New Species, New Combinations and New Synonymies Towards a Treatment of Acanthaceae for the Manual de Plantas de Costa Rica

In preparation for the publication of the Manual de Plantas de Costa Rica, new species, names, combinations, and synonymies are provided in six genera of Acanthaceae: Anisacanthus, Chamaeranthemum, Dicliptera, Justicia, Ruellia and Stenostephanus. The new species are A. grace-woodiae, J. altior, J. lithophila and S. chavesii. A new name at the species level, R. leonardiana, is provided for R. tubiflora var. hirsuta. With Habracanthus, Hansteinia, Kalbreyeriella and Razisea being subsumed within Stenostephanus, the new combinations S. blepharorhachis, S. citrinus, S. leiorhachis (= Razisea spicata non S. spicatus), S. strictus, S. ventricosus, S. villosus and S. wilburii are formalized. Seven new synonymies are presented for species of Chamaeranthemum, Dicliptera, Justicia and Stenostephanus, as well as lectotypifications in the first and latter two genera.

introduction Fieldwork that has supported the monumental effort to completely document the flora of Costa Rica in the form of the Manual de Plantas de Costa Rica (Hammel et al. 2003+) continues to reveal new species and new country records of Acanthaceae. This is true, despite the fact that the country's flora has been studied by botanists for more than 150 years, and its Acanthaceae have been treated in three major previous works (i.e., Lindau 1898;Leonard 1938;Durkee 1986). A number of new species have been discovered and described since the last countrywide treatment (e.g., Daniel 1993;Gómez-Laurito and Hammel 1994;Durkee and McDade 1996;McDade and Tripp 2007). Recently, focused effort to complete the Acanthaceae treatment for Volume IV of the Manual de Plantas de Costa Rica has resulted in a number of additional novelties and the need for several new combinations and synonymies.  Hammel et al. 19243 (CR).-7. Stenostephanus chavesii, Zamora et al. 9411 (CR). (Fig. 1-7, scale bar = 10 µm).
corolla and from A. tetracaulis by the lobes of the lower lip being longer and strap shaped (vs. ovate).
Known only from Cerro Brujo, Parque Nacional Diriá, Santa Cruz, Guanacaste, plants of this species have been collected in flower and fruit during the dry season, January through March and in December. Etymology.-We are happy to name this species for Grace Wood, who first discovered this charismatic plant and whose collecting has helped make clear that the Cerro Brujo area of Costa Rica warrants considerable additional fieldwork. Grace and her husband Monty have visited Costa Rica almost annually for many years, Grace collecting plants while Monty collects flies. A gifted amateur botanist, Grace has a very good eye and it has been a joy to have her prowling the herbarium in Santo Domingo, carefully identifying her own collections, during those visits.
Among Costa Rican Acanthaceae, this species may be recognized by its densely racemose inflorescences with one flower per node, bilabiate (effectively 4-lobed), bright red flowers that are presented with the filaments clasped within the ± horizontal median lobe of the lower lip and the anthers exserted beyond that lobe, and at anthesis mostly with the lateral lobes of the lower lip held at divergent angles to the corolla tube and the minutely bilobed, vertical upper lip also held at a divergent angle to the corolla tube.
