Pandanaceae of the island of Yapen , papua ( West new Guinea ) , Indonesia , with their nomenclature and notes on the rediscovery of Sararanga sinuosa , and several new species and records

Eleven species of Pandanaceae are recorded for Yapen Island, Papua, Indonesia, seven of Pandanus, three of Freycinetia, including two new ones, and the rediscovery of Sararanga sinuosa. Except for the latter all others are new records for the island.


InTRoduCTIon
Yapen is one of the islands in the Cenderawasih (Geelvink) Bay in the Indonesian Province of Papua, West New Guinea.The island is about 2 400 km², of which approximately 3/4 is still covered with lavish lowland tropical rainforest, and an area of about 780 km² is protected as the Yapen Tengah Nature Reserve.Despite the magnificent landscape, Yapen compared to its neighbouring island Biak has remained little explored.Since a short visit by Beccari in 1875 (Solms-Laubach 1883) no further explorations on the pandan flora of the island have been made, thus the pandan flora remains largely unknown.This paper describes the results of the most recent exploration in Yapen (Keim et al. 2006a).11 species of Pandanaceae are recorded, of which 3 belong to Freycinetia Gaudich., 2 are new, 7 to Pandanus L.f., and Sararanga sinuosa Hemsl.Except for the latter (Solms-Laubach 1883, Guppy 1887, Hemsley 1894, Stone 1961) the rest of the species are new records.dESCRIpTIon of SpECIES freycinetia 1. Freycinetia allantoidea A.P. Keim, Robustus scandens; infructescentia terminalis vel lateralis, plerumque lateralis, ternatus vel quaternatus; syncarpio allantoido, glaucescenti; stigmata 6. -Typus: A.P. Keim 808 (BO), Indonesia, New Guinea, Papua, South Yapen District, Yapen Island, Sarawandori, 10 October 2006.
Etymology -The epithet name refers to the sausage-shaped appearance of the cephalia.
Habitat -Lowland tropical rainforest at about 100 m altitude.
Vernacular name -Not recorded.
Uses -Not recorded.
However, the possession of the sausage-shaped corky-warted cephalium of this species is unique.The corky-warted surface of the infructescence refers to the appearance of each berry that is rostrate with a rigid and succulent apex.So far no such structure has been found in any other members of the section.Furthermore, the obvious bright yellow bracts straightforwardly distinguish F. allantoidea from F. funicularis and F. rhodospatha, which both have conspicuously red to reddish orange bracts (see Rumphius 1743, Ridley 1916).
Freycinetia lauterbachii is known only from male collections (Warburg 1900a, b).Warburg (1900b) mentioned that this species has many bracts, the outer ones green and the innermost ones pink.Stone (1969) regarded it as a synonym of F. funicularis.
Although known only from immature collections F. pleurantha shares many morphological characters with F. allantoidea, such as the obviously glaucous abaxial surface of the leaf, both terminal and lateral (axillary) infructescences, pale glaucous unripe berries, 6 -8 stigmas (see Merrill & Perry 1940).However, there is no information about the colour of the bracts and the structure of the berries, especially the pileus.
With the dimension of the cephalium (9 by 3 cm), non-succulent ovoid berries, and the 5 or 6 stigmas (Warburg 1900b), F. papuana is morphologically most similar to F. allantoidea.However, there is no information regarding the leaf or infructescence.
Warburg described the infructescence as "syncarps cylindric 9 cm long 3 cm wide, peduncled, peduncle nearly terete, margin little rugulose, 4 cm long 3 mm wide, berries not at all connate and not succulent, ovoid, c. 1.5 cm long, in the middle 1/2 cm wide, apex pyramid-shaped acuminate, below the seeds very much filled".Martelli (1910) stated that the type of F. papuana (Hollrung 218a) consisted of half a syncarp that without doubt belonged to F. lauterbachii.Stone (1969), however, regarded F. papuana as illegitimate as the type would be a mixed collection, partly belonging to F. funicularis.Up to 1966 the Code ruled that names based on discordant elements (mixed collections) must be rejected, from 1972 on this Article (70) has been deleted, and the combination F. papuana is legitimate.Freycinetia funicularis is now the correct name for F. lauterbachii (fide Stone 1968).The reason for this statement is unclear as Warburg (1900b) did not provide any figures and the type consisting of infructescence and fruits in B were destroyed during WW II.No duplicates are known to exist.
