Four new species of Andean Pilea (Urticaceae), with additional notes on the genus in Venezuela

Abstract Four new species of Pilea (Urticaceae) from the Andes of Venezuela are described and illustrated: Pilea matthewii sp. nov., Pilea miguelii sp. nov., Pilea nicholasii sp. nov., and Pilea nidiae sp. nov. The affinities of these species and their positions within the informal classifications of Pilea proposed by Weddell and Killip are discussed. Notes on other species of Pilea found in Venezuela also are presented.


Introduction
Pilea Lindl. (Urticaceae), a large genus of 700 or more species, is found worldwide in tropical, subtropical, and temperate areas although it is absent from Australia, New Zealand, and Europe (Monro et al. 2012). Southeast Asia is believed to be the center of morphological and phylogenetic diversity for the genus, while the center of species diversity is in the Caribbean and Andes (Monro 2006). Field and herbarium work focused on producing a fl ora of Guaramacal National Park (Portuguesa and Trujillo states), which protects part of the Venezuelan Andes, convinced us that the following four species of Pilea from the Andes of Venezuela should be described as new.
Th ere has been no critical examination of the genus Pilea in the northern Andes (including the Coastal Cordillera of Venezuela) since Killip ( , 1939) published his regional revision and it is not surprising that undescribed species are found. Th e most recent enumeration of the genus for Venezuela (Romaniuc Neto 2008) recognized 28 species of which 12 are reported from Andean states. In addition to the four species described here, we believe Romaniuc Neto (2008) overlooked two species reported from Venezuela, recognized another two species that do not occur in the country, and listed two species that might not occur in Venezuela. Th us, by our count there are at least 32 species of Pilea in Venezuela, the majority occurring in the Andes and the Coastal Cordillera. Our modifi cations to the enumeration of the species of Pilea known from Venezuela (Romaniuc Neto 2008) are summarized in the fi nal section of this paper.
We are aware that the classifi cation of Pilea proposed by Weddell (1869) and modifi ed by Killip ( , 1939 is artifi cial, but Killip's informal classifi cation especially is the only current practical way to group Andean species. A world-wide monograph of the genus is unlikely to be prepared anytime soon although Monro (2006) has proposed a phylogenetic framework for revising the genus based on cpDNA, nrDNA, and morphology. He did not fi nd support for Weddell's (1869) classifi cation but did fi nd a strong geographical signal in his molecular phylogeny. Th is led Monro (2006) to conclude that a combination of morphologically and geographically circumscribed groups may provide a pragmatic way to identify monophyletic units for an eventual global revision of Pilea.

Methods
Th e new species are based principally on our collections and those of our collaborators, which were made as part of the Flora of Guaramacal project (PORT-US). We also examined collections from throughout Venezuela and adjacent Colombia that are deposited in MO, NY, PORT, US, and VEN (herbarium abbreviations follow Index Herbariorum, http://sweetgum.nybg.org/ih). Th e US collections were particularly useful because Killip was based at the U.S. National Herbarium (US) when he published his revisions of the Andean species of the genus (Killip , 1939. Sheet numbers are cited for the holotypes deposited in PORT. Barcodes are cited for isotypes deposited in US. Identifi cation numbers (sheet numbers and/or barcodes) are not available for the remaining type material collected by us and our collaborators, which will only be distributed upon publication of this paper.
A morphological species concept was adopted and descriptions were modeled on those of Monro (2001Monro ( , 2006 and Monro et al. (2012) in order to facilitate comparisons. Material was examined and measured using an Olympus SZH binocular microscope.
Conservation assessments were undertaken using IUCN (2001) criteria. However, the only available data for our new species concern the geographic range of these species: IUCN criteria B1 (extent of occurrence) or B2 (area of occupancy). We have no data with respect to population size or dynamics (viz., whether or not populations are declining or expanding).

