The use of demersal trawling discards as a food source for two scavenging seabird species: a case study of an eastern Mediterranean oligotrophic marine ecosystem

1.   Introduction

The avifauna of Australia and New Guinea is still rich in challenges for research. Some de novo discoveries of undescribed taxa likely await, especially in montane New Guinea, as suggested by recent discoveries, e.g., a honeyeater (Meliphagidae; Beehler et al., 2007) and a berrypecker (Melanocharitidae; Mila et al., 2021). Species limits among many of the region's birds are contentious (reviewed in Joseph, 2021). New subspecies are still described (e.g., Black et al., 2020) and entirely new populations of some birds are still found (Nielsen, 2015; Jackett et al., 2017; Murphy et al., 2018). Still more gaps remain in knowledge of the natural history and ecology of many species especially in New Guinea (Beehler and Pratt, 2016; Gregory, 2017). Movement ecology, which addresses the ecological drivers of the spectrum of migration strategies used by birds (reviews in Dingle, 2004, 2008), well illustrates ecological challenges.

Here we show that each of the challenges just outlined are embodied in one species found in Australia and New Guinea, the Black-winged Monarch (Monarcha frater, Sclater, 1873; Passeriformes: Monarchidae). The species is widespread in foothill and lowland rainforests between 400 and 1600 ​m above sea level (asl) down occasionally to sea level across New Guinea (Diamond, 1972; Coates, 1990; Freeman and Freeman, 2014; Diamond and Bishop, 2015; Beehler and Pratt, 2016; Gregory, 2017). It also occurs in Australian rainforests primarily in far North-East Cape York Peninsula (Schodde and Mason, 1999; Higgins et al., 2006, Fig. 1).

Figure  1.  Map of New Guinea and northern Australia highlighting Cape York Peninsula and showing the broad distribution of the Black-winged Monarch (Monarcha frater; modified from Birds of the World (https://birdsoftheworld.org/bow/species/blwmon1/cur/introduction; accessed 20 October 2022) and subspecies taxonomy at commencement of the study; the circled area shows the localities of specimens newly reported here and identified as M. f. frater. Question marks indicate presumptive but essentially unsampled parts of the range. Maroon arrows and question marks indicate possible directions of migration of M. f. canescens to its unknown non-breeding range. White and black dots indicate bicoloured and wholly black lores, respectively, as well as specimen localities; for full details of specimen localities and numbers see Appendix Table S1. Illustrations are based on specimens as follow: M. f. frater: AMNH 654264; M. f. kunupi: AMNH 302156; M. f. periophthalmicus: AMNH 420522; M. f. canescens: AMNH 654177. Key localities mentioned in the text are shown. AUS: Australia; IND: Indonesia; PNG: Papua New Guinea.

DownLoad: Full-Size Img PowerPoint

The four consistently and currently recognized subspecies and their distributions as currently understood (see Fig. 1 for distributions and key localities mentioned in the text; Appendix Table S1 also includes geocoded localities) provide a framework for further describing the challenges we will address.

(1) M. f. frater Sclater, 1873. Restricted to the Bird's Head Peninsula at the western end of New Guinea (Beehler and Pratt, 2016);

(2) M. f. periophthalmicus Sharpe, 1882 (first named at species rank) and M. f. kunupi Hartert and Paludan, 1934 (first named at subspecies rank). Both were treated without reasons as subspecies of M. frater by Hartert and Paludan (1934) and Mayr (1941). These names have been assigned to populations across the rest of New Guinea east of the Bird's Head Peninsula; they have never been well differentiated from each other and so their putative ranges are similarly ill-defined (Sharpe, 1882; Hartert and Paludan, 1934; Rand and Gilliard, 1967; Beehler and Pratt, 2016; Gregory, 2017). M. f. periophthalmicus is generally assigned to populations from all of Papua New Guinea (PNG) and central and mid-western Indonesian West Papua. M. f. kunupi is usually restricted to populations from the Weyland Mountains in far western sectors of New Guinea excluding the Bird's Head Peninsula;

(3) M. f. canescens Salvadori, 1876 (first named at species rank). Assigned to migratory populations breeding in Australia (Schodde and Mason, 1999; Higgins et al., 2006). M. f. canescens was confused with and sometimes treated as conspecific with the Black-faced Monarch (M. melanopsis); only since the 1970s has it been consistently placed within M. frater (cf. Keast, 1958; Storr, 1973, 1984; Mees, 1982; Schodde and Tidemann, 1986; Schodde and Mason, 1999);

(4) The only other species-group names applied to the M. frater complex are M. kurandi Mathews, 1915 and M. canescens claudia Mathews, 1917. They were poorly diagnosed relative to M. f. canescens (see Mathews, 1915, 1917) and both have long been unremarkably regarded as synonyms of that form (LeCroy, 2008).

It is well-established that M. f. canescens is migratory and largely vacates its Australian range, the southern extent of which is ill-defined (see Higgins et al., 2006 and Discussion below), between approximately April and October (Draffan et al., 1983; Storr, 1984; Schodde and Mason, 1999; Higgins et al., 2006; Menkhorst et al., 2017; Joseph et al., 2019a). Remarkably, and apart from the few non-breeding birds remaining in Australia, details of the non-breeding range of M. f. canescens remain completely unknown (Schodde and Mason, 1999; Joseph et al., 2019a). Although the non-breeding range of M. f. canescens is reasonably assumed to be in New Guinea, no records of the species in New Guinea have ever been attributed to M. f. canescens (e.g., Rand and Gilliard, 1967; Coates, 1990; Higgins et al., 2006; Beehler and Pratt, 2016; Gregory, 2017).

The present paper has two aims. The primary aim was to solve the mystery of the non-breeding range of the Australian birds assigned to M. f. canescens. Our strategy was to use detail of plumage traits, localities, and dates of collection from specimens in collections as well as conventionally published and on-line photographs and recordings of vocalizations. This necessitated a second aim of reviewing the putatively diagnostic phenotypic characteristics of the four currently recognized subspecies of M. frater by which we could address their validity and distribution. If validity of M. f. canescens could be confirmed, then our strategy of assessing New Guinean specimens to detect non-breeding individuals of M. f. canescens would be feasible. Accordingly, we first report the results of that review and then address the non-breeding range of M. f. canescens.

2.   Materials and methods

2.1   Museum specimens

We contacted curators and collection managers of each of 25 ornithological collections with specimens of M. frater listed online in the Global Biodiversity Information Facility (GBIF; http://www.gbif.org), the Atlas of Living Australia (ALA; http://www.ala.org.au) and VertNet (http://vertnet.org); six other collections were contacted or checked online and had no specimens (Appendix Table S1; Acknowledgements) and attempts to contact two other collections yielded no reply. The collections with specimens are well-known as repositories of Australo-Papuan specimens due to histories of both European colonization generally and of 19th and 20th century ornithological exploration.

