The lost jewel of the Atlantic Forest: Kinglet Calyptura Calyptura cristata (Aves: Platyrinchidae) specimen inventory and plumage variation

. Kinglet Calyptura Calyptura cristata is one of the great enigmas of the South American avifauna. Endemic to an ap‑ parently tiny area of south‑eastern Brazil, in the Atlantic Forest biome, the species was not definitely seen between sometime in the second third of the 1800s and 1996, when it was briefly rediscovered in submontane forest north‑west of Rio de Janeiro. Since then, C. cristata has been reported several times, but without documentation and always by single observers. It is cur‑ rently considered Critically Endangered by BirdLife International, and various authors have speculated that the species might already be extinct. Given the extreme paucity of knowledge of this species, we provide a complete inventory of museum ma‑ terial for Kinglet Calyptura – more than 100 specimens are listed, the majority held in European collections, almost doubling previous estimates made in the literature. Several are held in relatively small institutions, thereby suggesting that yet more specimens might still be identified or found. In addition, with the benefit of this large sample of material, we discuss morpho‑ logical variation in the species and we hypothesise particularly about the appearance of male, female and juvenile plumages.


INTRODUCTION
The genus Calyptura, Swainson, 1832, com-prises just a single species, Kinglet Calyptura C. cristata, whose phylogenetic relationships have taken until the 21 st century to resolve.Traditionally, Kinglet Calyptura was treated as a member of the Cotingidae (Sclater, 1888;Hellmayr, 1929;Ames, 1971;Snow, 1973Snow, , 1979Snow, , 1982Snow, , 2004;;Kirwan & Green, 2011), based on its tarsal scutellation (pycnaspidean, not exaspidean like manakins and tyrannids) and foot structure, with the toes free, not more or less united (as in manakins; cf., Snow, 1982: 39).In contrast, Olalla (1943) suggested that Iodopleura (the purpletufts; now Tityridae) and Calyptura might form a family apart, but this proposal never acquired support, and much earlier Sclater (1888: 394) had argued that "I have little doubt that Iodopleura is not its remote ally", as well as noting, perspicaciously, that Calyptura has "Tyrannine plumage".In the light of accumulating evidence that several genera long considered to be cotingas actually belong in other families, especially the Tityridae (see, e.g., Ericson et al., 2006;Ohlson et al., 2008;Tello et al., 2009), Ohlson et al. (2012) finally demonstrated that Calyptura cristata is most closely related to the genera Platyrinchus and Neopipo.Furthermore, these three genera constitute a deep branch within the clade sometimes recognised as the family Rhynchocyclidae (tody-tyrants and flatbills).Ohlson et al. (2012) proposed to recognise the clade formed by Platyrinchus + Neopipo and Calyptura at family level, the Platyrinchidae.Nevertheless, most reference works have maintained all of these taxa within the Tyrannidae (Dickinson & Christidis, 2014;del Hoyo & Collar, 2016;Gill et al., 2021;Remsen et al., 2021).In other words, the size and structure of Calyptura -rather similar to Platyrinchus and Neopipo but abnormally small for a cotinga -while not necessarily taxonomically informative, provided better clues as to its relationships than might have been expected.The work of Ohlson et al. (2012) also reinforced knowledge that tarsal scutellation is not phylogenetically informative in respect of many Tyrannides.
