Post-Fledgling Survival of the Cooperatively Breeding Puff-Throated Bulbul (Alophoixus Pallidus)

Abstract.

In the tropics, passerines are expected to have survival rates higher than those in temperate zones, though their post-fledging survival has rarely been quantified. Furthermore, because of increased care by adults, cooperative breeding species should have even higher rates of juvenile survival, but few data are available on such species also. Over three breeding seasons we examined factors relating to post-fledging survival of the cooperatively breeding Puff-throated Bulbul (Alophoixus pallidus) in an evergreen forest in northeastern Thailand. These factors included the presence of helpers, fledgling age, sex, body mass prior to fledging, date of fledging within a year, and variation by year. The probability of surviving the 8-week period of dependency was 0.61 ± 0.09 (SE), higher than most reports from the temperate zone. Survival probability was strongly associated with age though not with the presence of helpers. Weekly survival was lowest during the first week post-fledging (0.73 ± 0.08), while average weekly survival during the 7 subsequent weeks was constant at 0.97 ± 0.01. The other factors tested appeared to have little or no effect on survival. Understanding factors influencing post-fledgling survivorship provides insights to this limiting life-history stage, which is typically characterized by high mortality and affects population dynamics. The survivorship of fledglings we report adds to the body of evidence that the post-fledgling survivorship of tropical passerines is generally higher than that of temperate-zone species. However, fledglings from cooperative breeders may not experience significantly higher survivorship than those from noncooperatively breeding birds in the same system.

Resumen.

Se espera que las aves paserinas tropicales presenten tasas de supervivencia mayores que las de las zonas templadas, pero su supervivencia después del abandono del nido rara vez ha sido cuantificada. Además, se predice que las especies que presentan cría cooperativa deberían exhibir tasas de supervivencia de los jóvenes aún mayores debido al mayor cuidado brindado por los adultos, pero también existen pocos datos sobre estas especies. A lo largo de tres temporadas reproductivas, examinamos los factores relacionados con la supervivencia posterior al abandono del nido en Alophoixus pallidus, una especie que cría cooperativamente en bosques siempreverdes del noreste de Tailandia. Los factores incluyeron la presencia de ayudantes, la edad de los volantones, su sexo, su peso corporal antes de abandonar el nido, la fecha de emplumamiento en un año dado y la variation entre años. La probabilidad de sobrevivir al período de dependencia de ocho semanas fue de 0.61 ± 0.09 (EE), un valor mayor que muchos de los que han sido documentados para la zona templada. La probabilidad de supervivencia se asoció fuertemente con la edad pero no con la presencia de ayudantes. La supervivencia semanal fue mínima durante la primera semana posterior al abandono del nido (0.73 ± 0.08), mientras que la supervivencia semanal durante las siete semanas subsecuentes se mantuvo constante en 0.97 ± 0.01. Los demás factores evaluados parecieron tener poca o ninguna influencia sobre la supervivencia. Entender los factores que afectan la supervivencia después del abandono del nido provee informatión sobre este estadío limitante de las historias de vida, que típicamente se caracteriza por una alta mortalidad y que afecta la dinámica de las poblaciones. La supervivencia de los volantones que documentamos se suma al cuerpo de evidencia que apoya la idea de que la supervivencia posterior al abandono del nido de los paserinos tropicales es generalmente mayor que la de las especies de zonas templadas. Sin embargo, los volantones de especies con cría cooperativa podrían no presentar una supervivencia mayor que aquellos de especies que no crían cooperativamente en el mismo sistema.

Introduction

Realistic estimates of juvenile survival are important for explaining fundamental life-history differences between tropical and temperate-zone species and for understanding population dynamics of birds in general (Sandercock et al. 2000). To date, most studies of avian survivorship have been conducted in temperate regions, while little is known about factors affecting survival and population sizes of tropical birds (Martin 1996). Information from temperate regions, however, is not necessarily appropriate for explaining population dynamics in the tropics because environmental factors generally vary less by season (Martin and Karr 1986) and life histories of tropical species are typically characterized by smaller clutch sizes, lower nesting success, extended parental care, and higher adult survivorship (Martin 1996, Arnold and Owens 1998, Russell et al. 2004, Brouwer et al. 2006).