Specimen seen.Only known from the type.Habitat -Lowland tropical rainforest to montane forest at 100-2650 m altitude.In Yapen Island F. beccarii is commonly found at 100 m altitude.
Uses -Not recorded.Note -This is a new record for Yapen.The result of this study recognises F. ellipsoidalis, F. elliptica, F. globiceps, and F. nervosa as synonyms of F. beccarii (Table 1).The important morphological features used as distinctive characters for the three species are slight and have proven to be continuous (i.e.not discrete).It is also in accordance with Martelli (1910) and Merrill & Perry (1939)
Uses -Not recorded.Notes -In appearance F. spinifera is extremely similar to the West Malesian F. rigidifolia Hemsl., especially regarding the habit, leaf dimension and the cephalium shape.The two species differ only on 2 morphological characters, the numbers of cephalia and stigmas (Table 2).
By the possession of spiny auricles F. spinifera seems to be a member of sect.Hemsleyella B.C.Stone, to which F. rigidifolia and F. pectinata Merr.& L.M.Perry also belong (see Stone 1968).
Until recently the only member of this section known from New Guinea and adjacent areas was F. pectinata of the Solomon Islands.Despite the same slender habit, spiny auricles and quaternate infructescences F. spinifera instantly differs from F. pectinata in the number of stigmas (Table 2).The number of stigmas in F. spinifera varies from 2 -3 (mostly 2); while in F. pectinata there are 4 -6(-12) (see Merrill & Perry 1939).A study of the isotype of F. pectinata (Brass 3247, BO) showed that F. spinifera is indeed distinct from F. pectinata.Habitat -Cultivated from sea level up to 2 000 m altitude.Never found in the wild.In Yapen Island, the red and yellow varieties are cultivated almost in every part of the South and East districts except in the Menawi area, where only the red variety is planted.
Uses -Leaves are used for mats.Vegetable fat extracted from the pericarp is used as sauce, medicine, and tonic.The cephalia are economically important and sold in the local market.Peekel (1984) and French (1986) under P. englerianus described a usage and method of preparation of the pulp extracted from the pericarp that is similar to the way P. conoideus is treated.
Notes -The presence of P. conoideus in Yapen Island is a new record.It is a good example of a species with a widespread distribution and an outstanding spectrum of morphological variation.The differences between P. conoideus and the numerous species listed above are slight and are merely in the dimensions of the leaves and drupes (Table 3) underrating the obvious similarities in the shape of the cephalia and stigmas.Despite the differences in the size all taxa listed above each have an obvious long cylindrical trigonally (triangular)-shaped yellow to
Uses -Leaves are used for mats and roofing.In Yapen the leaves are used as wrappers for steaming fish.This usage has never been reported before.In the Moluccas the leaves are used in preparing (i.e.cooking) the Sago (Heyne 1927), in which the starch is enwrapped with the leaves and then baked, a practise which is not found in Yapen.The cephalium is not eaten in Yapen, but it is in the New Hebrides (Stone 1967) and elsewhere in the Pacific region (Walter & Sam 2002).The flavour is much like that of coconut meat.
Notes - Martelli (1904Martelli ( , 1913) ) suggests that P. macrocarpus Vieill. is a mixed collection which he split up in P. macrocarpus s.s. and P. veillardii Martelli. Brongniart (1875) did not comment on this, but noted that the description of the fruits did not agree with what he had seen, while in CAEN there was a fruit of P. odoratissimus under this name.He attributed this to a mixup of labels.
The presence of P. dubius in the Andaman Islands and Yapen are new records.In Java in some phalanges in the apical part Table 3 Morphological comparisons on the sizes of leaf, cephalium, drupe and shape of cephalium between Pandanus conoideus and some species listed above as synonyms.
of the cephalium there may be only a single stigma (Keim et al. 2006c).Thus the difference is insufficient to keep the two species apart and P. andamanensium is submerged.Despite the difference in the number of stigmas in a phalange, the stigmas in these two species are arranged in a single row (Table 4).
