Drosera hirtella Saint-Hilaire (1826: 262)

Gonella, Paulo Minatel, Sano, Paulo Takeo, Rivadavia, Fernando & Fleischmann, Andreas, 2022, A synopsis of the genus Drosera (Droseraceae) in Brazil, Phytotaxa 553 (1), pp. 1-76 : 40-45

15. Drosera hirtella Saint-Hilaire (1826: 262) . Figures 7b, 14, 15a–e, 16

Lectotype: — BRAZIL. Minas Gerais: [one sheet with two different labels corresponding to two collection localities, which cannot be individually assigned to the seven mounted Drosera specimens] Marais desséchés près Formiga dans le Certão, Cap. des Mines et dans la Serra dos Pyreneos [desiccated marshes near Formiga ( Montes Claros de Formigas ) in the sertão, Minas Gerais State, and in the Serra dos Pirineus] + [on a second label] Serra dos Pyreneos, Cap. de Goyaz [Serra dos Pirineus, Goiás State], s.d., Saint-Hilaire B1-1762bis ( P-749158 !, first-step lectotype designated by Correa & Silva (2005) , second-step lectotype designated here: second specimen from the right, Fig. 16a) .

≡ D. montana var. hirtella (A.St.-Hil.) Diels (1906: 89) .

Seasonal geophyte perennial, rosetted, acaulescent. Roots succulent. Leaves decumbent, with geniculate-involute vernation, distinctly petiolate, spatulate, lamina obovate to suborbicular, red (greenish in shaded environments); stipules rectangular in outline, divided into several laciniae from near the base. Scape base strongly arcuate (often decumbent and ascending outside the rosette; Fig. 14a–f), scape red in color (greenish-red in shaded environments), densely eglandular-pilose on the basal half, reducing in density towards apex, eglandular trichomes red (Fig. 14f), 1–2 mm long, scape becoming densely glandular-pilose towards the apical half. Sepals exclusively glandular-pilose or also with sparse eglandular trichomes; petals pink; gynoecium 3-carpelate, styles bifurcated at the base. Seeds ovoid, black, testa reticulate.

Illustrations: —this work, Figure 15a–e.

Distribution: — Brazil (North: TO; Northeast: BA; Central-West: DF, GO; Southeast: MG), endemic. Drosera hirtella has two main disjunct distribution centres, each representing a more or less continuous range: one along the Espinhaço Range in Minas Gerais and Bahia, and another in central-northern Goiás (including the Distrito Federal), western Minas Gerais, western Bahia, and southeastern Tocantins (Fig. 7b).

Habitat: —Sandy to peaty soils in seasonally wet areas of campos rupestres, and also along the margins of veredas and springs in campo limpo vegetation in the Cerrado domain. Found at elevations between 600–1600 m.

Phenology:— Drosera hirtella has been collected in flower from March to July, corresponding to late wet season and mid dry season. From late July to November (mid dry season to early wet season) the plants usually lie dormant as roots, when only dry leaves and scapes may be seen. Slightly succulent roots have been observed.

Conservation status: —Least Concern (LC). Drosera hirtella is widespread and common (AOO= 364 km 2 and EOO= 586,293.847 km 2). It occurs inside the National Parks of Brasília (DF), Chapada Diamantina (BA) , Chapada dos Veadeiros (GO), Grande Sertão Veredas, Sempre-Vivas and Serra do Cipó (MG), as well as the State Parks of Biribiri, Rio Preto, Serra do Cabral, and Serra Nova (MG), and Serra dos Pirineus (GO).

Notes: — Drosera hirtella is most similar to D. lutescens (Fig. 14g) but is distinguished by the obovate to suborbicular lamina (Figs. 14a, e, 15b, c; vs. narrowly obovate to spatulate, Figs. 15g, 18c), scape base strongly arcuate, often decumbent and ascending outside the rosette (Figs. 14c, 15a; vs. slightly arcuate, Figs. 15a, b, 18b), scape red, with red eglandular trichomes (Fig. 14a–g; vs. scape yellowish green, with white eglandular trichomes, Figs. 14g, 18a–e).

