Breeding biology of two sympatric Nesospiza finches at Nightingale Island, Tristan da Cunha

Nesospiza finches are a classic example of a simple adaptive radiation, with two ecologically distinct forms confined to the Tristan da Cunha Archipelago, South Atlantic Ocean: an abundant, small-billed dietary generalist, and a scarce, large-billed specialist. These have segregated into two species at Nightingale Island, but there is still local introgression between the two forms at Inaccessible Island. We describe the phenology and breeding behaviour of the two sympatric species at Nightingale Island (2.6 km2): Wilkins’s Finch Nesospiza wilkinsi (Endangered) and Nightingale Island Finch N. questi (Vulnerable). The finch breeding season starts in late October-November but the onset of breeding varies by 4–5 weeks among years. The small-billed Nightingale Island Finch typically (two of three study seasons) starts breeding 1–3 weeks earlier than the large-billed Wilkins’s Finch, unlike at Inaccessible Island where the Wilkins’s Finches start breeding first. Laying of initial clutches was quite well synchronised, peaking 1–2 weeks after the first nests were found. Females constructed the nests, which were mostly (>90%) in dense Spartina arundinacea tussock grass stands and occasionally in ferns or sedge grasses. Clutches comprised one or two eggs, with no difference between Wilkins’s (1.66 ± 0.48) or Nightingale Island finch clutches (1.71 ± 0.46). Incubation periods averaged longer for Wilkins’s Finch (18.3 ± 0.5 d) than Nightingale Island Finch (17.7 ± 0.5 d), but this difference was not statistically significant. Females incubated the eggs, and were fed by the males. The difference in egg volume within two-egg clutches was 2–13% for Wilkins’s Finches (mean 5.9 ± 3.3%) and 1–19% for Nightingale Island Finches (mean 8.4 ± 5.3%). At least 31% of pairs re -laid after their first breeding attempt failed but there was no evidence of double brooding. Repeat nests were 0–20 m (mean 5.6 ± 4.9 m) from the initial nest site and inter-seasonal nest sites for 38 known pairs were 0–33 m apart (mean 12 ± 9 m). No inter-species pairs or hybrid birds were seen, but two instances of inter-species fledgling provisioning were observed.