Because of limited flying skills and the variety of hazards naïve fledglings may encounter (Anders et al. 1997, Suedkamp Wells et al. 2007), birds typically face the highest risk of mortality of their life cycle immediately after leaving the nest (Wheelwright et al. 2003). This stage is also one of the most difficult at which to assess survivorship because juveniles are typically difficult to follow (McGowan and Woolfenden 1990, Magrath 1991, Suedkamp Wells et al. 2007), particularly for non-ground-dwelling species and in closed-canopy tropical forests on rugged terrain. Consequently, little is known about how post-fledging survival contributes to population dynamics overall, particularly in the tropics.

Several factors have the potential to influence juvenile survival. Parental care is clearly important for food provisioning, for learning foraging and predator-avoidance skills, as well as for protection from predators and the elements (Martin 1996, Griesser 2003). In cooperative breeding species it may be amplified by helpers or allofeeders (McGowan and Woolfenden 1989, Caffrey 2000, Doerr and Doerr 2006). Other factors that may be important include age, sex, body condition (mass), and timing of fledging (both intra- and interannual variation). Age may be particularly important; older, more developed, and experienced birds are more likely to survive than younger ones (Colwell et al. 2007). In addition to age, sex-biased survival of offspring is likely to result in biased sex ratios of adults and may have several ecological and evolutionary consequences (Promislow et al. 1992), although it has rarely been demonstrated in fledglings (but see Green and Cockburn 2001). Limited available data suggest that body mass at fledging is another possible determinant of post-fledging survival, with heavier birds' survivorship increased (Tinbergen and Boerlijst 1990, Magrath 1991). Additionally, studies have demonstrated that the timing of fledging, early or late in the breeding season (Krementz et al. 1989, Verboven and Visser 1998, Naef-Daenzer et al. 2001) as well as the year of fledging (Colwell et al. 2007), correlates with juveniles' survivorship.

We test whether helpers (more than two adult birds caring for young), age, sex, body mass of nestlings prior to fledging, fledging date within a year (Julian date), and year of fledging play a significant role in post-fledging survivorship in the Puff-throated Bulbul (Alophoixus pallidus). We evaluated ecological models to explain patterns of survival after fledging over 3 years in Khao Yai National Park, Thailand. We focus on this species because it breeds both cooperatively and noncooperatively at a single site, is relatively common in evergreen forest, and individuals can be followed. We are fundamentally interested in understanding the population dynamics of its unusual breeding system.

Methods

Study Area and Study Species

The study was conducted in Khao Yai National Park, northeastern Thailand (14° 26′ N, 101° 22′ E) on the 30-ha Mo-Singto Long-term Biodiversity Research Plot (Brockelman 1998). The plot is in a mature, seasonally wet evergreen forest with undulating ridges and valleys at 723–817 m in elevation (Brockelman 1998). The average annual rainfall is 2554 mm, most of which falls between May and October; the dry season occurs from November to February (∼7% of days with rain). The vegetation is dominated by evergreen trees. The canopy layer is approximately 30–35 m tall, with emergent trees of more than 40 m. The lower story is dominated by Polyalthia evecta (3–5 m), the ground-cover layer with dense patches of Strobilanthes sp. (1–2.5 m). Rattans (Daemonorops spp. or Calamus spp.) are abundant along wet gullies and distributed through most of the plot.

The Puff-throated Bulbul (family Pycnonotidae), classified in a genus recently split from the African genus Criniger (Pasquet et al. 2001, Moyle and Marks 2006), is a regular though not an obligate cooperative breeder (Pierce et al. 2007). Of the 141 species of Pycnonotidae worldwide, only seven have been documented as cooperative breeders, although an additional 13 species are either suspected or inferred to be cooperative (Cockburn 2006). Puff-throated Bulbuls live in social groups of two to seven individuals, defend territories year round, and breed both cooperatively and noncooperatively on the Mo-Singto plot (Pierce et al. 2007). We never observe more than two breeding females occupying a single territory. However, all adults participate in foraging and territory defense. Groups can comprise single or plural breeding pairs. In cases of plural breeders, the birds nest separately within a single territory throughout the breeding season.

