Two species of seed finch (Sporophila, formerly Oryzoborus) have been documented on the island of Trinidad (Trinidad & Tobago, South America). The populations of both species have been depleted or extirpated due to persecution by cagebird enthusiasts (ffrench 2012). The smaller, dark-billed species is Chestnut-bellied Seed Finch Sporophila angolensis Linnaeus, 1776. A larger species with a proportionately heavier bill, pale in the male, is documented on Trinidad by three specimens at the American Museum of Natural History, New York (AMNH) and three at the Natural History Museum, Tring (NHMUK), all initially identified as Large-billed Seed Finch S. crassirostris J. F. Gmelin, 1789, without trinomial. The species identity of this larger form, however, has been controversial, with some authorities concluding that it is Great-billed Seed Finch S. maximiliani Cabanis, 1851, which is globally Endangered (BirdLife International 2019). S. crassirostris and S. maximiliani are morphologically very similar, with large, extremely deep bills. Adult males are black with a white patch at the base of the primaries. Females are brown, paler below, with a blackish bill.

History

Ubaid et al. (2018) presented the rather complex nomenclatural history of Sporophila maximiliani, postulating its type locality to be the Jucu River, Espírito Santo, south-east Brazil. The type locality for S. crassirostris has been designated as Cayenne (= French Guiana) (Berlepsch & Hartert 1902). P. L. Sclater (1862) was first to report a larger species of seed finch on Trinidad, listing a male and a female ‘Oryzoborus crassirostris' from Trinidad in his collection. These were the only specimens he recorded as O. crassirostris, but he also possessed a pair of ‘Oryzoborus maximiliani' from ‘Brazil'. Sharpe (1888) listed Sclater's Trinidad specimens, by then in the British Museum, as O. crassirostris along with specimens from Venezuela, Colombia and Guyana, and restricted O. maximiliani to ‘southern Brazil'. Léotaud's Oiseaux de l’île de la Trinidad (1866) provided measurements and descriptions of ‘Spermophila crassirostris - Gml’. Subsequently, Chapman (1894) collected a single ‘Oryzoborus crassirostris’ on Trinidad, now at AMNH.

For Hellmayr (1938) species identity of the Trinidad larger form was uncontroversial, as he placed all large pale-billed forms of Oryzoborus in species crassirostris. Hellmayr defined allopatric ranges for all of these subspecies, restricting O. c. maximiliani to ‘southern and interior Brazil’, and attributing all specimens from Trinidad, Venezuela, the Guianas, northern Brazil, eastern Colombia and north-east Peru to O. c. crassirostris. He wrote, however, that ‘I have not been able to correlate certain variations in dimensions and in size of bill with geographic areas’ and ‘(males) from Trinidad and the Orinoco Valley are larger, the tail in particular longer, and have much bulkier, heavier bills, some being fully as large billed as O. c. maximiliani’.

Phelps & Phelps (1950) focused on this problem in more detail, and concluded that two taxa occurred in close proximity in northern South America. They described a new subspecies O. c. magnirostris from the Orinoco Delta (including the type locality, Misión San Francisco de Guayo, Territorio Delta Amacuro), the lower Caura River, Caicara on the Orinoco River, and Trinidad. The type specimen is part of the Phelps Collection, although it is on deposit at the AMNH along with other Phelps types—and it is not ‘inaccessible’ as claimed by Ubaid et al. (2018). Six other specimens at the AMNH for which Phelps & Phelps (1950) gave sex and locality (though not registration numbers), including the three Trinidad specimens, are paratypes. Their labels have a pencil annotation ‘magnirostris', without initials. Phelps & Phelps (1950) diagnosed O. c. magnirostris as differing ‘from O. c. crassirostris in having a much larger bill and longer wing and tail’. They concluded that the three specimens from Trinidad at NHMUK also are magnirostris based on wing and culmen measurements provided by J. B. Macdonald, then curator. They remarked that ‘the known ranges of the new form and O. c. crassirostris approach each other closely’, and speculated that non-breeding flocks may ‘approach or even temporarily invade the territory occupied by the neighboring form’. The name magnirostris has since been replaced by parkesi (Olson 1981) because magnirostris is preoccupied.

Paynter & Storer (1970) recognised ‘O. crassirostris magnirostris’ and included Trinidad within its distribution (‘rare’). However, Herklots (1961) did not follow Phelps & Phelps (1950), instead identifying the taxon on Trinidad as O. c. crassirostris without discussion. He had no familiarity with the species in life, and given that he wrote that the male’s bill colour is black, it is doubtful that he had even seen specimens.