We assessed the dorsal, lateral and ventral patterning of each specimen with regard to the few characters cited in the literature as diagnostic among the subspecies: facial pattern; extent of black on head and throat; colour of inner secondaries and longest upper tail-coverts; shades of dorsal and ventral grey, and tibial feathering; reviewed in Results). It quickly emerged that the most discrete and repeatably observable differences concerned characters in the patterning of the frons, crown and lores. They could be assessed without risking damage to sometimes very old and fragile specimens, and we focussed on them.

Concerning shades of dorsal and ventral grey and ventral chestnut, which have been reported as diagnostic between M. f. frater and M. f. periophthalmicus (e.g., Finsch and Meyer, 1886), we did not quantitatively test these subjective and poorly supported claims. They were based on few specimens in older literature, and we urge caution in any analysis of these characters on specimens collected over more than 100 years and curated under widely varying curatorial conditions in different collections. Similarly, colour of minute tibial feathers had also been cited by Sharpe (1879) as diagnostically different between M. f. frater (orange) and M. f. canescens (black). Schodde and Mason (1999) noted that this difference in tibial feathering was not obvious in material available to them. Further, differing modes of preparation, especially placement of labels and degrees of loss, wear and damage, render tibial feathering very difficult to assess. Its repeatability is questionable, and it is certainly difficult to assess without risking damage to many specimens. Accordingly, we report our assessment of these traits only where specimens could be examined without damage.

We located 194 specimens of M. frater in 26 collections. We gathered basic metadata (locality, date) on all 194 specimens. For 189 of the specimens, we either had photographs of dorsal, ventral and lateral views often including close lateral views of the head. LJ personally examined and measured the study skin specimens held in AM, AMNH, ANWC and SAMA (museum abbreviations in Acknowledgements). We received photographs of some spirit specimens. Photographs of free-living birds were examined on eBird (http://ebird.org) and in Coates (1990). Elevational data are reported in the text below in metres above sea-level (asl); for original label data in feet asl or associated with specimens in the literature see Appendix Table S1.

2.2   Morphometrics

Appendix A contains morphometric data based on recorded mass on labels (grams) and wing length (flattened chord in millimetres), which are here deemed the most repeatable and reliable measurements and those incurring least risk of damage to specimens, made by LJ on specimens in AM, AMNH, ANWC, and SAMA. Unpaired t-tests were conducted among plumage-defined groups introduced in Results using a simple online-calculator, GraphPad (https://www.graphpad.com/quickcalcs/ttest1/; accessed May 17, 2023).

2.3   Literature review

A search for Monarcha frater in the Biodiversity Heritage Library (http://biodiversitylibrary.org) generated 53 links to the species and all were checked. Many were inclusions in an index or references to locality data for one or more specimens. Major ornithological checklists and global compendia and reviews (e.g., Mees, 1982; Schodde and Mason, 1999; Beehler and Pratt, 2016; del Hoyo and Collar, 2016) consistently cite the same literature relevant to diagnostic traits of subspecies and any details of their distribution.

2.4   Photographs and vocalizations

Photographs were searched in online (eBird; https://ebird.org) and traditional literature. Recordings of vocalizations were searched for online at eBird and at the xeno-canto website (https://xeno-canto.org). Comparisons of vocalizations are at present difficult because of difficulties in ensuring homology among vocalizations in different recordings so we consider these only briefly.

3.   Results

3.1   Overview of specimens

Appendix Table S1 lists details of 194 ​M. frater specimens we located: 29 specimens from the Bird's Head Peninsula, 147 from the rest of New Guinea (141 skins, 6 in spirit or skeletons), and 18 from Australia (11 skins, 7 skeletons, egg clutches or in spirit). Eight further specimens (all study skins) had been identified as M. frater but were misidentified specimens of Black-faced Monarch (M. melanopsis) and one MNHN specimen is either a misidentified Island Monarch (M. cinerascens) or an immature M. frater (Appendix Table S1). Ninety-two percent of the specimens were collected before 1970 and one has been collected in 36 years since 1987. Twenty-four specimens were collected between 1870 and 1899, 17 between 1900 and 1909, 29 between 1910 and 1919; 25 between 1920 and 1929, 30 between 1930 and 1939; 3 between 1940 and 1949 (all in 1949), 9 between 1950 and 1959, 41 between 1960 and 1969, 4 in the 1970s, 7 in the 1980s, and one in the 2010s. Four specimens had no date. Only four specimens, three from the 1980s and one from 2014, have accompanying frozen tissues (AM O.59034, ANWC B26710 and B27265; AMNH 840033; see 'Notes' field in Appendix Table S1). Forty percent (n ​= ​77) of all M. frater specimens are held at AMNH.

Across the species' total geographical range, sampling is reasonably (perhaps surprisingly) even but has been densest in easternmost PNG. There, both numbers of localities and specimens per locality are highest e.g., Mafulu (n ​= ​11), Wau (9), Karimui and Sattelberg areas (both 8); (see Appendix Table S1). Sampling in western PNG north and south of the Central Cordillera and westwards into West Papua is much sparser. There are only 17 specimens from between 144° E and the PNG-Indonesian border at 141° E. Sampling in Indonesian New Guinea has been greatest on the Bird's Head Peninsula (n ​= ​29) whereas just 12 Indonesian specimens from five localities are within West Papua east of the Bird's Head Peninsula. Eleven of the 18 Australian specimens are from "Claudie River" and one, notably, is from Torres Strait where it was found dead beneath a lighthouse.

Photographs of the species from Australia and the Bird's Head Peninsula are abundant online (e.g., https://ebird.org) but we know of only two photographs from New Guinea anywhere east of the Bird's Head Peninsula: one in Coates (1990) and one at eBird (https://ebird.org/checklist/S47105297; accessed 30 August 2022). Two recordings of vocalizations are from the Bird's Head Peninsula, five are from four localities in PNG, and three are from Australia.

3.2   Subspecies of M. frater: two phenotypic groups

Among all specimens of M. frater, two groups, hereafter Groups 1 and 2, were evident and readily distinguishable through plumage characters of the head and throat (Fig. 1). Group 1 comprises M. f. frater and M. f. canescens and 13 other New Guinean specimens from east of the Bird's Head Peninsula discussed separately below. The lores are bicoloured (white or pale grey closest to the eye, black closest to the bill). The black of the crown does not extend beyond the forecrown but is continuous with the black of the anterior part of the lores and face. A black periocular ring of feathers when present is usually very narrow (< 3 ​mm).