Identification of helpers is based upon a number of cues. Adults that are present but not nesting separately within a group or territory are defined as helpers when they contribute to nestling and fledgling care. Such behavior is verified through direct observation during nesting and fledging. In territories with plural breeders, birds may help the other pair if their own nests fail. In such cases we defined these birds as helpers also. In the absence of DNA fingerprinting by which actual parents of offspring could be identified, we assume parents on the basis of behavior of males around females known to be responsible for a given nest and/or offspring. Only females contribute to nest building, incubating, and brooding of young, while males guard their females closely. Females build open cup-shaped nests, generally in the understory (0.5–6 m high), although a small proportion (<5%) of nests are built >15 m above ground (Pierce et al. 2007). Most females attempt a second brood when the young from their first brood are 3–5 weeks out of the nest. Breeding males and helpers join the females in provisioning the young during both the nestling and fledgling stages. All adults in the group share in defending the territory, alarm calling and predator mobbing, and hence are considered helpers if they are not the presumed parents of fledglings. During foraging groups move widely between the understory and subcanopy layers.

From 2006 to 2008 we monitored Puff-throated Bulbuls occupying territories within and adjacent to the 30-ha permanent plot. We monitored each territory at least once per week through the breeding season (February–July) and monthly in the nonbreeding season (August–January). We walked the entire 30-ha study area exhaustively to ensure no new groups or breeding birds were missed. Twenty-eight groups were present and monitored in 2006, 31 in 2007, and 32 in 2008.

Since 2003, Puff-throated Bulbuls have been banded and censused regularly on the Mo-Singto plot, though not all individuals were banded. We banded captured birds with a unique combination of two or three colors and one Royal Thai Forest Department aluminum band with a unique identifying number. Using binoculars, we followed and identified individuals from at the beginning of the 2006 breeding season to record their foraging locations and encounters with other groups. We mapped encounter locations as territory boundaries (e.g., Woolfenden and Fitzpatrick 1984, Carmen 2004). To determine group size we recorded the number of individuals and unique color band combinations detected at each encounter. We considered individual birds as part of a group after detections and resightings were consistent for >2 months.

We searched for active nests of all breeding individuals throughout the entire breeding season during each year of the study (minimum monthly effort 370 observer-hours) (n = 98, 152, and 74 nests in 2006, 2007, and 2008, respectively). We monitored all known breeding females at least once per week to determine their nesting stage (i.e., building, laying, incubating, or nestling), and we located a majority of all active nests and a minimum of 95% of all successful nests (those fledgling at least one young) each breeding season. We checked active nests regularly (every 2 or 3 days during the egg stage; every 3 days during the nestling stage) until failure or fledging. We banded and weighed nestlings at an age of 8 or 9 days (2 or 3 days before fledging). During the nestling stage, we observed feeding at the nests to determine whether adults that were not presumed parents were helping to provision. Over the 3 years of the study, eight nests fledged young from heights inaccessible for banding chicks, nine nests were found after fledging, and for three fledged broods no nest was found prior to fledging.

We defined fledging date as day 0 for survival history. After day 0, we monitored fledglings once per week until the end of week 8. Each week we searched for fledglings of each group for up to 2 hr per group by searching throughout the territory. We recorded the birds' histories with 1 for observed and 0 for not seen. During our monitoring, we recorded if the fledglings were fed by adults to confirm helping by extra adults. Although fledglings started to feed partly independently at 3–4 weeks, they are fed by adults until weeks 7–8. After 8 weeks, we assumed juveniles to be feeding independently, even if they remained with groups of adults within the parental territory.

In cases where juveniles were unbanded, during the 8 weeks of dependency they could be distinguished from adults by bill and plumage color. The bill of juveniles is pale yellow; that of adults is grayish. Juveniles' plumage is browner than that of adults, which are uniformly olive-brown. Juveniles begin to molt at approximately 4 weeks, so contrast between the olive new feathers and the brown old feathers is apparent well beyond 8 weeks. We constructed capture—resighting histories for 87 fledglings from 41 broods (23 fledglings in 2006, 37 in 2007, and 27 in 2008).

Because male and female Puff-throated Bulbuls are visually indistinguishable, sexing in the field is not possible. Therefore, when possible, we collected approximately 10–15 µL of blood from the tarsus vein or brachial vein of nestlings or fledglings. Blood was stored in 1 mL lysis buffer for DNA extraction and subsequent polymerase chain reaction and analysis to determine sex (Griffiths et al. 1998).

