Breeding Ecology of the Tylas Vanga Tylas Eduardi in Southeastern Madagascar

Ornithol Sci 8: 15–22 (2009)

SPECIAL FEATURE Birds of Madagascar Breeding ecology of the Tylas Vanga Tylas eduardi in southeastern Madagascar

Masahiko NAKAMURA1,#, Mitsuteru TATENO1 and Hajanirina RAKOTOMANANA2

1 Laboratory of Animal Ecology, Department of Biology, Joetsu University of Education, 1 Yamayashiki-machi, Joetsu-shi, Niigata 943–8512, Japan 2 Department of Animal Biology, Faculty of Science, University of Antananarivo, Antananarivo 101, Madagascar

Abstract Although the family Vangidae provides one of the most striking examples ORNITHOLOGICAL of adaptive radiation in the avifauna of Madagascar, basic information on the breed- SCIENCE ing biology of each species is lacking. To examine the breeding system of the Tylas © The Ornithological Society Vanga Tylas eduardi, a species endemic to Madagascar, we studied the contributions of Japan 2009 made by males and females of two pairs (involving four nests) to nest building, incubating, brooding, and feeding the young. The study was conducted from October 2007 to January 2008 at Ranomafana National Park, situated in southeastern Mada- gascar. Males had dispersed territories and paired with single females. Copulation took place between females and males on whose territories they nested. During the nest-building stage, males and females provided nest materials. We observed parental care at two nests. Both sexes participated in incubating and brooding at one nest, whereas only the female incubated and brooded at another nest. During the nestling period, both sexes delivered food (mainly caterpillars, moths, butterﬂies and crickets) to the nestlings and no helpers were observed. These results suggest that the Tylas Vanga is socially monogamous. We discuss the sex determination based on plumage color and mating system of Tylas Vanga.

Key words Biparental care, Diet, Madagascar, Monogamy, Plumage color

The family Vangidae is endemic to Madagascar (Mizuta et al. 2001). and consists of 14 (Langrand 1990) or 19 species The Tylas Vanga Tylas eduardi belongs to the fam- (Yamagishi et al. 2001). The vangids are best known ily Vangidae and it has two subspecies, T. e. eduardi for their adaptive radiation, which is comparable to of eastern Madagascar, and T. e. albigularis of west- that of the Galapagos finches or of the Hawaiian hon- ern Madagascar (Langrand 1990). The species is eycreepers (Grant & Grant 2008). Yamagishi and strictly forest dependent, and the nominate, eastern Eguchi (1996) observed the foraging behavior of 13 subspecies, T. e. eduardi, occurs in humid evergreen vangid species, and documented the differences in forest at elevations between sea level and 1,800 m their foraging niches. In contrast to foraging ecology, (Langrand 1990). It is sometimes found alone or in little is known about the breeding ecology of each pairs, but is usually observed within mixed-species species (Langrand 1990; Schulenberg 2003), except feeding flocks (Langrand 1990; Morris & Hawkins for the Rufous Vanga Schetba rufa (Yamagishi et al. 1998). They feed on small insects mainly taken by 1995; Eguchi et al. 2002), the White-headed Vanga gleaning in the upper shrub layer (Yamagishi & Artamella viridis (Nakamura et al. 2001), the Helmet Eguchi 1996). Although their feeding ecology is well Vanga Euryceros prevostii (Rakotomanana et al. known, their basic breeding ecology remains un- 2000; Marca & Thorstrom 2000), the Hook-billed known. Only sparse or incidental observations have Vanga Vanga curvirostris (Rakotomanana et al. 2001) been made of two clutches and a cup-shaped nest and the Van Dam’s Vanga Xenopirostris damii (e.g. Rand 1937; Appert 1970; Langrand 1990). The aims of this study were therefore: (1) to pro- (Received 28 March 2009; Accepted 15 April 2009) vide the first detailed information on the breeding # Corresponding author, E-mail: [email protected] ecology of Tylas Vangas, and (2) to determine the

15 M. NAKAMURA et al. type of their breeding system by evaluating the parental investment of males and females throughout the reproductive cycle.

