Drosera graminifolia Saint-Hilaire (1826: 269)

Gonella, Paulo Minatel, Rivadavia, Fernando & Sano, Paulo Takeo, 2012, Re-establishment of Drosera spiralis (Droseraceae), and a new circumscription of D. graminifolia, Phytotaxa 75 (1), pp. 43-57 : 45-49

Drosera graminifolia Saint-Hilaire (1826: 269) . Figs. 2 , 3 .

Type:— BRAZIL. Minas Gerais: “Crescit in summis montibus dictis Serra-da-Caraça; alt. circiter 6000 ped. Florebat Februario”, Saint-Hilaire s.n. (holo P!) .

Drosera graminifolia var. major Eichler (1872: 396) . Type:— BRAZIL. Minas Gerais: Serra do Caraça , December 1830, Sellow 1300 (holo B image!).

Illustrations:— Saint-Hilaire (1826: t. XXV); Diels (1906: 87, fig. 31M, N).

Perennial rosetted herbs, acaulescent or forming short stems up to 6(–15) cm long, covered by the persistent dead leaves. General indumentum consists of white eglandular hairs 1.5–3.5 mm long (longest on leaves), translucent-yellow short-stalked multicellular globose (“TSG”) trichomes 0.1–0.12 mm in diameter, and sessile glands 0.03 mm in diameter. Leaves 110–240(–295) mm long, linear, with regular circinate vernation, green in color, erect to semi-erect, patent when old, apex narrowly acute, ending in a single tentacle; petioles 30–45(–50) mm long, 1.5–2.4 mm wide, abaxial and adaxial surfaces covered with TSG trichomes, sessile glands and eglandular hairs, hairs sparser on the adaxial surface (relative to abaxial surface), very narrowly transversely elliptic (flat) in cross section; lamina 80–195(–245) mm long, 1.2–3(–3.5) mm wide, adaxial surface covered with numerous translucent red, carnivorous, capitate tentacles, sessile glands and TSG trichomes, abaxial surface covered with eglandular hairs, sessile glands, and TSG trichomes; stipules 7–14 mm long, 6.5–10 mm wide at the base, triangular, membranaceous, bronze-gold in color, apex acute and fimbriate. Scapes 1–2 per plant, (180–) 240–380 mm long (including inflorescence), 1.5–3 mm in diameter at the base, base erect; inflorescence a scorpioid cyme, often bifurcate, bearing 9–23(–35) flowers, indumentum of the scape, pedicels, abaxial surfaces of bracts and sepals consisting of eglandular hairs, sessile glands, and TSG trichomes; bracts 2–5.5 mm long, lanceolate, usually absent; pedicels 2–7.5 mm long, inserted 3–15 mm apart from each other; sepals 5, 6.5–9.5 mm long, 1.8–2.3(–3.5) mm wide, oblong-lanceolate to oblongobovate, united at basal 1/4–1/5 of length; petals 5, 7– 10 mm long, 4–7 mm wide, obovate, light pink-lilac in color; stamens 5, 3.5–4.5 mm long, anthers 1–1.3 mm long, bithecate, yellow; ovary 1.3–1.6 mm in diameter at anthesis, 3-carpellate, globose to globose-ellipsoid, slightly 3-lobed in outline; styles 3, forked at the base, 3.5–4.5 mm long (including stigmata), stigmata flabellate or bilobed, pink-lilac in color. Fruit a dry capsule, 3–3.5 mm long, ellipsoid, 3-valvate. Seeds oblong-fusiform, 0.65–0.8 mm long, c. 0.2 mm wide, testa reticulate, black.

Distribution and ecology: — Brazil, Minas Gerais, southern Cadeia do Espinhaço highlands, endemic to the Serra do Caraça, between the municipalities of Catas Altas and Santa Bárbara ( Fig. 1 ). Drosera graminifolia is also doubtfully reported from the Serra do Gandarela (Serra do Retiro) by a single specimen from the 19 th century (Damazio s.n., s.d. (OUPR 4308)), mislabeled Serra do Betim by Silva (1994) , Silva & Giulietti (1997), and Correa & Silva (2005) .