Meyer de Schauensee (1970a) asserted that nominate crassirostris and parkesi (‘magnirostris' therein) ‘actually occur together in Venezuela (El Llagual...) and in Guyana (Annai)'. His list of specimens examined included three crassirostris and three ‘magnirostris' from ‘El Llagual', all attributed to AMNH—they are not there, but he was probably referring to a series from El Yagual at the Carnegie Museum, Pittsburgh. He listed four crassirostris from Annai at AMNH, and in a postscript (‘Since writing this paper...') referred to two male ‘magnirostris' from Annai, and a male ‘magnirostris’ from ‘French Guiana, the type locality of O. crassirostris’, the latter three specimens at the Museum National d’Histoire Naturelle, Paris. The ‘additional’ specimens mentioned in the postscript, seen in Paris and also NHMUK, were not included in his list of material examined. Meyer de Schauensee (1970a) determined that nominate crassirostris was smaller, with a smaller bill, and a proportionately shorter tail than all other forms then included in crassirostris (e.g., by Paynter & Storer 1970), and therefore treated it as a monotypic species, with the remaining subspecies united under the oldest name maximiliani. This split between S. crassirostris and S. maximiliani, based on sympatry, has been recognised by subsequent authors (e.g., Ridgely & Tudor 1989, Dickinson & Christidis 2014, del Hoyo & Collar 2016). Note that many authors, e.g., Dickinson & Christidis (2014) and del Hoyo & Collar (2016), have followed the suggestion of Ridgely & Tudor 1989 to place the western South American subspecies occidentalis within crassirostris rather than maximiliani.

Meyer de Schauensee (1970a) listed just one Trinidad specimen, a male at AMNH, which he identified as ‘magnirostris’. He did not note whether he confirmed the identity of the Trinidad specimens at NHMUK. Included in his paper is a map of localities for specimen records of relevant taxa, though ‘grandirostris’ is substituted for magnirostris in error. It is ‘grandirostris’ to which Trinidad records are attributed on this map.

Specimen review by Phelps & Phelps (1950) and Meyer de Schauensee (1970a) would appear to have established that the Trinidad population belonged to S. maximiliani parkesi, not S. crassirostris. Despite this, taxonomic placement of the former Trinidad population in the literature since 1970 has been checkered, with many authors reverting to S. crassirostris without giving reason or expressing dissent (see Table 1).

Meyer de Schauensee (1970b) treated crassirostris and maximiliani as separate species in his A guide to birds of South America, but listed both as occurring on Trinidad. Inexplicably, Meyer de Schauensee & Phelps (1978) included Trinidad in the range for O. crassirostris alone, not O. maximiliani.

In their list of specimens examined, Ubaid et al. (2018) listed only one of the parkesi paratypes at AMNH as S. m. parkesi, and the others, including the three AMNH Trinidad specimens, under S. crassirostris. They did not map either species on Trinidad. The mensural data and morphological analysis by Ubaid et al. (2018) is restricted to comparison of S. m. parkesi with S. m. maximiliani, although it is unclear how specimens were assigned to S. m. parkesi prior to the analysis. Their conclusions therefore appear unsupported. Ubaid et al. (2018) discussed the distribution of nominate maximiliani in detail, but the other forms not at all.

Methods and Materials

Fifty specimens at AMNH of likely Sporophila c. crassirostris, S. m. maximiliani and S. m. parkesi, including the Phelps holotype, and the six El Yagual specimens from the Carnegie Museum (CM; see appendix), were measured and compared with the goal of discerning diagnostic characters for S. c. crassirostris and S. m. parkesi, and determining to which species Trinidad specimens belong. Measurements were taken with dial callipers, except tail length, which was taken with a ruler. Culmen was measured from the base, at the skull, to the tip. Wing, tail and tarsus were rounded to the nearest 1 mm, and bill measurements are accurate to 0.1 mm. Incomplete tails were not measured, and depths were not taken of open bills. Of these 56 specimens, 33 are black (adult) males. The remaining 23 specimens are brown. Nine of the brown individuals were labelled female, 12 as male (some denoted ‘juv’ or ‘imm’). Due to uncertainty of methods of ageing and sexing, the brown birds were compared as groups.

Results and Discussion

Measurements.—My mensural data (Tables 2–3) are closely similar to those of Phelps & Phelps (1950), an expected result since the sets of specimens are broadly overlapping. They are also similar to Meyer de Schauensee (1970a), although his tail measurements are smaller (and therefore his tail / wing indices different). The tail / wing proportions found here in parkesi are not as different from crassirostris as reported by Meyer de Schauensee (1970a). Culmen measurements in Ubaid et al. (2018) are much smaller, and perhaps represent exposed culmen rather than culmen from skull, but other measurements are closely similar.

