Abstract: The confusion that existed with regard to the correct identity of Puya humilis Mez and P. tunarensis Mez is clarified. A region near the Cotacajes river, in the Department of Cochabamba, Bolivia, between 1,300 and 1,400 m, is identified as the type locality of P. tunarensis (and as well as of P. kuntzeana). Ecology and geographical records of both species are presented. The relevant characters are explained supporting that both taxa really are distinct species. A distribution map, ecogram and photographs of both species are provided. P. butcheriana H. Luther is proposed as a new synonym of P. humilis Mez.
Introduction
Since the discovery of Puya tunarensis and P. humilis by Otto Kuntze, in 1892, confusion has existed with regard to the correct identity of both species. One has been called the other and vice versa. This confusion involved many botanists and collectors who tried to promote the knowledge on Bolivian bromeliads: e.g., Smith & Downs (1974), Luther (1995), Kromer et al. (1999) and lbisch & Vasquez (2000).
In their key, Smith & Downs (1974) clearly separated both species as being distinctive with regard to the floral bract that is erect in P. humilis and recurved in P. tunarensis. However, this character is not consistent in P. humilis which may show recurved scape bracts and basal floral bracts like P. tunarensis. Thus, Smith and Downs (1974) determined the specimens Cardenas 2372, 3580, Cardenas et al. 7639 and Foster 2555 as P. tunarensis - although they belonged to P. humilis.
On the one hand, among others, they did not take into account habitat characteristics like the altitude of the localities of the types of both species, and of course they were not familiar with the geography of the Cochabamba region. On the other hand, the original description of the type locality of P. tunarensis is confusing - leading to a somewhat misleading epithet.
Both holoypes are deposited at the New York Botanical Garden and were revised by the first author in November of 2001. The herbarium label of the holotype of P. humilis says: "Bolivia 3,000 m plain near Sacaba, 13-21 April 1892." This site can be identified easily. Sacaba, some 12 km to the east of the city of Cochabamba, is surrounded by mountains that are higher than 3,000 m. There the species is found with abundant populations. In the case of P. tunarensis the voucher label indicates as locality: "Bolivia Tunari mountains 1,300 m May 92 ( 1892)." It is this locality that created confusion because the city of Cochabamba is located at an altitude of about 2,500 m, and the Tunari Mountains, northwest to northeast of the city, reach an altitude of more than 5,000 m (Tunari summit). Thus, the question was: Where in the Tunari Mountains could a site exist of about 1,300 m as indicated by Kuntze? Or was he just wrong about the altitude?
The answer was found during a trip of the first author, in June 2001, participating in a botanical expedition to the Cotacajes valley that is the border between the Departments of Cochabamba and La Paz. The Cotacajes valley is one of the rather isolated dry inter-Andean valleys that are inserted in the humid forests of the Yungas ecoregion. Recently, an endemic bromeliad was discovered in this valley: Fosterella cotacajensis Kessler, Ibisch & Gross (Kessler et al. 1999).
Going from Cochabamba, this valley is reached crossing the Tunari Mountains northwestwards. And there, in the vicinity of the village of Cotacajes, on steep rocky escarpments, we found two species of Puya: the first rather large, to 1.5-2 m high, with large green leaves and reddish tips, erect scape, covered by large bracts, and a compact and multiflowered inflorescence. This species turned out to be P. kuntzeana, also named by Mez in honor of its first collector Otto Kuntze! Obviously, Kuntze had traveled to the Cotacajes valley. The second species was smaller and presented clear green leaves, cylindrical inflorescences with recurved apices of the floral bracts. These characters fit well to P. tunarensis. The fact that these two species occurred together - at an altitude of more or less 1,300 m (!) - confirmed to us that we had arrived at the type locality of both P. tunarensis and P. kuntzeana, not in the Tunari Mountains as vaguely indicated by Kuntze, but on their other side, northwestwards of the Tunari summit. This interpretation is confirmed by the label of the P. kuntzeana type: "Bolivien 1300 m, Tunari 4/5/1892". The three species P. kuntzeana, P. tunarensis and P. humilis were discovered by Otto Kuntze in 1892. Mez identified them as new species in 1894 and described them in 1896.
