Revisions of Ruizodendron and Pseudephedranthus (Annonaceae) including a new species and an overview of most up-to-date revisions of Neotropical Annonaceae genera

Abstract We present revisions of the Neotropical genera Ruizodendron and Pseudephedranthus (Annonaceae). Ruizodendron includes a single species R. ovale. Pseudephedranthus now comprises two species, with the description of the new species P. enigmaticus sp. nov. extending the range of the genus beyond the Upper Rio Negro region of Brazil (Amazonas) and adjacent Venezuela (P. fragrans), to include Guyana, Suriname, and the Brazilian state of Pará. An overview is provided of current revisions of Neotropical Annonaceae genera that will aid in accessing proper species information for this frequently encountered tropical rain forest family.


Introduction
Annonaceae are frequent components of tropical rain forests worldwide (Gentry 1993;Tchouto et al. 2006;Punyasena et al. 2008;Sonké and Couvreur 2014), for which it is vital to have a good overview of the most recent knowledge of the ca. 2450 spe-cies (Rainer and Chatrou 2006). Revisions and monographs are an important tool to document this knowledge, be it online or in a more traditional printed format. This documentation is not a trivial exercise. For instance, Ter Steege et al. (2016) wanted to estimate how many tree species occur in the Amazon basin. They used, amongst many other sources, the most recent revision of the Annonaceae genus Guatteria (Maas et al. 2015) to check their preliminary taxa list. Since Maas et al. (2015) reduced 62 Amazonian tree species into synonymy and described 10 new Amazonian species the check of Ter Steege et al. reduced the taxonomic error for this group substantially.
With respect to Neotropical Annonaceae, renewed revisionary efforts have been underway since the early eighties (Maas 1984;Chatrou 1999;Oliveira and Sales 1999). To date, 29 out of 34 Neotropical genera (Maas et al. 2011) have been revised, of which 11 published after 2000 (Table 1). Revisionary work on three further genera (Desmopsis Saff., Ephedranthus S.Moore, and Sapranthus Seem.) is at an advanced stage, and the final two, Annona L. (including Rollinia A.St.-Hil.) and Xylopia L., are subject of ongoing research. The Duckeanthus R.E.Fr. revision (Fries 1934) is the oldest still in use and together with a more recent addition by Maas et al. (1993) provides an adequate summary of all knowledge available on this monospecific genus. However, not all recently revised genera might be monophyletic (e.g. Desmopsis, Klarobelia Chatrou, Oxandra A.Rich. and Stenanona Standl.) and some of the taxonomic treatments will need to be modified in the light of future molecular phylogenetic evidence.
Two genera so far still awaiting revision were the monospecific genus Ruizodendron R.E.Fr. and the non-related genus Pseudephedranthus Aristeg.

Taxonomic history of Ruizodendron
The monospecific genus Ruizodendron occurs from Amazonian Brazil to Bolivia, Peru, Ecuador, and Colombia ( Fig. 1). It is readily recognized by the asymmetrical transversely ellipsoid monocarps (Fig. 2b, Fig. 3d, e, f ). The genus name consists of the Greek word for tree (dendron) and part of the name of the Spanish botanist Hipolito Ruiz López . Ruiz undertook a field trip throughout South America from 1779 to 1788 together with J. Pavón after which they published Florae Peruvianae et Chilensis Prodromus (1794, reprinted 1797) and Flora Peruviana et Chilensis (1798).
Ruizodendron was not originally described as a separate genus. During the field expedition of Ruiz and Pavón in South America in 1778-1788 they collected a specimen near Pozuzo in Peru that they described as Guatteria ovalis Ruiz & Pav. (Fries 1936;Maas et al. 2011;Maas et al. 2015). The specimen bore solely fruits and the description was therefore incomplete: "G. foliis oblongis ovalibusque. Arbor quadriorgyalis." The epithet "ovalis" was appropriate for their material since the type has oval, on both sides rounded, leaves (Fig. 3a) of ca. 10 cm long (Fries 1939, p. 544; although we now know that leaves can be found that are less typically oval). Fries studied this type specimen (B) and found that vegetative parts and structure of the monocarps differed from that typical of Guatteria, concluding that this specimen should not be attributed to that  Maas et al. 2011) and their most up-to-date revision. A more elaborate overview of taxonomic literature of these taxa can be found in Erkens et al. (2012).