The measurements of black males (Table 2) support Phelps & Phelps' (1950) assertion that there is a population in northern South America that is larger and larger billed than S. c. crassirostris, and these are from within the range of crassirostris, in some localities co-occurring with it (e.g., El Yagual, Venezuela). These birds, S. m. parkesi, are similar mensurally to S. m. maximiliani except in being shorter tailed. Two S. c. crassirostris (‘Bogota’ and ‘near Granada’) are much larger than the rest, and are responsible for the close approach or slight overlap of measurement ranges between S. c. crassirostris and S. m. parkesi (as noted by Phelps & Phelps 1950).

Figures 1–2.

Lateral and dorsal views of (top to bottom) males of Sporophila c. crassirostris, Guyana (AMNH 514238); S. maximiliani parkesi, Trinidad (AMNH 59130); S. maximiliani parkesi, Misión San Francisco de Guayo, Territorio Delta Amacuro, Venezuela (Phelps Collection 48301, holotype); S. m. maximiliani Goiás, Brazil (AMNH 514255) (Dale Dyer)

TABLE 1

Identification of the large species of seed finch on Trinidad by authors since Meyer de Schauensee (1970a).

Measurements of AMNH males from Trinidad are very similar to S. m. parkesi from Venezuela (Table 4) and larger than S. c. crassirostris. This is true also for measurements given in Phelps & Phelps (1950) of the NHMUK Trinidad males. Those for the NHMUK female are smaller, but see below. Leotaud's (1866) wing measurement also fits S. m. parkesi.

Mean measurements of brown birds (Table 3) are smaller than for black males, but show the same pattern of size difference between S. m. maximiliani and S. c. crassirostris. I am unable to determine if sex, maturity, or both are responsible for smaller size in brown birds. Mention should be made here of CM 33264, an ‘imm male’ crassirostris from El Yagual, which has the longest wing measured herein. Its tail / wing index and bill measurements are consistent with identification as S. crassirostris. Its tag annotation states ‘Wing too long for crassirostris but [illegible]’.

Colour.—Adult male S. c. crassirostris are black with a noticeable blue-green sheen. The three male S. m. maximiliani are browner black with much slighter blue-green reflections. Most S. m. parkesi are similar to nominate maximiliani, but the type has a blue-green sheen as strong as some S. c. crassirostris males. Blackness and sheen may be related to condition of the plumage. The amount of white at the base of the primaries, as well as on the underwing-coverts and axillaries, varies individually.

The bills of male specimens of S. c. crassirostris have most of the maxilla and the sides of the mandible grey, with the tomia and underside of the mandible cream or buff. On photos of live birds the bill appears silver-white (e.g., Angehr 2011). Bills of specimens of S. m. maximiliani and parkesi are cream with a small but variable dark area at the base. Photos of live S. m. maximiliani show a very pale pink bill (e.g., Ubaid et al. 2018). I do not see the difference in bill texture that Meyer de Schauensee (1970a) reported (‘shiny’ in crassirostris vs. ‘bonelike’ in maximiliani).

TABLE 2

Measurements (in mm) of black males, given in the format: range (mean, number (n) of specimens) for each variable. Trinidad specimens are included in S. m. parkesi.

TABLE 3

Measurements (in mm) of brown birds, given in the format: range (mean, number (n) of specimens) for each variable.

TABLE 4

Comparison of Trinidad specimens with Venezuela S. m. parkesi (black males).

The colour of brown birds is very variable, dull brown above, but ranging from deep tawny-rufous to pale greyish buff below. Individuals of S. m. maximiliani are darker, redder and more saturated than S. crassirostris, but immatures may be darker than adults, and worn birds are palest. Some S. crassirostris marked ‘male imm’ show a few black feathers, and some white on the primaries. Bills of all are dark.

Conclusion

Adult males of S. m. parkesi are diagnosable from S. c. crassirostris by larger measurements in all characters, especially bill and tail. Specimens of large seed finches from Trinidad are S. m. parkesi, not S. c. crassirostris. Criteria in Ubaid et al. (2018) for distinguishing S. maximiliani from S. crassirostris were not stated, nor was rationale given for their placement of several S. m. parkesi paratypes (including those from Trinidad) in S. crassirostris (p. 569), and without of examination of the S. m. parkesi holotype. Thus their morphometric comparison of S. m. maximiliani and S. m. parkesi is flawed, and the conclusion that they ‘consider S. m. parkesi (Olson, 1981) (= Oryzoborus maximiliani magnirostris Phelps & Phelps, 1950) as a synonym of the nominate taxon’ unsupported.

S. maximiliani is Endangered (BirdLife International 2019). Any effort to restore its population on Trinidad would probably require reintroduction. Although such a project is unlikely, as the threat is not habitat loss but continued trapping (ffrench 1985, White et al. 2015), restoration of the historical Trinidad population would depend on introduction of the correct species.