Anisacanthus grace-woodiae was included in a phylogenetic study of the Tetramerium lineage (Justicieae) by McDade et al. (2018), as Anisacanthus sp. 26865. Results showed that it is sister to the other two Central American species of Anisacanthus, A. tetracaulis and A. nicaraguensis; these last two are sister taxa. All three species have bright red flowers with an apparently 4-merous corolla with the stamens and style held adjacent to the middle lobe of the lower lip at anthesis. The newly described and only Costa Rican species of this genus is readily distinguished from its closest relatives by the following key: Chamaeranthemum Nees, a New World genus with only four species, is very tenuously-if at all-distinct from pantropical Pseuderanthemum Radlk. ex Lindau. Both genera have corollas with a narrow tube that expands near the mouth and a subactinomorphic limb, two bithecate stamens and two staminodia (very rarely in Costa Rican Chamaeranthemum, one or both of the "staminodia" have been found fertile with the anther monothecate). Neither genus has been revised and the results of McDade et al. (2000;in progress) indicate that these genera are part of a clade referred to by these authors as the "Pseuderanthemum lineage," a group that remains poorly known and requires additional phylogenetic and nomenclatural study. Pseuderanthemum with ca. 120 spp. is the younger, nomenclaturally. Costa Rican material of Chamaeranthemum may be separated from that of Pseuderanthemum principally by the secund flowers (mostly opposite in the latter). However, whereas at least two Costa Rican species of Pseuderanthemum are heterostylous, all Costa Rican material of Chamaeranthemum we have seen appears to be homostylous. However, Lindau indicated that C. tonduzii Lindau had a style 15-16 mm long and a corolla with a 10 mm long tube. None of the type material, nor any other we have examined, confirms a style longer than ca. 9 mm. Lindau's anomalous measurements are consistent with the long-styled form of a heterostylous plant, or a fairly extreme case of approach herkogamy. Neither of these seems to apply to extant Costa Rican material, but field observations are sorely lacking and more extensive collections are needed. LAM has studied numerous specimens (including type material of C. durandii) of Costa Rican plants identified as the only two spp. known (also described) from the country, C. durandii and C. tonduzii. The former was said to be distinct from the latter by its hirsute leaves and inflorescence rachis. Indeed, one can separate specimens fairly consistently into those with glabrous leaves and a puberulous inflorescence rachis versus those with both leaves and inflorescence rachis sparsely pilose. A few specimens include material with both pubescence forms mounted on a single sheet. Also, plants of both pubescence forms rarely have one or both of the lower stamens fertile, the anther monothecate. Given these patterns, we consider this to be a species with intraspecific variation in trichome and androecial features, and here synonymize C. durandii with the older C. tonduzii. The latter name had two syntypes, both here considered to belong to this one species. We have chosen Tonduz 8567 as the lectotype because it has more known duplicates (the one at CR being the one we have studied in formulating these understandings) and with more flowers than the remaining syntype. Furthermore, the image at F of the B duplicate (presumably destroyed) has sketches on it, most likely in the hand of Lindau, showing many of the very same floral measurements used by him in the protologue. This verifies that material of the chosen lectotype was seen by Lindau, and served as an important source for his original description. As per specimens at CR and US, "Tonduz 8367" is actually Cordia lucidula I.M. Johnst. (Boraginaceae-APG Cordiaceae).

Dicliptera
Three of eight Costa Rican species of Dicliptera Juss. have deep magenta-purple flowers > 25 mm long, borne in long pedunculate cymes and are subtended by green, relatively small bracts: D. iopus Lindau, D. skutchii Leonard and D. trifurca Oerst. Study of type material of these three species indicates that D. trifurca and D. iopus are conspecific and refer to plants that are less than 1 m tall (except when supported by other plants), have corollas 33-36 mm long that are resupinate through 180°a nd markedly bent 4-5 mm above the base resulting in a semipendent presentation, and capsules 12-14 mm long. Unfortunately, for some decades, the epithet D. trifurca has been applied to a different species that remains unnamed pending geographically comprehensive study of this group. Plants of the undescribed species are larger in stature (to 3 m tall), have longer corollas ([4.8]  Because the undescribed species lacks full characterization in the literature, we provide a key to these three taxa below. In addition to the traits used in the key, LAM and CAK have observed habit differences among these three species. Plants of D. sp. A are self-supporting, sparsely branched shrubs to 3 m tall; those of D. skutchii can attain 5 m in height but are not self-supporting and must scramble on other plants. Finally, plants of D. trifurca rarely exceed 0.75 m in height unless sustained by other plants. Habit traits are not reliably described on herbarium specimen labels such that we do not use this trait in the key. However, we urge collectors to take note of habit when gathering specimens of these plants. 