Kinglet Calyptura is endemic to an ostensibly very restricted range just north of the city of Rio de Janeiro, in south-east Brazil; the type is stated to be from 'Rio de Janeiro' (Hellmayr, 1929), which has generally been interpreted as meaning the immediate environs of the city itself (e.g., Pinto, 1944;Sick & Pabst, 1968;Snow, 1982).Ruschi (1953) listed the species for the state of Espírito Santo, but this was doubted by King (1978King ( -1979) ) and rejected firmly by Collar et al. (1992) and Pacheco & Bauer (2001).A specimen, purported to have been collected somewhere in the state of São Paulo between May 1819 and April 1820, was discovered in the Museum für Naturkunde, Berlin, in 2007 (Stopiglia et al., 2009); however, its true provenance has not been satisfactorily established (Rego et al., 2013).Given the number of specimens mentioned in the previous literature, the species was not uncommon, even in secondary forested habitats, in the early to mid-19 th century (Collar et al., 1992;Kirwan & Green, 2011), but virtually the only available information concerning its habits/ecology is the testimony of Descourtilz (1852).The latter author mentioned C. cristata as feeding on fruit and insects, usually in pairs in the midstorey, maintaining contact with surprisingly loud vocalisations.A trickle of specimens continued to reach overseas museums with the last said to have been collected around 1890 (Snow, 2004: 88, repeated by Kirwan & Green, 2011), but as we show in this paper several were accessioned even later than this, although we generally lack robust knowledge of when they were actually collected.Thereafter the species went unrecorded until two individuals were seen in the Serra dos Órgãos, at Guapimirim, in the environs of Teresópolis, by multiple observers on several days in late October 1996 (Sick, 1997;Pacheco & Fonseca, 2000, 2001).Most of the few specimens with reasonably precise locality data come from this general region.There have been no reliable records since 1996 despite searches, mainly in Rio de Janeiro state, for example in the Reserva Ecológica Guapiaçu, the Teresópolis area, the foothills of the Serra do Mar, and between Nova Friburgo and Sumidouro, during September-November 2006 (Lambert & Kirwan, 2010) and in October 2016 (https://ebird.org/news/king-letcalyptura2016?tagId=128).Nevertheless, there have been several claimed records of the species, all by single observers and from the broad environs of Ubatuba in the state of São Paulo, in July 1990, March 1997, March 2006and September 2008(Sigrist, 2006;Lambert & Kirwan, 2010;Kirwan & Green, 2011).Earlier, D.F. Stotz (in Ridgely & Tudor, 1994) had speculated that the species might eventually be found in this region.Whilst BirdLife International (2023) currently treats the species as Critically Endangered, with a population expected to number fewer than 50 individuals, some authors have speculated that the lack of definite records in the last c. 25 years suggests that C. cristata is likely to be extinct (Lees & Pimm, 2015).
Given that virtually our entire knowledge of this species is based on the specimen record, it is pertinent to assemble a complete inventory of such material and its provenance.Various authors have commented on the number of specimens in natural history museums.For example, Snow (1982: 39) stated that "there exist only a handful of specimens in a few museums", whereas Collar et al. (1992: 726) mentioned that there were more than 45 specimens, preserved in AMNH, ANSP, NHMUK, RBINS, NML-VZ, MCZ, MNHN, NMW, UMZC, USNM, ZMB and ZMUC (for museum acronyms, see Methods).The figure of 45 was repeated by Pacheco & Fonseca (2000), but a year later the same authors stated that the total was approximately 50 (Pacheco & Fonseca, 2001).Presumably drawing on the latter, Snow (2004: 88) repeated the c.50 assertion.Tobias et al. (2006) did not speculate on the actual number of specimens, but suggested that their number did provide a hint as to the species' former abundance; Krabbe (2007) also reported that there are "nearly 50 specimens".Subsequently, Lambert & Kirwan (2010), Kirwan & Green (2011: 578) and Ohlson et al. (2012) all mentioned the existence of c. 55 specimens; this revised total was based on work conducted by GMK's own studies in museums additional to those mentioned by Collar et al. (1992), as well as specific publications in the interim (e.g., McGhie, 2005).Finally, Hume & Walters (2012: 354) noted the presence of specimens in various museums of which those in Florence (MZUF), in Italy and Kiel (ZMK), in Germany, were extra to the inventory of Collar et al. (1992) and those additional institutions visited by GMK.
Here, we report the existence of 104 specimens and provide as complete as possible an inventory of museum material, with details on its provenance and dating, based on research conducted by ourselves, separately and collaboratively.We also present a review of the species' morphological (sexual and age-related) variation based on photographs and personal examination of the specimens located to date.We also present photographs of all of these specimens, as well as historical information concerning their provenance.

MATERIAL AND METHODS
Building on the brief history concerning our collective knowledge of specimens in the world's museums outlined above, we used a combination of personal contacts and consultation, published information (e.g., Fisher, 1981;Lima, 2005;McGhie, 2005;Ghiraldi & Aimassi, 2019), as well as web searches of online museum catalogues and broader compendia (e.g., VertNet, GBIF, etc.), plus appeals on internet fora, e.g., the eBEAC mailing list.
In particular, we sought information from smaller museums in France, Germany, Italy and Switzerland, given the evidence already furnished by Hume & Walters (2012) and Waldeck (2018) that such institutions harboured previously unnoticed specimens of C. cristata.Our requests for information were coupled with solicitations for photographs of any specimens in order to assess plumage variation in the species.These specimen images originated from many museums, each one of which produced their photographs in different ways, ranging from professional standard, with excellent quality artificial lighting and in high resolution, to photographs made with mobile phones under natural light conditions.
