Breeding of White-tailed Tropicbirds ( Phaethon lepturus ) in the western South Atlantic

Basic information on natural history is crucial for assessing the viability of populations, but is often lacking for many species of conservation concern. One such species is the White-tailed Tropicbird, Phaethon lepturus (Mathews, 1915). Here, we address this shortfall by providing detailed information on reproductive biology, distribution and threats on the Fernando de Noronha archipelago, Brazil – the largest colony of P. lepturus in the South Atlantic. We assessed reproduction from August 2011 to January 2012 by monitoring tropicbird nests and their contents. A population estimate was obtained through a combination of active searches for nests and by census at sea between 2010 and 2012. Breeding success was calculated by traditional methods. The growth curve of chicks and life table were also calculated. Additional information on nest and mate fidelity and on age of breeding birds was obtained from the banded birds. Our results indicate that the unusual nest form (limestone pinnacles) and predation by crabs may be responsible for the observed patterns of hatching and fledging success. Although the Fernando de Noronha population appears to be stable (at between 100-300 birds), a long term monitoring program would be desirable to assess fluctuations in this globally important population. Conservation strategies should focus on controlling predation by land crabs and tegu lizards.


Introduction
Tropicbirds (order Phaethontiformes) are mid-sized seabirds distributed in tropical and subtropical regions (Orta, 1992).They have a complex life-cycle, using pelagic areas for feeding and oceanic islands for breeding (Orta, 1992).Tropicbirds are represented by a single genus, Phaethon Linnaeus, 1758, currently composed of three species: red-billed tropicbird, Phaethon aethereus Linnaeus, 1758, red-tailed tropicbird, P. rubricauda Boddaert, 1783 and white-tailed tropicbird, P. lepturus Daudin, 1802 (Orta, 1992).In South Atlantic, there are records of all three species, but only the red-billed tropicbird and the white-tailed tropicbird have been observed to breed on Brazilian islands (Couto et al., 2001;Piacentini et al., 2015).White-tailed Tropicbirds (hereafter WTTBs) breed on two Brazilian archipelagos, Fernando de Noronha and Abrolhos (Figure 1).The former is the largest colony of WTTBs in the South Atlantic outside of the Caribbean sea with the species less abundant on the Abrolhos archipelago (Schulz-Neto, 2004).
Among birds, colonial seabirds are particularly prone to predation due to their high breeding densities within small areas (Butchart et al., 2004) combined with their limited mobility on the ground (Ricklefs, 1996;Jones et al., 2008).These vulnerabilities make pelagic seabirds one of the most threatened groups of birds (Croxall et al., 2012), and the reproductive period is the most vulnerable period of their life cycle.As in other seabirds, the breeding cycle, breeding site selection and nest location of WTTBs are influenced by climate, food availability, density and other biotic an abiotic factors (Phillips, 1987;Ramos et al., 2005;Catry et al., 2009).Breeding success in colonial seabirds is particularly sensitive to fluctuations in food availability (Weimerskirch, 2001) and predation intensity (Whittam and Leonard, 2000;Nordström et al., 2004).In the case of Tropicbirds, predation by exotic species (mainly rats -Gaston and Jones, 1998;Sarmento et al., 2014) and native species (mainly large crabs -Phillips, 1987;Schaffner, 1991) is the main cause of mortality during breeding.
Althoughthe WTTB is not considered as globally threatened species (BirdLife International, 2012), in some localitiespopulations are declining (Bermuda, Cuba, the Cayman Islands, Puerto Rico, and Jamaica) and face a range of threats (Orta, 1992;Lee and Walsh-McGehee, 2000).In Brazil, the WTTB is listed as a nationally threatenedspecies due to its small population size (Efe, 2008;BRASIL, 2003).Like other seabird species, effective conservation of the small, isolated Brazilian sub-populations is handicapped by a lack of basic information on their natural history (Bartholomew, 1986) and reproductive success (Dearborn et al., 2001).Such information is crucial for accurately assessing the viability of these isolated populations (Weimerskirch, 2001).Knowledge about the breeding biology of Brazilian WTTBs is especially scarce (Oren, 1982;Schulz-Neto, 1995;Alves et al., 1997;Alves et al., 2004;Schulz-Neto, 2004).Here, we address this shortfall by providing detailed information on the reproductive biology, distribution and threats faced by WTTBs on Fernando de Noronha, the largest breeding colony in Brazil.

Study area
Fernando de Noronha is a volcanic archipelago located 400 km off the Brazilian coast (Figure 1; 3° 54'S and 32° 25' W).It is composed of 21 islands and islets and occupies a total area of 26 km 2 (IBAMA, 1990).The average annual temperature is 27ºC and annual rainfall is around 1400 mm.The mean water surface temperature ranges from 28ºC and 30º C and the surface salinity ranges between 35.0 and 37.0% 0 (Macedo et al., 1988).The climate is tropical and has two distinct seasons: a rainy season, from January to August and a dry season for the remainder of the year.The archipelago is influenced by the intense and constant alisios winds, blowing from east to southeast (Batistella, 1996).The archipelago contains cats, rats and the large lizard (Tupinambis merianae) that are known to predate seabirds (Efe, 2008).

