Galactia purshii Desv., Ann. Sci. Nat. (Paris) 9: 413. 1826

Franck, Alan R., 2017, Typifications for Galactia purshii and G. volubilis (Fabaceae), PhytoKeys 85, pp. 11-26 : 12-16

Galactia purshii Desv., Ann. Sci. Nat. (Paris) 9: 413. 1826 Figs 1 , 2

Galactia purshii Desv., Ann. Sci. Nat. (Paris) 9: 413. 1826. Galactia glabella DC., Prodr. 2: 238, 1825 nom. illeg. (Art. 53.1) non G. glabella Michx. Fl. Bor.-Amer. (Michaux) 2: 62. 1803 nom. illeg. (Art. 52). Lectotype (designated here): Carol. [Carolina] mer. [meridionale], Fraser s.n. (G [G00726366]). = Galactia floridana Torr. & A.Gray var. longeracemosa Vail, Bull. Torrey Bot. Club 22: 505. 1895. Lectotype (designated by Nesom 2015): USA, Florida, 1889, Simpson s.n. (US; isolectotypes, MU, US). syn. nov. = Galactia michauxii A.R.Franck, Phytologia 99: 148-149. 2017. Type: USA, Florida, Palm Beach Co., W side of US 1, Juno Beach area, 21 Apr 1962, Lakela 24958 (holotype, USF; isotype, FSU). syn. nov.

Remarks.

One species of Galactia found along the Atlantic Coast and Florida peninsula, USA, that still lacks stable nomenclature is characterized by its strigose stems that are prostrate to occasionally twining or climbing, petioles usually shorter than the terminal leaflet, leaflets often drying to a darkened or brownish color with conspicuous reticulate venation adaxially and prominent secondary venation abaxially, flower buds usually acuminate at the apex, and non-villous flowers 10-18 mm long not drying reddish that are often congested together near the apex of the inflorescence. The earliest applicable name for this species is G. purshii , a name chiefly ignored, and the recently introduced G. michauxii A.R.Franck is a synonym. This species had previously gone under several misapplied names, including G. glabella Michx., nom. illeg. ( Pursh 1814, Nuttall 1818, Elliott 1824, Candolle 1825, Torrey and Gray 1838-1840, Chapman 1860, Britton 1881, Duncan 1979), G. regularis (L.) Britton et al. ( Britton et al. 1888, Vail 1895, Small 1903, Small 1933, Rogers 1949, Long and Lakela 1971, Wunderlin 1982, Gleason and Cronquist 1991, Isely 1998,), G. volubilis ( Ward and Hall 2004, Wunderlin and Hansen 2011), and G. brachypoda Torr. & A.Gray ( Nesom 2015, 2017). Due to the scarcity of specimen citations, it is often difficult to know if the species concepts of these authors were wholly equivalent to G. purshii or had conflated G. purshii with other species.

The first description of the taxon here referred to as G. purshii may be attributable to Michaux (1803), who supplied a brief and somewhat insufficient description. He introduced the name G. glabella , probably partly based on one of his own specimens (P [P00680461], Nesom 2015: fig. 6, Franck 2017a: fig. 15), and gave its distribution as Carolina and Georgia. This sterile specimen has narrowly ovate and brownish leaflets with conspicuous reticulate venation, leafy curvaceous stems, and one leafless twining stem. The twining stem is a disconnected fragment twining around the leafy stems. Assuming all stems are of the same plant, the Michaux specimen is consistent with G. purshii ( Duncan 1979 [as G. glabella ], Franck 2017a [as G. michauxii ]). The Michaux specimen is similar to other specimens of G. purshii that have some moderately twining stems such as Daoud 49 (USF) from North Carolina, Kral 11078 (USF) from Virginia, and Seymour 91 7 20 (USF) from Virginia, each of which had been previously annotated with three different names: G. glabella , G. regularis , and G. volubilis . The Michaux specimen had also been identified as G. volubilis partly because of the twining found on the specimen ( Nesom 2015). However, G. volubilis usually has profusely twining stems and leaflets that dry to a light green and have obscure, inconspicuous reticulate and secondary venation unlike the Michaux specimen. The leaflet shape and stem vestiture also do not seem consistent with G. mollis Michx. or G. regularis , which also occur in the Carolinas and Georgia.