Data Analysis and Model Development

We developed models to test hypotheses on the relative importance of (1) presence/absence of helper(s), (2) age, (3) sex, (4) year, together with two continuous covariates (nestling body mass and fledging date) on the probabilities of survival and resighting. Survival probability, φ, is defined as the probability of a fledgling surviving from one week to the next, whether or not it was resighted. Resighting probability, p, is the probability that a bird will be resighted, given that it is alive. We estimated weekly survival probabilities to fit Cormack—Jolly—Seber models in program MARK (White and Burnham 1999, Cooch and White 2008) to generate maximum-likelihood estimates of survival probabilities. We used nine encounters (week 0 through week 8) to analyze post-fledging survival. Because not all birds were sexed, our data set was restricted to 29 females and 25 males, limiting our ability to analyze several variables simultaneously, so we pooled data where appropriate. Specifically, in MARK, we tested the six variables separately against time in relation to survival. This test indicated that annual variation had the lowest likelihood weight (1%), whereas the other variables had likelihoods >5%, and that the survival probabilities did not differ by year. Therefore, we pooled the three years of data for subsequent analyses.

For the initial analysis we used the global model, which incorporates the most parameters: φ(g × h × t) p(g ×h × t) (Lebreton et al. 1992), where g, h, and t refer to probabilities dependent on sex, presence/absence of helper(s), and time (week), respectively. We tested goodness of fit by parametric bootstrapping (1000 simulations) from the fully time-dependent global model. The variance-inflation factor (ĉ) was calculated as the observed deviance (global model's deviance) divided by the expected deviance simulated from parametric bootstrapping (Cooch and White 2008). A ĉ ≤ 3 suggests that there is little overdispersion in the data (Lebreton et al. 1992). Values of Akaike's information criterion corrected for small samples (AIC_C) were adjusted to allow for the extent of overdispersion, measured by ĉ, to derive a quasi-likelihood (QAIC_c) (Cooch and White 2008).

Because of data limitations, we reduced the number of possible models by fixing the best-fitting resighting model. We selected the most parsimonious resighting model by running the survival models in saturated form, that is, φ(g × h × t) (Githiru and Lens 2006). The most parsimonious model is the one with the lowest QAIC_c or ΔQAIC_c of zero (Lebreton et al. 1992), and models with a difference (ΔQAIC_C) ≤2 are as parsimonious as the best-fit model (Burnham and Anderson 2002). Nested models with ΔQAIC_c ≤ 7 were further tested with likelihood-ratio tests (LRT) in MARK. We used the LRT to test the significance of various effects on the survival and resighting probability with contrasting pairs of nested models, one containing the effect of interest with one where it is omitted (Lebreton et al. 1992). A difference in QAIC_C of >7 is generally considered as strong evidence that the models in question are different (White and Burnham 1999).

We developed and compared survival models on the basis of one or two variable hypotheses focusing on the five main effects (helpers, age, sex, mass, and fledging date) and the interaction between age and each of the four other variables. The survival models were run with the most parsimonious resighting model selected by the methods above. We used logit-link functions and standardized the values of the covariates. We modeled potential age effects (where survival probability increases with age) by grouping the eight 1-week intervals into several possible categories. The time categories we tested included all 8 weeks (full-time dependence), week 1 against weeks 2–8, weeks 1 and 2 against weeks 3–8, and so forth. The best-fitting age-class model separated week 1 (a₁) from weeks 2–8 (a₂). Therefore, we added (a₁, a₂) into the model set as reduced time parameters and modeled these against the other variables of interest. We used the body mass of nestlings when banded at an age of 8 or 9 days (range 18.0–28.5 g) and tested for sex differences in mass with the Mann-Whitney U-test. We excluded from the analyses nestlings weighed at day 10, day 11, or after leaving the nest. We used a Spearman rank correlation to further test for a possible correlation between fledging date and nestling mass.

Survival probability during the dependency period was estimated as a product of the eight weekly survival estimates (cumulative survival), and the variance of the cumulative 8-week survival estimate was obtained by the Delta method (Seber 1982, Cooch and White 2008).