The acronyms of those museums wherein specimens of Calyptura cristata were located are as follows:

Inventory
To date 104 specimens of C. cristata have been identified in 47 museums, the majority in European collections (Tables 1 and 2).Of this material, 86 specimens are held in a total of 39 different European institutions, 17 at seven collections in the USA, and one specimen is held in Brazil (a second, currently ZMB 2000.12102, is due to be sent to MZUSP in the forseeable future; P. Eckhoff in litt., 2023).Among the specimens listed herein, ten are in eight different smaller museums, considered part of the European 'B' list of institutions each harbouring fewer than c. 4,000 skins, or c. 5,000 bird items in total (Roselaar, 2003), or not mentioned in the latter inventory (e.g., MAB).An additional 12-15 specimens (Table 3) are known or are currently believed to be lost (see footnotes 7 and 8 to Table 3, and note 45 to Table 2 ♂ = male, ♀ = female, U = unknown sex, ad = adult, imm = immature); Name -collector, donor or purchaser (and at least in one case taxidermist) on label; Trade?
-'Yes' = traded (exported) skins, 'No' = skins known to be obtained directly by a naturalist that visited Brazil; SM -indicates how the specimen is currently stored: Sk = skin, Ms = mounted specimen; Source -publication that mentioned the specimen; Photo -corresponding figure in the text.
No.  ¹ Maximilian Collection/Maximilian Alexander Philipp, Prinz zu Wied-Neuwied (1782-1867), German explorer, ethnologist and naturalist, who led an expedition to south-east Brazil (in the states of Rio de Janeiro, Espírito Santo and Bahia) in 1815-1817, initially with Sellow (see note 51), and subsequently published a travelogue, Reise nach Brasilien, containing descriptions of new birds, as well as his Beiträge zur Naturgeschichte von Brasilien.His zoological collections, among them many avian type specimens, but also including material from North America (which he visited in 1832-1834), were sold in Paris three years after his death, and purchased by AMNH (Allen, 1889).This specimen was incorporated into Wied's collections after his trip to Brazil.Despite his having remained in Rio de Janeiro, the species is not mentioned in his works.
³ Henry Whitely (1844-1893), English naturalist and explorer, who visited Japan in 1864, and subsequently, in 1867, travelled in Peru and across northern South America, eventually residing in what is now Guyana until his death; there is no evidence that he visited south-east Brazil, so the ultimate provenance of this specimen is quite unknown (Sclater,1893).
⁴ Boucard Collection/Adolphe Boucard (1839Boucard ( -1905)), French ornithologist and trader in specimens who collected extensively in Middle America, concentrating on hummingbirds, selling bird skins to museums and private individuals such as P.L. Sclater, as well as supplying the plume trade; he spent the last few years of his life in England (see Kofoid, 1923).
⁵ Rivoli Collection/François Victor Masséna, Duke of Rivoli (1799-1863), an amateur French ornithologist, who amassed a collection of c. 12,500 bird specimens, many of them from the New World.He sold his collection in 1846/Thomas Bellerby Wilson (1807-1865), natural history collector and influential patron of ANSP, purchased the Duke of Rivoli's collection and donated it to the Philadelphia museum.The many types included therein were studied first by Stone (1899).
⁶ Malmazet Collection/Jean-André Malmazet (1808-1877); for an obituary, see Mulsant (1878).We have no information as to where he may have acquired this specimen.⁷ Henry Perkins Bryant (1820-1867), American physician and naturalist, who collected in the USA and on various islands in the West Indies, and travelled to France in 1865 to purchase the Lafresnaye Collection on behalf of the Boston Society of Natural History, who he had served as Curator of Ornithology since 1854 (for an obituary, see Abbot, 1867)/ex-Lafresnaye Collection 2210/Baron Nöel Frédéric Armand André de Lafresnaye (1783-1861), French aristocrat, ornithologist and collector, amassed a collection of more than 8,000 bird specimens.Following its purchase by Bryant, these birds eventually passed to MCZ in 1914.