Group 2 comprises M. f. periophthalmicus and M. f. kunupi. The lores are completely black and this black colouring is continuous over the crown to and behind the eyes. Group 2 also has a complete and conspicuously thicker periocular ring (> 3 ​mm) that often has a postocular streak of black extending towards the occiput from the "top-most" (coronal) side of the ocular ring itself. These traits give Group 2 a distinctive "masked" appearance (Figs. 1 and 2). For brevity, Appendix Table S1 notes the lores as Bicol (= bicoloured white and black; Group 1) or black (Group 2).

Figure  2.  Lateral views of representative specimens of the four nominal taxa within M. frater at the commencement of the study. Arrows highlight the extent of black on frons and crown, loral colour, and post-ocular white streaks in frater (pronounced) and canescens (reduced, absent). Specimens shown and whether assigned in the text to Group 1 or 2 based on these patterns are M. f. frater (Group 1): AMNH 293999, AMNH 654264; M. f. kunupi (Group 2): AMNH 302898, AMNH 302156; M. f. periophthalmicus (Group 2): AMNH 654284, AMNH 420515; M. f. canescens (Group 1): AMNH 654214, AMNH 654177. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

DownLoad: Full-Size Img PowerPoint

Distinctions are less clear, however, between the two subspecies within both Groups 1 and 2. Concerning Group 1, Fig. 3 shows living examples of M. f. frater and M. f. canescens. Differences between them that have been cited have mostly been vague and subjective (Table 1). Sharpe (1879) and Schodde and Mason (1999) claimed differences between them that could be traced to explicitly described analyses of individual characters in specimens that were also clearly cited. Mostly, however, putative differences between them have been uncritically repeated when mentioned in reference books and compendia (e.g., Rand and Gilliard, 1967; del Hoyo and Collar, 2016; Beehler and Pratt, 2016; summarized in Table 1). Peer review of these details has likely been far less granular than had they been in primary research papers. We find that in M. f. canescens, vagaries of specimen preparation notwithstanding, the black of the frons reaches neither the forecrown nor the most anterior point of the eye. In M. f. frater, however, it reaches the forecrown but not as far as the mid-point of the eye closest to the crown (Fig. 2). We agree with Schodde and Mason (1999) that the black throat plumage in M. f. canescens (1) can extend to the upper chest and so is the most extensive of all subspecies of M. frater, and (2) likely more extensive in males than females (e.g., SAMA series of two males and two females) but, we add, not consistently so (e.g., female specimen AMNH 654177). We further note that the black of the throat appears narrower than in all other populations. MV and AM skin specimens of M. f. canescens are all males (n ​= ​3) and show the extent of gular black in males. Lastly, the post-ocular streak of white evident in M. f. frater is reduced or absent in M. f. canescens (Fig. 2; LJ, pers. obs. of AM, AMNH and SAMA specimens).

Figure  3.  Map of the approximate total distribution of M. frater showing photographs of living M. f. frater and M. f. canescens to emphasize their extreme but rarely remarked similarity. See text for discussion of differences.

DownLoad: Full-Size Img PowerPoint

Table  1.  Comparisons between M. f. frater and M. f. canescens.

Statements of comparison between M. f. frater and M. f. canescens	Reference
M. f. canescens: thighs black (M. f. frater–orange; see text); breast and abdomen deep chestnut (frater: orange as for the thighs); grey of upperparts and chest very pale-light pearly grey; entire throat black (frater: only chin and upper throat black; lower throat grey like the chest)	Sharpe (1879)
No remarks. Only noted uncertainty between Australian and eastern New Guinean populations as known at that time.	Keast (1958)
"yet to differentiate" in Australia i.e., undifferentiated from New Guinean populations	Keast (1961)
No remarks	Rand and Gilliard (1967)
Black mask of canescens and nominate frater does not reach the eye; there may be some black above the eye in latter	Mees (1982)
"canescens has marginally less extensive area of black (about the head) than nominate (= M. f. frater), not reaching the eye (black more extensive on throat than above bill)	Clement (2006) in del Hoyo et al. (2006); Present study
Black of face more extensive in M. f. frater than M. f. canescens "not including eye but extending to above eye (Mees, 1982)"	Higgins et al. (2006)
M. f. canescens is markedly sexually dimorphic, facial mask of males slightly greater than females and reaching black orbital ring (of M. f. periophthalmicus) and more extensive on forecrown than chin. No explicit comparison with M. f. frater. Tibial feathering difference cited by Sharpe (1879) not evident in specimens examined. Tail noted as disproportionally long but not stated relative to which other Monarcha taxa	Schodde and Mason (1999)
Illustrated showing loral white in M. f. frater and loral grey in M. f. canescens	del Hoyo and Collar (2016)
Cited Schodde and Mason (1999) and mainly relative to M. f. periophthalmicus	Beehler and Pratt (2016)
Postocular white streak more pronounced in M. f. frater but reduced or absent in M. f. canescens; reduced black frons in M. f. canescens relative to M. f. frater	Present study (LJ pers. obs.)

 | Show Table

DownLoad: CSV

Concerning Group 2, Hartert and Paludan (1934) based M. f. kunupi on two specimens from the Weyland (Kobowre) Mountains in western Indonesian New Guinea (Fig. 1). They diagnosed it relative to M. f. frater and M. f. periophthalmicus, the latter being represented by an unspecified number of specimens of M. f. periophthalmicus from Sattelberg in the Huon Peninsula region ~1000 ​km eastwards in present-day PNG; we have located eight specimens from Sattelberg (Appendix Table S1). On this scant basis, they diagnosed M. f. kunupi as having completely black inner secondaries not broadly lined with grey, more extensive black on the crown, darker grey on the dorsal surface and black, instead of grey, longest upper tail coverts (no measurements given). Rand and Gilliard (1967) and Beehler and Pratt (2016) essentially repeated some of these purported differences (extent of black on crown, secondaries, and on the longer upper tail coverts in M. f. kunupi) but indicated no specific examination of specimens. LJ examined the holotype and noted it as partially melanic (e.g., scattered black feathers in the grey chest). Further, variation in whether the inner secondaries are black or grey is not sex-related and not geographically structured although we cannot rule out an age-related component, and thus far more likely individually based. For example, AMNH 302157 is nominally M. f. kunupi based on its Weyland Ranges locality but it has grey inner secondaries (Fig. 2). The ANWC series of males (n ​= ​6) and females (n ​= ​3), all from within the range of M. f. periophthalmicus in PNG, shows inner secondaries varying from grey to black in both sexes. YPM 89252 (male) and YPM 89253 (unsexed) from the Baiyer River also show the two extremes. Two series of specimens in NHMUK from PNG (n ​= ​9) and the Utakwa River in West Papua (n ​= ​4) closer to the nominal geographical range of M. f. kunupi also show the full range of variation between the two extremes. Similarly for the other three characters listed by Hartert and Paludan (1934: tone of dorsal grey; grey or black longest upper tail-coverts; extent of black on the crown), the same series of specimens as just cited shows variation in these traits to be unstructured geographically and almost certainly individually based and susceptible to loss during specimen preparation. We find that each character on which M. f. kunupi was based is more likely to be individually variable across its nominal geographical range (Weyland Mountains) and that of M. f. periophthalmicus. Again, we cannot rule out an age-related component to variation in each of these traits; Finsch and Meyer (1886) stated baldly that in younger birds, the back of the head and neck is spotted with black. We reject this based on three clearly immature birds (brownish primaries, reduced facial black): AMNH 293998, 654287, 654288. We confirm that Group 2 does have more extensive black on the crown than Group 1.