Genus:
Most up-to-date revision: Anaxagorea A.St.-Hil. Westra 1984, 1985  Ongoing work genus. On the basis of the axillary flower buds with imbricate petals, and stamen and carpel morphology he concluded that this species should be placed close to Cremastosperma R.E.Fr. (Fries 1936(Fries , 1939. In 1925 the same species was collected by Steinbach in Santa Cruz in Bolivia, but also this specimen solely bore fruits. Only with the finding by Klug of a flowering specimen in the Upper ("oberen") Río Huallaga region (San Martín; not far from the type locality of Guatteria ovalis) the placement of the genus within Annonaceae could be studied for the first time. From the flower it was clear that this indeed was not Guatteria. Fries (1936) even thought that the flowers were quite aberrant for Neotropical Annonaceae: they had very thin, long and narrow petals ("linear lanzettlich"; Fig. 2a, Fig. 3b). Based on the position of the articulation on the pedicel and the position of the bracts Fries (1939) stated that Ruizodendron was related to Cremastosperma. However, the fruiting pedicel was placed in the middle of the long side of the monocarp ("…indem der Same bei Guatteria ovalis quer gestellt und in horizontaler Fläche ausgestreckt zu sitzen kommt mit dem Stiel des Monokarpiums mitten an der Längseite befestigt"; Fig. 2b, 3d, f ). Based on this fruit type Fries concluded that Ruizodendron was unique among American Annonaceae ("Wir erhalten dardurch ein Früchttyp, der allein dasteht unter den amerikanischen Annonaceen"). Furthermore, based on dried material he judged that the specimens had fleshy fruit in vivo. This was not similar to his dried material of Cremastosperma species, which had a dry, fragile pericarp. All this led Fries to describe Ruizodendron ovale (Ruiz & Pav.) R.E.Fr.
More recently, Van Setten and Koek-Noorman (1992) used fruit and seed features to place Ruizodendron as part of a group with six other Neotropical genera (be it together with other non-Neotropical genera): Cremastosperma, Ephedranthus, Malmea R.E.Fr., Oxandra, Pseudephedranthus and Pseudoxandra R.E.Fr. This is in line with an earlier grouping of Walker (1971) based on pollen characters and a concurrent conclusion by Van Heusden (1992) based on flower morphology. One year later, Keßler (1993) included Ruizodendon with Oxandra, Pseudoxandra, Cremastosperma, and Ephedranthus in his "Oxandra group", characterised by axillary flowers, imbricate sepals and petals, sulcate pollen grains, a single basal or lateral ovule, and free monocarps.

Taxonomic history of Pseudephedranthus
Until now, Pseudephedranthus consisted of a single species: P. fragrans (R.E.Fr.) Aristeg. known only from Brazil in the state of Amazonas (Upper Rio Negro Region, Rio Cauaburí), and adjacent Venezuela (Piedra de Cucuy; Fig. 1). In 1993, Maas collected a specimen (Maas et al. 6878,Fig. 6, 7) of P. fragrans during a botanical expedition in the Upper Rio Negro Region of Brazil and adjacent Venezuela and reported that the species was fairly common in the forested hills at the base of Piedra de Cucuy. Newly described here, P. enigmaticus Maas & Westra is distributed across Guyana, Suriname, and the Brazilian state of Pará (Fig. 1).
Pseudephedranthus (Aristeguieta 1969) can be easily recognised vegetatively. Its leaves exhibit some quite distinctive features (fig 4a): 1) the lamina is shiny on the upper side; 2) the primary vein is both raised and ribbed at the upper side of the lamina; 3) relatively few, widely spaced secondary veins (5-9) are present, and 4) the upper side shows a reticulate venation. The latest revision dated from 1999 (Oliveira and Sales 1999), written in Portuguese. Maas et al. (1993) hypothesised that Pseudephedranthus is closest to Oxandra by its elongate connective appendages, and to Ephedranthus by its androdioecious flowers and its extremely large seeds (25-30 × 13 mm, Fig. 5b). Pirie et al. (2006) already showed that Pseudephedranthus is nested within Klarobelia. More recent molecular phylogenetic work (Chatrou et al. 2012) confirmed this and demonstrated as well that Pseudephedranthus is part of the same clade as Ruizodendron together with Ephedranthus, Klarobelia, Mosannona, Oxandra, and Pseudomalmea (in tribe Malmeeae Chatrou & R.M.K.Saunders).