Actually, Kurz (1869) himself already noted that P. andamanensium is very similar to P. dubius and would differ only by the number of stigmas per phalange, 1-2, while P. dubius would have 2 -3.Although when comparing these two species Kurz cited the protologue of P. dubius, the possession of 2 stigmas in a linear row in P. dubius was not discussed.Warburg (1900b) mentioned that P. andamanensium is very much the same as P. dubius and differs only in the size of the phalange and numbers of stigmas but still kept them as two distinct species.Backer & Bakhuizen van den Brink Jr. (1968) reported that phalanges from individuals found in Java regarded to belong to P. andamanensium were observed with 2 -3 stigmas in a row; thus fitting P. dubius.
The same argument is also applied when placing P. compressus in the synonymy of P. dubius.The arrangement of 3-4 stigmas in a linear row per phalange here (Martelli 1905) is slightly different from P. dubius.Stone (1967) noted that P. compressus is very similar to P. dubius and differs particularly in the number of stigmas (Table 3), and seeds (usually 2-3, while usually solitary in P. dubius), and the position of the seed (basal, while central in P. dubius).In the present study these characters only vaguely distinguish P. compressus from P. dubius.Although it rarely happens P. compressus can also possess a phalange with 1 stigma (Stone 1968), so its range overlaps that of P. dubius.On the other hand, although the phalanges of P. dubius are usually found with a single seed, sometimes they may have more, thus agreeing with P. compressus.
The result of my study indicates that P. bidoer is also a synonym of P. dubius, which is in accordance with the opinions of Warburg (1900b), Koorders (1911Koorders ( , 1913, see also Koorders-Schumacher 1913) and Stone (1972a) but not with Backer & Bakhuizen van den Brink Jr. (1968) who treated P. bidoer as a distinct species.
The main differences between P. bidoer and P. dubius would be in the dimensions of the leaf (200 by 40 cm against 93-500 by 11-14 cm in P. dubius) and phalanges (Table 4), which are regarded here as insufficient.The length (and corresponding width) of the leaves in a single individual of P. dubius may vary from less than 100 up to 500 cm.The size of phalanges from an individual of P. dubius (identified by 2 -3 stigmas in 1 linear row) found in Java can reach 10 -14 by 5 -7 cm (Keim et al. 2006c).Thus P. bidoer cannot be distinguished.
Pandanus yamagutii is also treated here as a synonym of P. dubius.The straightforward similarity with P. dubius is in the number of stigma per phalange (Table 4).The smaller size of the phalange described for P. yamagutii is due to the immature nature of the cephalium (Kanehira 1936).The rest of the characters match P. dubius and even Kanehira mentioned that the species was a member of a section that is recorded chiefly from the Asiatic continent and the Philippines (see Kanehira 1936).Stone (1975a) regarded P. andamanensium as a variety of P. leram, P. leram var.andamanensium.He believed that it represents the wild form of the cultivated P. leram.I am not in accordance with this.Pandanus leram has several straightforwardly distinctive morphological characters, especially the number of stigmas in a row (Table 4).Furthermore, it is a native of the Nicobar Islands.It is only in Ceylon that this species has never been found in the wild.Until data from molecular study become available, the 'polyploidy' theory involving the 'wild' P. andamanensium and the 'cultivated with doubled number of stigmas' P. leram as was suggested by Stone (1975a) is regarded here as tentative, thus best avoided.Based on the morphological data available at present P. andamanensium is undoubtedly the same as P. dubius.Specimen seen.IndonesIa, Papua, Yapen Island, South Yapen District, Menawi, beach, 30 Sept. 2006, A.P. Keim 788 (BO!).Habitat -Mangrove, lowland swampy up to submontane forests from 0 up to around 1600 m altitude.In Yapen abundantly found at around 100 m altitude.
Uses -In Yapen the cephalium is eaten.The usage and method of preparing the fatty substrate extracted from the pericarp is similar to that of P. conoideus Lam.Indeed, in the other areas in New Guinea, P. krauelianus is used as a substitute to P. conoideus (Stone 1992).The leaves are used for mats.
