Translator Disclaimer


Nineteen days of surveys at four sites between 2,117 m and 3,200 m elevation in the Kaijende Highlands of western Papua New Guinea detected 102 species of birds. The Long-bearded Melidectes (Melidectes princeps) was recorded for the first time west of the Mount Hagen massif. It presumably ranges westward at high elevations to the Strickland Gap. The Ribbon-tailed Astrapia (Astrapia mayeri) was common from 2,117 m at Lake Tawa to 3,200 m at Omyaka Camp. We did not encounter Stephanie's Astrapia (Astrapia stephaniae). Its western-most range extension must lie east of Porgera valley in the north, although it extends west to the Doma Peaks/Tari Gap area in the south.

The Crested Bird of Paradise (Cnemophilus macgregorii) was not encountered on this survey, but one local informant at Omyaka Camp stated that it was present and it was collected during a 1989 survey (Kula 1989). It appears that this species is rare or absent between the Hagen/Giluwe area and the Star Mountains of Papua (Indonesia). This provides an intermediate-stage example of Diamond's “drop-out” phenomenon (Diamond 1972). The Kaijende Highlands support a rich upland bird fauna that might best be conserved through the creation of a large contiguous community-managed reserve that encompasses uninhabited traditional hunting lands. This could serve as a useful biodiversity offset for the Porgera mine operation.


Study of the geographic distribution of birds in Papua New Guinea has reached the phase where most field surveys provide fine-tuning of the elevational and geographic distribution of a few interesting species endemic to Papua New Guinea, while also confirming the excellent state of our knowledge of many of the more common and widespread species (Coates 2001). This rapid assessment of the bird fauna of the Kaijende Highlands produced some interesting distributional records that help to further clarify the pattern and process of differentiation of the montane forest birds of New Guinea.

Few surveys of the western highlands regions of Papua New Guinea have been undertaken, but most of these were individually comprehensive (see Frith and Beehler 1998, Appendix 2). To the southeast of our area of focus, the Mount Hagen massif was surveyed repeatedly for birds between 1938 and 1956 by Shaw Mayer, Blood, Gilliard, and Bulmer, and Mount Giluwe was worked between 1961 and 1973 by the Schoddes, Clissold, Sedlacek, and Mirza (in Frith and Beehler 1998, Appendix 2).

To the south and west of our area of focus, surveys were conducted as follows: The Tari Gap/Doma Peaks were surveyed between 1983 and 1991 by Mackay, the Friths, a PNG Department of Environment and Conservation (DEC) team, and Laska (in Frith and Beehler 1998, Appendix 2). The Victor Emanuel and Hindenburg Ranges were surveyed between 1954 and 1997 by Gilliard, Bell, Mirza, Wanga, Murray, and Gregory (in Frith and Beehler 1998, Appendix 2).

The Enga highlands of Kompiam (east of the Kaijende Highlands) were surveyed by Whiteside in 1994 (in Frith and Beehler 1998, Appendix 3), and the Kaijende Highlands above the Porgera mine were surveyed by a DEC team in the 1980s (G. Kula pers. comm.). Here we report the results of a brief but intensive survey of the Kaijende Highlands conducted at elevations between 2,117 m a.s.l. and 3,200 m a.s.l.


Field methods

Birds were surveyed during daily walking censuses and these data were supplemented by mist-netting (at Lake Tawa only) and through interviews with knowledgeable local naturalists who accompanied and assisted our field parties. ‘Walking Censuses’ were conducted by the senior author in the early morning, mid-morning, and again in the afternoon. Early morning surveys were typically three hours long and the later surveys were usually two hours long. During each survey the observer walked slowly and quietly through a selected habitat, and noted all birds heard and seen. An attempt was made to count individuals once per walk. The main objective of the surveys was to record all bird species present in each habitat, with a secondary objective being to obtain a general indication of relative species abundance.

Mist-netting at Lake Tawa followed standard procedures (Beehler and Mack 1999). Nets were set in places where high capture rates could be expected, including thickets, ridgecrests, and areas of regenerating low vegetation. They were operated and tended from dawn to dusk, and in some instances nets were opened at night to capture bats and night birds. Birds were weighed and measured and then marked by cutting the tip of an outer tail feather. Recaptured birds were released from the net without being measured again.

Study sites

Intensive surveys were conducted at the four sites described below. Additional details on the vegetation and habitat characteristics of each site can be found in Chapter 1.

