Is the Poo-Uli a Hawaiian Honeycreeper (Drepanidinae)? ’
Total Page:16
File Type:pdf, Size:1020Kb
The Condor94:172-180 01 ‘he Cooper Ornithological Society 1992 IS THE POO-ULI A HAWAIIAN HONEYCREEPER (DREPANIDINAE)? ’ H. DOUGLAS PRATT Museum of Natural Science,Louisiana State University,Baton Rouge, LA 70803-3216 Abstract. The Poo-uli (Melamprosopsphaeosoma), discovered on Maui in 1913, was originally believed to be a Hawaiian honeycreeper(Drepanidinae). Doubts as to the validity of this classificationprompted an investigation of the defining charactersof the subfamily and the possible position of M. phaeosomawithin it. The Drepanidinae are monophyletic with a suite of certain and possiblesynapomorphies that clusterthe group,but Melamprosops lacksall of these characters.Hawaiian honeycreepershave a distinctive odor which the Poo- uli lacks. Its tongue has prominent rearward projections whereas drepanidine tongueslack “lingual wings.” Most drepanidines lack the usual passerine predator-responsebehaviors but M. phaeosomaexhibits them. Vocalizations of the Poo-uli do not resemblethose of any of the three vocal groupingsof Hawaiian honeycreepers.The color pattern of M. phaeosoma is unique among native Hawaiian birds. Phenotypic charactersthus provide no basis for inclusion of Melamprosopsin the Drepanidinae; its relationships are unknown. Key words: Poo-uli; Maui; Drepanidinae; drepanidine odor; avian tongues;Hawaiian honeycreepers;Melamprosops phaeosoma. INTRODUCTION unlikely to become available and any assessment The Poo-uli (Melamprosops phaeosoma) is a re- of the systematicposition OfMelamprosops must cently discovered Hawaiian native bird (Casey be based on existing specimensand observations and Jacobi 1974) with a restricted range in the of live birds. rainforests of the northeastern slope of Halea- Casey and Jacobi (1974) initially placed this kala, Maui (Scott et al. 1986). The bird is a small genusand speciesin the Hawaiian honeycreepers passerine (length ca. 14 cm) with a short, thick, (Drepanidinae), apparently on the basis of sup- and vaguely finchlike bill that superficially re- posedoverall resemblancesand probably also on semblesthat of Ciridops anna (Casey and Jacobi geographic probabilities. Melamprosops was 1974), a Hawaiian honeycreeper, and those of considered an offshoot of Amadon’s (1950) ge- some small tanagers (pers. observ.). Two study nus Loxops (Casey and Jacobi 1974), which Pratt skins with tongues are preserved (Casey and Ja- (1979) regardedas a polyphyletic assemblagenow cobi 1974). They are dark brown above, buffy dispersedamong five genera (AOU 1983). Pratt below, with rufous-tingedflanks and a broad black (1979) and Berger(198 1) consideredthe position mask (see illustration in Pratt et al. 1987). Both of Melamprosops among the Drepanidinae un- skins are now believed to be in immature or certain but Amadon (1986), without giving any subadult plumage. Breeding adults observed in supporting evidence, saw “no good reason to hes- 1986 were plain gray above, white below, with itate to place Melamprosops in the Drepanidi- a sharply defined and prominent black mask and nae.” The following discussionspresent several lacked most of the rufous and brown tones (A. reasons for such hesitation. Engilis, P. Ching, pers. comm.). The Poo-uli has MONOPHYLY OF THE DREPANIDINAE been considered endangered since its discovery (Casey and Jacobi 1974), but its population has Any argument as to whether Melamprosops is a now declined to critically low numbers that in- Hawaiian honeycreeper is compromised unless habit a tiny remnant of a range that was already the group can be shown to be monophyletic. As restricted at discovery (Mountainspring et al. in Darwin’s finches, another oft-cited avian ex- 1990). Thus, additional anatomical material is ample of adaptive radiation, the case for mono- phyly of the Drepanidinae is weak (Baptista and Trail 1988). Perkins (1893) was the first to sug- I Received 30 May 1991. Accepted 23 September gest that the Hawaiian native finches and hon- 1991. eycreepersbelonged to a single taxon. His hy- ~1721 POO-ULI SYSTEMATICS 173 pothesis was widely (and rather uncritically) accepted, although Bryan and Greenway (1944) CHARACTER REVIEW considered the group possibly diphyletic. Ama- Drepanidine odor. Perkins’ (1893) suggestionof don (1950) stated that monophyly was “evident, monophyly of the Hawaiian honeycreeperswas chiefly as a result of Perkins’ field work” but basedalmost entirely on a characterthat has been ignored the primary character of depranidine denigrated or ignored by recent workers. He not- odor, upon which Perkins’ (1893, 1903) hypoth- ed that the native Hawaiian finches possessthe esis was based. Anatomical studies by Beecher same distinctive and peculiar odor characteristic (1953) Bock (1960) James et