⁸ Karl Hermann Konrad Burmeister (1807Burmeister ( -1892)), a German naturalist of wide-ranging interests and a professor of zoology at the Martin Luther University of Halle-Wittenberg; he visited Brazil between 1850 and 1852, principally the area around Nova Friburgo (Rio de Janeiro), where he acquired specimens from another German, C.H. Bescke, and further north (outside the known range of C. cristata) around Lagoa Santa (Minas Gerais), where he met Lund and Reinhardt (Sick, 1997: 54).Following a visit to Argentina during 1857-1860, he returned to the latter country in 1861, making a significant contribution to the development of the natural sciences in his adopted home.
⁹ Reinhold Ferdinand Sahlberg (1811-1874), a Finnish naturalist, participated in a circumnavigatory expedition between 1839 and 1843, visiting Brazil (for three weeks, in 1839), Chile and Alaska, and returned to Brazil in November 1849-January 1851, during which period he was based in Petrópolis and visited Rio de Janeiro, Sumidouro, Cantagalo, Boa Sorte and Ouro Preto (Lima, 2005).His collections reside in LUOMUS and NRM.
for reasons to consider that the total of "lost" specimens is 12, rather than 15).One of the specimens added to the list in a small Swiss museum (MAB) was previously misidentified as a species from North America, a Regulidae, the Rubycrowned Kinglet, Corthylio calendula (Linnaeus, 1766), labelled: "Regulus calendula ♂, le Roitelet, Amerique du Nord".It was discovered by Julien Mazenauer who recognised the specimen in the museum's public exhibition as a Kinglet Calyptura.
One of the two RBINS specimens (RBINS 10126A) was removed from the list, because it is not a C. cristata, but rather a Tyrannidae, possibly a Yellow-crowned Tyrannulet Tyrannulus elatus (Latham, 1790), despite its label declaring it to be C. cristata.
Of the 104 extant specimens, just eight (IZH-V 3260, NHMUK 1888.1.20.972,NMB 2041, NMB 2042, NMSG 5741, SMNG A07725a, ZMUC 105508 and ZMUC 105507) possess locality information that is any way specific ("Nova Friburgo", "Cantagalo" or "Rosário").These three localities lie within 45 km of each other.Nova Friburgo (also Neu Freiburg or Novo Fribourgo) is at c. 850 m elevation in the Serra dos Órgãos, in north-central Rio de Janeiro state (22°16′S, 42°32′W).This general area was visited by several relevant collectors, including Peter Wilhelm Lund, Jean-Théodore Descourtilz, Karl Hermann Konrad Burmeister and, but Paynter & Traylor, 1991).Rosário (sited at c. 22°16′S, 42°32′W, c. 855 m, according to Paynter & Traylor, 1991, but 22°06′S, 42°25′W, c. 655 m, by Krabbe, 2007) was another farm in the environs of Nova Friburgo; alternatively, JFP and others have visited a locality that is still called Rosário, at the border of the municipalities of Bom Jardim and Duas Barras, at c. 22°08′S, 42°29′W, c. 1,110 m, and just 15 km north-east of Nova Friburgo.Krabbe (2007) indicated that Lund was at Rosário almost permanently between 8 February 1827 and late June 1828, and that the fazenda was somewhere between Nova Friburgo and Cantagalo, but not as close to the former as the coordinates given by Paynter & Traylor (1991) suggested.Krabbe (2007) also noted that Lund may well have taken many of the birds labelled Rosário on a forested mountain named "Morro Queimado" ("Burnt Hill"), perhaps as far as several hours walk from the farm.A "Morro Quemado" (presumably the same locality) was also mentioned by another contemporary observer, Descourtilz. However, D. Miller (in Lambert & Kirwan, 2010) ascertained that Morro Queimado was the headquarters of a fazenda of the same name depicted in the background of an aquatint produced in the late 1820s, and that "Morro Queimado"  was sometimes used synonymously with "Nova Friburgo" when George Gardner visited in 1840 (Miller et al., 2006).According to Miller, "Morro Queimado" refers to the area around the farm headquarters (probably the present base of Anchieta College, Nova Friburgo; https://pt.wikipedia.org/wiki/Fazenda_do_Morro_Queimado), and the probable elevation of the original forest in this area would have been 600-900 m (Lambert & Kirwan, 2010).