STUDY AREA AND METHODS This study was conducted in the Sahamalaotra area (21°13S, 47°23E; 1,141 m above sea level) of Ra- nomafana National Park, southeastern Madagascar, from October 2007 to January 2008. The period from May to October is the dry season, and from Novem- ber to April is the rainy season in the park. This mon- tane tropical rainforest is composed of primary forest with a high floristic diversity and secondary growth consisting mainly of the introduced Psidium cat- tleianum. The understory is very diverse with many shrubs. T. e. eduardi is common in Ranomafana Na- tional Park (Langrand 1990). To examine home ranges, we walked along trails at random from 0500 to 1000 almost every day during November 2007. When we located individuals, we recorded their positions and behaviors on a map. Home ranges were delineated with curved lines connecting the outermost observation points of each individual. During the study, we found a total of 12 individuals, identified six home ranges, and found four nests. We were able to classify the 12 individuals as belong- ing to two color types: one with a black head, nape, Fig. 1. Male (a) and female (b) Tylas Vanga Tylas eduardi and chin, and white underparts (Fig. 1a), one with a eduardi at nest D (Photos by Masahiko Nakamura). black head, nape, and chin, a white band across the upper chest, and orange underparts (Fig. 1b). We con- (6) prey items delivered to nestlings. An incubation sider the six birds with white underparts to be males session was defined as having started when a focal in- and the six birds with orange underparts to be fe- dividual began incubating eggs, and was considered males based on an observation of copulation in which to have ended when that individual left the nest. Sim- a white-bellied bird mounted an orange-bellied bird. ilarly, a brooding session was defined as having We were able to recognize four individuals attending started when a focal individual began brooding the their nests based on variation in their plumage color nestlings, and was considered to have ended when (color of throat, cheek and tail tip) and the shape of that individual left the nest. If a parent stayed for the bill tip. more than two minutes in the nest with its hatched Four nests (A to D) were found and their heights young, we defined this behavior as brooding. above ground were measured. We identified the sci- Nests were visited at two-day intervals to examine entific name of the nesting trees. The diameter at their status. Direct observations (for a total of 186 h) breast height (DBH) of each of the nest trees was also were made using a 20 spotting scope, and indirect measured. To examine the relative contributions of observations (for a total of 253 h) were made with a the sexes to reproductive efforts, we studied the fol- SONY video Hi8 camera (CCD–TRV96), from a lowing activities: (1) nest building, (2) egg laying, (3) vantage point that allowed a good view of the nest time budgets of incubation (percentage of time that (5–15 m). Direct and indirect observations were usu- individuals spent incubating), (4) time budgets of ally made between 0600 and 1200, sometimes from brooding (percentage of time that individuals spent 0600 to 1800. Data from both direct and indirect ob- brooding), (5) feeding frequency (times/h/brood) and servations were combined for the analysis.

16 Breeding ecology of Tylas Vanga

Calls were recorded at three vanga territories during the study period. All recordings were taken from a distance of 5–10 m using a Sony TCM-500 cassette tape recorder and a Victor MZ-200 shotgun micro- phone. We digitized recordings using Roland R- 09HR recorder (sampling rate 44.1 kHz and sample size 16 bits). Sound files were analyzed using Raven 1.2 (Cornell Laboratory of Ornithology 2004, Ithaca, NY, US). Analysis was conducted on a spectrogram display type ‘Hann’ (filter bandwidth: 124 Hz; size: 512 samples, time grid: overlap: 50%; grid resolu- tion: 5.8 ms, 86.1 Hz, DFT 512 samples). We chose the best example of each call type (i.e. the least de- graded recording with the least background noise) to measure in our study. We calculated the time budgets of song (percentage of time that individuals spent singing). Statistical analyses were performed with StatView 5.0 (SAS 1998). We used a significance level of 0.05 Fig. 2. Sonograms of four calls of Tylas Vanga Tylas ed- and, unless otherwise stated, the reported P-values uardi eduardi. a: territorial call, b: contact call (type A), c: are based on two-tailed tests. In cases, when data contact call (type B), d: alarm call. were not normally distributed, we used nonparamet- ric tests. Mean values in the text are reported with standard errors (SE).