1. Leaf blades with 4-6(7) secondary veins per side; sepals connate less than 1/3 of total length; corolla usually strongly bent about 5 mm from base to become nearly pendent, tube 1.  Daniel 2011Daniel , 2016, and at least 15 additional genera (Kiel et al. 2017). Justicia is globally widespread in the tropics and subtropics with species richness centered in the New World. Although the genus is highly polyphyletic, New World Justicia along with five smaller New World genera, Cephalacanthus Lindau, Clistax Mart., Harpochilus Nees, Megaskepasma Lindau and Poikilacanthus Lindau, are together monophyletic (Kiel et al. 2017(Kiel et al. , 2018. Justicia can be characterized by flowers with (1) a zygomorphic corolla, (2) a rugula (i.e., stylar furrow), and (3) two stamens with bithecate anthers modified in various ways (e.g., connective tissue may be expanded and thecae may be of unequal size, inserted at different levels on the filament, displaced and/or appendaged). Here, we provide descriptions for two new species and make three new synonymies. The total number of species now known from Costa Rica is 40. Study of original material of Justicia bitarkarae reveals that it is a mixed collection of flowering material of J. costaricana and fruiting material of Stenostephanus silvaticus (Nees) T.F. Daniel. Lectotypification of J. bitarkarae is thus required. The flowering element is here chosen to lectotypify J. bitarkarae; in both the lectotype (CR) and the isolectotype (USJ), that element has leaves attached (the fruiting element has no leaves) and it also accounts for most of the description in the protologue. On the CR specimen at the lower right hand side there is also a piece of stem, with leaves only; that too belongs with the flowering element. Although the protologue indicates isotypes at F and MO, no such specimen has been located at either of those institutions. Gómez-Laurito (1991)  In his description of Justicia parvibracteata, Leonard (1938) mentioned affinities to J. pittieri but noted the former could be distinguished by bract width (1 mm wide or slightly wider vs. 2 mm, respectively). Leonard also claimed, in his key and descriptions of these two species, that the calyx lobes of the former are equal or subequal while in the latter they are distinctly unequal, with one lobe smaller than the others. On examination of the type of J. parvibracteata and other similar material from near the type locality, we have found the calyx to be indistinguishable from that of J. pittieri. It is not clear to us that Durkee studied type material of J. parvibracteata and J. pittieri when he prepared the treatment of Justicia for the Flora Costaricensis (Durkee 1986). His key brings these two out together claiming that the first has "bracts apically obtuse; leaves narrowly elliptical oblong" and the latter "bracts apically acute; leaves lanceolate elliptic." However, even looking only at the types, it is clear that both have bracts with variably shaped apices, from acute to rounded, and both have very similarly shaped, elliptic or lanceolate elliptic leaves. His descriptions of these two don't clearly distinguish them and he doesn't compare them in notes after either. Durkee (1986) noted that J. pittieri was only known from two collections: the type, Pittieri 8642, from the upper Río Yorkín, Talamanca, and Skutch 4652 from the vicinity Pejivalle in Cartago but the type of J. parvibracteata, Standley & Valerio 47001, is also from wet forests near Pejivalle. Durkee's concept of J. parvibracteata (1986) was clouded by having included specimens of the species newly recognized here, J. lithophila, a species with features of the inflorescence and seeds more like J. candelariae. Our study of the type material and recent collections from all along the Atlantic slope from where the two were originally described indicates that there are no traits that unambiguously separate these two species and we here synonymize J. parvibracteata with the older J. pittieri.

Justicia altior
Known from the Pacific slope of the eastern Cordillera de Talamanca (faldas del Cerro Amuo), Fila Costeña, in tropical wet forests, from 800-1400 m elevation, this species has been collected in flower from January to March. The pollen of J. altior most closely resembles that of Poikilacanthus macranthus Lindau (Kiel et al. 2018: Fig. 2I-J) and P. skutchii D.N. Gibson, which are 6-and 5-aperturate, respectively, and densely covered with insulae. However, in contrast to these, pollen of J. altior have insulae that are loosely spaced across the grain and are crateriform vs. planar. Poikilacanthus macranthus and P. skutchii are members of the "core Brandegeeana lineage" which also contains four endemic Costa Rican taxa: J. brenesii (Leonard)  Etymology.-The epithet "altior" refers to this species occurring at higher elevation than the outwardly similar Justicia aurantiimutata.