Results

Seventeen of 23 fledglings (74%) in 2006, 24 of 37 in 2007 (65%), and 22 of 27 (82%) in 2008 survived to independence, resulting in an actual survival rate of 72%. The goodness of fit indicated only minor overdispersion of the data (ĉ = 1.67), suggesting our models fit the dataset adequately (Lebreton et al. 1992). On the basis of the 54 individuals used in the analysis, post-fledging survival probabilities were similar each year. The cumulative survival probabilities for years 2006, 2007, and 2008 from model averaging were 0.61 ± 0.10, 0.61 ± 0.09, and 0.60 ± 0.08, respectively. Thirty-three of the 54 (61%) birds of known sex survived the 8 weeks to independence, equal to the cumulative 8-week survival probability (± SE) estimated in MARK (0.61 ± 0.09).

Resighting probability. The best-fitting model of resighting probability was associated with the interaction between presence/absence of helper(s) and time (model 1; Table 1). This model was more supported than the helper- or time-dependent models alone or other tested interactions based on the QAIC_c weights (Table 1). The LRT also suggest the interaction between presence/absence of helper(s) and the time-dependent model was significantly different from the other resighting models tested (Table 1).

Presence/absence of helper(s). The presence/absence of helper(s) alone (model 8; Table 2) had no support in the survival models. On the basis of QAIC_C weights a helper effect contributed 9% to an interaction with age (model 4; Table 2), but on the basis of the LRT this model was not different from the most supported model of age alone (model 1; Table 2). The cumulative survival probabilities by the end of the dependency period were 0.62 ± 0.10 and 0.60 ± 0.13 for fledglings in groups with helper(s) and no helpers, respectively.

Time-dependent effect on survival. The greatest effect on fledglings' survival was related to time since fledging (age) (models 1–6; Table 2). Survival estimates strongly supported the age-effects model (model 1; Table 2), in which the average survival during the first week (a₁) was 0.73 ± 0.08, while the average weekly survival probability for weeks 2–8 (a₂) was constant at 0.97 ± 0.01. This result suggests that mortality was more likely during the first week after fledging. On basis of the QAIC_C weights this age-effects model was >1.4 times more supported than the full time-dependent model (model 2; Table 2).

Sex. An effect of sex alone was not supported (model 12, Table 2). In combination with age, sex received a moderate level of support (model 6, Table 2), but according to the LRT this model did not differ from the age model alone. Therefore the model's likelihood was probably the effect of age rather than that of sex. The survival probabilities during the first week after fledging were 0.76 ± 0.11 and 0.70 ± 0.11 for males and females, respectively. The cumulative survival for the 8-week post-fledging period derived from model averaging suggested that males tended to survive better than females, 0.68 ± 0.10 and 0.56 ± 0.12, respectively.

Individual covariates. Neither of the individual covariates had clear effects on survival. The body mass of nestlings contributed approximately 6% of the QAICc weights, but only when associated with age (model 5; Table 2). Body mass alone received no model support (model 10; Table 2). The model of age effects associated with fledging date contributed 11% to the relative weights (model 3; Table 2); however, fledging date alone received no model support (model 9; Table 2). The LRT suggested no significant difference between the age model (model 1; Table 2) and the age—mass and age—fledging date models (models 3 and 5; Table 2). We found no correlation between fledging date and body mass of nestlings (r_s = 0.15, P = 0.16). The body mass of male nestlings (mean ± SD = 23.11 ± 2.14; n = 19) did not differ from that of females (mean ± SD = 23.36 ± 2.30, n = 21; Mann—Whitney U-test, z = -0.34, P = 0.73).

Discussion

During the 8 weeks after fledging the Puff-throated Bulbul's probability of survival was relatively high (61%). The cumulative survival probability by the end of the period of post-fledging dependency is similar to that of the Seychelles Warbler (Acrocephalus sechellensis; 0.61; Brouwer et al. 2006), White-throated Thrush or Robin (Turdus assimilis; 0.67; Cohen and Lindell 2004), and Brown Thornbill (Acanthiz apusilla; 0.64; Green and Cockburn 2001), all of which are tropical or South Temperate Zone species. This higher survivorship is notable in comparison to post-fledging survivorship rates detected in passerines of the North Temperate Zone: 0.42 in the Wood Thrush (Hylocichla mustelina; Anders et al. 1997), 0.36 in the Lark Bunting (Calamospiza melanocorys; Yackel Adams et al. 2006), and 0.13 in the Great Tit (Parus major; Greño et al. 2008), though these studies relied on different criteria and different time scales (i.e., daily, weekly, or whole dependency period). However, not all studies in the temperate zone indicate lower post-fledging survival rates during the dependency period (Vega Rivera et al. 1998).