RESULTS 1) Vocalizations Tylas Vangas gave four different types of calls during the breeding season. The first call type was a territorial call usually consisting of three (sometimes two) whistled notes (Fig. 2a), diminishing in volume. The second call, a contact call (type A), was a quiet ‘mobiru mobiru mobiru’ (Fig. 2b), given while the pair moved through the forest. The third call, also a contact call (type B), was a whistled ‘phu’ (Fig. 2c), sometimes a double ‘phu phu’, given by both sexes when changing over at the nest for incubation. The fourth call, a quiet ‘fififi’, was an alarm call (Fig. 2d) repeated by both sexes.

2) Space use We found six home ranges along the trails (Fig. 3), each of which was shared by one male and one female. None of the home ranges overlapped (Fig. 3). Fig. 3. Distribution of nests and home ranges of Tylas Van- Seven confrontations were observed at the periph- gas. Ellipses show home ranges. Triangles with capital letters eries of the home ranges (Fig. 3). All confrontations (A–D) represent nest sites; stars indicate the locations of observed confrontations between individuals, solid circles show involved males chasing males. the song posts used for territorial calls and open circles indi- We found four nests (nests, A, B, C and D). One cate where individuals were observed. Lines connecting two pair built nests A, C and D in a home range (Fig. 3), territorial calls denote countersigning. Thin lines indicate but failed their breeding attempt at nests A and C. trails.

17 M. NAKAMURA et al.

Another pair built nests B in a home range (Fig. 3). 4) Nest building We could not find the nest in other home ranges al- During 119 h of observations, we noted nesting though pair seemed to breed. materials being brought by both males and females to When other bird species (Ashy Cuckoo-shrike nests A, B and C. At nest A, the female brought ma- Coracina cinerea, Long-billed Green Sunbird Nec- terial on 19 (70%) occasions, whereas the male tarinia notata, Madagascar Bulbul Hypsipetes mada- brought material on eight (30%) occasions within gascariensis, Long-billed Greenbul Phyllastrephus 32 h (data pooled over four days). At nest B, the fe- madagascariensis, Madagascar Magpie-Robin male delivered material on 96 (74%) occasions, Copsychus albospecularis and Nelicourivi Weaver whereas the male brought material on 34 (26%) occa- Ploceus nelicourvi) approached the nest, the pair sions, within 29 h (data pooled over three days). At chased them away using alarm calls. nest C, the female brought material on 243 (86%) oc- A Pollen’s Vanga Xenopirostris polleni visited nest casions, whereas the male brought material on 63 B at 0753 on 7 November and sat on the nest. The (14%) occasions, within 58 h (data pooled over eight bird (probably a female) crouched on the nest, fluffed days). At nests B and C where the sample sizes were her body feathers, and fluttered her wings and tail, in large, the female was more engaged in nest building a manner resembling the begging behavior of the than the male: assuming that male and female young of Pollen’s Vanga. She continued to flutter her brought nesting material in equal proportion, there wings while sitting on the nest, but flew away at was a significant bias towards the activities of the fe- 0755. We did not know why the Pollen’s Vanga did male at nests B and C (nest B: G15.9, P0.0001, this. nest C: G59.3, P0.0001). We have no precise data on the duration of nest 3) Nest building. Construction of nest B had already begun The nests were cup-shaped, and each was built in by 5 November and the pair completed building the the fork of a tree (see Fig. 1), usually located 4.2– nest on 12 November. Construction of nest C had al- 6.5 m above the ground (Table 1). Nests A and C ready begun by 10 November and the pair completed were blown down by strong winds, so these were col- building the nest on 20 November. Thus, they spent lected. The exterior nest diameter of both nests was at least 8–11 days constructing their nests. 95 mm. The interior diameter of nest A was 75 mm During the nest building stage, male Tylas Vangas and that of nest C was 70 mm. The exterior depth of sang, and repeated territorial calls several times at nest A was 60 mm, while that of nest C was 50 mm. regular intervals from established song posts (Fig. 4). The interior depth at the center of both nests was 45 However, they stopped singing when their females mm. began incubating their eggs (Fig. 4). Nest material consisted mainly (about 90%) of the petioles of dead leaves, but also included hairs from 5) Copulation and courtship display Palmae leaves, and Asparagus sp. leaves. The exteri- At 1002 on 14 November, during the nest-building ors of the nests were composed of spider webs and mosses (Muscinees), while the interiors were lined with the inflorescences of Panicum sp.