Distribution of Puya humilis and Puya tunarensis
Puya humilis is endemic to Bolivia but has a rather large range, in altitudes between 2,500 and 3,900 m, more or less associated with the upper headwaters of the Rio Grande river system ( departments of Cochabamba, Chuquisaca and Potosi). P. tunarensis is restricted to the isolated and narrow Cotacajes valley, between the Cochabamba and La Paz departments, at an altitude of 1,200 to 1,400 m . The species are separated by the at least 4,500 m high Cochabamba/Tunari Mountain chain.
Recorded specimens
Puya humilis Mez, DC. Monogr. Phan. 9: 498. 1896.
Puya werdermannii Harms. Notizbl. Bot. Gart. Berlin 10: 793. 1929. Tipo: Bolivia: NE sierra de Cochabamba, Incachaca, Chuquechampi to Cuchicancha, Jul. 1926, Werdermann 2006 (B fotog. 11423, F).
Puya butcheriana H. Luther, Selbyana 16(2): 235-238. 1985. Syn. nov. Type: unknown origin, possibly Bolivia or Peru, 24 January 1992, D. Butcher s.n. (Holotype: SEL).
COCHABAMBA: Prov. Arani: Toralapa, January 1946, Ccirdenas 3580 (US); ibid.: between Vacas and Rodeo, approximately 17°42'S, 65°45'W, 3,400m, on rocky crevices of dry, alpine scrub vegetation, 20 April 1987, G.S. & U Varadarajan & J Brandbyge 1451 (LPB); ibid.: 1.5 km, road from Alalay to Mizque, 17º43'S, 65º39'W; 3,750 m, 23 July 1998, P. & C. Ibisch 98.0103 (LPB); ibid.: 1.5 km, road from Arani to Mizque, 15 km after turn-off to Vacas, 17°42'S, 65°39'W, 3,900 m, 7 August 1993, P. & C. Ibisch 93.0563 (LPB); ibid.: Chulku Mayu Community, 3,700 m, Polylepis forest, montane forest, 2 March 1991, I. Hensen 1223 (LPB); ibid.: Koari, km 90 old road between Cochabamba and Santa Cruz, 3,400 m, 28 December 1999, R. Vasquez 3500 (VASQ); Prov Carrasco: Quirusillani to Totora, April 1943, Cardenas 2372 (GH); ibid.: Siberia, between Comarapa and Pojo, 11 January 1965, Vogel 468 (US), ibid.: nearby Montepuncu, on the road to Sehuencas, 17º33'S, 65º 17'W; 2,800 m, 27 June 1993 (flowers in cultivation, 1 October 1993), P. & C. Ibisch 93.0486 (LPB). Prov Chapare: 3,500m, 17º24'S 65º52'W, 23 Oct. 1985, JC. Solomon 14509 (MO); ibid.: between Melga and Waganki 3,240 m, terrestrial on rocky slopes, 5 February 1999, R. Vasquez 3160 (VASQ). Prov. Esteban Arce: from Anzaldo to Tarata, 9 January 1943, Cardenas, Cutler & Gandarillas 7639 (US). Prov Mizque: Cocha (Cocha-Cocha) to Vila Vila, 15 November 1948, Foster 2555 (US). CHUQUISACA: Prov. Oropeza: Guerraloma, Sucre, February 1949, Cardenas 4128 (US); north of Presto, 3070 m, "ckayara-kayanta ", 29 August 1980, G. Muhlbauer s.n. (LPB); ibid.: track from La Recoleta to the "Cristo", Eucalyptus plantation, approximately 2,900 m, 25 July 1991, P. & C. Ibisch 25072 (LPB). POTOSI: Tagua Comuno, 15 km southwards from Potosi, 25 March 1936, West 6370 (UC); ibid.: Prov. Charcas: Toro Toro, El Vergel canyon, 18º08'S, 65º45'W, 2,500 m, 22 July 1998, P. & C. Ibisch 98.0076 (LPB).
Type: BOLIVIA: COCHABAMBA: Monte Tunari, 1,400 m, 4 May 1892, Kuntze sn. (Holotype: NY).