Drosera graminifolia grows in vegetation intermediate between campo rupestre and campo de altitude, in very narrow habitats mostly defined by altitude, in high-montane regions at elevations of (1700–) 1800–1950 m. It forms small and scattered populations at the Serra do Caraça near the summits of the four highest peaks (Canjerana, Carapuça, Inficcionado, and Sol), growing in a peat-sand soil mixture, or on a thin layer of this same soil over sandstone, or more rarely in cracks of bare sandstone, commonly among mosses such as Sphagnum L. spp. ( Sphagnaceae ) and often partially shaded by short shrubs ( Fig. 3A ).

Drosera graminifolia has been observed to grow vigorously the whole year round. Even during the dry season its habitat does not dry out completely, due to the regular water condensation that occurs at night on the mountain summits, keeping the soil constantly moist just under the surface. This species grows sympatrically with Genlisea violacea Saint-Hilaire (1833: 431) , Utricularia laciniata Saint-Hilaire & Girard (1838: 870) , U. reniformis Saint-Hilaire (1830: 224) , and U. subulata Linnaeus (1753: 18) , of the Lentibulariaceae family, all of which are also carnivorous.

Phenology: — Drosera graminifolia flowers early in the wet season, from January to April. A secondary and minor flowering period between August and September was reported by Rivadavia (1996).

Conservation Status: —The Serra do Caraça is located in the Quadrilátero Ferrífero (‘Iron Quadrangle’), a region that bears a unique flora but has long suffered from intense mining of its rich mineral deposits. Unfortunately, mining has resulted in the recent degradation of several localities of native vegetation in this region, including preserved areas such as the Serra do Caraça (SC). As a direct consequence of intense mining activities at the eastern and southern base of the SC, just outside the park, a significant decrease in population size of D. graminifolia was observed during the 1990’s, reduced to a single adult plant on Mt. Carapuça in 2000 (F.Rivadavia, pers. obs.). During this period, the fragile and unique flora native to the SC highlands deteriorated, possibly due to particulate matter pollution from the adjacent mining activities, leading to increased accumulation of dead vegetation—which in turn fueled a large fire in September of 1997. This fire was reported as the main cause of local extinction of the Lycopodiaceae Huperzia rubra ( Chamisso & Schlechtendal 1833: 389) Rothmaler (1944: 60) , as well as many other species at the SC massif, and was the facilitator for the invasion of the highland areas by exotic grasses that compete with the local species for light and space ( Vasconcelos et al. 2002), and also increase the flammability of the campo rupestre vegetation ( Alves & da Silva 2011 ).

Fortunately, the highland vegetation of the SC has made a surprising recovery over the past decade— including D. graminifolia , which seems to have recolonized most of the areas where it was observed to grow in the early 1990’s, before its near extinction at the turn of the millennium. However, it is hard to estimate if and how many other species suffered extinction at the SC as a result of this tragic incident, and how many other species are indirectly doomed to extinction due to a population crash (and inescapable genetic bottle neck) similar to that suffered by D. graminifolia .

According to the criteria of the IUCN Red List (2001), D. graminifolia is considered as Critically Endangered (CR) due to the small and isolated populations of this species, its restricted distribution area, the drastic reduction of population size over the past 20 years, and the possible continuous decline in quality of the habitat caused by further side effects of mining, including the increase of invasive species of grasses, especially Melinis minutiflora Palisot de Beauvois (1812: 54) , regionally known as capim gordura.

Notes: —The type of D. graminifolia var. major was examined and clearly represents only a robust individual of D. graminifolia . The characteristics cited by Eichler (1872) —longer leaves (over 20 cm in length), plus denser and longer indumentum on leaves and scapes—are within the range of variation reported here for D. graminifolia .