Notes -Prior to the present study P. krauelianus was only known from the eastern part of mainland Papua New Guinea and the Bismarck Archipelago (Stone 1992).As it has now also been found in Yapen this is a new record.Several names have turned out to be synonyms.Except for relatively slight differences in the sizes of leaf, cephalium, drupe and its colour, there appear to be no significant morphological differences with P. krauelianus (Table 5).Stone (1992) mentioned the semidwarf habit of P. kosteri (1.7-2 m high), the red colours of bracts and cephalium as diagnostic characters for the species.Indeed, compared to the other members of sect.Maysops H.St.John P. kosteri is the smallest one.However, the discovery of 2-3 m high individuals straightforwardly belonging to P. krauelianus as is indicated by the salmon pink to orange cephalia covered with layers of green, bright yellowish green to cream orange bracts in Yapen undermined the recognition of P. kosteri as a distinct species and it is therefore reduced here.This has a further consequence that P. krauelianus is now to be recognised as a widespread species occurring from the Moluccas through the mainland of New Guinea and its adjacent islands to the Solomon Islands and the northern part of Australia with a wide spectrum of morphological variation.Species with such a great variability are not uncommon in Pandanus as can be seen in P. conoideus, P. odoratissimus L.f., and P. polycephalus Lam.
The presence of P. krauelianus in the Moluccas, the western side of mainland New Guinea, Solomon Islands, and northern part of Australia are new records.
Uses -Local people in Yapen Island use the fibres extracted from the prop-roots of P. papuanus as source material for making strings, handicrafts (bags), and mats.The cephalium is not eaten.Heyne (1927) reported that P. papuanus in Halmahera can only be found in cultivation.Purwanto (2006 pers. comm.) reported that in Halmahera Island the young leaf of a taxon that closely resembles P. papuanus is used as a medicine.
Notes -Although in the protologue of P. biakensis St. John (1960) stated that the type was made from a tree planted in a forest edge and was said to have been imported from the neighbouring Yapen, prior to this study no collection had been made in Yapen; thus mine confirms this and is to be regarded as a new record.surface green, venation green, adaxial ventral pleats absent; abaxial surface pale green, venation green with no spines, recurved spines obvious; leafsheath yellowish white to white.Infructescence in a spike, 30 -35 cm long, consisting of 4-7 compactly arranged and sessile cephalia; peduncle 13-18 cm long, glabrous.Cephalium globose, slightly depressed, 7 cm long, 4.5 -4.8 cm diam (14 -15 cm circumference), consisting of numerous drupes.Drupe elongated globose to slightly ellipsoid, green when young turning to bright yellow then to bright reddish orange or red when mature.
Distribution -Celebes, Philippines, Moluccas, New Guinea and adjacent islands including Yapen Island, Bismarck Archipelago, Solomon Islands, and Yap Island in the Caroline Islands (Micronesia).
Habitat -Coastal, beaches, mangrove, and swamps at seaside to few m altitudes.
Uses -No use of this species has been recorded for Yapen.In the Moluccas the young leaf and bract are used as vegetable and eaten raw (Heyne 1927).The leaves are also used as medicine against food poisoning, especially from seafood (such as fishes and crabs).Leaves mixed with betel nut (Areca catechu) and pepper leaves (Piper betle) are chewed and act as a mild narcotic (Heyne 1927) and also believed as medicine for strengthening the teeth or masticatory (Walter & Sam 2002).Older leaves are used for mats.Beccari (see Solms-Laubach 1883) reported that the fresh cephalium is used in Sulawesi to cause abortion.
Notes - Warburg (1900b) probably because of the reference to Rumphius and the use of the same epithet treated P. humilis Lour.(1790) as a synonym of P. polycephalus, but the latter species does not occur in Indochina.Because the citation of Rumphius is too a pre-Linnaean name, Loureiro's name is not necessarily typified by it, and is the correct one for a Indo-Chinese species.Martelli (1913Martelli ( , 1937) ) regarded it as a synonym of P. pierrei Martelli, a species native to Cambodia and South Vietnam.It is the other way around: Loureiro's name is the correct one, and P. pierrei is a synonym of it.As Merrill (1917) already wrote: "The type of Loureiro's species is manifestly the Cochinchina plant described, not the Rumphian synonym".
The presence of P. polycephalus in Yapen Island is a new record.Warburg (1900b) assumed that P. polycephalus would not be native to the Philippines and that individuals identified as such in the Philippines had actually been introduced, presumably from somewhere in Indonesia (then the Dutch East Indies, most likely from the Bogor Botanic Garden).Merrill (1904) identified two collections made by Copeland from Davao, Mindanao, as P. polycephalus.Martelli (1908), however, disagreed and regarded the specimens from Davao as "quite distinct" and to represent an undescribed species, P. brevispathus.Merrill (1922) accepted this.