This species is very similar to J. aurantiimutata; notably plants of these two species share remarkable trichomes that dry to a reddish-orange color (see Fig. 11). However, in pollen morphology, the two species are quite different and we predict that they are not closely related among New World 'justicioids.' Justicia aurantiimutata occurs at lower elevations (0-450 m) and the two species are readily distinguished as indicated in the following key: This species is similar to Justicia candelariae, also with mostly spicate inflorescences with strongly imbricate and similarly shaped bracts, flowers with an unequally five-lobed calyx and relatively small, white or lavender corolla, and seeds with hemispherical papillae. It differs by its preference for rocks along streams; plants also have smaller bracts and smaller seeds. It is also similar to J. pittieri, which differs most notably by its more spicate-fasciculate inflorescences and seeds with sharply conical papillae.
Known from both slopes of the Cordillera de Guanacaste and from the Atlantic slope of the Cordillera Central, from 450-1200 m elevation; this species has been collected in flower from January to April. One collection (Godfrey 66536, MO), supposedly from ca. 120 m elevation on "steep forested slope, vicinity of Rincón, Península de Osa," is so far out of range for the species that a label mix-up is suspected. In fact, Tropicos ALISO McDade,Hammel and Kiel Fig. 11. Justicia altior. Scan of the holotype, Hammel et al. 26182 (CR); note trichome color on this dried specimen. Costa Rican Acanthaceae Novelties (www.tropicos.org) indicates that Godfrey's next number (Godfrey 66537) was from 700 m on "boulders...Rió María Aguilar" on the Caribbean slope of Alajuela, exactly where one could expect to find J. lithophila. The fact that the label data indicate that Godfrey 66537 was collected two weeks earlier than Godfrey 66536 is further evidence of a label mix-up. On our inquiry, staff at Florida State Univ. were unable to locate the original field notes. For these reasons, we prefer to exclude Godfrey 66536 from the formal list of specimens examined. Etymology.-The epithet refers to this species' frequent occurrence on rocks (in streams).
This species is distinctive for its plants of small stature (often growing on rocks along streams) with relatively small, narrow leaves, and for its densely bracteate, spicate or spicatefasciculate inflorescences with one or two flowers per bract, relatively small, white or lavender flowers with an unequally 5lobed calyx. In habit and habitat it can be very similar to plants in some populations of J. comata (L.) Lam. that, however, have spicate inflorescences arranged in elongate panicles, often with whorls of spikes at lower nodes, and equally 5-lobed calyces. In inflorescence and floral details the present species is more similar to J. candelariae. It is also somewhat similar to J. pittieri, which differs by its flowers in fascicles of 2-4(5) per bract, larger fruits and sharply (vs. hemispherically) papillate seeds. The new species can be distinguished from these other two as indicated in the following key:

Ruellia
Genus Ruellia L. can be characterized by its flowers with a subactinomorphic or weakly zygomorphic and ± campanulate or funnel-shaped corolla, with four bithecate stamens, and seeds with hygroscopic trichomes. The fruits are capsular, with the retinacula persistent on the fruit wall except that species formerly treated in Blechum P. Browne have placentae and attached retinacula that split away from the fruit wall at dehiscence. The recent description of four new species and finding that Blechum (two species from Costa Rica) is nested within Ruellia (McDade and Tripp 2007;Tripp et al. 2009) yields a total of 22 species in Costa Rica. Here, we provide a description of an additional species, which has never been fully characterized. Hammel, nom. nov. (Fig. 6, 13) Herb, suffrutex or shrub 0.5-2(-3) m tall; stems erect, quadrangular, lanate. Leaves with the petiole 1-3.5 cm long; blade 10.5-19 × 4-9 cm, elliptic, cuneate to attenuate at the base, acute to (more often) acuminate at the apex, subentire to (more often) shallowly crenate, velutinous (often subapressed) on both sides, with patelliform glands ± conspicuous on the lower surface, apparently lacking on the upper. Inflorescences terminal, dense (but few-flowered), sessile or with a peduncle up to ca. 0.7 cm long, subspicate-capitate, the rachis null or up to ca. 0.3 cm long; bracts ± involucrate-imbricate, 30-90 × 5-50 mm, foliaceous, elliptic, acute to acuminate at the apex, pubescent like the leaves; bracteoles not seen. Flowers sessile or subsessile with the pedicel to ca. 1 mm long, velutinous; calyx with the tube 0.5-1 mm long, the equal or subequal segments (4-)6-8 mm long, narrowly deltate to lanceolate, pilosulose and ciliate, often also with a few longer trichomes; corolla white, the tube 46-72 mm long, the proximal unexpanded part 20-37 mm long, the distal expanded part 21-35(-42) mm long, the mouth 15-18 mm wide, the lobes 8-10 × 6-10 mm, irregularly crenate-erose to emarginate, externally puberulous to tomentose; anthers included, 5.5-6 mm long; pollen spheroidal, 3-porate, surface verrucate, coarsely reticulate, reticulum homobrochate and psilate; style 31-55 mm, approximately equal in height to the anthers, the stigma to ca. 4 mm, bilobed and somewhat laminar, with one lobe smaller. Fruits ca. 1.2 cm long, ellipsoid, minutely puberulous, the placenta not fracturing; seeds potentially 6 (often only 2 develop), ca. 4 mm in diam., suborbicular, flattened, apparently glabrous.