Fieldwork
Reproduction of WTTBs in the Fernando de Noronha archipelago was monitored between August 2011and January 2012.Monitoring took place on Chapéu Island, located on the south side of the main island because it had a high concentration of nests.All actives nests were located, labeled, geo-referenced and visited weekly.Nests were monitored carefully to avoid abandonment by adults.Adults, chicks with birth dates recorded and eggs were measured (X ̅ ± SD) with calipers (± 0.01mm) and their weight was measured using a Pesola scale (± 5g).Blood samples of adults was collected and stored on an FTA card for the molecular determination of sex.DNA was extracted according to Boyce et al. (1989) protocol.Molecular sexing was performed using P2-P8 primers designed originally by Griffiths et al. (1998).Amplifications were performed following Nunes et al. (2013).Females were characterized by two bands on the gel, while maleswere characterized by the presence of only one band (see details in Nunes et al., 2013).All adults were marked with numbered metal bands from the CEMAVE/ICMBio (Brazilian Banding Center).
The population was estimated by exhaustivecounts of active nests between August 2011 to January 2012and through census at sea with boats circulating around the islands of the archipelago.Three counts using binoculars were performed in 2010 (6 August -3h15,9 August -2h40 and 25 November -1h).Data was also obtained during CEMAVE/ICMBio expeditions to colony-site of the Fernando de Noronha between 25 November and 05 December 2012 and to Abrolhos archipelago as part of a Monitoring Program coordinatedby AVIDEPA (Brazilian NGO) in August, September and November 2011 and in April, June and July 2012.

Statistical analysis
Breeding success was calculated by dividing the number of hatched eggs by the number of eggs laid (hatching success), and by dividing the number of chicks that departed nests by the number of chicks hatched (fledging success).Average incubation and fledging times were assumed to be 41 and 71 days (Schaffner, 1991), respectively.Nests where the egg was not incubated or which exceeded the maximum incubation time (43 days) were recorded as abandoned.Nests where the eggs or chicks disappeared before the minimum incubation and fledging times (40 and 65 days, respectively, Schaffner, 1991) were considered depredated.
The growth curve of chicks was obtained using the mean values of bill length (measured in a straight line from the tip of the beak to where the feathering starts on the bird's forehead), tarsus length (measured from the inner bend of the tibio tarsal articulation to the base of the toes), wing chord (measured from the bend of the wing to the tip of the longest primary feathers with natural curvature) and weight of chicks with known hatching date.A life table based on 17 WTTB chicks (followed from birth to death) was created using the methods described by Brower and Zar (1998) and Krebs (1999).Such a table can be used to describe mortality rates (or survival) and reproduction by age, and to identifythe most vulnerable ages.Information on nest fidelity (the tendency to return to a previously occupied nest), mate fidelity (tendency topair with their previous mating partner) and age was obtained from birdsbanded in previous years by CEMAVE staff.

Breeding biology
In most islands on Fernando de Noronha archipelago, birds lay only one egg in caves and burrows in cliffs.However, on Chapéu Island, the nests are on exposed ground among limestones pinnacles.On Abrolhos archipelago, WTTB nests were located in burrows on cliffs or burrows directly on rocks.Eggs (n = 43)measured 54.4 ± 2.1 mm in length, 38.1 ± 1.9 mm in width and 41.2 ± 4.5 g.Breeding was continuous, and nests with eggs and chicks were present throughout the study period (Figure 2).
A total of 35 nests were monitored between August 2011and January 2012.Of these, seven were utilized more than oncein the same season.Hatching success was 40.5% and fledging success was 35.3%.Overall breeding success (hatching x fledging) was 14.3%, while 50.0% of the nests were depredated in the egg stage, 9.5% were abandoned in the egg stage and 26.2% were depredated in the chick stage.Incubation time, based on three eggs (with laying date and hatching date known) was 42 days.A single chick (with hatching date and departure date known) remained in the nest for 71 days.During nest monitoring on Chapéu Island, only one predation event was observed -that of a 34 days old chick by a land crab (Johngarthia lagostoma (H.Milne Edwards, 1837)).Feces of the exotic tegu lizard (Tupinambismerinae (Durimél & Bibron, 1839)) were also found on the Island.
The life table indicates that chicks had very high survival at 0-5 days old (93.8%) and 6-10 days old (100%), and that the critical period (q x = 0.273) is between 31 and 35 days old (Table 1).The growth curve of chicks indicates stabilization of tarsus and bill growth and a practically continuous growth of wing and weight until 60 days old (Figure 3).
During November and December 2012, 14 nests were active (with egg or chick) out of the 35 nests that had been monitored the previous year.In addition to these, 11 new nests were recorded.On Abrolhos, three adult birds and two chicks were captured in three nests, in 2011.In 2012, two adult birds and two chicks were captured and one adult bird was recaptured, totaling two active nests (with chick).
In 2011, 10 banded birds were recaptured on Fernando de Noronha archipelago: five males and five females.Of these, one male and one female had been banded as chicks in 2009 and weretherefore breeding at two years old.The remaining eight recaptured birds had been banded as adults in 2009 (n = 4), in 2010 (n = 2) and in 2011 (n = 2).All recaptures occurred at the same locality where the birds had originally been banded, demonstrating some breeding site fidelity.
Between November and December 2012, six birds were recaptured on Fernando de Noronha.Five had been banded in 2011 when they were breeding, and one had been banded in 2009 as a chick.Of these, four were located in the same nests they had previously used.On Abrolhos, one adult bird banded in 2011 was recaptured in the same nest in 2012.