Materials and methods
Herbarium material was investigated from the following herbaria: A, AAU, CGE, E, F, G, GB, GH, INPA, K, L, LPB, MICH, MO, P, NY, RB, S, U, UC, US, WAG, and WIS. Dried herbarium material was used for measurements, colour indications and descriptions of surface structures. If measurements were done on material kept in spirit this is indicated by curly brackets { }. We have indicated the density of hair cover according to the following gradations: densely, rather densely and sparsely.
Distribution Notes. This species is easily recognizable by a combination of oval and on both sides rounded leaves with an olive green colour, long and narrow petals (often drying blackish), and stipes and apicules that are excentrically placed. Maas et al. 8773  Inflorescences axillary, 1-4-flowered with 2 nd order flowers originating from axils of lower bracts (or possibly also through accessory buds), often persisting on older leafless branchlets, pedicels with articulation in lower part and with 3-5 bracts, the uppermost bract above the articulation. Flowers actinomorphic, bisexual or staminate (androdioecious), 3-merous, perianth consisting of one whorl of sepals and two whorls of petals; sepals 3, valvate, basally connate, much shorter than the petals; petals 6, imbricate, elliptic, free, subequal; staminate flowers: torus conical, stamens numerous, extrorse, filament very short, apical prolongation of connective broadly ovoid in basal stamens to discoid in distal stamens; bisexual flowers: torus slightly raised, stamens numerous, but less so than in staminate flowers, apical prolongation of connective broadly ovoid; carpels numerous, spirally arranged, free, ovary 1-locular, with 1 basal ovule, style absent, stigma ovoid, papillate. Fruit apocarpous, composed of few, indehiscent monocarps, these ellipsoid, distinctly stipitate. Seed 1, pale brown, rumination lamellate in 2-4 parts, raphe a distinct groove.
Distribution. Two species in the Amazon regions of Venezuela and Brazil, and in Guyana and Suriname. Description. Tree, 3-15 m tall, 12-20 cm diam.; young twigs glabrous. Leaves: petioles 3-5 by 1-2 mm; lamina narrowly elliptic, 12-22(-26) by 4-6(-9) cm (index 2.8-4), chartaceous, pale gray to greenish gray above in sicco, somewhat bullate above in vivo, greenish brown to pale brown below in sicco, base acute, apex acuminate (acumen 5-10 mm long), primary vein raised above, secondary veins 6-10 on either side of primary vein, raised above, smallest distance between secondary veins and margin 4-7 mm, tertiary veins raised, rarely flat above, reticulate. Only staminate flowers seen, Inflorescence axillary, 1-2(-several)-flowered, pedicels 3-12 mm by 0.5-2 mm, rather densely to sparsely covered with erect to appressed, brown hairs to ca. 1 mm long, soon glabrous; bracts 4-5, depressed ovate, 1-2 mm long, outer side rather densely to sparsely covered with erect to appressed, brown hairs; flower buds ellipsoid; sepals shallowly ovate-triangular, ca. 2 by 2-3 mm, outer side rather densely to sparsely covered with erect to appressed, brown hairs; petals white, tinged with pale green in vivo, oblong-elliptic to narrowly so, 7-12 by 3-6 mm, outer side of outer petals densely to rather densely covered with appressed, brown hairs, inner side densely covered with whitish or greyish-white, curly hairs except for the glabrous base, outer side and apical part of inner petals densely covered with curly, white hairs; staminate torus conical, 2-2.5 mm long, ca. 1 mm diam. at base; stamens ca. 50, 2-2.5 mm long, apical prolongation of connective discoid, broadly elliptic. Monocarps 3-15, green in vivo, black in sicco, ellipsoid, 12-32 by 7-15 mm, glabrous or sparsely covered with appressed hairs, apex rounded, wall 0.2-0.5 mm thick, stipes 1-4 mm long, 1-1.5 mm diam.  Notes. Material of this species had previously been filed in herbaria under different generic names such as Cremastosperma, Guatteria, Klarobelia, Malmea, Oxandra,and Rollinia (which is quite aberrant!). The confusion is aptly expressed in the epithet "enigmaticus". This new species fits quite well, however, within the genus Pseudephedranthus (segregated from Ephedranthus by Aristeguieta in 1969), among others by the leaf venation, fruit and seed structure, and the strong similarity of the flowers. We acknowledge the fact that Pirie et al. (2006) demonstrated that Pseudephedranthus fragrans is nested in Klarobelia. From a morphological point of view (leaves and venation; flower morphology) this is quite surprising, given that overall morphology of Klarobelia is homogenous, and Pseudephedranthus is deviant from the general Klarobelia morphology. Therefore, we prefer to describe this new species in Pseudephedranthus to reflect the morphological similarity to P. fragrans. P. enigmaticus is distinct from P. fragrans by shorter petioles and, particularly, by the much denser indument of small curly hairs on most of the inner side of the petals (versus mostly small straight hairs). Also, seeds in P. enigmaticus are ellipsoid and 12-19 by 7-10 mm in contrast to P. fragrans were they are ovoid and larger (25-30 by 13-15 mm).
Distribution. Restricted to the Upper Rio Negro region of Brazil and adjacent Venezuela. Fig. 1.
Habitat and ecology. In lowland rain forest (in forested hills at the base of Piedra de Cucuy, Maas, pers. comm.). At elevations of 50-600 m. Flowering: April, December; fruiting: October, December.