Dioscorea (Dioscoreaceae) is described from wet to very wet areas on both Pacific and Caribbean slopes of Costa Rica, from the Monteverde area to Tapantf National Park. The species' confusion with the related D. lepida C. V. Morton is discussed, and a key is provided to distinguish it from that species and the other two closely related Costa Rican species, D. racemosa (Klotzsch) Uline and D. standlyei C. V. Morton. During the course of work on a treatment of Dioscorea for the Manual de Plantas de Costa Rica, material of three possible new species was set aside for further study. Careful examination and review demonstrated that one of them would not be separated from the recently described Panamanian D. davidsei 0. Tllez, and that the other corresponds to D. lepida. The third, here described as D. natalia, had previously been identified as D. lepida and with that species, is closely related to two other Costa Rican species, D. racemosa and D. standleyi. Dioscorea Hammel, sp. nov. TYPE: Costa Rica. Cartago: Parque Nacional Tapantif, Sendero Oropendula, 1400 m, 26 Sep. 1997 (stam. & pist. fl, fr), B. Hammel & S. Troyo 21068 (holotype, INB; isotypes, CR, F, MEXU, MO). Figure 1. Species subtilis D. racemosae affinis, a qua imprimis differt inflorescentiis brevioribus, paucior-floribus rachidibus exilioribus. Stems delicate, twisting to the right but also very rarely to the left. Leaves alternate, ovate to narrowly triangular-cordate, 4.5-6(-9.5) X 2.5-3.5(-5) cm, 7(9)-nerved, glabrous, very faintly glandular-dotted on the lower surface. Staminate inflorescences 48(-10) cm, simple racemes to racemose, usually 1 to rarely as many as 5 per axil; the rachis + verruculose, very delicate, up to only twice the diameter of the pedicels; the cymes of 1 (rarely 2) flowers, subtended by a pair of narrowly triangular bracts 1-1.5 mm; pedicels 1.5-2(-4) mm. Staminate flowers with tepals appearing nearly valvate in bud, broadly triangular, 0.8-1.1 mm, patent, pale yellowish green or cream; stamens 3, connivent in the center of a + hexagonal torus; filaments ca. 0.1-0.15 mm, united only at the very base; anther cells ? ovoid, ca. 0.15-0.2 mm, touching those of adjacent anthers, dehiscing upward; staminodia none. Pistillate inflorescences 5-8 cm, spikes. Pistillate flowers with tepals as in the staminate ones but to 1.25 mm; styles 3, united in a conical stylar column 0-0.25 mm, the free portion, including stigmas, ca. 0.4 mm; stigmas 3, simple; staminodia lacking. Fruits perpendicular to the rachis, elliptic, 1.2-2.0 X 0.5-0.9 cm, apically acute; seeds 1-1.4 X 0.3-0.4 cm, winged on the posterior end. Paratypes. COSTA RICA. Alajuela: Reserva Forestal San Ram6n, 1100 m, Apr. 1991 (fl), Bittner 972 (CR); between San Miguel and La Palma de San Ram6n, Nov. 1926, Brenes (278) 5124 (CR); Monteverde Reserve, Rfo Pefias Blancas, 900 m, Aug. 1988 (fl), Bello 248 (CR). Guanacaste: Monteverde, 1 km al N de Las Nubes, 1300 m, Sep. 1988 (stam., pist. fl, fr), Haber & Zuchowski 8675 (INB, MO); road to Las Nubes, 1400 m, Nov. 1988 (stam., pist. fl), Haber 8749 (INB). Puntarenas: Monteverde Reserve, 1 km SW of Station, 1500 m, Feb. 1992 (fl), Ingram & Ferrell 1298 (CR); Monteverde vicinity, Upper Rio Negro valley, 1500 m, Sep. 1985 (fl, fr), Haber ex Bello 3020 (CR); Monteverde vicinity, San Gerardo Biological Station, 1200 m, Feb. 1995 (pist. fl, fr), Penneys 167 (CR, F, INB, MO, US). San Jose: Rio Hondura, 1100 m, Feb. 1974 (fl), Lent 3791 (CR). Cartago: Cafi6n del Rio Grande de Orosf, Oct. 1983 (fl, fr), Chacdo'n et al. 1483 (CR--6 mounted duplicates); Parque (Refugio) Nacional Tapantf, 1250 m, Aug. 1991 (stam. fl), Hammel et al. 18324 (INB, MO); 1400 m, Dec. 1992 (fl), Herrera 5813 (CR), June 1994 (stam. fl), Morales & Carnevali 2873 (INB); Valle del Reventaz6n, Moravia de Chirrip6, 1100 m, Sep. 1993 (stam. fl), Campos & Campos 186 (INB). Distribution. This species, apparently endemic to Costa Rica, is known from some of the very wettest parts of the country, especially P. N. Tapantif. However, it is also known from the Las Nubes area of Monteverde, which is relatively dry. So far it is known only from 900-1500 m, from Monteverde and Tapantf, and several intervening localities. Etymology. It was my pleasure to defer to Silvia Troyo, illustrator of this new species, the honor of choosing its name. When she came across the species for the first time in the field, she was struck by the same joy of discovery as any taxonomist upon finding such a new species, so tiny and precious. For all good reason, she chose the name Natalia, one often given here to first-born daughters, and that of her own. The epithet natalia is used NovoN 10: 378-381. 2000. This content downloaded from 207.46.13.151 on Fri, 25 Mar 2016 09:23:00 UTC All use subject to http://about.jstor.org/terms Volume 10, Number 4 2000 Hammel Dioscorea from Costa Rica 379 ~~4*. . ..~?~ / : .. . .. ..** '' '' '----7'' ::';?I