Ruellia leonardiana
Known from wet forest at 100-400(-800+) m elevation on the southern Pacific slope of Costa Rica (Fila Costeña and Golfo Dulce region), with flowering collections from Jan, Feb and Oct through Dec. Also known from Colombia (Meta: near Villavicencio, 500 m).
Etymology.-The epithet chosen is in keeping with the botanical tradition of honoring the author of a replaced name. In the case of Leonard, prolific student of New World Acanthaceae (author of more than 500 epithets in the family), the wonder is that he had yet to be so honored in Ruellia.
Following the taxonomy of Tripp and Luján (2018), the treatment of Ruellia for the Manual de Plantas de Costa Rica (Hammel et al. 2003+) will again recognize Ruellia tetrastichantha Lindau [syn. R. tubiflora var. tetrastichantha (Lindau) Leonard] at the level of species. Likewise, R. tubiflora var. hirsuta Leonard is easily recognized and we here elevate it to species level. Both of these species have pure white corollas, whereas R. tubiflora, not known from Costa Rica, has corollas with a purple spot in the throat (corolla lobes may also be suffused with purple). Ruellia leonardiana may be distinguished from R. tetrastichantha by its very different pubescence (leaf blades velutinous vs. glabrous or at most pilosulose in R. tetrastichantha), and by its consistently terminal inflorescences (vs. both terminal and axillary) with persistent bracts (vs. soon deciduous) that thus apparently do not produce the notably naked condensed rachiscone-like in fruit-that is so characteristic of R. tetrastichantha. We note that extreme variation is reported in the length of the corolla tube in both of these species, as well as in the lengths of the unexpanded basal portion and the expanded distal portion (see McDade and Tripp [2007] for explanation of terms). This variation, based apparently on open flowers on herbarium specimens, merits further examination, ideally including populationlevel samples of fresh flowers. The new species can be confused with R. odorata E. Tripp & McDade, which has the leaf blades at most strigose and appressed puberulous along the principal veins, smaller bracts (10-23 × 2-5 mm), longer calyx lobes (10-14[-18] mm long), and is known only from the nearby, but biogeographically distinct, Osa Peninsula. The Costa Rican species of Ruellia with pure white flowers may be distinguished by the following key: Key to the Costa Rican species of Ruellia with corollas uniformly white:

Stenostephanus
The genera Habracanthus Nees and Hansteinia Oerst., both with Costa Rican species, have been accepted on morphological grounds as synonyms of Stenostephanus Nees, with some new combinations already made. Kalbreyeriella Lindau and Razisea Oerst., also with Costa Rican species, have been suspected as possibly also better placed in Stenostephanus (see, e.g., Daniel 1999). More recent molecular phylogenetic studies (e.g., Kiel et al. 2006) have supported the inclusion of all four genera within a thus monophyletic Stenostephanus, characterized by bilabiate, more or less tubular corollas lacking a rugula, two stamens with monothecate anthers, these held, with the style, adjacent to the upper lip of the corolla, staminodia lacking, and capsular fruits with the retinacula remaining attached to the fruit wall. These plants also share pollen that is banded or "girdled" (i.e., "Gürtelpollen" sensu Lindau 1895). We here provide the remaining new combinations, a few new synonymies and one new species description that are necessary to accommodate the Costa Rican species of Stenostephanus. Note that we use "T.F. Daniel ex" in cases where our colleague Tom Daniel has proposed the new combinations via Tropicos as well as annotations on specimens at various herbaria (e.g., K, MO, US), but the changes have not yet been published. This brings the total species known for the country to 12. ( BEH has examined numerous specimens of this Costa Rican endemic and has studied and collected material at one population in the field. The type specimen-with no known duplicates, no paratypes cited, nor other material known to have been used-is represented only in photos produced by McBride's Berlin negatives, made available online at F: http://emuweb.fieldmuseum.org/botany/search_berlin.php.

