Eupherusa, Gould, 1857

Specimen evidence View in CoL

Study skins of [ T.] ridgwayi and three of four Eupherusa species (all except Stripe-tailed Hummingbird E. eximia ) are rare in collections because of their restricted geographic ranges and conservation statuses: [ T.] ridgwayi is currently listed as Vulnerable and Blue-capped Hummingbird E. cyanophrys as Endangered ( BirdLife International 2018a,b). The Delaware Museum of Nature & Science (DMNH, formerly Delaware Museum of Natural History), Wilmington, contains approximately 50% (15/30) of the global total of publicly accessible (databased) study skins of [ T.] ridgwayi , 33% (34/102) of E. cyanophrys and 23% (10/44) of White-tailed Hummingbird E. poliocerca Elliot, 1871 (Vertnet.org, accessed 19 March 2023). The DMNH collection also contains a sizeable sample (27/602) of available E. eximia ( Delattre, 1843) specimens and a single male of Black-bellied Hummingbird E. nigriventris (1/95). Most of this material was collected in the 20th century and is in excellent condition, making the DMNH collection well suited for a study of this kind.

Morphological criteria for generic classification

Genera defined by morphological characters are typically composed of (1) multiple species that share a suite of ‘unique synapomorphies’ (i.e., shared derived characters not present in other closely related clades) or, in the case of monotypic genera, (2) one species that possesses a suite of ‘unique autapomorphies’ (i.e., derived characters not shared by closely related species). Therefore, to demonstrate that [ T.] ridgwayi is sufficiently divergent from all Eupherusa species as to warrant the erection of a monotypic genus ( Dicranurania ), one would need to identify a suite of unique autapomorphies in [ T.] ridgwayi not found in the closest relatives of the [ T.] ridgwayi + Eupherusa clade (i.e., the genera Goldmania Nelson, 1911 , and Microchera Gould, 1858 ; see McGuire et al. 2014). Here, I applied this standard to scrutinise the diagnostic characters used by Sangster et al. (2023) to distinguish Dicranurania from Eupherusa .

Red tertials and secondaries.— Sangster et al. (2023) wrote that the ‘Secondaries and tertials lack red [in Dicranurania , whereas the red coloration is] present in Eupherusa ’. However, Goldmania bella ( Nelson, 1912) , a member of the likely sister group of the [ T.] ridgwayi + Eupherusa clade ( McGuire et al. 2014), also has red tertials and secondaries ( Nelson 1912, Schulenberg & Sedgwick 2021). Therefore, this character is not uniquely synapomorphic to Eupherusa , as Sangster et al. (2023) claimed when they wrote that ‘ Eupherusa can be distinguished from all other emeralds [tribe Trochilini ] by the presence of red on the tertials.’ Red tertials were either ancestral to the Goldmania + Eupherusa ( sensu lato) clade (i.e., not derived in the [ T.] ridgwayi + Eupherusa clade, and therefore unfit for use in diagnosing genera) or evolved independently more than once (i.e., also unfit to be diagnostic characters). Sangster et al. (2023) were adamant that any character that ‘has evidently evolved multiple times independently in emerald hummingbirds carries little weight in grouping taxa into genera.’ Therefore, by their own criteria, since red tertials are also lacking in Violet-capped Hummingbird Goldmania violiceps Nelson, 1911 , Snowcap Microchera albocoronata ( Lawrence, 1855) , Coppery-headed Emerald M. cupreiceps ( Lawrence, 1866) , White-tailed Emerald M. chionura ( Gould, 1851) , and apparently all other emerald species, the absence of red tertials in Dicranurania must also be ancestral, or else it evolved multiple times independently. It is not a ‘unique autapomorphy’ of [ T.] ridgwayi , as defined above, and therefore not valuable for diagnosing genera.

Tail shape.— Sangster et al. (2023) claimed that all four Eupherusa species have ‘square’ tails, and this formed the basis of their argument that the ‘forked’ tail of [ T.] ridgwayi is a unique autapomorphy that supports its placement in a monotypic genus. However, they apparently overlooked that E. nigriventris has a ‘more strongly rounded’ tail than other Eupherusa species ( Ridgway 1911: 399)—decidedly not square—as demonstrated by Gould & Sharpe’s (1887) original plate ( Fig. 1 View Figure 1 ), the lone E. nigriventris study skin in my sample ( Fig. 2 View Figure 2 , DMNH 59857) and photos of live birds taken from an appropriate angle (e.g.., Macaulay Library, ML 449382291). Furthermore, within the clade containing Eupherusa and its nearest relatives, the forked tail is not unique to [ T.] ridgwayi —both Pirre Hummingbird Goldmania bella and G. violiceps have forked tails—and forked tails also occur in other clades of emeralds, as well as in more distant clades within Trochilidae . Therefore, this homoplastic character ought to ‘[carry] little weight in grouping taxa into genera’ ( Sangster et al. 2023: 63).