Other records from Bolivia:
COCHABAMBA: Prov. Ayopaya: between Pujiuni and Cotacajes, 1,370 m, terrestrial on rocky slopes, 20 June 2001, R. Vasquez, G. Navarro, M. Fernandez F:Miranda & H. Rocha 4125 (VASQ, LPB).
Morphological comparisons
The relevant characters of Puya humilis show clear morphological differences when compared to P. tunarensis, strongly supporting that they are different species ( see Table I ).
Table I. Comparison of relevant characters of Puya humilis and P. tunarensis (based on the revision of living and herbarium material).
Characters P. humilis Mez P. tunarensis Mez
Height Flowering to 60 cm. Flowering to 1.50 cm.
Leaves Gray, narrowly triangular, 20-25 Clear green, triangular, to
cm long X 10-15 mm wide, 60 cm long X 4-5 mm wide,
densely cinerously lepidote, in scarcely cinerously lepidote
longitudinal stripes abaxially, abaxially, glabrous adaxially,
cinerously lepidote adaxially, serrate with clear or dark
laxly serrate, with yellowish, brownish, to 5 mm long spines.
3 mm long spines.
Sheaths Ovate, 1.5 cm long. Suborbiculate, 3 cm long.
Scape 6-8 cm long, 10 mm in diameter, 6-12 cm long, 12-15 mm in
white tomentellous. diameter, slightly cinerously
lepidote.
Inflorescence Simple, strobilate, fusiform or Simple, cylindric, 20-40 cm long,
cylindric, 20-25 cm long, 2-3 cm about 3 cm in diameter, axis
in diameter, axis densely white- lepidote.
tomentellous.
Floral bracts Erect, the basal ones not recurved Erect, recurved even before
or slightly recurved after anthesis, anthesis, imbricate, strongly
imbricate but not strongly adpressed before anthesis, less
adpressed, elliptically lanceolate, strongly after anthesis, elliptic to 20-28 long X 10-12 mm wide, elliptically ovate, apiculate apex,
longer than the petals, adpressed, 20- 25 mm long X 12-15 mm
rose, nerved, acuminate apex, wide, powdered-lepidote, rose
pungent, white tomentellous, violet, dry after anthesis, margin
margin microscopically serrate entire.
or entire.
Pedicel 2-3 mm long. 2-3 mm long.
Sepals Subtriangular, acute, 12-13 mm Obovate-acuminate, 13-14 mm
long X 3-5 mm wide, carinate, long X 4-5 mm wide, coriaceous,
slightly lepidote, glabrous towards slightly pubescent.
the apex.
Petals 15-20 mm long X 10-12 mm wide, 18-20 mm long X 10-12 mm
dark violet (to almost black) with a wide, dark violet with a greenish
white base, apex mucronate. base, apex acute.
Stamens Not exserted. Not exserted.
A new synonym
In 1995, Harry Luther published the description of Puya butcheriana, based on plants cultivated by Derek Butcher in Australia. Butcher had bought seeds from Karel Knize collected in Bolivia or Peru (Luther 1995). Knize had been an active trader of cacti and bromeliads and made several trips to Bolivia. It is probable that the seeds sold to Butcher indeed came from Bolivia.
Luther differentiated P. butcheriana from P. humilis for being taller, a narrower inflorescence and for a diminutely serrate margin of the floral bracts ¬characters, however, that can be observed as well in P. humilis. Using the key from Smith & Down (1974), for the recurved scape bracts the plant seemed to be close to P. tunarensis. However, as shown above, this is not a trustworthy character. Revising the description of Luther ( 1995) and the photographs of D. Butcher and L. Colgan published in the web (www: bsi.orglbromlinfo gallerylpuya) we see that P. butcheriana clearly looks like P. humilis as we know it from Bolivia, and thus, we suggest that it should be regarded as a synonym of P. humilis. In the past, it had been supposed by others that P. butcheriana might be P. humilis; (e.g. see Len Colgan on Charlie's Web: "Puya butcheriana is named after Derek Butcher but some wonder if it is a form of Puya humilis") (http ://www.charlies-web.comlbromeliads-alphalistltex562 .html, consulted in July 2002). —SeeVasquez & Ibisch 2002p. 52(4): 158-167