Some specimens lack any indication of locality, while others are erroneous (e.g., AMNH 5156, labelled "Guiana, Suriname"; NML-VZ 3029, "Bogotá"; RMNH AVES.172607, "México"), but most are labelled simply Brazil (Brésil, Brasilia, Brasilien, Brasile) (n = 57), "southeast Brazil" (n = 6), or occasionally "Rio de Janeiro" (n = 11).Others are, inaccurately, attributed more generally to the Americas (CCECL 41007379, MHNVT R1-E6-C151-002).Two (MSNTP AV3081, AV3083) are labelled "Nuova Olanda" (= New Holland), in an apparent (but clearly erroneous) allusion to the area of northeast Brazil, stretching from Sergipe to Maranhão, and centred on Recife, which was formerly administered by the Dutch, until it was ceded to the Portuguese crown in 1661 (Lockhart & Schwartz, 1983).One specimen (ZMB 2306) stemming from Friedrich Sellow mentions São Paulo ("San Paulo") on the label, and Stopiglia et al. (2009) contested that this provided the first incontrovertible proof that C. cristata is not endemic to the state of Rio de Janeiro.Rego et al. (2013), however, advocated caution in placing too much faith in the label data because Sellow sent multiple shipments of specimens to ZMB during the relevant period, and because of gaps in the original documentation in the museum's archive concerning these arrivals.
Only a small percentage of the material can be more or less accurately dated as to when it was collected, all of which was probably acquired during the first twothirds of the 19 th century, for example IZH-V 3260, MCZ 75787, MNHN 3137, MSNT uncatalogued, ZMB 2305 and 2306, and ZMUC 105507 and 105508 (for the latter two specimens, see Krabbe, 2007).For the majority of traced specimens, the date of registration in a European or North American institution, or the date of acquisition, donation or exchange makes identifying when they were collected extremely difficult, probably impossible, especially given a lack of knowledge of the collectors involved.Nevertheless, where archival material might be available, we encourage curators and researchers at the museums concerned to pursue such avenues of investigation as may be open to them.Although some material was certainly not accessioned until the last third of the 19 th century, e.g., NMBE 1033790 (sometime in the 1870s) and all of that at NHMUK (in the 1880s and 1890s), and a few specimens (MM B6327, MNHN ZO- MO-2000-2153and MNHN ZO-MO-1931-1285) were not registered in other European collections until the first third of the 20 th century, in none of these cases do we possess definite evidence that they were collected so late.For example, the NHMUK specimens came to the British Museum via P.L. Sclater (n = 1), John Gould (n = 1), the Salvin-Godman collection (n = 1) and Alexander Fry (n = 2).That originally in Sclater's private collection must have been collected prior to 1862, given that it was listed in his own catalogue published in that year (Sclater, 1862: 247), and the specimen reached him via the natural history specimen dealer, James Argent, about whom very little seems to be known, but the British Museum purchased a total of 500 specimens directly from him between 1843 and 1854, including a batch of 30 from Brazil in 1846 (Sharpe, 1906: 300-301).We lack any knowledge concerning the provenance of Gould's specimen, which was purchased by the British Museum in 1881, on his death; neither a dealer nor a collector is mentioned on the museum's label (no original label is attached to the specimen).A trawl of Gould's correspondence (reproduced by Sauer, 1998aSauer, , 1998bSauer, , 1999) ) has failed to located any reference to this bird, but Sauer's volumes only cover his life up to 1845.Future researchers might wish to study the Gould correspondence held in the UK National Archives (https://discovery.nationalarchives.gov.uk/details/r/55f98d6c-8dac-4163-bafe-70414ef1c494) for information pertaining to this bird.The specimens from the Salvin-Godman collection came via John Youds, another person about whom very little is known, especially as the family (originally from Cheshire, in north-west England, but emigrated to Bahia, Brazil, around the 1820s) numbered multiple members named John, at least one of whom was involved in the slave trade.Ihering (1900) reported many species/specimens from the vicinity of Nova Friburgo attributable to Youds, but gave (or knew) no further details about him.Given the complexities, it may prove impossible to elucidate (even to the decade) when this specimen was collected, especially as Youds presumably was responsible only for trading it.Fry lived in Brazil between 1838 and 1854, and it seems most likely that his specimens were collected during this 16-year period, but he did continue to visit the country in the years thereafter (see Table 2, note 28).Even Snow's (2004) suggestion that the species' last known specimen was collected around 1890 appears to be based more on assumption than definite knowledge (and his basis is unfortunately unknown).Indeed, it is unclear whether any of the 100+ specimens was taken later than the 1860s.One of the last that can be pinpointed with any certainty is NRM 90127687, which was acquired by Gunnar Olof Hyltén-Cavallius sometime between 1860 and 1864 during his tenure in the Swedish consulate in Rio de Janeiro (see Table 2, footnote 37).