Uppertail colour.— Sangster et al. (2023) claimed that all Eupherusa species have ‘black’ tails, evidently referring to the dorsal surface of the central rectrices, because the outer rectrices are mostly white in Eupherusa . On this basis, Sangster et al. (2023) argued that the ‘mostly blue’ tail of [ T.] ridgwayi is a unique autapomorphy that distinguishes it from all Eupherusa species and supports the erection of a monotypic genus. This argument was apparently based on a misunderstanding because no Eupherusa species has a black tail.

In E. cyanophrys and E. poliocerca , adults have ‘bright metallic bronze-green’ tails ( Fig. 3 View Figure 3 ; Ridgway 1911: 398). In E. eximia , the tails are ‘very dark greenish-bronze’ ( Elliot 1879: 213), ‘dark bronze-green or greenish bronze’ ( Ridgway 1911: 394), and, in the subspecies E. e. nelsoni Ridgway, 1910, ‘dark dull bronzy, sometimes more dusky terminally’ ( Ridgway 1911: 396). In some specimens of E. eximia (e.g., DMNH 24959), there is even a dark bluishgreen hue that resembles the tail of [ T.] ridgwayi ( Fig. 3 View Figure 3 ). The tail of E. nigriventris has been described as ‘dusky bronze or dull blackish glossed with bronze’ ( Ridgway 1911: 400) and ‘purplish-black’ ( Elliot 1879: 211), but the adult male in my sample (DMNH 59857) has a dark greenish-bronze tail—certainly not ‘black’ ( Fig. 3 View Figure 3 )—and the birds depicted by Gould & Sharpe (1887, Pl. 54) also had bronzy-green tails ( Fig. 1 View Figure 1 ).

By reducing the colourful variation in Eupherusa tails to ‘black’, Sangster et al. (2023) gave the false impression that the ‘mostly blue’ tail of [ T.] ridgwayi was a unique autapomorphy that supported the erection of a monotypic genus. This error cannot be attributed solely to a lack of access to study skins of the relevant taxa, because the correct range of tail colour variation in Eupherusa was available in literature (e.g., Elliot 1879, Ridgway 1911), which Sangster et al. (2023) did not cite.

Undertail pattern.— Sangster et al. (2023: 64) stated that ‘females of [ T.] ridgwayi do not differ mainly in tail pattern from Eupherusa ’. This seems to have been a typographical error (i.e., the word ‘not’ was mistakenly inserted). However, their assertion that the absence of white on the outer rectrices of [ T.] ridgwayi males is a unique autapomorphy, which supports the erection of a monotypic genus, does not withstand scrutiny. Males of Goldmania bella and G. violiceps also lack white in the tail, and they do not form a clade with [ T.] ridgwayi to the exclusion of Eupherusa ( McGuire et al. 2014, Stiles et al. 2017). Therefore, the absence of white is either ancestral (i.e., not derived and therefore unfit for diagnosis) or independently evolved more than once, in which case it is not appropriate for diagnosing genera ( Sangster et al. 2023: 63).

The lessons of Callipharus Elliot, 1879

Sangster et al. (2023) neglected to mention that, until Peters (1945: 76–77) placed E. nigriventris in Eupherusa , it was formerly classified in the monotypic genus Callipharus Elliot, 1879 , because of its considerable divergence from E. poliocerca and E. eximia . Ridgway (1911: 400) was so convinced of the need to separate E. nigriventris from Eupherusa , that he wrote: ‘Considering all its peculiarities, I feel sure that Callipharus should stand alone in any arrangement of the family that does not involve a very great reduction in the number of genera.’ In addition to the diminutive size of ‘the miniature Eupherusa ’ ( Elliot 1879: 211), the tail of E. nigriventris is ‘more strongly rounded [than other Eupherusa species, and] … [the] underparts [are] velvety black’ ( Ridgway 1911: 399). However, phylogenetic data suggest that E. nigriventris is nested within a clade of other Eupherusa species ( McGuire et al. 2014), rendering the monotypic Callipharus unnecessary, and thereby expanding the morphological heterogeneity of Eupherusa .

Sangster et al.’s (2023) omission of this literature is notable because Callipharus set the precedent for the degree of morphological divergence that would be expected, if one were to attempt to carve out a monotypic genus for any species in the Eupherusa ( sensu lato) clade. On morphological grounds, the case for recognising Callipharus (for E. nigriventris ) is arguably stronger than the case for recognising Dicranurania (for [ T.] ridgwayi ).

Conclusion

The purportedly diagnostic characters used to distinguish Dicranurania from the genus Eupherusa ( sensu stricto) were based on a mischaracterisation (homogenisation) of Eupherusa phenotypes, which made the characters of [ T.] ridgwayi appear to be exaggerated. Sangster et al. (2023) also overlooked the precedent of the monotypic genus Callipharus Elliot, 1879 , which further weakens their argument.

Therefore, I propose that [ T.] ridgwayi be included in a morphologically heterogenous Eupherusa , following Stiles et al. (2017), and that Dicranurania be demoted to the synonymy of Eupherusa .