3.3   Previously unrecognized Group 1 populations

We located and LJ examined 13 specimens of the Group 1 phenotype from a part of New Guinea east of the Bird's Head Peninsula where this phenotype has never been recorded (Fig. 4). All are easily recognizable and distinguishable from Group 2 by their bicoloured lores and black limited to the frons and forecrown. They are from two widely disjunct areas within what has been presumed to be the nominal geographical range of M. f. periophthalmicus (e.g., Rand and Gilliard, 1967; Gregory, 2017). They were tentatively assigned to M. f. periophthalmicus by Beehler and Pratt (2016) but they are clearly not that form (Fig. 4). Nine of the specimens (BBM 109825, 109912; AMNH 829531–7) are from three localities in far north-western PNG (Mt Nibo and Mt Menawa in the Torricelli and Bewani Mountains, respectively; Agpo Creek ​= ​13 ​km ​E Utai). Notably, AMNH 829531-7 have all been labelled as Monarcha f. frater. The other four (AMNH 341159–61; YPM 75920) are from two localities in Indonesian West Papua ~500 ​km to the southwest (Bernhard Camp and valley of Doendre River; Fig. 4). All 13 share with M. f. frater its slightly more extensive black forecrown relative to the more restricted black frons of M. f. canescens (examples in Fig. 5). Phenotypically, we find them essentially indistinguishable from M. f. frater (see Discussion). We found no literature reference to the specifics of the plumage of any of these specimens; this includes two admittedly preliminary reports on the avifauna of known localities such as the Torricelli and Bewani Mountains (Diamond, 1967, 1968) or feasible localities (Diamond, 1972; Beehler et al., 2012).

Figure  4.  Locations and representative examples of the hitherto almost totally unremarked Group 1 specimens (bicoloured lores) from east of the Bird's Head Peninsula (see Beehler and Pratt, 2016). Registration numbers and the locations of the relevant collections are given: AMNH, American Museum of Natural History, New York; BBM, Bernice P. Bishop Museum, Honolulu; YPM, Yale Peabody Museum, New Haven.

DownLoad: Full-Size Img PowerPoint

Figure  5.  Comparison of M. frater (left, AMNH 654264), three Group 1 specimens from Mt Menawa and the Torricelli mountains of PNG (i.e., east of the Bird's Head Peninsula; centre, AMNH 829533, 829534, 829535), and one M. f. canescens (right, AMNH 654177) to show subtle but slightly greater extent of black to the forecrown in the four at left relative to its lesser extent barely reaching the eye if at all in M. f. canescens.

DownLoad: Full-Size Img PowerPoint

3.4   Elevation and phenotype

We found no correlation between elevation as recorded in label data of specimen localities and the extent of black on the face and head in Group 1 and 2 phenotypes (Appendix Table S1). Australian specimens (Group 1) range from near or close to sea level to a potential maximum of ~520 ​m at Mt Tozer, one of the highest peaks within the Australian range (Higgins et al., 2006). For New Guinea, label data of two Bird's Head Peninsula Group 1 specimens, ANSP 132605 and 132606, record elevations as ~700 ​m and 990 ​m asl, respectively (see Mayr and Meyer de Schauensee 1939). Group 1 specimens reported above from east of the Bird's Head Peninsula are from 960 ​m to 792–1463 ​m asl (e.g., AMNH 829531–7, BPBM 109825, BPBM 109912). Similarly, label data for Group 2 specimens (full black lores; extensive black on crown and frons) span all but the lowest recorded elevational range of the species e.g., 800–880 ​m asl (NHMUK 1911.12.20.1476–1479 from Camp 6a, Utakwa River, and RMNH 136340, from Heuvelbivak, West Papua) while most other specimens are from between 900 ​m and 1280 ​m. Sight records confirm Group 2 ​at lower elevations closer to or at sea level, e.g., Coates's (1990) observation, and a photograph of a typical M. f. periophthalmicus in Varirata National Park (https://ebird.org/checklist/S47105297; accessed 30 August 2022), which spans elevations of 630–830 ​m asl (https://www.papuanewguinea.travel/; accessed 30 August 2022).

3.5   Sexual dimorphism in ventral and tibial feathering?

Sexual dimorphism in the subjectively assessed depth of ventral chestnut (richer in males, lighter in females) is apparent in the two pairs comprising four SAMA specimens of M. f. canescens (SAMA B1540, B1541, B21590, B21591). A similar range of intensity in ventral chestnut, however, is within six male M. f. periophthalmicus at ANWC and there is no consistent difference in its intensity between the six males and three females. Similarly, we find no difference in this trait in a pair of M. f. frater (FMNH 280438, 280439) or in all AMNH specimens.

After examining specimens in AM, AMNH, ANWC and SAMA, LJ concluded that some specimens retain longer, outer tibial feathers that are chestnut and which in life fall above the minute blackish feathers with very pale tips. If these outer feathers are lost or displaced in collection and preparation, the tibial feathering will appear black on quick examination. This reinforces our concerns stated in Methods about the repeatability and utility of the tibial feathering to diagnose population differences. With this proviso, we report the following observations. Black or dark tibial feathering with pale chestnut tips is evident in some male specimens of M. f. canescens (e.g., the holotype MSNG 14166; SAMA B21590; MV R6492; AM O.59641) whereas apparently orange or chestnut colouring concolorous with the belly is in at least one M. f. frater (AMNH 654264). Very pale feather tips to these otherwise minute black feathers are often apparent e.g., in two female M. f. canescens (SAMA B1541, B21591). The ANWC series of six male and three female M. f. periophthalmicus, however, has mostly chestnut tibial feathers with blackish bases, one female (ANWC B27377) having these feathers almost wholly chestnut save for miniscule black bases; a similar description applies to male AM O.39430 and female O.2582. Sexual dimorphism across the M. frater complex rather than between subspecies differences may explain these observations but we again stress the unreliability of this character.