Omyaka Camp (5°31′ 37″ S, 143° 03′ 23″ E, 3,200 m elevation) was situated in the forested verge of an open grassland valley. The main walking track from Porgera to the Mount Kare mine traverses the center of the valley. A forest mosaic dominates the slopes and the bases of the slopes around the valley, whereas alpine tussock grass and scattered treeferns dominate the valley floor. The forest was low (typically no more than 12 m), closed-canopy upper montane/subalpine forest. Trees were moss-encrusted and trunks were twisted and irregular. The tree flora included broadleaf taxa (e.g., Decaspermum, Eurya, Ilex, Prunus, Quintinia, Symplocos, Syzygium, Vaccinium, Xanthomyrtus, fide Takeuchi, Chapter 1, this volume) and gymnosperms (e.g., Libocedrus, Podocarpus, Phyllocladus, Dacrydium, fide Takeuchi, Chapter 1, this volume). These gnarly elfin woodlands were edged by a shrubland (with Vaccinium, Melicope, Olearia, fide Takeuchi, Chapter 1, this volume) that gave way to grasslands.

Lake Tawa Camp (5°35′43″S, 142°50′26″ E, 2,117 m) was set in a remarkable closed montane basin whose catchment drained exclusively through one or more major sinkholes. The basin features several small inter-connected lakes, whose level rose and fell with the rainfall because of low drainage capacity of the sinkholes. A narrow band of lake-associated grassland encircled the lakes, and Pandanus-dominated forest bordered the grasslands. The forest interior was lower montane closed forest containing Syzygium, Symplocos, and Lithocarpus (fide Takeuchi, Chapter 1, this volume). Flat ridgetops that we visited exhibited a Pandanus scrub with scrambling ferns dominating the ground layer.

Paiela Road (5°30′12″S, 143°5′6″ E, 2,800–2,900 m). We surveyed this road daily while based at the mine camp at Suyan. The road traverses the northeastern face of Mount Kumbepara, extending from the Waile Creek Road junction in a mainly north-westerly direction. It cuts through relatively undisturbed upper montane forest that afforded excellent conditions for bird survey.

Mine Environs and Suyan Forest (5°29′12″ S, 143° 9′E, 2,150 m). We observed birds in and around the Suyan mine camp, and on one full day we observed birds in the adjacent Suyan forest. Accessible habitats were a mix of open mid-montane valley habitats (gardens, grasslands, casuarinas groves, re-growth, remnant forest). The Suyan forest gave access to a lower montane bird fauna that was absent in other sites outlined above. We only imperfectly surveyed this elevational zone because our main focus was the more pristine uplands. We suggest that the zone below 2,200 m merits additional ornithological survey.


Survey results

Nineteen days of surveys at four camps between 2,117 m and 3,200 m elevation in the Kaijende Highlands of western Papua New Guinea detected 2,378 individuals of 102 species of birds between 21 August and 9 September 2005. Sixty-two individuals of 26 species were mist-netted at Lake Tawa. Walking censuses at the three principal sites produced comparable numbers of individuals based on similar total search efforts, ranging from 23 to 36 individuals per hour. The summary results of the walking censuses are presented in Table 3.1, and the total numbers of individuals of each species on each census are presented in Appendix 3. Because of the random and fragmentary nature of the Mine/Suyan observations they are not comparable with those from the three primary survey sites, but these results are incorporated into the species list presented in Appendix 4. The discussion that follows focuses mainly on the results obtained from the three major survey sites.

Table 3.1.  Results of ‘walking censuses’ of birds in the Kaijende Highlands region, Papua New Guinea.i978-1-934151-08-2-45-1-4-t04.gif

Over the elevational span extending between 2,117 m and 3,200 m there is a linear decrease in species richness with increasing elevation (Figure 3.1). This inverse relationship between elevation and species richness has been reported elsewhere in New Guinea (Diamond 1972, Beehler 1982). Although it is neither new nor counter-intuitive, this is a robust result that can be demonstrated with minimal field effort. It is one of the ornithological “laws” that governs the structuring of bird communities in montane New Guinea (Diamond 1972, Beehler 1982). Given our ability to demonstrate this with a low-investment survey methodology, we suggest the entire elevational range from sea level to 4,000+ m should be surveyed rapidly using walking surveys once there is a compendium of bird recordings for all species along the transect. Such a complete transect has never been attempted. Twenty-five hours at each elevational zone should be adequate to generate a robust dataset. This can then be used to tease out additional environmental patterns, especially the question regarding the putative “spike” in species richness in the hill forest zone, which is likely to be a sampling artifact (Kikkawa and Williams 1971).