Drosera hirtella and the infrataxon D. hirtella var. lutescens were described simultaneously by Saint-Hilaire (1826), who claimed that the latter differed from the autonymous variety by the “smaller leaves, obovate laminae [vs. obovate to round-obovate] which are more frequently glabrous abaxially, and eglandular trichomes on a more rigid and yellowish scape [vs. eglandular trichomes on a more flexuous and red scape]”.

However, subsequent treatments on the genus (Eichler 1872, Diels 1906, Hamet 1907, Correa & Silva 2005 ) denied both taxa as being distinct. Leaf shape has been widely disregarded for small, rosetted Brazilian species (see Rivadavia et al. 2014), and the color of trichomes on scapes is often very difficult to discern in historical, dried specimens, although very conspicuous in living specimens. For these reasons, D. hirtella has been historically considered a synonym under a broadly circumscribed D. montana for over a century. This wide taxonomic concept of the rosulate Brazilian Drosera started with Eichler (1872), who placed D. hirtella var. lutescens under synonymy of D. hirtella , and was then followed by Diels (1906), who included D. hirtella (including D. hirtella var. lutescens ) as an infrataxon of D. montana , namely D. montana var. hirtella . The latter taxonomic concept was strictly followed by all subsequent taxonomic works of the 20th Century, with the exception of Santos (1989), who corroborated the original circumscription of both D. hirtella var. hirtella and D. hirtella var. lutescens sensu Saint-Hilaire (1826) adding further morphological data. Finally, Correa & Silva (2005) placed both taxa, along with all other small, red, rosulate Brazilian Drosera , under synonymy of a single taxon, D. montana , with no varieties.

Such a broad taxonomic concept of D. montana , however, had already been revealed to be artificial in the phylogenetic reconstructions by Rivadavia et al. (2003): D. hirtella belongs to a clade of diploid, mostly Neotropical species (D. section Drosera, sensu Fleischmann et al. 2018a , b), while D. montana was retrieved from a clade of tetraploid, predominantly Brazilian species (D. section Brasilianae ; Fleischmann et al. 2018a, b). This led Rivadavia et al. (2014) to exclude D. hirtella and allies from the D. montana complex, but a detailed treatment for this species had been lacking until now.

Based on extensive field and herbarium studies, both D. hirtella var. hirtella and D. hirtella var. lutescens are here considered two taxa very distinctive from each other, and from all other Drosera species, being re-established (sensu Saint-Hilaire 1826) following several previous works that have already pointed out the necessity of recovering the two names (e.g. Santos 1989, Rivadavia et al. 2003, 2014, Rivadavia 2008, Rivadavia & Gonella 2011, Gonella et al. 2012, 2014). In contrast to Saint-Hilaire (1826), but in accordance with contemporary generic concepts in Drosera (e.g. Gonella et al. 2014, Lowrie 2014), these two distinctive, sometimes sympatric and even syntopic taxa (Fig. 14g) are distinguished here at species rank, not as infrataxa; hence D. hirtella var. lutescens is elevated to species rank (see D. lutescens ).

As the first step in this process, it became necessary to properly identify and circumscribe the autonymous D. hirtella , because in-depth examination of Saint-Hilaire’s type material (Fig. 16) found it to be a mixed gathering of different plants from at least two different localities separated by over 530 km: Formiga (referring to Montes Claros de Formigas, current Montes Claros, N Minas Gerais; see Saint-Hilaire 2011 —although Saint-Hilaire had also been to the town of Formiga in S Minas Gerais on the route to Goiás, his B1 notebook gathers specimens he collected during his travels in N Minas Gerais, e.g. D. ascendens , D. sessilifolia , and D. spiralis ), and Serra dos Pirineus, in central Goiás. To make matters even more complicated, both D. hirtella and D. hirtella var. lutescens occur sympatrically and even grow syntopically, occasionally hybridizing, on the Serra dos Pirineus (see ‘Natural Hybrids’ below).