3.6   Chestnut on lower back and rump in M. f. canescens?

Some specimens and photographs of living birds show what appears to be a band of chestnut feathering on the lower back or rump (e.g., Anon, 2018). In almost every specimen where this was apparent, it was due to an artefact of specimen preparation resulting in feathers of the lower belly having been 'wrapped' or tucked under a wing and so visible on the dorsal surface. Possibly, one lower dorsal feather in AMNH 654269 was truly chestnut and AM O.59641 does have some grey lower dorsal feathers with a terminal band of chestnut. Accordingly, we reject the claim (Anon, 2018) that a "rufous patch on the rump" is diagnostic of M. f. canescens but we acknowledge relevant individual variation.

3.7   Morphometrics

Differences in mass and wing length were not significant between all New Guinean Group 1 (bicoloured lores) and Group 2 specimens (p ​ > ​0.065 and p ​ > ​0.69, respectively). Among all pairwise comparisons of sexed specimens within and between New Guinean specimens of Groups 1 and 2 for wing length and mass, four comparisons, each between sexes, were significant (Table 2): Group 1 males (n ​= ​9) were significantly heavier than Group 1 females (n ​= ​7) (means: 22.20 ​± ​0.99 vs 20.44 ​± ​1.92 ​g; t ​= ​2.39, p ​= ​0.032) as were Group 2 males (n ​= ​11) (means 22.44 ​± ​1.92 vs 20.44 ​± ​1.92 ​g, t ​= ​2.15, p ​= ​0.047); Group 2 males (n ​= ​35) were significantly longer winged than either Group 1 females (n ​= ​8) (means 85.97 ​± ​2.92 vs 83.5 ​± ​2.20 ​mm; t ​= ​2.25; p ​= ​0.03) and or Group 2 females (n ​= ​25) (means 85.97 ​± ​2.92 vs 83.28 ​± ​3.43 ​mm; t ​= ​3.27, p ​= ​0.002). Wing length was significantly shorter in M. f. canescens (mean ​= ​81.14 ​± ​3.02 ​mm, n ​= ​7), however, whether relative to all New Guinean Group 1 (mean ​= ​85.04 ​± ​3.56 ​mm, n ​= ​24), "new" Group 1 specimens (85.31 ​± ​4.05 ​mm, n ​= ​13), or Group 2 (mean ​= ​84.71 ​± ​3.39 ​mm, n ​= ​65; t ​= ​2.62, p ​= ​0.0136; t ​= ​2.37, p ​= ​0.0288; t ​= ​2.66, p ​= ​0.0096, respectively). Male M. f. canescens (n ​= ​4) were significantly shorter winged than Group 2 males (means 81.75 ​± ​4.03 vs 85.97 ​± ​2.92 ​mm; t ​= ​2.65, p ​= ​0.012), and female M. f. canescens (n ​= ​3) were significantly shorter winged than New Guinean Group 1 females (means 80.33 ​± ​1.15 vs 83.5 ​± ​2.20 ​mm; t ​= ​2.31; p ​= ​0.046). These last comparisons are at best indicative given the small sample sizes of sexed M. f. canescens specimens.

Table  2.  Summary of pairwise comparisons of wing length (above diagonal) and mass (below diagonal) within and between sexes and within and between Groups 1 and 2 in New Guinea.

		Group 1			Group 2
		Males	Females		Males	Females
Group 1	Males	–	NS		NS	NS
	Females	*	–		*	NS
Group 2	Males	NS	*		–	*
	Females	NS	NS		NS	–
See Appendix Table S2 for raw data. NS: not significant; * significant, see text for details.

 | Show Table

DownLoad: CSV

3.8   M. f. canescens: non-breeding and breeding ranges

We found no specimens identifiable as M. f. canescens from New Guinea. At least three lighthouse-killed birds have been reported from Booby Island, Torres Strait: two seen (but evidently not preserved) in August 1975 (Stokes, 1983), and one specimen from 1 April 1979 (QM O17650; unsexed). MacGillivray (1918) reported a bird entering his tent and then leaving among flocks of seabirds at Raine Island on 8 December (? year, probably 1913). A sighting in eBird (https://ebird.org/checklist/S101248770) with no accompanying detail lists the species from Dauan Island, Torres Strait immediately south of the PNG mainland from 4 December 1988. Draffan et al. (1983) note the species as a passage migrant on western islands in Torres Strait, citing three specific islands either with supporting data (Booby (based on QM O17650) or without such data (Badu (= Mulgrave), and Tukupai (= Clarke)).

Specimens of M. f. canescens from mainland Australia are from the northern tip of Cape York Peninsula (e.g., Piara (= Paira), AMNH 654177) south to the Coen River (SAMA B1540, B1541) but most are from the intervening Claudie River area (Appendix Table S1). Records further south are reviewed in Higgins et al. (2006). Some are unsubstantiated and suggestive of confusion with Black-faced Monarch M. melanopsis (e.g., Werren and Barwell, 1987) whereas other records have been accompanied by more detail (e.g., Magarry, 1983). That M. f. canescens and M. melanopsis were considered possibly conspecific (see Discussion) likely caused some confusion among field observers. We note some paler M. melanopsis specimens with dark primaries and tails from near Cairns e.g., AMNH 654195 (the holotype of M. m. pallida Mathews, 1915), and AMNH 654186 and 654196. These specimens are distinguishable from M. frater and thus M. f. canescens by their sharply demarcated black throat plumage not tapering further on to the upper chest (Appendix Fig. S1). We discuss them further below.

4.   Discussion

The 194 specimens of the Black-winged Monarch (Monarcha frater) we have located in the world's collections are surely a majority if not all specimens in existence of the species. When coupled with surveyed online and traditional literature, they still provide no evidence as to the whereabouts of the non-breeding range of populations that breed in far north-eastern Australia, besides a few individuals remaining in Australia. We now review the quality of sampling provided by the specimens we have examined and then present a taxonomic reassessment looking at distribution and characteristics of all four currently recognized subspecies. Lastly, we probe details of distribution and migration.

