Two new combinations, lectotypifications and a new name for Costa Rican Palicourea s.l.

Abstract Species of the complex and diverse genera Psychotria and Palicourea are common but little-known elements in many tropical forest ecosystems. DNA-phylogenetic studies and a re-evaluation of morphological characters have recently shown that species of Psychotria subg. Heteropsychotria are nested within Palicourea s.l., which was traditionally separated by exhibiting a bird-pollinated (vs. insect-pollinated) pollination syndrome. In order to render both genera monophyletic groups, species of subg. Heteropsychotria need to be transferred to Palicourea s.l. For Central American species, most of the necessary combinations have already been made. In the course of ongoing research on the phytochemical characterization of species and clades of Costa Rican Palicourea s.l., the nomenclature of Mesoamerican species was revised. As a result, two new combinations and a new name are proposed here: Palicourea horquetensis (Dwyer & Hayden) A. C. Berger & C. M. Taylor is based on Rudgea horquetensis Dwyer & Hayden, Palicourea tonduzii (K. Krause) A. C. Berger is based on Cephaelis tonduzii K. Kraus and Palicourea longiinvolucrata A. C. Berger replaces Psychotria hispidula Standl. In addition, two lectotypes are designated.


Introduction
Species of the complex and diverse genera Psychotria L. (1759: 929) and Palicourea Aubl. (1775: 172-175) are prominent but little-known elements in tropical forest ecosystems. Both have long been considered closely related, and Palicourea was differentiated from Psychotria by characters associated with hummingbird rather than insect pollination. Species of Palicourea are typically found in the understory of rainforests and are especially frequent in high elevation habitats where Psychotria and other related genera are less speciose (Taylor 1996(Taylor , 1997.
Both genera were traditionally classified in the tribe Psychotrieae. Recently, however, DNA-phylogenetic studies and a re-evaluation of morphological characters have shown that species of Psychotria subg. Heteropsychotria Steyerm. (1972: 484) are more closely related to Palicourea. Consequently, views shifted towards a narrower concept of Psychotria and Psychotrieae which peaked in the ongoing segregation of hundreds of species and the establishment of the sister tribe Palicoureeae (Nepokroeff et al. 1999, Razafimandimbison et al. 2014, Robbrecht and Manen 2006.
Species of Psychotria subg. Heteropsychotria and Palicourea cannot be distinguished by vegetative or fruit characters. In addition, both groups show similar accumulation of alkaloids (e.g., Berger et al. 2012Berger et al. , 2015Berger et al. , 2017, flavonoids (e.g., Berger et al. 2016) and a group of defensive peptides termed cyclotides (Koehbach et al. 2013). As traditionally defined, both groups deviate only in a suite of traits associated with pollination syndromes: Flowers in subg. Heteropsychotria are arranged in open, somewhat grouped to densely capitate inflorescences with inconspicuously colored inflorescence axes, though some species have inflorescences subtended by showy bracts. Flowers are usually sessile or subsessile and have small, white, to greenish or yellow corollas with short and straight tubes in bee-pollinated or white and long-tubed corollas in moth-pollinated species. (e.g., Steyermark 1972, Taylor 1996.
By contrast, species of Palicourea are hummingbird-pollinated, frequently have longpedunculate and open inflorescences, colored inflorescence axes, large and long pedicellate flowers and vividly colored corollas with well-developed tubes. Corollas have a gibbous, nectar-accumulating swelling at their base that is protected by an internal ring of hairs. In many plant groups, bird-pollinated flowers have repeatedly evolved in groups of bee-pollinated ancestors and are not phylogenetically informative at the generic level (e.g., Castellanos et al. 2004, Fenster et al. 2004, Pirie et al. 2016. Similarly, it was hypothesized that pollinator shift has occurred multiple times in Palicourea s.l. and that bird pollinated species (i.e., the traditional concept of Palicourea) repeatedly evolved out of bee-pollinated ancestors (i.e., the traditional concept of Psychotria subg. Heteropsychotria) or vice versa (Taylor 1996(Taylor , 1997. In order to render both Palicourea and Psychotria monophyletic groups, most species of Psychotria subg. Heteropsychotria have to be transferred to Palicourea s.l. The combined group includes more than 800 species, is variable in flower characters, but is supported by vegetative and fruit characters as well as by DNA phylogenetic data (Razafimandimbison et al. 2014). The process of transferring species of subg. Heteropsychotria was started with the publications of Taylor et al. (2010), Taylor (2015aTaylor ( , 2015b and Taylor and Hollowell (2016), which provided combinations for species belonging to newly defined sections within Palicourea s.l. Finally, Mexican, Mesoamerican and Venezuelan species were transferred by Borhidi (2011Borhidi ( , 2017aBorhidi ( , 2017b, and species occurring in the Guianas were transferred by Delprete and Kirkbride (2016).
In the course of ongoing research on the phytochemical characterization of species and clades of Costa Rican Palicourea s.l., the nomenclature of Mesoamerican species was revised and the need for a new name and two new combinations became apparent. These are proposed here. In addition, a complete synonymy, an enumeration of type specimens and two lectotypifications are provided for these three species.