Figure 3.1.  Elevational trend in species richness of birds in the Kaijende Highlands.i978-1-934151-08-2-45-1-4-f04.gif

Distributional notes

A series of notes on the distribution of lesser-known bird species, and those with particular distributional anomalies is presented below.

Clytoceyx rex (Shovel-billed Kingfisher)

This widespread but rarely observed species was common and vocal at Lake Tawa. It was recorded on 8 of 13 surveys and several birds were heard singing at dawn every morning. We also heard individuals singing at dusk. It was observed at close range on several occasions. A high count of six individuals was made on 30 August.

Petroica bivittata (Mountain Robin)

This uncommon upper montane species was encountered on two surveys at Omyaka Camp and was also observed there during non-survey periods.

Melidectes princeps (Long-bearded Melidectes)

This rarely seen and striking high-elevation honeyeater was previously known only as far west as the Hagen Massif. It was observed on six surveys at Omyaka Camp, mainly foraging in shrubbery at the interface between forest and grassland where it was greatly outnumbered by the smaller but similar Sooty Melidectes (Melidectes fuscus). On our 12 censuses we counted 11 Long-bearded Melidectes and 70 Sooty Melidectes. Our record of this species represents a north-westward range extension of 100 km. The Long-bearded Melidectes and Short-bearded Melidectes (M. nouhuysi) are sister taxa and their relative ranges had stood out as peculiar until recently — with an unexplained gap of several hundred kilometers between the two species. The range of the Short-bearded Melidectes is currently known to extend from Papua Province into the Star Mountains of far-western Papua New Guinea (Gregory and Johnston 1993). Given the pattern exhibited by the ranges of other high cordilleran bird species (e.g., Chestnut Forest-Rail, Splendid Astrapia) the Short-bearded Melidectes presumably ranges eastward to the Strickland Gap. We suggest that the Long-bearded Melidectes ranges westward to that same lowland barrier. This species-pair is hypothesized to have speciated across this major barrier for montane birds.

Melidectes rufocrissalis (Yellow-browed Melidectes)

The lesser-known twin of the Melidectes belfordi/rufocrissalis species-pair was found on the grounds of the Suyan Camp, inhabiting wooded copses set among gardens. It occupies a lower elevational zone than does the more abundant and better-known Belford's Melidectes, with which it readily hybridizes in areas of contact (Gilliard 1959). Belford's Melidectes was abundant at Omyaka Camp (mean of 14 individuals per survey) and along the Paiela Road (mean of 16 per survey), and was common at Lake Tawa (mean of 9 per survey). This hybridizing species-pair has caused taxonomic consternation for generations (see Mayr and Gilliard 1952, Gilliard 1959, Diamond 1967, 1972) mainly because of the complex and patchy distribution of M. rufocrissalis in the central range of New Guinea, combined with the apparent hybrid ancestry of some populations. Unlike the M. princips/M. nouhuysi species pair, this one is presumably considerably older, making the unraveling of its history more difficult. It is particularly unusual that the lower-elevation form (rufocrissalis) is confined to central New Guinea, whereas the higher-elevation form (belfordi) ranges from the Weyland Mountains of the far west to the far southeast of the Owen Stanley Range. The rather straightforward northwest-southeast vicariance that typifies Diamond's “drop-out” model (Diamond 1972) cannot explain this species-pair's distribution. Diamond (1967) noted that most records for rufocrissalis were from the northern watershed, and that it is closely related to the morphologically similar M. leucostephes of the Bird's Head region of far western Papua Province and M. foersteri from the uplands of the Huon Peninsula. In fact rufocrissalis is well distributed in the southern watershed, the Tari and Porgera valley populations being two examples. Molecular analyses of component taxa will be required to more fully understand the evolution of this remarkable group. The Melidectes belfordi/rufocrissalis species-pair offers a wonderful and challenging speciation/hybridization phenomenon whose study will provide an important insight into the geographic differentiation of montane forest birds of New Guinea.