Nevertheless, G. glabella is an illegitimate name because Michaux cited the earlier name Ervum volubile Walter in synonymy, thereby adopting the type of E. volubile ( McNeill et al. 2012: Art. 52, Franck 2017a). Ervum volubile (= G. glabella ) so far remains untypified. Within Galactia its precise application may be relatively inconsequential since its specific epithet is blocked by G. volubilis (L.) Britton. Additionally, the species of Galactia in the eastern USA are likely to remain in Galactia as they appear to be closely related to the type species of Galactia , G. pendula Pers., and part of a monophyletic clade based on a recent DNA phylogeny ( Queiroz et al. 2015).

Pursh (1814: 487) adopted the name G. glabella and expanded Michaux’s description by adding the Latin terms "prostrata, subvolubilis, foliis ternatis utrinque glabris, racemis axillaribus simplicibus abbreviatis paucifloris" [prostrate, partly twining, leaflets three both sides glabrous, raceme axillary singular short few-flowered], and "leguminibus villosis" [legume villous]. With the annotation “v.v.” he noted he had made field observations of this species. Pursh stated that the flowers were "extremely pretty, purple, red and white mixed." His description of purple flowers matches G. purshii but the red flower color and villous legume seem to pertain more to G. mollis ( Radford et al. 1968: 644), possibly indicating Pursh had included more than one species in his description. Pursh gave the distribution as New Jersey to Carolina and cited E. volubile of Walter and Dolichos regularis of Willdenow (1803) in synonymy. Willdenow (1803: 1049) had simply repeated the description of D. regularis from Linnaeus.

Nuttall (1818: 117) continued the use of G. glabella , with his description closely matching that of Pursh (1814). Nuttall added that the leaves were subcoriaceous and lucid, racemes pedunculate and a little shorter than the leaves, flowers pedicellate, and legumes smooth. Under G. mollis , Nuttall stated that "In Herb. Muhl. [ G. mollis was] confounded with G. glabella ," supporting the notion that Pursh may have also conflated G. mollis with his concept of G. glabella . Galactia glabella was also recognized by Elliott (1824: 239).

Candolle’s (1825: 238) Latin description of G. glabella was nearly verbatim of Pursh (1814), but added that the flowers were pedicellate as Nuttall (1818) had also described. Candolle mentioned that the legume was villous based on Michaux and Pursh but was glabrous based on Nuttall and his own observations. Anent this discrepancy he stated "An duae spec. confusae?" Since Michaux never described the fruits in the protologue, perhaps Candolle observed a Michaux specimen labeled G. glabella with villous fruits, a character which would be more like G. mollis . In synonymy Candolle listed E. volubile Walter and D. regularis L.

In Desvaux’s (1826) account of Galactia , he included descriptions for five species. For his first species he introduced the name G. purshii , validated solely by the description of G. glabella given by Candolle (1825: 238). I could find no explanation for its etymology, but it presumably honors Frederick Pursh. Desvaux considered G. glabella and its listed synonyms ( Candolle 1825: 238) to be misapplied to the newly coined G. purshii . Desvaux excluded the synonyms listed by Candolle ( E. volubile and D. regularis ) with the abbreviation "excl. syn." By excluding E. volubile , Desvaux excluded G. glabella since it is a superfluous name homotypic with E. volubile . Furthermore, for his second species, Desvaux (1826) listed and provided a separate description for G. glabella of Michaux, and included G. pilosa as its synonym. Galactia pilosa is currently considered a synonym of G. mollis and the ambiguously described E. volubile (= G. glabella ) might also be conspecific with G. mollis ( Franck 2017a). Desvaux’s description of G. glabella could fit the current concept of G. mollis or G. volubilis . Desvaux errantly cited page 64 instead of page 62 for the protologue of G. glabella Michx.

After Desvaux’s (1826) treatment, G. purshii was abandoned from usage, treated as a synonym, or considered illegitimate. The name G. glabella continued to be utilized ( Torrey and Gray 1838-1840: 287, Chapman 1860: 109, Britton 1881: 27), although it is difficult to ascertain if its taxonomic concept was completely equal to the concept of G. purshii here. Galactia glabella was then considered a synonym of G. regularis without mention of G. purshii ( Britton et al. 1888: 14, Jackson 1893: 987, Small 1903: 650, Small 1933: 719, Long and Lakela 1971: 493, Gleason and Cronquist 1991: 305, Isely 1998: 569). Vail (1895) listed both G. glabella and G. purshii as synonyms of G. regularis . Duncan (1979) resurrected the use of G. glabella , including G. purshii in synonymy and separating it from G. regularis , and some specimens at USF were annotated by Nesom as G. glabella . Ward and Hall (2004) included G. glabella as a synonym of G. volubilis .