The increased survivorship of juvenile birds in the tropics and South Temperate Zone is often explained by the life-history traits of smaller clutch sizes and prolonged parental care (Martin 1996, Russell et al. 2004). A growing body of evidence suggests that tropical and South Temperate Zone birds have life histories “slower” than those in the North Temperate Zone, i.e., smaller clutch sizes, extended parental care, and higher adult survival rates (Russell et al. 2004); the Puff-throated Bulbul clearly fits this paradigm. In addition, Martin (2002) suggested that adult survival is the primary driver in explaining increased parental care in southern and tropical birds. Accordingly, we have found high rates of survival rates of adult Puff-throated Bulbuls (Sankamethawee, unpubl. data).

Studies of cooperatively breeding birds report both positive (Magrath 2001, Conner et al. 2004, Woxvold and Magrath 2005, Valencia et al. 2006) and negative effects (Brouwer et al. 2006) of group living on reproductive success and survival. In the Apostlebird (Struthidea cinerea), fledgling production and annual recruitment appear to result from the presence of active helpers rather than being associated with group size per se (Woxvold and Magrath 2005). The effects of helpers are probably not restricted to food provisioning (Woolfenden 1978) but include increased protection from predation by alarm calling, mobbing, and warning of approaching predators for fledglings to find better cover (McGowan and Woolfenden 1989). Our results indicate that the presence of helpers or allofeeders in the family group resulted in no discernible advantage with respect to post-fledging survival. Although helper/allofeeder Puff-throated Bulbuls feed fledglings more frequently than they do nestlings (P. Wonkson, unpubl. data), and additional food may improve fledglings' health, the contribution may be difficult to detect and is likely to be minor compared to predation risk, particularly during the first week after fledging.

Sex-dependent survival during the post-fledging period has rarely been demonstrated, and nearly all studies that have found sex biases in survival addressed nestlings (Røskaft and Slagsvold 1985, Arroyo 2002) or adult birds (Grüebler et al. 2008). Our data add to a rarely reported aspect of the post-fledging period. However, we found that sex was at best weakly associated with post-fledging survival. Although sex alone had no model support, survival during the first week as well as the cumulative survival for the entire 8 weeks of dependency appeared to be slightly higher for males than for females. Sex-associated effects were also found in one of the few other studies of this topic: male Brown Thornbill fledglings, which are generally heavier, are more likely to survive to independence than lighter female fledglings (Green and Cockburn 2001). In the Puff-throated Bulbul, however, we found no significant difference between the body mass of male and female nestlings, also indicating that nestling mass alone was not strongly related to survival. Behavioral differences could account for survival difference, as Vergara and Fargallo (2008) suggested that in the Eurasian Kestrel (Falco tinnunculus) male fledglings are more competitive during the post-fledging period, which may allow them to obtain relatively more food and in turn enhancing their survival. Other factors could influence dominance and survival, but we currently have no data regarding within-group dominance hierarchies, territory quality, or hatch order.

Our model results suggested no clear relationship between nestling mass and fledgling survival. Others (Garnett 1981, Tinbergen and Boerlijst 1990, Monrós et al. 2002), however, have shown that heavier nestlings survive at higher rates during the post-fledging period. Nestling body mass may reflect body condition from the nestling stage through to the post-fledging stage (Green and Cockburn 2001, Vergara and Fargallo 2008). Similarly, Yackel Adams et al. (2006) reported that predation played an important role in reducing survival of the lightest young. Although our study yielded no strong result, 46%) of fledglings lighter than the mean weight survived to independence whereas 73% of those heavier than the mean survived.

Reproductive success is often influenced by the timing of breeding (Young 1994). Like Magrath (1991) and Tinbergen and Boerlijst (1990), we did not find a clear difference in survival related to fledging date, although others have demonstrated such a link (Krementz et al. 1989, Verboven and Visser 1998, Naef-Daenzer et al. 2001, Wheelwright et al. 2003). However, 70% of the individuals that fledged earlier than the mean Julian date survived to independence, whereas 51% of those fledged after the mean survived, hinting at a possible effect. Puff-throated Bulbuls can renest up to seven times in a single breeding season, as we have observed, and, in general, adult birds presumably spend significant energy on multiple nesting attempts because of high rates of nest predation (Pierce and Pobprasert 2007). If effort toward renesting reduces the breeding performance of adult birds as the breeding season progresses (Monrós et al. 2002, Borboroglu et al. 2008), it would not be surprising if young produced later in the breeding season were in poorer condition and hence suffered reduced survivorship.