Population estimates
Based on active nests, the population of WTTB in Fernando de Noronha was estimated at 174 adult birds in 2010 and 128 adult birds in 2011.By census at sea, the largest count was 108 individuals recorded in Fernando de Noronha on the 06 August 2010 (other counts were 52 individuals on 09 August 2010 and 18 on 25 November

Discussion
Previous studies indicate that breeding success of colonial seabirds is mainly influenced by climatic variations (Ancona et al., 2011), food availability (Hamer et al., 1993;Dearborn et al., 2001), introduction of exotic species (Russel and Le Corre, 2009) and intra/interspecific competition (Coulson, 2001;Dobson and Madeiros, 2010).In cavity-nesting seabirds, cavity orientation can ameliorate micro-climate effects (e.g.Conner, 1975;Stauffer and Best, 1982).For WTTBs, breeding success has been observed to be affected by nest abandonment, intraspecific combats and predation by rats (Rattus rattus (Linnaeus, 1758) (Garnett and Crowley, 2000;Sarmento et al., 2014) and crabs (Gecarcinus spp.and Ocypode spp.) (Schaffner, 1991;Phillips, 1987).In this context, the unusual nest form (recorded previously only in Christmas Island -Stokes 1988) which exposes birds to both predators and adverse weather conditions may be responsible for the low hatching and fledging success observed in this study.
Estimated hatching and fledging success in Fernando de Noronha was lower than Puerto Rico, Ascension Island and Cousin Island (Table 2).In all these areas nests have traditional characteristics, being formed in burrows and crevices and therefore providing some protection from predators and adverse weather conditions (Stonehouse, 1962;Phillips, 1987;Orta, 1992).It is well known that in some colonies reproductive success varies as a function of habitat and nest-site choice, and that climatic events can affect the breeding success (Hamer et al., 2001).An open nest exposes the egg to predation and inclement weather, and may therefore decrease the probability of hatching (Mallory et al., 2009;Hoegh-Guldberg and Bruno, 2010).
Nest predation is an important and common cause of reproductive failures (Ricklefs, 1969;Yang et al., 2014).We observed several chicks to unexpectedly 'disappear' from their nests on Fernando de Noronha, possibly as a consequence of predation by crabs (Shealer and Burger, 1992).In the study area, chicks were more likely to die after the first month of life -a period in which we observed one chick being predated by a crab (J.lagostoma).Nest defense is often related to stage of breeding cycle (Lack, 1968;Burger,1984;Kilpi, 1987), and the observed increase in vulnerability after the first month of life can be explained by adults spending shorter periods at the nest and longer periods engaged in prey capture (Sommerfeld and Hennicke, 2010), leaving the chicks alone and potentially exposed to crab predators.However, starvation -a known cause of chick mortality (Boersma and Stokes, 1995) -cannot be ruled out, since this period also coincides with asudden increasein growth parameters (see Table 2).
The egg dimensions observed in this study were similar to those described in other isolated populations such as Cocos-Keeling Islands and the Aldabra Atoll (Table 2).
Seabirds usually reach sexual maturity after several years (Hamer et al., 2001).This extended 'pre-reproductive' period may be needed to develop efficient methods of feeding and locating prey (Irons, 1998), for development of mating behavior and claiming a territory (Harrington, 1974;Nelson, 1978;Hudson, 1985), or for attaining full physiological maturity (Ainley, 1978).Previous studies have estimated a very low probability of first breeding of P. rubricauda within the first one to two years of life, increasing in three and four year old birds (Doherty Junior et al., 2004).There are records of WTTBs breeding atfive and six years old (Harris, 1979).In contrast, our results indicate that WTTBs in Brazil can start breeding by the second year of life.Such early breeding may also be a contributing factor to the observed low reproductive success, since young breeding adults may lack breeding and/or foraging experience.
The recaptured individuals in our study indicated some philopatry and nest fidelity.In Fernando de Noronha and Abrolhos all recaptures occurred at the same site as the initial banding, and five individuals were found in the same nest on two consecutive years.Philopatry has been described in many seabirds, especially in monogamous birds with high longevity (Stenhouse and Robertson, 2005;Weimerskirch et al., 2005) such as the WTTB.The absence of islands with adequate sites to breed near these archipelagos (Coulson, 2001) may reinforce this behavior by limiting opportunities to establish new nesting sites.
Population size estimates of WTTBs are problematic due to the difficulty of accessing colonies, finding nests and counting birds, and because it is difficult to determine the optimal timing of visits; seasonally and in relation to the time of day (Lee and Walsh-McGehee, 2000).The WTTB population of Fernando de Noronhahas been estimated at between 100 to 300 birds (Oren, 1984;Schulz-Neto, 1995).Despite the population being listed as stable (Schulz-Neto 2004) and the number of birds counted in our study being within the range of previous population size estimates, there may be considerable temporal (yearly or longer) fluctuations in the number of adults in this population.A long term monitoring program that develops periodic and standardized censuses would therefore be desirable to assess any such fluctuations and identify their potential causes.