Table 1. Summary of nest tree (scientiﬁc name), nest height (m above the ground), DBH (diameter at breast height) at four nests in Tylas Vangas.

Name Height DBH Nest tree of nest (m) (cm)

A Mascarenhasia arborescens 5.3 4.1 B Mascarenhasia arborescens 4.2 4.3 C Mascarenhasia arborescens 4.5 2.5 D Mascarenhasia arborescens 6.5 7.5 Fig. 4. Song time budgets of male Tylas Vanga at three nests.

18 Breeding ecology of Tylas Vanga period, we observed one copulation near nest C involving the nesting female and her mate. At first, the female courted the male by crouching on a branch near her nest, fluffing her body feathers, and fluttering her wings and tails in a posture resembling the begging behavior of the young, however, she was silent. The male was not observed to perform any sexual display, but rather appeared simply to watch the soliciting female, then approach her from the side and mounted her. No courtship feeding was observed. During the nest-building period, both males and females were seen giving the soliciting display (fluttering wings extended). At nest A, the male once and Fig. 5. Incubation time budgets of male and female Tylas the female twice. At nest B, the male gave the flutter- Vanga at two nests. Open bars represent males; hatched bars ing display, twice and the female once. At nest C, the represent females. male gave the fluttering display four times and the female twice. Both males and females performed the z4.86, P0.0001). The male did not give food to fluttering display at the nest or on a branch near the the incubating female. At nest D, we observed the nest. The display occurred when they changed over pair change over at the nest during incubation 105 during nest-building. times. On 16 of these 105 occasions, the change over was preceded by contact calls (type B) given by the 6) Egg laying and clutch size arriving or the leaving adult, nine times by the male During the egg-laying period, only the females and seven times by the female. brought nesting materials at nests B and D. When We observed the fluttering display when birds their females approached the nest with nesting mate- changed over during incubation at nest D. Out of 105 rials, the males associated closely with them. change overs the male performed the display seven We photographed the contents of nests B and D times and the female 13 times. Both male and female with a camera mounted on an extendible pole. Nests performed the display only when their mate was near B and D contained two eggs. The eggs were pale the nest. The display lasted 10.16.8 s (male; N10) blue, with small black spots. Assuming that interior and 11.04.4 s (female; N13); the difference be- diameter of nests B and D was 72.5 mm, the major tween the length of displays by males and females and minor axes of the eggs were 24.7–25.6 mm and was not significant (z1.78, P0.07). 17.4–17.7 mm, respectively. We were unable to deter- We did not observe the fluttering display at nest B, mine the clutch sizes of nests A and C because the and did not observe courtship feeding at either nest B pairs abandoned their nests before the female laid her or nest D. eggs (nest A) and just after the female incubated eggs The female at nest B began incubating on 13 No- (nest C). vember and the pair was observed delivering food to their nestlings on 29 November . Thus, their incuba- 7) Incubation tion period lasted for at least 16 days. The pair at nest At nest B, only the female incubated the eggs, D began to incubate on 3 December and was first ob- spending a total of 69.22.3% of her time incubating served delivering food to their nestlings on 21 De- (N5 days; Fig. 5). The mean incubation session was cember. Thus, their incubation period lasted for at 27.91.8 min (N57). Males did not bring food to least 18 days. the incubating female, but perched near the nest and watched the eggs during the female’s incubation 8) Brooding breaks (16 out of 57 incubation sessions). Nests B and D contained two nestlings. At nest B, At nest D, both the male and the female took turns the nestlings were brooded only by the female (Fig. incubating, spending a total of 70.54.3% of their 6), and both nestlings were found to be out of the nest time incubating (N10 days; Fig. 5). The female had on 9 December (when the chick was 11 days old). At significantly longer incubation sessions (32.82.1 nest D, the male brooded the nestling on just one day min, N64) than males (13.42.0 min, N16, (Fig. 6). Females spent less time brooding their