The specimens from Florence (MZUF) and Kiel (ZMK), cited by Hume & Walters (2012), could not be re-located.After contact by e-mail with Fausto Barbagli (University of Firenze), it proved impossible to confirm the presence of a specimen of the species in MZUF, only that the collection is currently stored in a warehouse due to renovation (Fausto Barbagli in litt., 21 October 2019).Neither could we confirm the existence of a specimen in Kiel; according to the curator the species is not represented in the collection (Malte Seehausen in litt., 20 March 2023).

Plumage variation
Although a much larger number of specimens preserved in museums is now known, the state of conservation of these specimens varies greatly, reflecting the conditions they have been exposed to over the last 150-200 years.
In the 19 th century it was common to publicly display birds in cases, with the result that many specimens have faded plumage, whereas others were quickly incorporated into scientific collections, remaining better protected inside cabinets unexposed to the light, but still others remain on display (e.g., LUOMUS 1573, MAB, ZISP 1819), despite the species being considered Critically Endangered and thus of special significance and very rare in natural history museums.Changes to original colours were caused not only by exposure to light, as Jon Fjeldså (in litt., 21 August 2019) at ZMUC, said: "The mounted specimen therefore is faded and stained grey by carbon particles, as the old exhibition rooms were heated by burning coal." To know a little more about the plumage of a recently prepared specimen (or a live bird), we consulted the original description of the species (Vieillot, 1818;pp. 528-529): "Le Pardalote huppé, Pardalotus cristatus, se trouve au Brésil, d'où il a été apporté par M. Delalande fils.La huppe qui orne sa tête est rouge et près l'occiput, comme dans le roitelei rubis; la gorge et toutes les parties inférieures sont d'un beau jaune, plus foncé sur le devant du cou et sur la poitrine; les pieds noirs; le bec est de cette couleur à sa base et à sa pointe, et couleur de corne sur le milieu et en dessous; la tête, le dessus du cou et du corps d' un vert olive tirant au jaune; les plumes du milieu de la tête, du front et de l' occiput, terminées de brunnoir; les petites couveriures des ailes moitié blanches à l'extérieur; les pennes brunes et bordées de vert-olive en dehors; celles de la queue du même vert et strê-courtes; taille à peu prés pareille à celle du pardalote pointillé".
Our translation: "The Kinglet Calyptura, Pardalotus cristatus, is found in Brazil, from where it was brought by the younger Mm.Delalande.The crest which adorns its head is red and close to the rear head, as in the roitelei rubis; the throat and all the lower parts are of a beautiful yellow, darker on the foreneck and on the breast; the feet are black; the bill is of this colour at its base and tip, and horn-coloured in the middle and below; the head, nape and upperparts are olive-green to yellow; the feathers on the middle of the head, forehead and rear head are brownish-black; the small wing coverts are half white on the outside; the remiges are brown and edged with olive-green on the outside; those on the tail are of the same green and very short; the size of the bird is about the same as that of pardalote pointillé".Roitelei rubis is a reference to the bird we now know as Ruby-crowned Kinglet Corthylio calendula (Linnaeus, 1766), whilst pardalote pointillé is the Spotted Pardalote Pardalotus punctatus (Shaw, 1792).
The holotype of Calyptura cristata (MNHN ZO-MO-2004-300) used in Vieillot's description was collected sometime in 1816 in Rio de Janeiro by Pierre Antoine Delalande, and is preserved in the Muséum National d'Histoire Naturelle, Paris, France (Fig. 1).It is possible to note clearly the effects of time, more than 200 years, on this specimen, which has faded coloration, mainly its yellow crest, as opposed to the red described by Vieillot.A feature that draws attention in the Kinglet Calyptura specimens (checked personally or via photographs) is the great variation in the colour of the crest, as well as the area or number of feathers with this coloration.The crest colour ranges from bright red (e.g., MHH 5665; MNHN ZO- MO-2000MO- -2153;;ZMB 2000/12103) through orange (e.g., LUOMUS 4949; MNRJ 3137; RMNH.AVES.172608;ZMB 2306) to yellow (e.g., AMNH 5156; MSNTP AV3082; NMBE 1033790; NRM 535288).Given that red was the original colour in the crest of the holotype, according to Vieillot's description, we might assume that other colours (orange and yellow) are only the result of discoloration over the intervening years, in line with research into the effects of time on the colour of museum specimens, perhaps especially carotenoid-based colours (such as reds and yellows) (McNett & Marchetti, 2005;Armenta et al., 2008;Riedler et al., 2014).However, this cannot be used to explain why a large number of specimens, all in excess of 100 years old, exhibit such variation in the colour of the coronal patch, especially as some specimens that are known to have been exposed to direct natural light for considerable periods of time, e.g., that in IZH, still have a bright red crown.However, we lack details of how every specimen has been stored throughout the last two centuries, and it is plausible that specimen storage and individual/sex-related differences could be among the factors to explain the variation that is currently observed.