Stenostephanus blepharorhachis
While these specimens do not themselves serve as types, they unquestionably help to understand the species. In this particular case, an illustration of floral parts (including a pollen grain!), most likely in Lindau's hand, is clearly visible and gives the very measurements used in the protologue description. One might consider that an uncited illustration that comprises part of the original material and use it as lectotype according to Article 9.12 of the Code (Turland et al. 2018). But since there is no longer any physical specimen, nor illustration, we chose a neotype. Although the protologue shows "Costa Rica et Veragua [i.e., Panama]" as the locality, the label itself says only "Costa Rica." Furthermore, the species is not otherwise known from Panama and was not treated in the Flora of Panama (Durkee 1978). The neotype we chose has numerous duplicates, some of which have been available to us for study and allow us to determine that it coincides with Lindau's protologue. The duplicate at US is scanned and available on the NMNH Botany Collections site: http://n2t.net/ark:/65665/39333371c-b932-4979-963d-1342f9eec6a1. Etymology.-We name this species in honor of former parataxonomist José Luis Chaves Chaves, who made the type collection (and many others) of this new species as well as many other interesting collections from the Cordillera de Guanacaste, especially on the slopes of the Tenorio and Miravalles volcanos.
Among the Stenostephanus species with the corolla tube gradually expanded from the base towards the mouth and spicate to narrowly racemose (flowers short pedicellate) inflorescences, S. chavesii stands out for its leaves with the blade rounded to often cordate at the base, and pilosulose on the upper surface. In leaf shape, it can be very similar to S. citrinus, which differs markedly by its widely paniculate (vs. narrowly racemose) inflorescences and bright yellow flowers. Likewise, S. wilburii can also have leaf blades that are basally obtuse, if not cordate, but that species has shorter corollas. The following key will distinguish among the red-flowered species of Stenostephanus with a narrowly racemose inflorescence and corollas with a tube that gradually expands from the base: We are grateful to Reinaldo Aguilar of Puerto Jiménez for his wealth of knowledge of plants of the Golfo Dulce region, especially the Osa Peninsula. He has been instrumental in the discovery of many new species or new country records. He assisted all three authors on one trip to the area and BEH on numerous. We are also grateful to Thomas F. Daniel for responses to numerous inquiries, to Michael Grayum for nomenclatural consultations, to illustrator José Alejandro Herrera, and to curatorial staff at CR, F, FSU, MO, RSA and USJ. We appreciate the careful work of two anonymous reviewers, which improved the manuscript. Funding from the Torrey Botanical Society (CAK) and from the U.S. National Science Foundation to LAM (DEB 9707693, DEB 0108589, DEB 0743178) and to CAK and LAM (DEB 1754845) was instrumental in accomplishing work on the Acanthaceae treatment for the Manual de Plantas de Costa Rica, as was the support of the Organization for Tropical Studies and Instituto Nacional de Biodiversidad (INBio).
literature cited