4.1   Sampling: strengths and gaps

Spatial sampling of M. frater across New Guinea and Australia is remarkably even despite involving just one or two specimens from some localities and small series from others. Potentially the most important sampling gaps from which there are no specimens but where the species is known or might be expected are in Australia between Cooktown and Cairns, in far western New Guinea (FakFak and Kumawa; see Gibbs, 1994; Diamond and Bishop, 2015), and in much of central New Guinea south of the Central Cordillera (see Fig. 1).

Group 1 specimens from at least two disjunct regions east of the Bird's Head Peninsula (Figs. 1 and 4) highlight remaining uncertainties of M. frater's distribution in New Guinea. For example, Diamond and Bishop (2021) did not record M. frater from the Van Rees Mountains to the west of where Group 1 specimens have been collected at Bernhard Camp; they considered it likely to occur in the Van Rees Mountains, however. Conversely, M. frater has been recorded above 300 ​m asl in the Foja Mountains, which are between the two areas where the "new" Group 1 specimens have been recorded (Beehler et al., 2012) but no details are available of this population's phenotype (B. Beehler, C. Milensky, pers. comm., March 2023). Further study is needed in the northern ranges i.e., east of the Bird's Head Peninsula, to determine whether Groups 1 and 2 are parapatric or, as we suspect on current evidence, either sympatric or replace each other geographically. Also needing study is whether the "new" Group 1 populations are disjunct or continuously distributed.

4.2   Taxonomy

Our findings suggest a need to reassess the now entrenched taxonomic view, which was never rigorously defended in terms we would expect today, that M. frater is one species comprising four subspecies M. f. frater, M. f. canescens, M. f. periophthalmicus and M. f. kunupi (see Introduction). We found no consistent differences supporting the separation of M. f. kunupi from M. f. periophthalmicus and we recommend their synonymy under M. f. periophthalmicus Sharpe, 1882. Continuing to recognize the epithet kunupi exemplifies what Prates et al. (2022) described as deference to the opinions of previous workers about population distinctiveness based on subspecies proposed decades ago and that are difficult to falsify and discard owing to typically vague morphological definitions based, in this case at least, on a handful of specimens. That the holotype of M. f. kunupi was partially melanic in a genus known for melanism (Uy et al., 2019) reinforces our view.

We have highlighted phenotypic differences in plumage of M. f. periophthalmicus relative to M. f. frater and M. f. canescens, the latter two being so extremely similar in plumage. We found no elevational correlation to the amount of black on the head and lores and so reject an environmental basis to its extent or pattern. Interpretation of morphometric data is constrained by low sample sizes but the results at best tentatively suggest that M. f. canescens is slightly smaller than all other measured populations and that within New Guinea males are larger than females. The earliest instance of which we are aware of periophthalmicus being included in M. frater was Hartert and Paludan (1934) who gave no basis for doing so. Therefore, we recommend that even in the absence of molecular data and on the grounds of its consistent phenotypic distinctiveness, i.e., the masked appearance of its facial and head pattern, M. f. periophthalmicus Sharpe, 1882 should be restored to species rank and that it is monotypic. It is endemic to most of New Guinea east of the Bird's Head Peninsula. Masked Monarch is an aptly descriptive English name.

Next is the question of whether M. f. frater (in which we here include the "new" Group 1 populations east of the Bird's Head Peninsula) and M. f. canescens should still be treated as the same or different species. Their slight but consistent differences negate the concerns of Prates et al. (2022) raised above. Five criteria question the merit of treating M. f. frater and M. f. canescens as conspecific and, conversely, suggest the merit of treating them both at species rank.

(1) They are widely allopatric but differentiated by relative extent of black on the head (limited to the frons in M. f. canescens; reaching forecrown in M. f. frater), greater extent of the black throat feathering, i.e., extending to upper chest in some individuals of M. f. canescens, reduced or absent post-ocular white in M. f. canescens, and, given limitations of our sample sizes, potentially smaller body size estimated here through wing length in M. f. canescens. Their marked similarity in plumage could have several explanations such as convergent evolution or strong stabilizing selection since having diverged from a common ancestor. It need not outweigh the evolutionary distinctiveness implied by their profoundly unusual biogeography and their subtle but consistent differences.

(2) If conspecific, the biogeography of one species having isolated populations in the Bird's Head Peninsula and Cape York Peninsula is apparently otherwise unknown. Within New Guinea alone, however, one taxon occurring on the Bird's Head Peninsula and north-western New Guinea east of the that Peninsula also seems unique, possibly being shared by Smoky Robin (Peneothello cryptoleuca) and Vogelkop Whistler (Pachycephala meyeri) (Beehler and Pratt, 2016);

(3) Their bicoloured lores are shared by the Black-faced Monarch (M. melanopsis), Island Monarch (M. cinerascens) and Bougainville Monarch (M. erythrostictus). Bicoloured lores are likely an ancestral or paedomorphic trait (or both) within Monarcha (see Andersen et al., 2015; Schodde and Mason, 1999) and need not necessarily indicate conspecificity of M. f. frater and M. f. canescens. Conversely, bicoloured lores are not inconsistent with treating the two taxa at species rank;

(4) Vocalizations of M. f. canescens are unlike those of New Guinean taxa but similar to those of M. melanopsis, as noted elsewhere by one of us (PG; https://xeno-canto.org/)

(5) M. f. canescens is migratory and while not necessarily a taxonomically distinctive trait, this is ecologically distinctive.

Of course, genetic analyses of relationships within and between two such closely similar taxa as M. f. frater and M. f. canescens are needed. These analyses should include M. periophthalmicus and the closely related M. melanopsis. Given the above, however, and while urging genetic work, we predict that treating frater, canescens and periophthalmicus at species rank will have equal if not greater merit than the status quo under any species concept. We advocate recognition of three species and suggested English names as follows.

(1) Arfak Monarch (M. frater) of the Bird's Head Peninsula and the "new" Group 1 populations reported here (seven of which had been perceptively labelled as M. f. frater);

(2) Pearly Monarch (M. canescens) breeding in Australia and of still unknown non-breeding range, and

(3) Masked Monarch (M. periophthalmicus) as a separate species in most of New Guinea east of the Bird's Head Peninsula.

Distinguishing between a "three monotypic species" taxonomic hypothesis and the "status quo" of one species (M. frater) and three subspecies (M. f. frater, M. f. periophthalmicus, M. f. canescens) should be tested with molecular data. An "-omics" study based extensively on toepad skin will be necessary (for examples see Billerman and Walsh, 2019; Ewart et al., 2019, 2020). This is because just four frozen tissue samples of the entire complex are available for genetic analysis, and all are from the PNG range of M. periophthalmicus (Appendix Table S1). We have commenced this work.