Methods
The present work is based on an extensive study of herbarium specimens, digital images and relevant literature including regional (e.g., Manual de Plantas de Costa Rica, Taylor 2014) and overregional floras (e.g., Flora Mesoamericana, Lorence and Taylor 2012). In addition, extensive fieldwork was performed in Costa Rica in 2010, 2013, 2015 and 2016. For all names, protologues were checked to verify or revise author and page citations, information on collectors and localities. Subsequently, the type category applying to each name was assessed in accordance with the ICN (Melbourne code, McNeill et al. 2012; see also McNeill 2014). Retrieved information was managed with the international JACQ herbarium database (http://herbarium.univie.ac.at/database) hosted at the herbarium WU. For citation of type collections, localities have been simplified and ecological and morphological details have been omitted. For all retrieved type specimens, herbarium acronyms and barcodes are given. For specimens seen either digitally or physically, their barcodes are followed by an exclamation mark.
For the three nomenclatural novelties, both possible author abbreviations of my name are preoccupied: Berger stands for Ernst Friedrich Berger (1814-1853) and A. Berger stands for Alwin Berger (1871Berger ( -1931. Therefore, I have adopted the abbreviation A. C. Berger that includes my second forename Christoph, which I have never used before in my publications.  Lorence (1999) and Correa et al. (2004). Lorence and Taylor (2012) were the first to exclude it from Rudgea, but did not suggest any further placement. Based on morphological characters, the species clearly belongs to the nocturnally flowering species group of Palicourea (Taylor et al. 2010) and is here treated as conspecific with Psychotria or Palicourea nebulosa. Rudgea horquetensis is the oldest available name for the taxon and a respective new combination is proposed here. Typification. The protologue of Coussarea nebulosa states that the holotype is located at MO. At the time of publication two sheets of the type collection have been accessioned at MO, making the holotype designation ambiguous. In 1975, Dwyer annotated both sheets as Coussarea nebulosa, but did not specify what sheet he intended to be the holotype. Hence, both specimens represent syntypes (ICN, Art. 40.2 & Note 1; see also McNeill 2014). Later, the specimens were annotated as sheet "1" and "2 of 2", respectively. Sheet 1 was also annotated as holotype by C. M. Taylor in 1988. Taylor (1995 cited the above-mentioned sheets as "holotype, MO 2162999; isotype, MO 4043108" which has to be considered a valid (though indirect) lectotypification according to the ICN (Art. 7.10, 9.9, 9.23; see also McNeill 2014). Later, Lorence (1999) cited the specimens as "Holotype MO 2162995; Isotype MO 4043108". The last digit of the numbering stamp on the corresponding sheet is hardly legible as a "9", possibly explaining the error in citation. In the last revision of the group (Taylor et al. 2010), the erroneous type citation of Lorence (1999) was repeated. In a repatriation project in 2001 (C. M. Taylor, pers. comm.), the lectotype (specimen 1) was deaccessioned and distributed to PMA. The sheet still bears a respective MO accession number stamp.

Taxonomy
A similar case of indirect lectotypification is found in Rudgea chiriquiensis. The species was described with reference to two collections at that time housed at MO, one of which was later distributed to PMA. Likewise, lectotypification (of the MO sheet) was achieved by Taylor (1995).
Distribution. Palicourea horquetensis is only known from few sites in Costa Rica and Panama. However, he overlooked that the earlier published name Cephaelis tonduzii K. Krause is still available under Palicourea and ought to have been adopted. The corresponding new combination is provided here. It is fortunate that the correct name for this showy species honors Swiss botanist Adolphe Tonduz (1862-1921), a long-term employee at the Museo Nacional de Costa Rica (CR) and one of the most prolific plant collectors in the country.
Typification. Cephaelis tonduzii was described with reference to an entire gathering by Tonduz. Krause did not cite a particular herbarium specimen and all possible duplicates are therefore syntypes (ICN, Art. 40.2 & Note 1; see also McNeill 2014). In addition, he was based at B and his working herbarium was destroyed during World War II. Collections made by Tonduz are widely distributed and more or less complete sets are found in CR and US (Stafleu andCowan 1976-1988). Nevertheless, the only known original material of Cephaelis tonduzii is a fragment at F. The fragment originates from the type at B, as shown by information given on the label. It consists of a capsule with a leaf and part of an inflorescence, which is mounted together with a photograph ("Berlin negative") of the original B specimen. The removed leaf is clearly recognizable on the photograph providing a definite link between both. Although not the best choice, the fragment at F appears to be the only extant original material and is here designated as lectotype.
A similar situation is found in Cephaelis discolor, which was based on a gathering of H. Polakowsky. Likewise, the only known original material is a fragment at F that originated from B. It consists of a leaf, some bracts and fruits in a capsule, which is mounted together with a photography of the B specimen, likewise destroyed. The fragment at F is here designated as lectotype.