As the type specimen of D. graminifolia var. major is only numbered as "1300" on a small separate handwritten sheet, but no lowercase letter in front of the number, this collection must have been made on Sellow’s 6 th (and last) Brazilian journey, when he travelled from eastern São Paulo state to Minas Gerais state between 1829 and 1830 ( Urban 1893). Sellow made c. 1500 collections on this trip, thus specimen 1300 must have been collected towards the end. According to Urban (1893), Sellow was not able to continue to northern Minas Gerais and further west to Goiás state as was the plan for the end of that expedition. Therefore, he went on to Ouro Preto in November 1830, from where he made various smaller expeditions to Itacolumi (currently Itacolomi) and, in the second half of December 1830, to the Serra do Caraça (where we suspect the last collections were made, with numbers up to c. 1500). This D. graminifolia specimen must have been one of Sellow’s very last collections, as he returned to São Paulo state in early 1831, where he made his testament in March 1831, and died only shortly after (at the age of 42) when swimming in the Rio Doce ( Urban 1893).

The name D. brasiliensis was first cited by Martius in his hand-written field notes and later considered by Eichler (1872) as a synonym of D. graminifolia . Since the name was not effectively published by Martius, but cited and considered only as a synonym by Eichler (1872), the name must be considered invalid according to Art. 34.1. of the ICBN ( McNeill et al. 2006), which states that “a name is not validly published […] (c) when it is merely cited as a synonym; […]”.

The specimen Martius 1287 (voucher for D. brasiliensis ) unfortunately could not be found at neither M nor BR (where specimens from Martius are held). It was allegedly lent to Diels at B for his treatment of Droseraceae for the Pflanzenreich, and was probably destroyed during World War II. However, taking into consideration the location data cited by Eichler—“Inter Cidade Diamantina et Bandeirinha” (Bandeirinha is a district of the city of Diamantina)—this specimen would best fit as conspecific with D. spiralis , because it lies within the geographic range of this species.

Line drawings of D. graminifolia in Eichler (1872), Silva (1994) , Silva & Giulietti (1997), and Correa & Silva (2005) illustrate D. spiralis instead. The latter three publications also refer to the adaxial surface of the sepals as being ciliate, but this character was not observed in any of the specimens examined of either D. spiralis or D. graminifolia .

Specimens Examined: — BRAZIL. Minas Gerais: Município de Catas Altas, Serra do Caraça, Pico do Inficcionado , 16 February 2000, Vasconcelos s.n. ( BHCB 52556 , SPF 148831 ) ; RPPN Santuário do Caraça, Pico da Carapuça , 17 February 2009, Oliveira & Arruda 322 ( BHCB) . Município de Santa Bárbara, Serra do Caraça , 25 January 1971, Irwin et al. 29062 ( NY, UB) ; Parque Natural do Caraça, Pico da Carapuça , 21 July 1972, Emygdio et al. 3552 (R), 02 March 1976, Windisch & Ghillány 508 ( HB) , 05 March 1992, Rivadavia 122 ( SPF) , 09 April 2011, Gonella & Siniscalchi 412 ( SPF) ; Trilha para Campos de Fora, Serra do Canjerana , 08 March 1982, Hensold et al. s.n. (K, R, SP, SPF 22406) ; Parque Natural do Caraça, subida do Pico Canjerana (ou do Retransmissor ), 01 July 1995, Rivadavia 430 ( SPF) . Without attribution to administrative area: Serra do Caraça , March 1892, Ule 2447 (R), s.d., Baeta s.n. ( OUPR 4309 ) ; Morro da Carapuça , 11 June 1884, Glaziou 14485 (R). No location data, no date: Baeta s.n. ( OUPR 4306 ) ; Baeta s.n. ( OUPR 4307 ) . Dubious location: Alto da Serra do Retiro , caminho S. Gandarela, s.d., Damazio s.n. ( OUPR 4308 ) .

Figures

FIGURE 1. Distribution of D. graminifolia (square) and D. spiralis (circles).

FIGURE 2. Drosera graminifolia (based on Gonella & Siniscalchi 412). A, habit. B, stipule. C, petiole, adaxial surface. D, leaf apex, abaxial surface. E, side view of the calyx. F, detail of the scape indumentum. G, seed. Drawn by P. M. Gonella.

FIGURE 3. Drosera graminifolia at Serra do Caraça, central Minas Gerais state. A, high-montane habitat with open shrubby vegetation. B, habit. C, base of a rosette, showing the short stipules and long petioles. D, young leaf showing the regular circinate vernation. E, flower (A–D by P.M. Gonella; E by Leonardo Desordi, used with permission).