How might molecular data inform this test? The present taxonomy, albeit with three not four subspecies within M. frater, would find some support if the three subspecies comprise a clade or monophyletic group regardless of whether each is itself monophyletic within that clade. They would then be each other's closest relatives within Monarcha again regardless of whether each is monophyletic with respect to the other two. If, however, genomic analysis reveals outcomes such as M. canescens being more closely related to M. melanopsis than to both M. frater and M. periophthalmicus, or M. periophthalmicus being sister to the other three, then the case for breaking M. frater as recently construed into three species would be reinforced.

The suggested vernacular names are either geographically informative (Arfak is very familiar in natural history for species inhabiting the Arfak Mountains of the Bird's Head Peninsula where this taxon was first described), descriptive (Pearly; Masked), or have been occasionally used (Pearly for canescens e.g., Slater, 1974; Higgins et al., 2006). Other Bird's Head Peninsula endemics having small parts of their range outside that area are known by its older name of "Vogelkop" (see above). Elimination of "Black-winged" would discourage field observers in Australia and southern New Guinea from focussing on grey vs black in the wings and tail when differentiating between any co-occurring M. canescens and M. melanopsis. These characters can be difficult to distinguish in an active, small bird high up in the shaded or dappled light of rainforest canopy. Instead, the suggested names encourage a focus on more easily discernible shades of grey of most of the plumage–light pearl grey, almost blueish grey of canescens and a darker grey in M. melanopsis.

4.3   Non-breeding range of M. canescens

Due to their superficial similarity to M. canescens, we return to and explicitly address the reasonable question of why we do not treat the "new" Group 1 populations as M. canescens from its non-breeding range: (1) black on the head is more extensive than in M. canescens i.e., reaching the forecrown, and so typical of M. frater (Fig. 5); (2) given low sample sizes, wing length is less in M. canescens, (3) AMNH Bernhard Camp specimens are from February when they might reasonably be expected to be absent if they were M. canescens, and (4) it is extremely unlikely that monarch flycatchers would cross New Guinea's Central Cordillera twice a year on migration. The latter two points can of course be countered as "possible", but we consider them inherently unlikely.

We affirm that M. canescens is a passage migrant south and north across Torres Strait between Cape York Peninsula and New Guinea before and after breeding, respectively, in Australia (MacGillivray, 1918; Draffan et al., 1983; Stokes, 1983). The whereabouts of the non-breeding range of M. canescens therefore remain a mystery beyond reasonably being presumed in New Guinea. Application of GPS loggers to birds in Australia and recovering them after return migration to Australia may be the simplest solution (e.g., Jiguet et al., 2022; Raz et al., 2023). The restricted latitudinal range within which the birds move is entirely within the tropics. When coupled with habitat of dense rainforest, however, this method's utility may be impaired (J. Fox, S. Deans, pers. comms., January 2023). Birds banded at breeding sites in Australia are unlikely to be recovered in sparsely populated, remote southern New Guinea. Stable isotope analyses in the absence of a well established isoscape are also unlikely to be of use.

The sampling gap we have described above in southern central New Guinea, particularly in PNG's south-western sectors (Fig. 1), might nonetheless include the non-breeding range of M. canescens. We stress that although there are sightings of the species in this last-named region (https://ebird.org; Gregory, 1995), none from eBird are accompanied by photographs or any reliable descriptive detail. Within that region in and around Tabubil, however, one of us (PG) has recorded only M. periophthalmicus e.g., three records from 750 ​m asl in August and September, and one from February at 1025 ​m asl (see Gregory, 1995). Further, he has not seen M. canescens there or at lower elevation in less suitable habitat near Kiunga (50 ​m asl) in the same region in M. canescens's non-breeding months. He cautions that ornithological tour groups mainly visit in July–August and only M. periophthalmicus, not M. canescens, has been recorded there by them.

Field observers should pay careful attention to the head pattern of any individuals in southern central New Guinea. Completely black lores and a masked appearance indicate M. periophthalmicus. Alternatively, bicoloured lores and black reduced to the frons if seen in birds in this area in approximately April–October most likely indicate M. canescens. Black-faced Monarch (M. melanopsis) is not expected in this area but caution is needed because of its similarly bicoloured lores. Despite no available evidence that this part of southern New Guinea in both PNG and Indonesia is the non-breeding range of M. canescens, it is a region that warrants closer searching. Much of that very remote region is directly north of the Cape York Peninsula breeding range of M. canescens. Therefore, there are potentially simple, straight-line migratory routes across Torres Strait between lowlands in these areas.

4.4   Breeding range of M. canescens

Broadly, localities of Australian specimens and most sightings in Australia of M. canescens are from the tip of Cape York Peninsula south to the Claudie River area, but the species has been reliably reported south to Cooktown (Fig. 1). Numerous reports from Cooktown south to Cairns might suggest the species occurs to Cairns at least occasionally. Pursuing this level of distributional detail in the Australian range is beyond our scope here but we agree with Schodde and Mason (1999) in urging caution, however, when distinguishing the species from M. melanopsis. Inexperienced observers especially may confuse the two species whether in adult or immature plumages.

4.5   Hybridization between, or conspecificity of M. canescens and M. melanopsis?

One of us (PG) notes that some Australian observers have raised the question of whether M. canescens and Black-faced Monarch (M. melanopsis) interbreed on Cape York Peninsula. Above, we have noted unusually pale-chested, black-winged and black-tailed specimens of M. melanopsis at AMNH from the Cairns area resembling but separable from M. canescens (e.g., Appendix Fig. S1). Notably, LJ observed that none of 24 AMNH specimens of M. melanopsis from further north on Cape York Peninsula where M. canescens routinely occurs suggested hybridization. Genetic research is warranted into these questions across Cape York Peninsula, however, because the phenotypic signal of hybridization can disappear within a few generations (e.g., Joseph and Moritz, 1993; Toon et al., 2012; Joseph et al., 2019b). Also, extreme caution in field identification of Monarcha in the Cairns region as M. frater or hybrids is warranted for two reasons. First, in M. melanopsis the extent of black in the wings and tail varies throughout the species' range. It can be unexpectedly extensive (e.g., black in wings of ANWC B34215 from Coffs Harbour in New South Wales) and is difficult to score accurately under field conditions and even in photographs. Second, and given the presence of melanism in other Monarcha species e.g., M. erythrostictus (Uy et al., 2019), it is not unreasonable to expect occasional, partial melanism in all of these monarchs. The holotype of M. f. kunupi exemplifies this, as noted above.