Conservation status
The WTTB is listed as threatened by the Brazilian Red List (BRASIL, 2003) due to its low breeding success, to the limited number of breeding islands and the small population size.Interestingly, Johngarthia lagostoma -the crab observed preying on WTTB chicks -is also listed as threatened (BRASIL, 2003;Coelho and Melo, 2008), so any proposed conservation strategies should ideally avoid negative impacts on either species.The tegu lizard, by contrast, is an exotic species that is known to prey on eggs and young birds (Chiarello et al., 2010), and may therefore represent a high threat (mainly on Chapéu Island where WTTBs lay their eggs on the ground).Thus, we suggest immediate actions to confirm tegu predation and, if verified, to eradicate this species from Chapeu Island.
Other exotic species (Rattus ratus, R. norvegicus (Berkenhout, 1769), Felis catus Linnaeus, 1758) are known to prey on WTTB eggs and chicks and are a threat to nests on the main island of the Fernando de Noronha archipelago.WTTBs were once considered common on this island (Oren, 1984), but are now restricted to a few nests.This illustrates the need for better control and eradication of exotic species throughout the archipelago.Moreover, a continuous monitoring program of WTTBs in Fernando de Noronha should be immediately implemented to detect long term trends and to provide early warning of population decline.(Schaffner, 1991) 42.0-55.9%,n = 69-34 79.0%, n = 26 Ascension Island (Stonehouse, 1962) 48.1%, n = 821 63.0%, n = 395 Cousin Island (Phillips, 1987) 61.2%, n = 273 58.8%, n = 167 Aldabra Atoll (Diamond, 1975) 53.2 ± 3.3 39.0 ± 1.7 Cocos-Keeling Islands (Gibson-Hill, 1950) 53 37-39 grants (to GRL).MAE projects was supported by the CNPq (#474072/2010-0).We would like to thank Centro Nacional de Pesquisa e Conservação de Aves Silvestres (CEMAVE/ICMBio), Parque Nacional Marinho dos Abrolhos and Marinha do Brasil for authorizations for work.Fieldwork in Abrolhos archipelago was supported for C.M. Musso and Associação Vila-Velhense de Proteção Ambiental (AVIDEPA) staff to which we are extremely grateful.For discussion and suggestions we thank L. Bugoni and T. Mott.

Figure 2 .
Figure 2. Variation in the number of eggs and chicks of white tailed tropicbird (Phaethon lepturus) while breeding in the Chapéu's island, Fernando de Noronha archipelago, Brazil between August 2011 to January 2012.

Figure 3 .
Figure 3. Growth curves with mean values and error bar (SD) of bill, tarsus, wing and weight of White-tailed tropicbird (Phaethon lepturus) chicks in breeding season 2011, Chapéu's Island, Fernando de Noronha archipelago, Brazil.

Table 1 .
Life table for White-tailed tropicbird (Phaethon lepturus) chicks, born in 2011 in the Fernando de Noronha archipelago, Brazil.n x = Number surviving to age x; l x = Proportion of surviving to age x; d x = Number dying between ages x and x+1; q x = Probability of dying between ages x and x+1; s x = Probability of surviving between age x and x+1.
2010).In Abrolhos, six adult birds were recorded between 2011 and 2012, four of them in three different nests and two flying over the archipelago.

Table 2 .
Egg dimensions (X ̅ ± SD), hatching and fledging success obtained in this and other studies of White tailed tropicbirds (Phaethon lepturus).