Cnemophilus macgregorii (Crested Bird of Paradise)

Although the cnemophilines are no longer considered to be birds of paradise (Cracraft and Feinstein 1999), they remain a fascinating New Guinean cordilleran lineage of uncertain provenance. In the absence of a new taxonomic disposition or group-name, for convenience we retain the original nomenclature in our discussion. The range of the Crested Bird of Paradise is unusual. Like the Blue Bird of Paradise (Paradisaea rudolphi) it is confined primarily to mainland New Guinea east of the Strickland Gap. However whereas the Blue Bird of Paradise is entirely confined to territory east of the Strickland, the Crested Bird of Paradise shows a less clear-cut range. It is only found regularly and commonly east of the Gap, but there are a few records from west of the Gap, ranging as far west as east-central Papua. Most strange of all, is that the range of Cnemophilus macgregorii exhibits a major hiatus between Mount Hagen/Doma Peaks and the Star Mountains of Papua (see Frith and Beehler 1998, page 184). We searched for this species in vain in the Kaijende uplands, but at least one of our informants reported the species to be present and Kula (1989) collected voucher specimens from the region in 1989. Thus the Crested Bird of Paradise is common east of the Strickland Gap, but very rare or absent for several hundred kilometers west of the Gap. It then occurs in sparse numbers far west of the Strickland Gap, but apparently does not range as far west as the western terminus of the main cordillera (as does its sister form, Loria's Bird of Paradise). In sum, the species is common in the eastern two-fifths of the central cordillera, more or less absent in the central fifth of New Guinea, sparse in the west-central fifth of New Guinea, and absent from the westernmost fifth of the island.

Although the Strickland Gap seems to be important as a distributional barrier dividing young montane-dwelling species-pairs (e.g., Chestnut/Forbes' Forest-Rails, Splendid/Ribbon-tailed Astrapias, Short-bearded/Long-bearded Melidectes), most montane species (e.g., Loria's Bird of Paradise, Belford's Melidectes, Common Smoky Honeyeater, Red-collared Myzomela, Papuan Lorikeet, Great Wood-swallow, and many others) or species-pairs (Astrapia, Melidectes, Parotia) range from one end of the Cordillera to the other (Beehler et al. 1986, Frith and Beehler 1998).

The strange range of the Crested Bird of Paradise seems to exemplify an intermediate step in Diamond's “drop-out” phenomenon. Diamond invoked his “drop-out” model to explain how eastern and western sub-populations achieved geographic isolation along a continuous montane cordillera. The Crested Bird of Paradise seems to have achieved this range disjunction. The western population now has an opportunity to differentiate from the eastern population. This major range discontinuity is much as Diamond illustrated for the Papuan Treecreeper (Diamond 1972). One pressing question arises from the discontinuity of the Crested Bird of Paradise's range. What ecological or demographic process produces these discontinuities? Why is the species common in the east, very rare or absent in the middle of its range, and sparse in the west? Is it simply a stochastic process? Getting a clear answer to this question would provide an important insight into the mechanism of montane speciation in the birds of New Guinea.

The Crested Bird of Paradise's range offers some insight into the possible history of the currently truncated, east-New Guinea only range of the Blue Bird of Paradise. Indeed the range of the former species might closely resemble the Blue's current range if all of the western populations (already rare or very rare) become extinct. Both the Crested and the Blue Birds of Paradise appear to be “older” lineages (neither have extant close sister forms) and thus it is safe to assume that both, at some earlier time, occurred along the entire length of the central cordillera (as Loria's Bird of Paradise does today) but that the western segments of the population have declined — partially in the case of the Crested, and completely in the case of the Blue. These appear to be “senescent” ranges of aging lineages. A similar but reverse pattern can be found for the Yellow-breasted Bird of Paradise, a close relative of Loria's and Crested Birds of Paradise. In the case of the Yellow-breasted Bird of Paradise it is the far eastern populations of its cordilleran range that have become extinct, leaving just the central and western components.

Astrapia mayeri (Ribbon-tailed Astrapia) (Photos 75, 77)

The Kaijende Highlands are the heartland of the circumscribed range of this remarkable species, the last of the birds of paradise to be described to science (Stonor 1939). We encountered no populations of this species' eastern sister form, Stephanie's Astrapia (Astrapia stephaniae) on our survey, and we were told by informants that the “black-tails” were not present in the area. We know from detailed species-mapping of these two species (Frith and Beehler 1998) that Stephanie's Astrapia occurs 165 km to the southeast (Wabag, Kompiam), and 45 km to the south-south-west (Doma Peaks, Ambua). These two Astrapia species hybridize freely where they come into contact, but Astrapia mayeri generally occupies higher elevations than A. stephaniae. At Lake Tawa Astrapia mayeri was common (9 of 13 walking surveys). This is the lowest documented elevation that the species has been recorded (2,117 m) although Coates (2001) mentions a record as low as 1,800 m. In the Ambua/Doma Peaks region, where both species are present, Astrapia mayeri is found to occupy higher elevations (above 2,500 m). The western terminus of the distribution of A. mayeri is unknown, but it is probably the Strickland Gap. The Splendid Astrapia (Astrapia splendidissima) inhabits the high ranges west of the Strickland Gap. It will be interesting to determine whether A. mayeri and A. splendidissima meet and hybridize north of the Gap along the central range, north or northwest of the eastward bend of the Strickland in Hewa country.