The distinction between the sexes in C. cristata is based mainly on the yellow coloration in the forehead and forecrown in males and dark olive-green in the females, and the volume of red feathers in the crest, with males (e.g., MNHN ZO-MO-1854-440) having more red feathers, and bolder black lateral borders (Snow, 2004;Kirwan & Green, 2011).Examination of a larger number of specimens allowed us to pinpoint other plumage characteristics that might help to identify males and females, and distinguish between adults and immatures.We noticed that in males (e.g., NHMO 66671; NMB 2041; NHMUK 1895.4.1.730)the black band starts just in front of the eye and continues to border the red crest on the nape, whereas in females (e.g., NHMUK 1895.4.1.731;NMB 2042;ZISP 117202;ZMB 2306) there is only a small yellow spot behind and below the eye, and the black borders to the red/orange/yellow coronal patch are much less well defined.
Juveniles/immatures might be those individuals with few, or only a very small number of red feathers in the crest (e.g., RMNH AVES.172608), and which also lack yellow in the forehead and forecrown, perhaps with immature males possessing more yellow and red feathers compared to immature females.Analysing tail feather shape, which can be useful for separating adults and juveniles of passerines (being narrower with more pointed tips in the latter), is not always easy in very old museum specimens, rather than live birds, especially when attempting Pap. Avulsos Zool., 2023;v.63: e202363038 15/20 to base the identification solely on photos.Feathers can be lost, become frayed or otherwise changed in shape, however, RMNH.AVES.172608does appear to have narrower rectrices.Among other specimens that share the full suite of these immature characters are ZMB 2306, ZMB 2000, 12104, ZMUC 105507, and NMBE 1033790.

DISCUSSION
Our inventory almost doubles previous estimates of the number of Calyptura cristata specimens held in the world's natural history museums.We have identified at least 104 specimens, the majority of which are held in European collections.Given that several specimens were found in relatively small institutions, there is still the potential for additional material to come to light in collections subject to less intensive and detailed curation work.
To try and contextualise just how remarkably large this total is, one useful comparison to make is with another species endemic to a tiny area (estimated at 410 km²) of southeast Brazil, in the lowlands of Rio de Janeiro state, and which was also "lost" for more than a century, Blackhooded Antwren Formicivora erythronotos.Pacheco (1988) related the species' dramatic rediscovery in the environs of Angra dos Reis in September 1987, slightly less than a decade prior to the multi-observer record of Calyptura cristata near Teresópolis in October 1996.Unlike the latter, however, the Formicivora continues to be seen, some additional (albeit nearby) localities have been identified for it, and there are now literally hundreds of photographs of the species on citizen science databases, attesting to its local abundance.Specimens in museums are, however, exceptionally few in number: Pacheco (1988), repeated by Collar et al. (1992) postulated there are "about 20 nineteenth century skins in European and American museums", but in fact we have found evidence of just ten or 11: one or two of them lost, five extant in Europe, one in the USA, and three in Brazil (MNRJ 33300); see below, and a pair held at the Museu de Zoologia da Universidade de São Paulo, MZUSP 76678 and 76679; (L.F.Silveira in litt., 2023).Hartlaub's holotype, a male, is in the Zoologisches Institut und Zoologisches Museum, Hamburg (Bolau, 1898); another male from Burmeister is held in IZH (Pacheco, 1988;GMK pers. obs.), an adult male reported to have been collected in February 1879 (?) is in the Naturmuseum Senckenberg, Frankfurt am Main (SMF 24044), and there are two specimens (one male, one female) in NHMUK (Knox & Walters, 1994;H. van Grouw in litt., 2023).Another specimen, reported to have been either a female or immature male, was presumably formerly in IZH (e.g., Burmeister, 1856), but seems to be no longer present there.The specimen in SMF has apparently never been specifically referred to in the literature before; it was purchased from Gustav Schneider (1834-1900), the Basel-based taxidermist.At one time or another, the museum in Tring had five specimens (Sclater, 1890;Knox & Walters, 1994), the others being the male now in MNRJ (exchanged in 1984), another male, from Gould, that was one of the four listed by Sclater (1890) but is now lost, and a third male, originally in the Sclater collection, that was exchanged with AMNH in 1921.The female still at NHMUK was also originally in Sclater's private collection and came to him via a dealer named Warwick (it thus seems unlikely to be the same specimen once at IZH).Searches of VertNet and GBIF, as well as databases of several of the large North American museums (AMNH, FMNH, MCZ, NMNH), have revealed no additional material.