Galactia purshii is a legitimate name since its protologue unequivocally excluded D. regularis , E. volubile , and G. glabella . Since G. purshii is validated by Candolle’s description (1825: 238), any specimens seen by Candolle for his treatment should be considered original material ( McNeill et al. 2012: Arts. 7.7 and 9.3, note 3). Candolle’s (1825) annotation “(v.s.)” indicated he had seen specimens. There are two specimens in the Candolle herbarium together on one sheet, labeled in Candolle’s handwriting ( Burdet 2017) as " Galactia glabella Nutt. Michx." (Fig. 1 ). The specimen on the right (G00726367) is G. volubilis , twining and with ovate lightly glaucous leaflets drying pale greenish with inconspicuous secondary and reticulate venation. The John Fraser specimen on the left (G00726366, Fig. 2 ) is consistent with G. glabella sensu Candolle (1825); it has leaflets drying to a dark brown adaxially with conspicuous abaxial secondary and reticulate venation, and flowers ca. 13 mm long. The indumentum of the stem is retrorsely strigose and of the calyx abaxially antrorse but scant (L. Gautier, pers. comm.). A phrase written on the label "An Ervum Walter" indicated an association with the modern sense of Galactia . This Fraser specimen must have come from L’Héritier’s herbarium, which was purchased by Candolle in 1805 ( Gray 1889, Stafleu 1966, Brummitt 1972), and can be considered part of the original material. The Fraser specimen (Figs 1 - 2 ) is designated here as the lectotype of G. purshii . The specimens cited for G. michauxii by Franck (2017a: Appendix 1) are here identified as G. purshii , with G. michauxii being a later synonym.

Recent descriptions for G. fasciculata Vail, such as having strigose stems ( Nesom 2015, 2017), may apply to specimens here considered to be G. purshii . However, the indumentum of the type specimens of G. fasciculata appears more similar to the villous stems of G. floridana ( Isely 1998, Franck 2017a). Galactia fasciculata was described as prostrate or climbing high by Vail (1895), whereas Nesom (2015, 2017) described it as high-climbing with coiling stems. The holotype label stated "climbing on small shrubs." Additional study is needed to determine if a high-climbing habit is a reliable and distinctive character since it can only be confidently ascertained from living plants and field observations. Ward and Hall (2004) considered G. fasciculata very rare while Nesom (2015, 2017) considered it an endemic of central peninsular Florida.

The stem indumentum of the holotype of G. floridana Torr. & A.Gray var. longeracemosa Vail does not appear to be villous like G. floridana , but appears more like G. purshii . Galactia floridana var. longeracemosa is considered here to be a synonym of G. purshii . The holotype of G. floridana var. longeracemosa was probably collected by Joseph H. Simpson relatively near to Bradenton, Manatee Co., Florida where he had lived ( Small 1919; Harper 1948).

The name G. brachypoda was apparently misapplied ( Nesom 2015, 2017) to specimens here considered to be G. purshii . Galactia brachypoda was first described by Torrey and Gray (1838-1840), who indicated the habit as not twining with a two foot long flexuous stem, the calyx villous, and the flowers half as large as G. glabella ( G. glabella sensu Torrey & Gray probably being misapplied to G. purshii ). Chapman (1860), who collected the two type specimens of G. brachypoda (NY), described it as erect, 1-1.5 feet high and with a woolly calyx, noting his descriptions were "all my own, copying no one, when I knew the plant" ( Chapman 1839-1890: 4 Apr 1959). On the label of the presumed holotype Chapman wrote "seems to come between G. mollis & G. sessiliflora [= G. erecta (Walt.) Vail]" which was later crossed out by a different pen, possibly by Torrey who also added what appears to be " brevipedunculata n. sp." to the same label. Vail (1895) described its calyx as "clothed with spreading" hairs, the lower calyx lobes acutish, corolla 8-10 mm long, and the “vexillum” 7-8 mm long, which Small (1903) mostly repeated. All of these observations are consistent with the type specimens of G. brachypoda , none of which match the concept of G. purshii .

The habit of G. brachypoda has sometimes been described as similar to G. purshii . However, the descriptions of a decumbent ( Vail 1895, Small 1903, Small 1933), ascending or sprawling ( Isely 1986), procumbent ( Nesom 2015), or prostrate habit ( Nesom 2017) for G. brachypoda appear to be based on speculation from specimens and not field observations. While it was conjectured that it was impossible for G. brachypoda to be erect because its type specimens had stems to 37 cm long ( Nesom 2017), another collection identified as G. brachypoda (Anderson 15642 [FSU, GA]) with stems well over 40 cm long described on its label "robust, erect plants with limited twining" and was noted to be very similar to the type specimens of G. brachypoda ( Franck 2017a, Nesom 2017). Furthermore, stems of some specimens of the erect G. erecta can reach 32-36 cm long (e.g. Biltmore 3956a [NY], Horn 1032 [DUKE], Orzell & Bridges 14271 [USF], Rugel 150 [NY]). It does not appear to be impossible for G. brachypoda to be erect and have stems to 37 cm long. The characterization of G. brachypoda as erect by Chapman (1860) and Anderson 15642 is considered here to be accurate.

