Average clutch size was 4.7 (range = 3 to 7) and average brood size was 4.5 (range = 2 to 6)

10

Results

In year 2004, nine females who were all caught utilized nine nest-boxes; estimated breeding density (in terms of nest-box user) was 1.70 female/ha. From year 2002 to 2004, 59 clutches (47 broods) were observed. The Green-backed tit can breed one to three times per year. Average clutch size was 4.7 (range = 3 to 7) and average brood size was 4.5 (range = 2 to 6). We banded and collected blood samples from a total of 70 adults (36 males and 34 females) and 126 nestlings (63 males and 63 females) of 32 broods (Table 1). Sex ratios in both adults and nestlings did not deviate from 1:1 expectation (p = 0.9050 and 1.0 for adults and nestlings respectively; binomial test, two-tailed). For 22 complete broods produced by 16 parental combinations, both nestlings (n = 93; Table 1) and parents were sampled, which comprised the complete brood samples whose mean size was 4.2 (range = 2 to 6, Table 2). These complete broods were produced by 16 social pairs (combinations of fifteen males and ten females). Some females could change mate within year and between years.

Parentage

EPY only occurred in broods of clutch sizes of 4 and 5 (Table 2). In complete broods, we found four broods that contained at least one EPY (18% of broods) (Table 2, 3); seven offspring were sired by extra-pair males (7.5% of young) (Table 3). The number of EPY in each EPF brood ranged from 1 to 4 (20% to 100%). Frequencies of EPF might vary temporally as we found no EPY in broods of year 2003 (n = 5), but 25%

of broods in year 2004 (n = 12) contained at least one EPY. Extra-pair fathers were identified for three EPF broods.

Disassortatitve mating by genetic similarity

11

For mated pairs, genetic similarity between ALL PAIRS (mean= -0.11; n = 16) and NEPF PAIRS (mean = -0.16; n =13) appeared to be lower than that in EPF PAIRS (mean = 0.016; n = 3) and CUCKOLDED PAIRS (mean = 0.13; n = 3) (Figure 1A). Rxy

between various combinations of mated pairs seemed to be slightly less than that of any adult male-female dyad (mean = 0.03; range: -0.47 to 0.59; n = 516). In ALL PAIRS, the mean R_xy was lower, but not significantly different from that expected in random (p

= 0.067; Figure 2A). However, the mean Rxy of NEPF PAIRS was significantly lower than that of random pairs (p = 0.005) (Figure 2B). This result suggests that female Green-backed Tits who were not engaged in EPF may have mated disassortatively with genetically more dissimilar males. On the other hand, females engaged in EPF tended to have social mates who were genetically more similar to themselves- mean R_xy of CUCKOLDED PAIRS was significantly higher than that of random expectation (p = 0.02; Figure 2C), whereas the mean Rxy between females and their extra-pair partners was not (p = 0.255). Comparing R_xy between EPF females with their social partners and EPF partners, we found that females were genetically more related to their social mates in two of the three EPF broods (Figure 3).

Comparisons of heterozygosity between offspring

Heterozygosity appeared to be lower in WPY (mean within-brood SH =0.905;

mean within-brood IR =0.032; n = 3) than NEPY (mean within brood SH =1.038; mean within-brood IR = -0.096; n = 18) and EPY (mean within-brood SH =1.057; mean within-brood IR = -0.101; n = 4) (Figure 4A & 4B). When comparisons were made between EPY and WPY within the same broods, EPY also seemed more heterozygous than their within-pair half sibs (Figure 5A & 5B), although without significantly statistical support (p = 0.5; sign test, one-tailed). Females who participated in EPF

12

seemed to produce with their social mates offspring with lower heterozygosity, a phenomenon constitutive of inbreeding: in terms of both SH and IR, WPY were significantly less heterozygous than expected from broods without EPY in random (Both SH and IR, p < 0.001) (Figure 6A & 6B), whereas EPY were not (SH, p = 0.242;

IR, p=0.412) (Fig 6C & 6D).