4.6   A last historical puzzle

The MSNG collection, Genoa, holds the holotype of M. frater Sclater, 1873 (MSNG 14130, collected prior to 1874), and that of M. canescens Salvadori, 1876 (MSNG 14166, male, collected in 1875) (Arbocco et al., 1979; Appendix Table S1). It is puzzling that Salvadori (1876), who worked in Genoa and described M. canescens solely from MSNG 14166, made no mention of the extremely similar M. frater, which at that time had just been described by Sclater (1873). Specifically, the puzzle is why Salvadori (1876) named and diagnosed M. canescens relative only to the Black-faced Monarch (M. melanopsis). Had Salvadori seen the holotype of M. frater, and had he seen Sclater's (1873) publication of M. frater? We suspect that the answer to both questions is "yes". Sclater's (1873) holotype of M. frater came from a collection made by Luigi D'Albertis and this collection reached Genoa in the first half of 1874 (E. Borgo, pers. comm.). So, the timing is consistent with Salvadori having seen the holotype of M. frater before he wrote and submitted his Italian literature description of M. canescens. Next, Salvadori (1874: 82) had remarked that Sclater named the more notable and interesting new species from the pre-1875 D'Albertis collection but that he (Salvadori) reserved the right to do more work on D'Albertis's collections. Indeed, Salvadori based his description of M. canescens on another D'Albertis specimen that was collected in 1875. It is therefore plausible that Salvadori had seen both the description of M. frater and its holotype, and that he correctly guessed that an Australian population would differ from the Bird's Head Peninsula population. This could have underpinned his decision to name the Australian population from the single Australian specimen from 1875 available to him. Regardless, it is remarkable that he did not diagnose it relative to M. frater, again, if the latter's holotype were available to him.

4.7   Concluding remarks

We have initiated genomic analysis of the populations we have studied here to (1) determine relationships within and among them, and (2) test our predictions that a "three monotypic species" taxonomic hypothesis (i.e., M. periophthalmicus, M. frater for now including the "new" Group 1 populations, and M. canescens) has more merit than other arrangements. We hope also to test for phylogeographic structure and gene flow across New Guinea whether within M. periophthalmicus or within and between each of M. frater (Bird's Head Peninsula), "new" Group 1 populations we ascribe to M. frater, and M. periophthalmicus. Careful, quantitative analysis of variation in the shades of dorsal and ventral grey and chestnut, which are predominant colours in the plumage across the entire group, would best be done on entirely new and fresh material the acquisition of which would be a non-trivial exercise. It would be valuable, however, regardless of its outcome. The phenotypic traits of any populations in the sampling gaps we have highlighted should be established. Finally, the still astounding fact that the non-breeding range of M. canescens is unknown means that in the 21st century one of the world's birds is missing for half of every year. We urge more searches for the non-breeding range of M. canescens especially in central southern New Guinea in PNG's Western Province (Fig. 1) and nearby areas of Indonesian New Guinea. In the meantime, we note with apologies to Oscar Wilde, that to lose a species is tragic; to not know where it is for half of every year is careless.

Authors' contributions

PG initiated the study through discussions with LJ about hybridization between M. canescens and M. melanopsis and contributed to writing. LJ led all aspects of the study from design to data gathering, analysis, and writing. JT contributed to analyses and writing and prepared Fig. 1. PS contributed to writing and observations of AMNH specimens. All authors read and approved the final manuscript.

Ethics statement

This study has been done entirely on museum specimens and literature.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgements

This project would not have been possible without the generously given time of present-day museum curators and collection managers who answered our queries and who sent photographs of specimens. We are of course grateful to the efforts of past collectors many of whom were assisted by often unacknowledged local indigenous people. We are grateful for the time and efforts of all these people, past and present, to collect specimens or answer our queries including translations to English, and, in most cases, sending photographs of specimens. We are therefore extremely grateful to the following institutions (acronyms are as listed in the text) and the indicated colleagues: Europe: Museum National d'Histoire Naturelle, Paris (MNHN, Patrick Bousses, Jerome Fuchs); Museo Civico di Storia Naturale "G. Doria", Genoa (MCSN, Enrico Borgo); Naturhistoriska Riksmuseet, Stockholm (NRM, Per Ericson, Ulf Johansson); Senckenberg Gesellschaft für Naturforschung, Senckenberg Museum, Frankfurt (Gerald Mayr); Museum für Naturkunde, Berlin (ZMB, Pascal Eckhoff, Sylke Frahnert, Frank Steinheimer); Natural History Museum, United Kingdom, Tring (NHMUK, Hein Van Grouw); Royal Belgian Institute of Natural Sciences, Brussels (RBINS, Oliver Pauwels); Finnish Museum of Natural History, Helsinki (LUOMS, Alexandre Aleixo). Australasia: Australian Museum, Sydney (AM, Leah Tsang; Emily Cave); Queensland Museum, Brisbane (QM, Heather Janetzki); Museum Victoria, Melbourne (MV, Karen Rowe); South Australian Museum, Adelaide (SAMA, Maya Penck); Australian National Wildlife Collection, Canberra (ANWC, Alberto Venchi); Papua New Guinea National Museum and Art Gallery, Port Moresby (PNGNMAG, Bulisa Iova). Americas: American Museum of Natural History, New York (AMNH, Thomas Trombone); United States National Museum of Natural History, Washington DC, Smithsonian Institution (USNM, Chris Milensky); Field Museum of Natural History, Chicago (FMNH, John Bates); Academy of Natural Sciences of Philadelphia at Drexel University (ANSP, Nate Rice); Carnegie Museum, Pittsburgh (CM, Steve Rogers); Museum of Vertebrate Zoology, Berkeley (MVZ, Carla Cicero); Museum of Comparative Zoology, Cambridge (MCZ, Jeremiah Trimble); Yale Peabody Museum, New Haven (YPM, Kristof Zyskowski); Los Angeles County Museum of Natural History (LACM, Allison Shultz, Kimball Garrett); Bishop Museum, Honolulu (BBM, Molly Hagemann); Museu de Zoologia da Universidade de São Paulo, Sao Paulo (Luis Fabio Silveira). For assistance with literature and advice on a range of other matters we thank: B. Beehler, E. Borgo (MCSN), M. Cohen, S. Dean (Lotek), J. Diamond, A. Engilis, M. Fisher-Casey, J. Fox (MigrateTech), S. Kepuknai, G. Mayr (Senckenburg Museum), C. Milensky (USNM), and T. Pratt. LJ's visit to AMNH was entirely supported by CSIRO National Research Collections Australia through Andrew Young.

Appendix A. Supplementary data

Supplementary data to this article can be found online at https://doi.org/10.1016/j.avrs.2023.100122. References cited only in the Supplementary Material are listed below and indicated with an asterisk (^*)