19 M. NAKAMURA et al.

Fig. 6. Brooding time budgets of male and female Tylas Vanga at two nests. Open bars represent males; hatched bars represent females. Nestling age is given in days after hatching Fig. 7. Feeding frequency of male and female Tylas Vanga (day 1 is the hatching date). at two nests. Solid circles represent males, open circles represent females. Open squares indicate a male giving food to the young as the nestlings grew (Fig. 6). female, who then fed it to the nestlings. Nestling age is given in days after hatching (day 1 is the hatching date). We observed the fluttering display four times at nest D, but not at nest B. The male performed this display three times on 21 December (when the chick Table 2. List of prey items of male and female Tylas Vanga. Prey items were recorded during the nestling period at two was one day old, see Fig. 6). Before he delivered nests (nests B and D). Figures indicate the sample size, with food to the nestlings, the male perched on the nest percentages in parentheses. and performed the fluttering display. After the female finished brooding, she performed the display on the Prey Male Female Total nest on 23 December (when the chick was three days old, see Fig. 6). Whenever a bird performed this dis- Spider 7 (3.7) 4 (4.4) 11 (4.0) play, its mate was also near the nest. We did not ob- Cockroach 7 (3.7) 2 (2.2) 9 (3.2) Cicada 8 (4.3) 4 (4.4) 12 (4.3) served this display after the chicks had reached five Cricket 13 (7.0) 4 (4.4) 17 (6.1) days old. Grasshopper 1 (0.5) 1 (1.1) 2 (0.7) Dragonfly 1 (0.5) 1 (1.1) 2 (0.7) 9) Feeding frequency Mayfly 1 (0.5) 0 (0.0) 1 (0.4) Both members of each pair brought food to their Fly 7 (3.7) 3 (3.3) 10 (3.6) nestlings during the nestling period (Fig. 7). The Horsefly 6 (3.2) 3 (3.3) 9 (3.2) feeding frequency of females was lower than that of Moth/ Butterfly 14 (7.6) 7 (7.7) 21 (7.6) Caterpillar 38 (20.4) 15 (16.5) 53 (19.1) males early during the nestling period (Fig. 7). This Snail 1 (0.5) 0 (0.0) 1 (0.4) was because the female brooded the nestlings during Frog 1 (0.5) 0 (0.0) 1 (0.4) this period (Fig. 6). When the female was brooding Unidentified prey 82 (43.9) 47 (51.6) 129 (46.4) the nestlings, the male did not deliver food directly to Total 187 (100.0) 91 (100.0) 278 (100.0) the nestlings but gave food to the brooding female who then either fed herself or gave the food to the nestlings. Males sometimes gave food to brooding fe- 10) Nestling diet males. The feeding frequency increased with nestling Throughout 72 h of observation at nests B and D, age but decreased after the chicks had reached 11 we observed 278 prey items being fed to nestlings days old (Fig. 7). (Table 2). It was difficult to identify food items espe- The brooding female at nest D was observed pick- cially during the early nestling period, because the ing up an eggshell on 21 December 2007 and all of parents usually masticated the food before offering it the nestlings left the nest on 5 January 2008. Thus, to their nestlings, and because feeding was very the nestling period lasted at least 15 days. rapid. However, we were able to identify 149 (53.6%)