Reproductive success and survival may also have been the result of factors we did not measure, particularly of the quality of the adults or territory (e.g., Ligon and Ligon 1990, Cockburn 1998). These qualities are both difficult to assess and have rarely been demonstrated to correlate directly with reproductive success, most likely because of confounding factors (Cockburn 1998, Legge 2000, Woxvold and Magrath 2005). For example, we observed two female Puff-throated Bulbuls that raised three young successfully to independence in each year of the study—regardless of mate or assistance from helpers. On the basis of an average fledging success of 0.08, three young per brood, and post-fledging survival probability of 0.61, our data suggested that the probability of a female raising one young to independence in a given year was 0.21. Therefore, the probability of a female raising three young per brood three years consecutively was 0.0005, suggesting that only one out of 1863 females is able to do this by chance. Although quantifying the overall quality of adults or territories across a population through space and time is challenging, other studies and systems surely provide numerous examples of individuals that produce exceptionally high numbers of offspring.

The critical period for survival appears to be during the first week after birds leave the nest. This finding is supported by other studies in both the North Temperate Zone (Sullivan 1989, Davies and Restani 2006, Berkeley et al. 2007) and tropics and South Temperate Zone, where high mortality rates are attributed to limited powers of flight (e.g., Green and Cockburn 2001, Cohen and Lindell 2004) that result in higher rates of predation (Sullivan 1989, Keedwell 2003, Berkeley et al. 2007). In our study we seldom found direct evidence of predation (two juveniles killed across all three years). However, we suspect that predation is a key source of fledgling mortality because young birds are less capable of sustained flight and are naïve with respect to predators. In temperate regions starvation can also result in juvenile mortality (e.g., Anders et al. 1997, Yackel Adams et al. 2006), but at our site, from data on rainfall and fruit phenology (Sankamethawee, unpubl. data) fruit scarcity did not appear to be an issue during nestling and post-fledgling periods.

One limitation of our study was the relatively small number of fledglings, despite our finding >300 nests, due to characteristically high predation rates in the tropics (Stutchbury and Morton 2001). In addition, the undulating terrain and tall canopy make following birds challenging and may result in juveniles being missed during a particular week. However, given the amount of time spent in the field (>370 observer-hours per month) during the breeding season, we are confident we followed birds through to off-plot dispersal or disappearance.

Our research provides insights into the breeding biology and factors influencing post-fledging survivorship of tropical birds, particularly cooperative breeding species in the little-studied Asian tropics. In the Puff-throated Bulbul we observed small clutch sizes, low nesting success, extended parental care, and high juvenile and adult survival, consistent with traits reported for other passerines in the tropics and South Temperate Zone. Still unanswered are some aspects of this species' life history such as territory-turnover rate and lifetime reproductive success, which are likely correlated with adult survival, prolonged parental care, and juvenile dispersal. Furthermore, such detailed or even rudimentary survival data concerning nearly all other species sympatric with the Puff-throated Bulbul have never been recorded. Therefore, long-term studies of the same population with comparisons to other sympatric bulbuls would be particularly informative for understanding life-history tradeoffs in tropical passerines.

Acknowledgments

Research funding was provided by the Biodiversity Research and Training Program (BRT_T350006) and the Royal Golden Jubilee Ph.D. Program (PHD/0240/2547). A special thanks to J. Khoonwongsa for her full-time help in the field. We thank A. J. Pierce and P. D. Round for supervision and collaboration in bird banding. We also thank all who supplied additional field data, especially D. Khamcha, S. Tanasarnpaiboon, K. Pobprasert, and N. Sukumal. Many thanks to P. Wonkson for helping with DNA sexing. We thank the Department of National Parks, Wildlife and Plant Conservation for permission to conduct the research and the Khao Yai training center for accommodation. We thank D. Ngoprasert for helping with data analysis and T. Savini and V. Chimchom for comments and suggestions throughout this project's development. Finally, we thank D. Westcott, G. Harrington, A. J. Pierce, R. L. Curry, and an anonymous reviewer for constructive criticism and valuable comments on the manuscript.

Literature Cited