Fragment size and the demography of an area‐sensitive songbird

Abstract

Summary 1. Many songbird species are absent from small forest fragments. One possible explanation for this finding is that populations in small fragments decline to extinction because reproductive success is too low to compensate for adult mortality (λ < 1). I tested this hypothesis from 1995 to 1997 by measuring reproductive success, adult survival, and population viability for four banded populations of the Eastern yellow robin (Eopsaltria australis, White), an area‐sensitive songbird. The populations inhabited two small (55 ha, termed S1, S2) and two large (500–1000 ha, termed L1, L2) forest fragments set within an agricultural landscape. 2. Females produced more fledglings but a similar number of independent young per capita in the small compared to the large fragments. Local female survival did not vary with fragment size. 3. Population models were used to calculate the finite rate of increase (λ) from the number of young produced (female yearlings per female per year) and adult female survival rates. Combining data from all four populations, λ = 0·96, indicating that robins produced 4% fewer female young per year than were needed to compensate for adult losses. Combining data from the two small fragments, λ = 1·02, suggesting that populations in small fragments were viable on average. In contrast, λ = 0·85 for the two large fragments combined, indicating that, on average, 15% fewer young were produced than were needed to replace missing adults each year. Averaging across fragments within the same size class masked the variability in population dynamics that existed between the replicates. Considering each of the four fragments separately, there was one potential demographic source (S2, λ = 1·12; L2, λ = 1·23) and one demographic sink (S1, λ = 0·90; L1, λ = 0·46) within each of the fragment size categories. 4. Observed changes in population sizes were similar to those predicted by the population models. Combining all populations, recruitment replaced most but not all missing females, and the observed population size declined by 8% per year. In both the small and large fragments, observed recruitment was insufficient to replace missing females and population sizes declined 5% and 8% per year, respectively. Within each of the fragment size categories, there was one stable population (S2 and L2) and one population that declined in size (S1 and L1, respectively, declined by 14% and 20% per year). 5. Reproductive success, adult survival, and population viability did not vary with fragment size. To explain area‐sensitivity, I suggest that a threshold in fragment size exists below which populations are never viable, but above which populations may or may not be viable depending on factors unrelated to fragment size, such as nest predation rates.