Certainty of paternity in two coucal species with divergent sex roles: the devil takes the hindmost

Background Certainty of paternity is considered an important factor in the evolution of paternal care. Several meta-analyses across birds support this idea, particularly for species with altricial young. However, the role of certainty of paternity in the evolution and maintenance of exclusive paternal care in the black coucal (Centropus grillii), which is the only known altricial bird species with male-only care, is not well understood. Here we investigated whether the differences in levels of paternal care in the black coucal and its sympatric congener, the bi-parental white-browed coucal (Centropus superciliosus), are shaped by extra-pair paternity. Results We found that male black coucals experienced a substantially higher loss of paternity than white-browed coucals. Further, unlike any previously reported bird species, extra-pair offspring in black coucals represented mainly the last hatchlings of the broods, and these last hatchlings were more likely to disappear during partial-brood loss. Conclusion The results suggest that exclusive paternal care in black coucals is not maintained by male certainty of parentage, and extra-pair fertilizations are unlikely to be a female strategy for seeking ‘good genes’. Extra-pair paternity in black coucals may reflect the inability of males to guard and copulate with the female after the onset of incubation, and a female strategy to demonstrate her commitment to other males of her social group. Electronic supplementary material The online version of this article (10.1186/s12862-018-1225-y) contains supplementary material, which is available to authorized users.

Germany) and the DNA from eggs and tissue samples (5% of the samples) was extracted by using  PCR was performed using 1µl of purified DNA of 10-60ng/µl concentration, 5ul of Type-It PCR 28 master kit (Qiagen, Hilden, Germany), 1µl of primer mix containing 4-6 different fluorescent-29 colored microsatellite markers, and 3µl of deionized RNA-free water making a total of 10µl volume. 30 Three different primer mixes were used in black coucals and four primer mixes were used in white-31 browed coucals. In total, 16 microsatellite markers were used in black coucals and 20 makers in 32 white-browed coucals (Additional file1: Tables S1 and S2). These microsatellites included the 33 P2P8 sex marker [1], some microsatellites that had been previously developed for parentage analysis 34 in black coucals [2] and pheasant coucals [3], as well as some microsatellites of other birds [4-7] that 35 were tested and found to work well in coucals (Additional file1: Tables S1 and S2).

37
The PCRs were carried out on a GeneAmp 2700 PCR System (Applied Biosystems, Darmstadt, 38 Germany) with the following conditions: 5 minutes of initial denaturation at 95°C, 28-30 cycles of 39 30 seconds of denaturation at 94°C, 90 seconds of annealing at 50-57°C and 60 seconds extension at 40 72°C, followed by one 30 minute of final extension step at 60°C before the PCR products were 41 3 cooled down and kept at 4°C until they were removed from the machine. Note that the exact number 42 of PCR cycles and annealing temperature varied depending on primer mix and the coucal species. 43 The PCR products were stored in a fridge at 4°C until analysis.   In all clutches for which one or both social parents were sampled, we first performed parentage 62 analysis by colour-coding to check for matching and mismatching alleles between the offspring and 63 their social parent(s). In all cases we started by fitting in the mother (if known) and then the father. 64 4 In cases where we observed a mismatch of allele(s) between an offspring and one or both of its 65 putative parents we rechecked the alleles in the GeneMapper software to confirm whether the 66 mismatch was real or was due to an error during allele scoring. In almost all cases the mismatches 67 observed were real. Mismatches between offspring and their putative social mothers were rare and 68 few (≤ 2 loci), but multiple mismatches between some offspring and their putative social fathers 69 were common, particularly so in black coucals. However, in seven black coucal clutches the females 70 that were thought to be the social mothers from field observation were not the biological mothers of 71 the broods. This happened because of territory take-overs with the new females 'usurping' the nests 72 and males of the defeated females, or assignment of nests on the boundary of two territories to a 73 wrong female. This did not happen for male black coucals or both sexes in white-browed coucals 74 that where confirmed to attend the nests.  Following this initial step we excluded four microsatellite markers for black coucals and five for 81 white-browed coucals because they showed significant deviation from the Hardy-Weinberg 82 equilibrium or had relatively high error rates (≥5%). Although Cervus can handle genotyping errors 83 and deviations from the Hardy-Weinberg equilibrium, we decided to conduct the final parentage 84 analyses by using the microsatellite markers that agreed to the Hardy-Weinberg equilibrium and had 85 low errors rates (≤5%). We did so to reduce errors in parentage assignments, as recommended by 86 [8]. Also, the markers that agreed to the Hardy-Weinberg equilibrium and had low error rates 87 5 (Additional file1: Tables S1 and S2) were sufficient and powerful enough to resolve parentage with 88 high confidence (>99.9%).

90
The final parentage analyses in black coucals were conducted based on 11 microsatellites including 91 Cgr2, Cgr6, Cgr7, Cgr9, Cgr13, Cgr15, Cgr16, Cgr17, CAM13, CP9, and CP11 (for more details 92 regarding these microsatellites see [2, 3] and also Additional file1: Table S1). These markers had a  The comprehensive parentage analysis with Cervus was preceded by a simulation of parentage based 106 on 60% of candidate parents sampled for black coucals and 80% for white-browed coucals (field 107 estimates), 99% of loci typed, and 1% of loci mistyped. We first conducted a maternity analysis with 108 all known females as potential mothers and then a paternity analysis with the identified biological location of their territories enabled them to be the genetic fathers). For all offspring that we were not 121 able to identify one or both of their biological parents we repeated the analysis by using all sampled 122 fledglings from previous years as potential parents. In some nests we already knew that the social 123 mother or father was a ringed fledgling from previous years but we had failed to catch them.

124
Therefore, this approach allowed us to identify recruited parents from previously sampled fledglings 125 that we did not catch as breeding adults.   But in this species we had very few clutches for which we failed to sample the social fathers, and 169 extra-pair paternity was extremely rare. Also, we had sampled almost all the potential extra-pair sires 170 around the territories of the males that we failed to catch and therefore, we are confident that we did 171 not fail to detect any cases of extra-pair paternity in the clutches whose social fathers were not 172 sampled.

174
In clutches for which we failed to sample the social fathers (N=66 in black coucals; N=15 in white-175 browed coucals) and our sibship analysis suggested that the clutches had multiple paternities (N=25 176 in black coucals; N=0 in white-browed coucals), we were able to unequivocally identify the extra-177 pair offspring in 15 clutches. This is because the extra-pair offspring were sired by males that we 178 knew they were not the social fathers of the respective clutches. However, in 10 black coucal 9 clutches for which we found evidence of mixed paternity, the extra-pair sires could not 180 unequivocally be identified. In these clutches we considered the smaller number of offspring with 181 the same genetic father to be the extra-pair offspring, and the larger number of offspring with the 182 same genetic father to be the within-pair young. This represents a parsimonious approach, because it        The last column [P(β) >0] gives the posterior probability of the hypothesis that the effect is greater