Correa & Silva (2005) designated the specimen “ Saint-Hilaire B1-1762 ” as lectotype, but not only is the collection number incorrect (this number refers to a Melastomataceae specimen, according to Saint-Hilaire’s field notebook), but the specimen which they probably intended to select as the lectotype (B1-1762bis) is a mixed gathering of Drosera (seven individuals are mounted on the sheet P-749158: three flowering rosettes, four detached scapes) and one individual of a graminoid plant. Therefore, a second-step lectotype is here designated (ICN Arts. 9.11 and 9.17; Turland et al. 2018).

Furthermore, the selection of a second-step lectotype was deemed necessary in order to establish a circumscription of D. hirtella , since the specimens mounted on the first-step lectotype sheet (as designated by Correa & Silva 2005 ) turned out to be a mixed gathering of the two morphotypes of D. hirtella recognized here:

i. “Type morphotype” (sensu the lectotype as here designated; Fig. 16b): leaves with round-obovate lamina (Fig. 15c), scapes strongly curved at the base (Fig. 14a–c, e; usually decumbent at the base, ascending from outside the rosette), and densely covered with shorter and more flexuous eglandular trichomes.

We selected a specimen with the round-obovate leaf lamina as the lectotype, as this is the most widespread morphotype, being the only one growing allopatrically in the eastern range of the species (along the Espinhaço Range), as well as occurring sympatrically (often syntopic) with the other morphotype (and D. lutescens ) in the western range of the species, in Goiás.

ii. “Western morphotype” (represented in the sheet P-749158 by the individual immediately to the left as the lectotype here selected; Fig. 16c): leaves with narrowly obovate lamina (Fig. 15b), scapes notably curved at the base, but not as strongly and widely arcuate as in the “type morphotype” (scape ascending within the rosette), scapes entirely reddish, or reddish at the base and yellowish towards the apex, with slightly longer and more rigid, red, eglandular trichomes.

This morphotype occurs only in the western range of the species, from central Goiás to the Jalapão area, in southeastern Tocantins. In its southern and central range, this morphotype often occurs sympatrically and syntopically with the “ type morphotype”, as well as D. lutescens . However, in the northern part of its range the “western morphotype” occurs solitarily, allopatric from closely allied taxa (no “ type morphotype” D. hirtella , nor D. lutescens have been found in Tocantins). It is possible that the “western morphotype” is of hybrid origin, due to its somewhat intermediate morphology between the “ type morphotype” of D. hirtella and D. lutescens , but further studies (including a genomic analysis) would be necessary to test this hypothesis. What can be ascertained is that the “western morphotype” does not simply constitute a primary hybrid, as it occurs geographically separate (allopatric) from D. hirtella and D. lutescens in part of its range and breeds true from seed (pers. obs.). It could represent an amphiploid hybrid, requiring further investigation.

Although the “western morphotype” forms large and stable populations, the differences between this and the “ type morphotype” are very subtle and intermediate plants have been observed when these morphotypes occur syntopically. Consequently, until further studies can be carried out, they are both here considered within the natural variation of D. hirtella , and identified as “morphotypes’’, without assigning any taxonomic rank (mirroring the authors’ concepts of intraspecific morphological variation of D. latifolia and D. montana , see Gonella et al. 2014 and Rivadavia et al. 2014).