It would therefore appear that P. polycephalus does not occur in the Philippines.However, I have reduced P. brevispathus to it, especially because of the similarity of the structure of the infructescences (which was already noted by Martelli himself).
Pandanus aruensis and its variety, P. aruensis var.contractus are also treated here as synonyms of P. polycephalus because of the structure of the infructescence.The species therefore also occurs in the southern part of the Moluccas.The variety was invalidly published as at the time the species had not yet been described.When that happened (Martelli 1914) the variety was not mentioned.
The presence of P. polycephalus as a native of Java has been the subject of debate for a long time.Miquel (1859, sub P. humilis) wrote that P. polycephalus is native to the Moluccas, and that according to Hasskarl (1845) it would be known in Java as 'Pandan serengseng' and 'Harrassas leutik'.Koorders (1911) mentioned the presence of P. polycephalus in Java with a question mark.He believed that the specimens identified by Miquel as P. polycephalus actually belonged to either P. caricosus Kurz (non Spreng.) or P. atrocarpus Griff.Later he (see also Koorders-Schumacher 1913) identified specimens named 'Pandan serengseng' as such.However, P. atrocarpus has been reported for the Malay Peninsula, Sumatra, Bangka, and Borneo, but has never been found in Java.Backer (1925;see Backer & Bakhuizen van den Brink Jr. 1968) identified Koorders' P. atrocarpus as P. caricosus.Regarding P. polycephalus in Java Backer (1925) noted "has been erroneously reported for the montane forest of West Java, it might, however here and there grow on the Javanese beach".This is repeated in Backer & Bakhuizen van den Brink Jr. (1968).The result of this present study indicates that specimens on which Koorders based his record of P. atrocarpus in Java (Koorders 20798, 26957, 40265, 40266, and 40267) belong to P. kurzii Merr., a widespread species in Java.I therefore agree with them and Stone (1972a) that all records in BO and L for Java are based on individuals growing in the Bogor Botanic Garden with various provenances, but none from Java.The most recent exploration made in the Ujung Kulon Nature Reserve in West Java (Keim et al. 2006c) failed to prove the presence of this species.
Solms-Laubach (1883) reported a new species (but did not give it a name because of the absence of male flowers and mature fruits) based on a collection made by Beccari in July 1874 in Lepo-Lepo, Kendari Peninsula, South East Celebes.His description matches P. polycephalus, particularly regarding the clumping habit, moderate high stem (3-5 m), an infructescence in a spike of 8 red cephalia, each cephalium 5-6 cm long; thus it is regarded here as belonging to P. polycephalus, and thus marks the first record of the species for Celebes.Fagerlind (1940) described P. columniformis on a collection made from a 27 year old individual in the Bogor Botanic Garden believed to be of Celebes origin according to the ledger of acquisitions of the Garden.In BO there is a voucher specimen (Van Vuuren 268) collected by Noerkas (one of the collectors of the Van Vuuren Expedition).Noerkas mentioned that the cephalia were red.Based on the original description, photographs, and the voucher (Table 6) this is clearly a synonym.7) is also a synonym of P. polycephalus and a new record of P. polycephalus for Yap Island, Micronesia.

Pandanus japensis (Table
Pandanus aimiriikensis and P. peliliuensis are also treated here as synonyms of P. polycephalus.These two species have many similarities with P. polycephalus including the infructescences consisting of 5-8 compactly arranged sessile cephalia and the possession of both terminal and lateral infructescences. Despite the slightly larger size of the cephalia, P. macrojeanneretia is very much the same as P. polycephalus especially because of the 5-8 compactly arranged and aggregated sessile cephalia (Kanehira 1933), thus P. macrojeanneretia is reduced.A lateral infructescence is a distinctive character of P. polycephalus as was already observed by Rumphius (1743).
The placement of P. aimiriikensis, P. macrojeanneretia, and P. peliliuensis as synonyms has the consequence that P. polycephalus is now also found in the Palau Island; thus a new record.
The placement of the numerous species listed above has the biogeographical consequence that the distribution of P. polycephalus is recognised here as confined to East Malesia (including Celebes and the Philippines), the Pacific and beyond.