Diagnosis
Distribution and ecology. Known only from the Andes of Venezuela (Portuguesa and Trujillo states) where it is found in the understory of cloud forest; 1000-2600 m.
Etymology. Th e epithet recognizes Matthew Dorr who participated in a number of expeditions to Guaramacal in search of specimens for the Flora of Guaramacal project (PORT-US).
Conservation status. Using IUCN criteria (IUCN 2001) we could not identify a threat to Pilea matthewii. We are aware of 15-20 distinct populations in Guaramacal National Park, which protects an area of 225 km 2 . Although this area is relatively small, the species is frequently encountered and the number of known populations exceeds the number of locations deemed critical under IUCN criterion B2(a) for either Endangered (E) or Vulnerable (VU). In addition, the east-facing slopes of the Sierra Nevada de Mérida, which have similar habitat, are very poorly collected (Dorr et al. 2005) and might harbor additional populations of this species.
Distribution and ecology. Known only from the Andes of Venezuela (Lara, Mérida, and Trujillo states) where it forms colonies in the understory of cloud forest; 1490-2600 (-3210) m.
Etymology. Th is species is named for S. Miguel Niño, professor at UNELLEZ, Guanare, and valued collaborator in our investigations of the Andean fl ora. Discussion. Pilea miguelii belongs in the Heterophyllae species group of Weddell (1869) and the Centradenioideae species group of . Th e new species is easily recognized by the combination of the extreme diff erence in leaf laminae size at each node and the branching infl orescences.
One of the specimens (Steyermark 56365, US) that fi ts our concept of Pilea miguelii was identifi ed by Killip (in sched.) as P. losensis Killip, but the similarity is superfi cial.  included P. losensis in his Multifl orae species group and unlike our new species the leaf laminae at each node are ± equal in size (versus distinctly unequal), the laminae are narrowly elliptic to oblong-elliptic (versus narrowly ovate to ovate or obovate), and the apices are acuminate (versus long acuminate). In addition, the type of P. losensis, at least, is sparingly branched while most collections of P. miguelii are profusely branched.
Another specimen (Steyermark 55767, US) that fi ts our concept of Pilea miguelii was identifi ed by Killip (in sched.) as P. carnosula Wedd., also in his Multifl orae species group . Th is "new record" for Venezuela was reported by Steyermark (1957 as "carnulosa") and subsequently repeated by Romaniuc Neto (2008 as "carnulosa"). Th e two species are only superfi cially similar. Th e leaf laminae at each node are ± similar in size in P. carnosula and the major laminae of P. carnosula are smaller than those of P. miguelii (0.8-4 versus 5-11.5 cm long).