The Kaijende Highlands are spectacular, with remarkable monumental features and wilderness qualities worth conserving. The bird fauna adds to the value of this resource. At least four species of birds of paradise (Brown Sicklebill, Ribbon-tailed Astrapia, King of Saxony Bird of Paradise, Short-tailed Paradigalla) inhabit the uplands, and additional species (Superb Bird of Paradise, Black Sicklebill) inhabit forests at lower elevations. The New Guinea Harpy Eagle and Shovel-billed Kingfisher also add biodiversity value to these environments. A community-delineated and managed reserve, focused on uninhabited traditional hunting lands, might be an appropriate vehicle for conserving these remarkable sub-alpine and upper montane environments and their wildlife.

75. Ornithologist Bruce Beehler with a male Ribbon-tailed Astrapia (Astrapia mayeri). This spectacular species was common at Lake Tawa.i978-1-934151-08-2-45-1-4-f1076.jpg
77. A male Ribbon-tailed Astrapia (Astrapia mayeri) at Lake Tawa. This species, the last of the Birds of Paradise to be described, has a very small range that fortunately includes the proposed Mt. Kaijende Conservation Area.i978-1-934151-08-2-45-1-4-f1078.jpg



BeehlerB. M. 1982Ecological structuring of forest bird communities in New Guinea.In GressittJ. L. (ed.)Ecology and biogeography of New GuineaThe HagueJunkPp837861Google Scholar


BeehlerB. M. T. K.Pratt D. A.Zimmerman 1986Birds of New GuineaPrinceton University PressPrinceton, NJGoogle Scholar


BeehlerB. M. A. L.Mack 1999Constraints to characterising spatial heterogeneity in a lowland forest avifauna in New Guinea.In AdamsN. J. R. H.Slotow (eds.)Proceedings of the 22nd International Ornithological CongressDurbanPp25692579Google Scholar


CoatesB. J. 1985The Birds of Papua New GuineaVolume 1Dove PublicationsAlderley, QueenslandGoogle Scholar


CoatesB. J. 1990The Birds of Papua New GuineaVolume 2Dove PublcationsAlderley, QueenslandGoogle Scholar


CoatesB. J. 2001Birds of New Guinea and the Bismarck ArchipelagoDove PublicationsAlderley, QueenslandGoogle Scholar



DiamondJ. M. 1972Avifauna of the Eastern Highlands of New Guinea.Nuttall Ornithological Club. No. 12Google Scholar


FrithC. B. B. M.Beehler 1998The birds of paradise: ParadisaeidaeOxfordOxford University PressGoogle Scholar


GilliardE. Thomas 1959Ecology of hybridization in New Guinea honeyeaters (Aves).American Museum Novitates, Number 1937Google Scholar



GilliardE. Thomas M.LeCroy 1968Birds of the Schrader Mountain region, New Guinea: Results of the American Museum of Natural History expedition to New Guinea in 1964.American Museum Novitates, Number 2343Google Scholar


GregoryP. A. 1995The birds of the Ok Tedi area. Ok Tedi MiningTabubilPapua New GuineaGoogle Scholar




KulaG. R. 1989A preliminary fauna survey report of Pogera areaDepartment of Environment & Conservation Nature Conservation DivisionWaigani, PNG12pagesGoogle Scholar


MayrE. E.Thomas Gilliard 1952Six new subspecies of birds from the highlands of New Guinea.American Museum Novitates, Number 1577Google Scholar




Appendix 2 Plant species recorded in the Kaijende Highlands

Wayne Takeuchi

Species recorded 19 August – 9 September 2005.

P = present, documented by collection(s) and Pw = present, but without collection.


Appendix 3 Total number of birds counted on ‘Walking Censuses’ in the Kaijende Highlands

Bruce M. Beehler

See Appendix 4 for scientific names and distributional records of all bird species within the study area.

See next page for Appendix 3.


Appendix 4 Bird species documented at four sites in the Kaijende Highlands of Enga Province, Papua New Guinea

Bruce M. Beehler and Robert Sine


Back to Top