20 Breeding ecology of Tylas Vanga of the food items. Most prey items were smaller than 2) Breeding system the adult’s bill size, and both male and female gave The dispersion pattern of individuals is a factor in small insects and spiders to the nestlings (Table 2). determining the mating system of a given species, and monogamous birds usually hold a territory that is DISCUSSION shared by both sexes (Davies 1991; Ligon 1999). Tylas Vangas were found to occupy home ranges, to 1) Plumage color of male and female Tylas have individual males paired with single females, and Vanga the single copulation that we observed, occurred be- Langrand (1990) reported that the eastern nomi- tween a male and a female in their home range. The nate Tylas Vanga is sexually dimorphic in plumage home ranges of the six pairs that we observed did not coloration; with the male having a black head, nape overlap, and confrontations between males were ob- and chin, and the female a dark gray forehead and served only at the peripheries of their home ranges front of the head and a whitish chin (see Plate 38 in (Fig. 3). Thus, we interpreted the home ranges as Langrand 1990). He also reported that the male has being territories and these territories appeared to be gray-green upperparts and orange underparts with a defended mainly by males, who responded to the white band across the upper chest extending up the songs of other males. sides of neck. Sinclair and Langrand (2003) de- Our results suggest that male and female Tylas scribed males of the eastern race as having orange Vangas did not contribute equally to overall breeding underparts. They also reported that some individuals activity. During the nest-building stage, both males are almost entirely white on the underparts, with only and females provided nest materials. Both sexes par- a faint tinge of orange. In contrast, Morris and ticipated in incubating and brooding at one nest (Figs. Hawkins (1998) did not describe the plumage color 5 & 6), however, females more frequently engaged in of both sexes in their photographic guide, although parental task (Figs 5 & 6). Both sexes delivered food they inserted a color picture of Tylas Vanga. to the nestlings (Fig. 7). In our study, the plumages of both male and female The degree of investment by each parent differs Tylas Vanga (see Fig. 1) differed from those previ- considerably between social mating systems such as ously described by Langrand (1990), and by Sinclair monogamy, polygyny, and polyandry (Clutton-Brock and Langrand (2003). The plumage of the males in 1991; Davies 1991; Ligon 1999). In monogamous our study was similar to that described for the west- birds, biparental care is common, and males are more ern race T. e. albigularis by Langrand (1990) and by active in territorial defense (Clutton-Brock 1991). In Sinclair and Langrand (2003); i.e. with a black most socially monogamous bird species, females take crown, cheeks, and nape, a white band across the the larger share of brood care by doing most of the upper chest extending up the sides of the neck, and incubation and brooding of the young (Clutton-Brock white underparts . The plumage of the females in our 1991). Our results fulfilled these conditions. In both study was the same as that described for males by the Rufous Vanga (Yamagishi et al. 1995; Eguchi Langrand (1990) and by Sinclair and Langrand 2002 et al.) and the White-headed Vanga (Nakamura (2003). et al. 2001), both also endemic to Madagascar, In this study, we were able to confirm, by observa- helpers have been observed at the nest. However, tion of copulation, that the male (mounting) has white during our study of Tylas Vanga we did not observe underparts and the female (mounted) has orange un- helpers, thus, we conclude that its mating system is derparts. Based on our observations, the bird most ac- best described as socially monogamous, like the Hel- tive in nest building, incubating and brooding was the met Vanga (Rakotomanana et al. 2000; Marca & female (Figs 5 & 6), which is to be expected. It ap- Thorstrom 2000), the Hook-billed Vanga (Rako- pears that Langrand (1990) and Sinclair and Lan- tomanana et al. 2001) and the Van Dam’s Vanga grand (2003) mistakenly described the sexes of Tylas (Mizuta et al. 2001). However, two questions remain: Vanga. However, since there is individual variation in why do both sexes perform the fluttering display plumage color and as plumage varies with age in from the nest building to the nestling periods and birds (Dale 2006), further study is required to confirm why do males sing during the nest building period, our observations. even after they have acquired a mate? Further nest observations are necessary in order to more fully describe the social structure of Tylas Vanga.