Pilea nicholasii
Distribution and ecology. Known only from the Andes of Venezuela (Lara, Portuguesa, and Trujillo states) where it is found in the understory of montane and cloud forest; 1900-2800 m.
Etymology. Th is species is named for Nicholas Dorr who assisted with fi eld work in the Venezuelan Andes, but clearly prefers the rigors of Chichiriviche to those of the mountains.  Discussion. Th e majority of collections of Pilea nicholasii have either staminate or pistillate infl orescences on a stem. Several collections, including the type (Stergios & Dorr 20208), however, have both staminate and pistillate infl orescences on the same stem, and at least one collection (Cuello et al. 1416) has both staminate and pistillate infl orescences arising from the same leaf axil. Th is suggests to us that the species is monoecious rather than dioecious.
Sometimes the pedicels on staminate infl orescences are sterile. Th e cause of this is not clear: it may be that some male fl owers are caducous or, as suggested by one of the reviewers of this manuscript, the consequence of fungal infection. A number of the pistillate infl orescences, especially on specimens with conspicuous staminate infl orescences, are very cryptic with very short peduncles. Other pistillate infl orescences have pronounced peduncles. In any case, there appears to be a bias toward collecting specimens with either staminate infl orescences or infructescences probably because these plants are more visible and manifestly fertile.
Pilea nicholasii belongs in the Heterophyllae species group of Weddell (1869). Its placement in one of the species groups proposed by  is somewhat problematic as depending upon which pair of leaves at a single node are measured P. nicholasii falls into either Killip's Centradenioideae species group with major leaf laminae more than twice as long as minor leaf laminae or his Capitellatae species group with the major leaf laminae less than twice as long as the minor ones. Among species placed in the former group, P. nicholasii is similar to P. hydrocotylifl ora Killip, which was described from Colombia (Norte de Santander). However, the undersurface of the laminae is pruinose in the former and glabrous in the latter species. Th is makes the leaves of P. nicholasii look lighter below than above while those of P. hydrocotylifl ora are uniformly green. In addition, the major laminae of the former are markedly asymmetrical whereas in the latter they appear to be ± symmetrical.
Pilea nicholasii does not appear to have any close allies in the Capitellatae species group of . It keys to a group of three species that are monoecious, but none of these three species has the pruinose undersurface of the leaf laminae found in our new species.
Characters for distinguishing Pilea nicholasii from P. hydrocotylifl ora and P. pichisana are given in Table 3.
Conservation status. Using IUCN criteria (IUCN 2001) we could not identify a threat to Pilea nicholasii. We are aware of 15-20 distinct populations, the majority of which are in Guaramacal National Park. Th e extent of occurrence (EOO) is less than 5000 km 2 and the area of occupancy (AOO) is less than 500 km 2 , which might suggest that the species is Endangered (E) under IUCN criteria B1 or B2, but there are > 5 populations and as with P. matthewii we would expect the species to be found in similar habitat along the east-facing slopes of the Sierra Nevada de Mérida.
Diagnosis. Similar to Pilea fl exuosa Wedd. from which it diff ers by its asymmetrically elliptic to narrowly-elliptic or obovate (versus broadly ovate) laminae that are asymmetrically cuneate (versus rounded or cordate) at the base.  de Steyermark 55533) or almost black, glabrous, cystoliths punctiform or short fusiform, often clustered at nodes, internodes 0.8-3.5 × 1-4 mm (shorter distally), terete in cross-section, angulate when dry, nodes constricted (at least when dry). Stipules 6-11 mm long, narrowly triangular, drying pale brown or tan, persistent. Leaves petiolate, distichous; petioles at the same node unequal by a ratio of 1:11.5-13.5 (-23), canaliculate above, glabrous; major petioles 2.3-2.7 cm long; minor petioles ca 1-2 mm long; laminae of leaves at each node unequal by a ratio of 1:1.7-2.2, major laminae in a pair 6.5-9.5 × 1.5-3.2 cm, asymmetrically elliptic to narrowly-elliptic or obovate, membranous, 3-nerved with lateral nerves diverging from midrib 1-6 mm above the base, forming pocket domatia where the 3 nerves join, midrib and lateral nerves prominent below, slightly impressed (or not) above, lateral nerves visible almost the entire lamina length but disappearing below the apex, secondary nerves 8-14 pair, borne 70-90° to the midrib and then curved distally, upper surface dark green, drying dark brown, glabrous except for scattered, minute peltate scales, cystoliths fusiform or absent, lower surface pale green drying dark brown, glabrous except for scattered, minute peltate scales, base cuneate, asymmetrical, margin coarsely toothed entire length, apex long acuminate; minor laminae in a pair 3-5 × 0.8-1.5 cm, otherwise as major laminae. Infl orescences 1-10 per stem, unisexual, green suff used with maroon; bracts ca 2 mm long; bracteoles ca 1 mm long. Discussion. Pilea nidiae belongs in the Heterophyllae group of Weddell (1869). Its 3-nerved, toothed leaves that are unequal in size at each node and conspicuous, persistent stipules place it in the Flexuosae group of Killip ( , 1939. We have not encountered any other Andean species from Venezuela with stipules that are as large as those of P. nidiae. Th e leaf laminae of one of the paratype collections (Steyermark 55533) are narrower than in the type of Pilea nidiae and densely covered in cystoliths above and below (the type mostly lacks cystoliths). All other characters (leaf shape, venation, toothing, etc.) agree with our concept of this new species.
Conservation status. Using IUCN criteria (IUCN 2001) we tentatively consider Pilea nidiae to be Endangered (E). Th e known range of the species is less than 5000 km 2 (IUCN criterion B1) and there are only four known populations and of these only two are in a protected area (IUCN criterion B1(a)). We know nothing, however, about the dynamics of these populations and whether or not they are declining. Earlier Dorr et al. (2000) reported this collection as Pilea pubescens Liebm., with which it is allied, but from which it can be separated by the coarse serrations on the leaf margin. Also, according to Killip (1939) P. pubescens is a species found at low elevations in South America. Pilea arguta (Kunth) Wedd. -Th e authorship as given by Romaniuc Neto (2008) is corrected. Also, Killip (1939) expressed doubt as to whether or not this species occurs in Venezuela. Th e type locality is "prope Nova Valencia Caracasarum," but the species is otherwise known only from high elevations in Colombia and Ecuador and it has not been recollected near Valencia, Carabobo nor anywhere else along the Coastal Cordillera of Venezuela. Pilea carnosula Wedd. -Th is species should be deleted from the fl ora of Venezuela because the published record (Steyermark 1957;Romaniuc Neto 2008) of its occurrence is based on a misidentifi cation of a specimen (Steyermark 55767,US) here considered to be Pilea miguelii (see above). Pilea centradenioides Seem. -Although Romaniuc Neto (2008) stated that this species occurred in the Distrito Federal, it was not reported from Venezuela by Killip ( , 1939. Th e Coastal Cordillera record confl icts with what otherwise appears to be a range corresponding to the Chocó in Central and South America. Consequently, we Table 4. Diagnostic characters that distinguish Pilea nidiae and P. fl exuosa.

Characters
Pilea nidiae Pilea fl exuosa Stipule shape (length) narrowly triangular (6-11 mm) ovate orbicular (4-6 mm) Major leaf lamina size 6.5-9.5 × 1.5-3.2 cm 2-6 ×1.5-4 cm Leaf shape elliptic to narrowly elliptic or obovate broadly ovate Leaf base cuneate rounded or cordate Leaf apex long acuminate abruptly acute to acuminate placed P. rhombea, described from Mexico, in synonymy under P. parietaria, described from the West Indies. Monro (2001), however, did not mention a South American element. Killip (1939) thought that P. rhombea and P. alsinifolia Wedd. were both confused with the West Indian P. parietaria. He did acknowledge, however, that all three species were part of the same complex. Pilea rhombea (L.f.) Liebm. -Th is species should be deleted from the fl ora of Venezuela because it is a synonym of Pilea parietaria (see above).