephala clangula examined using protein fingerprinting

Abstract

Eadie (1989) developed a method based on variation between females in egg length, width and weight to detect conspecific brood parasitism in the field: using these three egg measures, Euclidean distance between all pairs of eggs within a clutch is calculated, and if maximum Euclidean distance (MED) between any two eggs exceeds a threshold value the nest is considered parasitized. The MED method has been tested in Finnish and Scottish common goldeneye Bucephala clangula populations but the results have been contradicting. Here we use protein fingerprinting to assess the validity of the MED method. Data comprised 35 clutches of which we knew, based on protein fingerprinting, how many different females laid the clutch (range 1-5 females). The mean MED of non-parasitized clutches (laid by 1 female only) was 1.470 (95% CI: lower 1.169, upper 1.771; n=21) and that of parasitized clutches (laid by 2 or more females) was 3.654 (95% CL: lower 3.083, upper 4.225; n=14). Using a MED>3.0 as a criterion to identify parasitized clutches 89% of all clutches were classified correctly either parasitized or non-parasitized when compared to the identification based on protein fingerprinting. Clutch size and the number of females (beyond 2 females) did not affect the clutch MED, whereas the status of parasitism did. Repeatability of egg length, width and weight were: 0.63, 0.76 and 0.80, respectively, implying that, variation in these egg measures occurs among rather than within females. Our new results confirm that the MED method is reliable enough to detect parasitism in common goldeneye. © 2009 J. Avian Biol.