al. (1989) Raikow of the drepanidine nectar-feeders and insecti- (1976, 1977a, 1977b, 1978), Richards and Bock vores. Virtually all collectors of Hawaiian birds (1973) and Zusi (1978) demonstrated a remark- noticed the scent, rather like that of old canvas able uniformity underlying the spectacularadap- tents, and their reports are remarkably concor- tive radiation of externally visible characters dant. By sniffing randomly arranged series of among Hawaiian honeycreepers. This unifor- specimens of various passerine birds placed in mity is consistent with monophyly (Raikow an opaque cloth bag, A. Engilis and I tested each 1977b), but none of these studies identified any other’s abilitv to detect the odor and found that characterthat clustersthe Hawaiian honeycreep- we could consistently identify drepanidine spec- ersalone, and none included Melamprosops. Bock imens on that basis alone. Because the scent is (1978) described several features of the tongue easily picked up on the fingers and transferred skeleton shared by M. phaeosoma with drepan- to specimensthat might not originally have pos- idines, cardueline finches, and some other pas- sessedit, investigations based on old specimens serines.The tongue musculature also has several are potentially misleading. However, abandoned “drepanidine features,” most ofwhich are “shared nests retain the odor indefinitely (Perkins 1903, with cardueline finchesand some with other nine- Bryan 1908, pers. observ.) and thus can be used primaried oscines” (Bock 1978). Bock claimed to investigate the distribution of the odor among that these “features not only support the inclu- species. Odor intensity varies within species sion of Melamprosops phaeosoma in the Drepa- (Perkins 1903, pers. observ.) and perhaps sea- nididae, but provide further support for the sonally (Henshaw 1902). It may be totally lacking monophyly of the family.” The similarities he in some individuals, especially in species that cites, however, must be regarded as primitive inhabit drv habitats (Fisher 1906). I have found characters that carry no phylogenetic informa- the odor to be most noticeable in specimensfrom tion. Zusi (in Amadon 1986, pers. comm.) found very wet areas.For example, specimensof Hemi- the interorbital septum of M. phaeosoma to be gnathus virens that I collected on the dry leeward of the general type found in cardueline finches slope of Mauna Kea on the island of Hawaii had and Hawaiian honeycreepers (Zusi 1978), but only a faint scent, whereas those of the same did not demonstrate that this type represents a speciestaken in the rainy Volcano area on the synapomorphy. same island had a pronounced odor. These ob- These anatomical studies plus behavioral and servationsare consistentwith the probability that ecological (Pratt 1979; van Riper 1980, 1987) the odor is contained in the uropygial secretions and biochemical research (Sibley and Ahlquist used in preening. In wet areas and during rainy 1982, Bledsoe 1988) produced a consensusthat seasons,birds would require more frequent ap- the Carduelinae are the honeycreepers’ closest plication of water-repellent oils. relatives. A recent genetic study (which did not Variations notwithstanding, the odor has been include Melamprosops) based on proteins pro- reported in nearly all drepanidine speciesbut not vided support for monophyly of the Hawaiian in any other native or introduced birds living in honeycreepers, but not for the cardueline/dre- the same habitats. Perkins (1893, 1901, 1903) panidine relationship (Johnson et al. 1989). Nev- reported “drepanidine odor” in Chloridops kona, ertheless, I herein consider that relationship to Rhodacanthis palmeri, Loxioides bailleui, Psit- be well establishedand show that most Hawaiian tirostra psittacea, Pseudonestor xanthophrys, honeycreepers possesstwo certain, and other Loxops coccineus, Hemignathus virens, H. par- possible, synapomorphies, but that Melampro- vus, H. sagittirostris, H. obscurus, H. munroi, sops and the enigmatic genus Paroreomyza do Palmeria dolei, and Drepanis jiinerea (taxonomy not share them. updated). From my own experience I can add 174 H. DOUGLAS PRATT Vestiaria coccinea, Himatione sanguinea, and 1986) where the odor would probably be pro- Oreomystis bairdi. A. Engilis (pers. comm.) de- nounced. Nor can its absencebe attributed to the tected the odor in 0. mana. The only supposedly immaturity of the two specimens (although the drepanidine genuswhose members definitely lack ontogeny of drepanidine odor has not been in- the odor is Paroreomyza (Perkins 1903, pers. vestigated)because young birds would likely have observ.), although it is very weak or lacking in picked up the scentfrom the adults during brood- most specimens of Telespiza (Fisher 1906, pers. ing or from the nest. Engilis and I found no trace observ.). of drepanidine odor in