21 M. NAKAMURA et al.

ACKNOWLEDGMENTS Ligon JD (1999) The evolution of avian breeding systems. pp 259–286. Oxford Univ Press, Oxford. We are grateful to the Government of the Republic of Mizuta T, Nakamura M & Yamagishi S (2001) Breeding Madagascar for providing permission to conduct research in ecology of Van Dam’s Vanga Xenopirostris damii, an the nature reserves, and to the staff of ANGAP at Ranomafana endemic species in Madagascar. J Yamashina Inst Or- National Park. This study would have not been possible with- nithol 33: 15–24. out the help of Jean de Dieu Rakotoniaina and Pierre Talata. Morris P & Hawkins F (1998) Birds of Madagascar: a Dr. Harison Rabarison helped identify the nest materials. We photographic guide. p 256. Pica Press, London. also express our thanks to Dr. Shigeki Asai and two anony- mous reviewers for their many constructive comments and Nakamura M, Yamagishi S & Nishiumi I (2001) Coop- suggestions. This study was supported by a Grant-in-Aid for erative breeding of the White-headed Vanga Lep- Scientific Research (B) of the Japan Society for the Promotion topterus viridis, an endemic species in Madagascar. J of Science (No. 17405008). Yamashina Inst Ornithol 33: 1–14. Rakotomanana H, Nakamura M, Yamagishi S & Chiba Eu- REFERENCES A (2000) Incubation ecology of Helmet Vangas ryceros prevostii, which are endemic to Madagascar. J Appert O (1970) Zur biologie der Vangawürger (Vangi- Yamashina Inst Ornithol 32: 68–72. dae) südwest-Madagaskars. Orn Beob 67: 101–133. Rakotomanana H, Nakamura M & Yamagishi S (2001) Clutton-Brock TH (1991) The evolution of parental Breeding ecology of the endemic Hook-billed Vanga, care. pp 131–152. Princeton Univ Press, Princeton. Vanga curvirostris, in Madagascar. J Yamashina Inst Dale J (2006) Intraspecific variation in coloration. In: Ornithol 33: 25–35. Hill GE & McGraw KJ (eds) Bird coloration, Vol. II. Rand AL (1936) The distribution and habits of Mada- Function and evolution. pp 36–86. Harvard Univ gascar birds. Bull Amer Mus Nat Hist 72: 143–499. Press, Cambridge. SAS (1998) StatView for Macintosh, version 5.0. SAS Davies NB (1991) Mating systems. In: Krebs JR & Institute Inc, Cary, North Carolina. Davies NB (eds) Behavioural ecology: an evolution- Schulenberg TS (2003) Vangidae, vangas. In: Goodman ary approach. pp 263–294. Blackwell Scientific Pub- SM & Benstead JP (eds). The natural history of lications, Oxford. Madagascar. pp 1138–1143. The Univ of Chicago Eguchi K, Yamagishi S, Asai S, Nagata H. & Hino T Press, Chicago. (2002) Helping does not enhance reproductive suc- Sinclair I & Langrand O (2003) Birds of the Indian cess of cooperatively breeding rufous vanga in Mada- Ocean Islands. p 160. Struik Publishers, Cape Town. gascar. J Anim Ecol 71: 123–130. Yamagishi S & Eguchi K (1996) Comparative foraging Grant PR & Grant BR (2008) How and why species ecology of madagascar vangids (Vangidae). Ibis 138: multiply: The radiation of Darwin’s Finches. pp 283–290. 124–125. Princeton Univ Press, Princeton. Yamagishi S, Urano E & Eguchi K (1995) Group com- La Marca GL & Thorstrom R (2000) Breeding biology, position and contributions to breeding by Rufous diet and vocalization of the Helmet Vanga, Euryceros Vangas Schetba rufa in Madagascar. Ibis 137: 157– prevostii, on the Masoala Peninsula, Madagascar. Os- 161. trich 71: 400–403. Yamagishi S, Honda M, Eguchi K & Thorstrom R Langrand O (1990) Guide to the birds of Madagascar. (2001) Extreme endemic radiation of the Malagasy pp 290–302. Yale Univ Press, New Haven and Lon- vangas (Aves: Passeriformes). J Mol Evol 53: 39–46. don.