The circumscriptions here adopted for D. hirtella and D. lutescens mostly follow those of Saint-Hilaire (1826), with the most reliable distinctive characters between the two taxa being color and eglandular indumentum of their scapes, although leaf shape and curvature of the scape are also diagnostic. Drosera hirtella presents red or reddishgreen leaves with obovate to suborbicular lamina, scapes usually red (sometimes green) with a very pronounced curve at the base (often with decumbent base running parallel to the ground, ascending outside the rosette) and densely covered with red eglandular trichomes. Drosera lutescens has vinaceous (sometimes green) leaves with narrowly obovate to spatulate lamina, yellowish-green scapes (sometimes reddish near the apex) with the base only slightly curved and densely covered with white eglandular trichomes.

Rare hybrids between D. hirtella and D. lutescens were observed at two sites where their ranges overlap in central Goiás, on the Serra dos Pirineus and near the town of Cristalina. The leaves of these hybrids are intermediate in shape and color, while the scapes are intermediate in their curvature, reddish near the base and orangish towards the apex, with intermediate-sized eglandular hairs that are reddish near the base, turning pinkish-white towards the apex (for specimens and photographs, see ‘Natural Hybrids’ below).

Although D. hirtella and D. communis can be found syntopically wherever their ranges overlap, hybrids between them are only known from one site in northern Goiás, at the Chapada dos Veadeiros. The hybrids are very abundant, and their morphology is intermediate between the parent taxa, the leaves being semi-erect and scapes with red eglandular trichomes near the base (see ‘Natural Hybrids’ below). Since both morphotypes of D. hirtella co-occur syntopically at that site, it is not known which one is the parent of the hybrids.

Drosera hirtella is also believed to hybridize with D. cayennensis at a site where both species grow syntopically near the town of Cristalina in Goiás. The suspected hybrids are morphologically intermediate, with leaves shorter than D. hirtella , scapes erect at the base, covered by red eglandular trichomes on the basal half, and seeds with reticulate testa (see specimen and photograph under ‘Natural Hybrids’ below).

Representative specimens examined: — BRAZIL. Bahia: Abaíra, subindo a Serra do Barbado , 14 July 2005, Rivadavia et al. 2008 ( SPF) . Drainage of the Rio Corrente , western Bahia, near Rio Piau , ca. 150 km SW of Barreiras, 14 April 1966, Irwin et al. 14845 (NY, UB) . Goiás: Alto Paraíso de Goiás, Jardim de Maytrea , 25 February 2014, Gonella & Andrino 650 (M, SPF) . Cristalina, BR-050 para Brasília , 30 April 1999, Rivadavia-Lopes & Sato 981 ( SPF) . Pirenópolis , Serra dos Pirineus, 21 March 1997, Rivadavia 670 ( SPF) ; ibid., 23 April 1999, Rivadavia & Sato 927 ( SPF) . Distrito Federal: Brasília, estrada Brasília-Buritis (DF-250), 18 March 1997, Rivadavia-Lopes 641 ( SPF) . Brasília, Fazenda Sucupira , 09 February 2007, Fontes et al. 25 ( CEN, UB) . Minas Gerais: Arinos, estrada para Chapada Gaúcha , 29 April 1999, Rivadavia-Lopes & Sato 964 ( SPF) . Diamantina , ao sul da cidade, 20 April 2010, Gonella et al. 275 ( SPF) . Itacambira, estrada para Montes Claros , 05 March 1997, Rivadavia 613 ( SPF) . Monte Azul, estrada para Sucuruiu / Riacho Grande / Brejo Grande , 09 September 2011, Gonella et al. 507 ( SPF) . Ouro Branco, Serra do Ouro Branco, 01 October 2009, N.G. Silva et al. 376 (R). Santa Bárbara, Serra do Caraça , 22 May 1997, MelloSilva et al. 1332 ( MBM, SPF) . Santana do Riacho, Serra do Cipó , 15 May 2008, Gonella et al. 82 ( SPF) . Tocantins: Mateiros, Jalapão , 11 June 2006, Rivadavia 2193 ( SPF) . Ponte Alta do Tocantins, Jalapão , 06 August 2012, Gonella et al. 568 ( SPF) .