Ferdinandusa includes at least 18 species of shrubs and trees found in a range of habitats from southern Central America to southeastern Brazil. It is characterized by its woody habit, generally elliptic to elliptic-oblong leaves that lack domatia, caducous interpetiolar stipules that are triangular and tightly convolute in bud, terminal cymose inflorescences, 4--5-merous flowers, white to green or red corollas with a well developed tube and the lobes convolute in bud, and distinctive, septicidal, elliptic-oblong capsules with rather large winged seeds.The corolla lobes are unusual in their emarginate tips, and these can separate unevenly in bud or split (at least on dried material) and this can complicate counting the number of lobes. Steyermark (1972, 1974) presented a very useful treatment of several Ferdinandusa species, but inaccurately described the capsules as loculicidal and overlooked the occurence of red flowers in some of the species. Some or most of the species of Ferdinandusa have the stamens inserted symmetrically in the corolla and the filaments not markedly unequal in length, but the anthers are grouped on side of the corolla; this arrangement is also found in some species of some other Rubiaceae genera (e.g., Palicourea, Condaminea). Anunciaço (2004) reported that the anthers in some species are sometimes barbate, but did not illustrate that. Several species of Ferdinandusa have bright red corollas that perhaps attract humminbirds, others have white corollas that resemble hawk-moth pollinated species in other Rubiaceae genera, and at least one species has externally green corollas and may be bat-pollinated. Some of the species are described on labels as having sweetly fragrant flowers. The most commonly collected species are, in general, Ferdinandusa chlorantha and Ferdinandusa speciosa.
Ferdinandusa has its center of species diversity in South America, in the eastern Amazon basin to eastern Brazil. Only Ferdinandusa panamensis out of South America, into southern Central America. Anunciaço (2004) commented that additional undescribed species may also be found in the Guianas and western Amazon basin. Some Ferdinandusa species are found in wet lowland forest, but many are characteristically found in seasonal formations with several characteristically in Amazonian and Guayanan scrub vegetation on nutrient-poor soils, often in areas that are seasonally inundated by blackwaters, and Ferdinandusa elliptica and Ferdinandusa speciosa frequently found in cerrado and campo rupestre vegetation.
Ferdinandusa was studied in a heroic and exceptionally knowledgable but unpublished thesis revision by Anunciação (2004; as Ferdinandea), who recognized 17 described species and three morphospecies. She detailed the pollen, leaf, and trichome anatomy, inflorescence arrangement and diversification, and seed coat surface, and identified a very large number of specimens. The thesis has some unresolved nomenclatural points but provides an excellent taxonomy for the most of the known species, and her illustrations are exceptional. Her taxonomy differed from that of Steyermark in synonymizing Ferdinandusa goudotiana with Ferdinandusa chlorantha, without discussion. Her morphospecies have not been further detailed or described but are (hopefully) better documented now.
Schumann (1889) presented the first detailed overview of Ferdinandusa, and separated two sections based on corolla form, the stamens included vs. exserted stamens, and stigma form. He included one species in his Sect. Pattalosia, Ferdinandusa speciosa, and he characterized this species as having included stamens and a subcapitate stigma. Steyermark (1972) next reviewed Ferdinandusa in a meaningful way, treating 13 species in northeastern South America. He noted that all of the Ferdinandusa species have exserted stigmas and bilobed stigmas, and did not recognize Schumann's sections. Anunciaço (2004) also documented the single species of Schumann's Sect. Pattalosia as having the anthers positioned in the corolla throat (fig. 31H) to exerted (31I) and the stigmas bilobed, and similarly did not recognize Schumann's sections. Steyermark designated a type for the genus, although it was not cited as such and thus is an inadvertent lectotypification.
This genus has several potentially confusing nomenclatural cases. The genus was first described by Pohl, who published two different name spellings for it: Ferdinandea (1827) and Ferdinandusa (1831). This problem was analyzed by Weddell in 1854, and he used the name Ferdinandusa (cf. his p. 70) and synonymized with it his own genus, Gomphosia Wedd. This genus has still been treated sometimes as Fernandea by various authors, including Anunciação (2004), but that name has also simuttaneously been regarded by other authors (e.g., Schumann, 1889, Steyermark 1972) as an illegitimate later homonym that is blocked by Ferdinanda Lag. (Asteraceae, 1816). This situation was addressed by the ICBN Nomenclature Committee for Vascular Plants, which concluded that Ferdinandea is an illegitimate later homonym in a binding decision (Applequist, 2013) so this genus now takes the name Ferdinandusa. Ferdinandea was later treated and expanded by Grisebach, who included several additional, newly described species from Cuba. These are distinct morphologicallly from Pohl's species, and were later separated by Borhidi in Suberanthus Borhidi, a genus more closely related to Rondeletia L. (Torres-Montúfar et al., 2020). The name Ferdinandusa goudotiana has been widely used for a species from northwestern South America, but this name was based on an earlier name and actually applies to Macrocnemum roseum.
The name Ferdinandusa goudotiana K. Schum. is also somewhat confusing nomenclaturally: this is a similar name to Gomphosia goudotiana Wedd., but these are nomenclaturally and Schumann's species seems to have no connection to Goudot, for whom it was named but who is not mentioned anywhere in the protologue other than the epithet. Ferdinandusa goudotiana described a species from Guyana in the northeastern Amazon basin, and its protologue included only one specimen collected by Schomburgk that thus represents its type. The name Gomphosia goudotiana has been considered a species of Ferdinandusa based on the collection by Goudot cited in its protologue by Weddell and deposited in P (e.g., Standley, 1930; Anunciação, 2004), but Weddell cited another name in synonymy here so his name is typified by that other name rather than his specimen. The name on which he based Gomphosia goudotiana is Cinchona dissimiliflora Mutis ex Humb., which is typified by Mutis's original materials. These are deposited in MA, and this name has not been formally lectotopified but Mutis's illustrations of Cinchona dissimiliflora, part of the materials that Humboldt inadvertently provided the valid names for, show both flowers and fruits and clearly represent Macrocnemum roseum, not a species of Ferdinandusa.
Ferdinandusa was long associated with the Tribe Cinchoneae based mainly on its capsular fruits and winged seeds. Andersson & Persson (1991) and Andersson (1995) found that the genera associated with Cinchona were not all closely related, and separated several include Ferdinandusa in the Calycophylleae L. Andersson C.H. Perss. based on morphological analysis. Later Kainulainen et al. (2010; as Condamineeae) found their tribe nested within the Tribe Dialypetalantheae based on molecular analysis, and related there to Simira, which has similar stipules and winged seeds along with loculicidal capsules.
Ferdinandusa is similar to and often confused with Ladenbergia, which has stipules that are help erect and flattened in bud and corollas with the lobes valvate in bud. Ferdinandusa is also similar to some species of Calycophyllum, which has leaves usually with domatia and smaller corollas with the lobes obtuse at the tip.
Author: C.M. Taylor
The content of this web page was last revised on 8 December 2022.
Taylor web page: http://www.mobot.org/MOBOT/Research/curators/taylor.shtml