The most detailed prior analysis of plumage variation in Calyptura cristata is that by Kirwan & Green (2011).This was based on the earlier literature and analysis of specimens by GMK in eight museums (IZH, MNHN, MNRJ, NHMUK, RMNH, UMZC, USNM and ZISP).These authors suggested that females differ from males by having the forehead and forecrown largely dark olive-green (rather than yellow), with paler orange-red restricted to the rear crown and any black not forming well-defined lateral crown-stripes, narrower white tips to the wing-coverts and tertials, and a paler yellow rump patch.Furthermore, in common with other authors (e.g., Snow, 2004), they noted that any non-adult plumages are undescribed, but that ZMUC 105507 (photos of which were analysed as part of the present study) said to be an immature male was illustrated (by J. Fjeldså in Krabbe, 2007) as having a pale throat and a greenish cast to the flanks (Kirwan & Green, 2011).This specimen does indeed appear to have the narrower and slightly more pointed rectrices that might be expected to characterise a younger bird.
Sex and age-related plumage variation in Kinglet Calyptura's closest relatives are not always well known.In Cinnamon Neopipo Neopipo cinnamomea, the sexes are similar, but the female's yellow coronal patch is smaller, and there is no published information concerning juvenile plumage (Farnsworth & Lebbin, 2004).Among the seven species of spadebills (genus Platyrinchus), the sexes are separable based on the colour and pattern of the coronal patch in three, Cinnamon-crested Spadebill P. saturatus, Golden-crowned Spadebill P. coronatus and White-crested Spadebill P. platyrhynchos (Hilty, 2003;Johnson & Wolfe, 2018), whilst less detailed but similar differences have been reported in coronal patch size and colour (smaller and often paler in females) of two other species, Stub-tailed Spadebill P. cancrominus and Whitethroated Spadebill P. mystaceus (Tello, 2004).In contrast, the sexes are reported to be alike in Yellow-throated Spadebill P. flavigularis and Russet-winged Spadebill P. leucoryphus (Tello, 2004), although this perhaps requires re-evaluation in light of the details presented by Johnson & Wolfe (2018).Among Platyrinchus, the juveniles of Yellow-throated and Russet-winged Spadebills are both undescribed, whilst some of the other species are reported to differ principally in that they lack any evidence of a coronal patch, with reduced facial markings (usually well marked in adults) and generally paler and duller underparts, but at least in the hand ageing is probably more reliably based on relatively pointed versus dull and rounded rectrices, molt limits in the wing-coverts, and subtle differences in the colour of some feathers of the wing (Tello, 2004;Johnson & Wolfe, 2018).It is also Pap. Avulsos Zool., 2023;v.63: e202363038 16/20 interesting to compare another, slightly more distantly related tyrannid (based on the Ohlson et al., 2012 phylogeny), Many-coloured Rush Tyrant Tachuris rubrigastra, in which females are characterised by generally duller colours and a smaller coronal patch, and juveniles are distinguished by their lack of wingbars, duller underparts, some yellow scaling on the green upperparts and a lack of blue in the facial mask (Clock, 2004).
We yet again confirm the vital importance of scientific collections in museums around the world, not only the large and famous, but also smaller and little-known institutions that house much smaller amounts of material.With a considerably larger number of specimens preserved in museums, C. cristata was evidently an even commoner species than was previously imagined, despite its seemingly restricted distribution, thereby making the reasons for its apparently imperilled conservation situation even more mysterious.

Table 1 .
, Distribution of Calyptura cristata specimens by country.

Table 2 .
Extant specimens of Kinglet Calyptura Calyptura cristata in the world's museums, listed alphabetically by institutional acronyms.Listed are: Institution (museum acronyms are listed in Methods);