Numerous authors noted a semblance of G. brachypoda with G. erecta ( Rogers 1949, Ward and Craighead 1990, Isely 1998, Ward and Hall 2004, Franck 2017a), while others also noted a similarity to G. mollis ( Chapman’s notes on the holotype of G. brachypoda , label notes of Anderson 15642, Franck 2017a). The acutely-tipped flower buds and relatively small reddish-drying flowers of the type specimens of G. brachypoda are features shared with G. erecta and G. mollis . The erect habit and elliptic leaflets with relatively long petioles of the type specimens of G. brachypoda are more similar to G. erecta . If the inflorescences of the type specimens of G. brachypoda are interpreted as immature ( Nesom 2017), the sizes of the flower buds and corolla (including the individually mounted petals of the holotype) are still rather small compared to G. purshii . The long stems, pedunculate inflorescences, and villous calyces of the type specimens of G. brachypoda are more similar to G. mollis . However, inflorescences of G. erecta can occasionally be pedunculate, with a peduncle to 14 mm long in Harper s.n. (NY [02569186]). Nesom (2017) characterized the calyx of G. brachypoda as "very sparse," dissimilar to other observations of the calyx as villous ( Torrey and Gray 1838-1840), woolly ( Chapman 1860), or "clothed with spreading" hairs ( Vail 1895). Another rather odd specimen (Duncan 17113 [GA]) seems to mix features of G. erecta and G. mollis in that it has subsessile inflorescences and long petioles like G. erecta and long, partly twining stems like G. mollis . Lastly, the left-most plant of a Chapman collection at MO (793008) appears erect like G. erecta but has shortly pedunculate inflorescences and indumentum more like G. mollis .

There are two known type specimens of G. brachypoda at NY (00008088 and 00008090), although there is a third specimen (NY [00008089]) that was labeled as G. brachypoda in Chapman’s handwriting. This third specimen is clearly G. erecta . It had been proposed that other authors were attempting to make two species out of G. erecta with the use of the name G. brachypoda through the study of this G. erecta specimen labeled as G. brachypoda ( Nesom 2015, 2017). This specimen consists of plants ca. 13 cm tall with subsessile inflorescences. Among authors who recognized both G. brachypoda and G. erecta , this specimen matches their concepts of G. erecta , and is clearly incongruent with their concepts of G. brachypoda ( Torrey and Gray 1838-1840, as G. sessiliflora Chapm., Chapman 1860, Vail 1895, Small 1903). The discordance of this specimen with Chapman’s (1860) concept of G. brachypoda suggests the possibility of a labeling error. Vail annotated the holotype of G. brachypoda , but not this G. erecta specimen. This G. erecta specimen was otherwise annotated only by Anita F. Cholewa in 1986, erroneously as a probable isotype of G. brachypoda . When Isely (1986) mentioned that G. brachypoda could be a "freak form" of G. erecta , he also stated that there were "two Chapman sheets [of G. brachypoda ] at NY" and that G. brachypoda had pedunculate inflorescences, unlike this G. erecta specimen. Ward and Craighead (1990) speculated G. brachypoda was "probably an aberrant form" of G. erecta , and later Ward and Hall (2004) also stated that G. brachypoda was "based upon two A.W. Chapman specimens (NY)." The evidence does not support the idea that this specimen (NY [00008089]) nor any other of G. erecta was used to formulate concepts of G. brachypoda . I concur with previous botanists that G. brachypoda is closely related to G. erecta and G. mollis , and numerous features associated with the type specimens of G. brachypoda (i.e., its erect habit, elliptic leaflets on a long petiole, acutely-tipped flower buds, villous calyx, and relatively small reddish-drying flowers) are inconsistent with G. purshii .

Figures

Figure 1. Sheet at G labeled Galactia glabella by Candolle. The specimen on the left is the lectotype of G. purshii (G 00726366). The specimen on the right is G. volubilis (G 00726367). Conservatoire et Jardin botaniques de la Ville de Geneve.

Fgure 2. Close-up of the lectotype of G. purshii (G 00726366). The stem fragment on the right appears to be from the specimen of G. volubilis. Conservatoire et Jardin botaniques de la Ville de Geneve.