An account made by Hooker (1894) of a possible presence of P. polycephalus in Burma is regarded here as pertaining to P. foetidus Roxb.This means that P. polycephalus does not occur anywhere in mainland South East Asia and thus supports the phytogeographical distribution of the species.Stone (1966) recorded the presence of P. polycephalus in the seashores of Malay Peninsula.However, neither detailed distributional records nor pictures were presented.It is assumed here that Stone might have misidentified P. labyrinthicus as P. polycephalus.Pandanus labyrinthicus shares a similar habit, some morphological features (including infructescences consisting of 3 -8 cephalia), and habitat with P. polycephalus.Indeed, excluding the style in each drupe the two species look very similar.Prior to this study, P. labyrinthicus was known only from the west coast of Sumatra (see Warburg 1900b).The presence of P. labyrinthicus outside Sumatra is now acknowledged as I have recently re-identified a specimen collected from Tarakan Island, East Kalimantan (Indonesian Borneo, W. Meijer 2567) as such; thus a new record.The specimen was previously identified by St. John as P. labyrinthicus, but later was misidentified as P. nitidus by Stone (he noted: "certe non labyrinthicus") underrating the fact that it is a coastal species with an infructescence consisting of 3 cephalia, in which each drupe has an obvious forked style, a morphological feature that is not possessed by P. polycephalus.Pandanus labyrinthicus, on the other hand, does.Habitat -Lowland tropical rainforest.In Yapen Island found at foothills at about 100 to 150 m altitude.
Uses -Not recorded.
Uses -In Yapen Island the fruits are consumed and the leaves are used for making mats.However, the species is not considered to be an economically important plant, and it is not cultivated.The fruits are reported to be eaten by birds as well.
Notes -The presence of S. sinuosa in Yapen Island was first reported by Beccari when he visited the island between 4-28 April 1875 during his great travels to Celebes and the then Dutch New Guinea.He landed in Ansus, an area within reach of Sarawandori on the western side of the island.Unfortunately, the plant that he saw was not in flower or fruit, thus he could only make a sterile collection (of a leaf, which is still kept in FI).Despite this, he believed without a doubt that the taxon that he saw represented a new genus of Pandanaceae.To satisfy his curiosity he revisited the island from 22 to 23 November in the same year, but failed to improve on his previous collection.Apparently due to the insufficient nature of the collection he refused to publish his finding.Nevertheless, he mentioned his discovery to Solms-Laubach, who then published the information (Solms-Laubach 1883).
12 years later Guppy (1887) found in the Solomon Islands a taxon that he believed was the same as what Beccari saw in Yapen Island.In contrast to Beccari, Guppy managed to obtain a complete (i.e.fertile) collection (Guppy 259).Based on this Hemsley (1894) described Sararanga and appointed it as the type of his S. sinuosa.As there has been no complete collection from Yapen after the one by Beccari Hemsley's publication raised the question whether the taxon in Yapen was indeed of the same genus and species.A complete collection of S. sinuosa from Yapen (Keim et al. 2006a) has now supported Beccari and ended the 130 years of uncertainty.
Habitat -Coastal and freshwater swamp.Vernacular name -Pandan pantai buah banyak (Malay).Uses -Leaves are used for making mats.The cephalium is not eaten.Note -The presence of P. kaernbachii in Yapen Island is a new record.Specimen seen.IndonesIa, Papua, Yapen Island, South Yapen District, Sarawandori, about 1 hour drive West of Serui, western side of the island, 10 Oct. 2006, A.P. Keim 801 (BO!).
Infructescences massive, branched, paniculate (each can weight up to 15 kg), terminal, bright green to brownish green; 10 infructescences can be found in 1 individual with branched stem; each branched to 3 orders, glabrous, 150-250 cm long; peduncle glabrous, c. 50 cm long, square shaped in cross sec-

Table 6
Morphological comparisons on the length and structure of the infructescences, the arrangement of the cephalia, and their colour when mature between Pandanus columniformis and P. polycephalus.tion;rachisglabrous,100 -200 cm long; rachillae numerous, glabrous, 36 -38 cm long.Fruits syncarpous, berries c. 100 per rachillae; kidney-shaped (reniform), pale green to red when mature, exocarp soft; seeds c. 60 per berry, triangle-shaped, flat and thin, pale brown.Distribution -Northern part of mainland New Guinea, Yapen Island, Manus Island, Solomon Islands.Habitat -Lowland tropical rainforest from sea level up to 200 m altitude.