Thenardia H.B.K. is a neotropical genus of vines endemic to rain forests of southern Mexico. The genus is distinguished within the Apocynaceae by the combination of watery sap (shared with Echites R. Br. subG. Pseudechites and selected species of Parsonsia R. Br.), fully exserted anthers (shared with Forsteronia G. Mey., Isonema R. Br., Pottsia Hook. & Arn. and selected species of Parsonsia), subumbellate inflorescences, fused follicles (shared with various species of Artia Guillaumin, Forsteronia, Mandevilla Lindl. and Parsonsia) and one species (T. floribunda) with coiled filaments (shared with selected species of Parsonsia and Dewevrella De Wild.). Because of its combination of convergent characters the relationship of Thenardia within the Apocynaceae has been under considerable dispute (see Generic Relationships). The recognition of species, however, is facilitated by each taxon's possession of discrete morphological characters which help to readily separate it from other members of the genus.

Taxonomic History

Kunth (1819) established the genus Thenardia as monotypic with the description of Thenardia floribunda. He named the genus in honor of his friend L. J. Thénard (1777–1857), chemist and plant physiologist. A second species of Thenardia (T. suaveolens) was described by Martens and Galeotti (1844), this species, however, was later reduced to synonymy under T. floribunda by Woodson (1936). Woodson's decision to not recognize T. suaveolens is followed here. Baillon (1890) described a third species of Thenardia (T. galeottiana), and discussed Thenardia's affinities with Parsonsia (see Generic Relationships).

Woodson (1936, 1938) provided the first taxonomic treatments of Thenardia, recognizing four species for the genus: T. floribunda, T. galeottiana, T. gonoloboides, T. tubulifera. The latter two species (T. gonoloboides and T. tubulifera) were described by him (Woodson 1932, 1936, respectively). The characters by which Woodson recognized these species, however, are regarded here as extremely variable and insignificant, and thus the species are regarded as synonyms under T. galeottiana and T. floribunda, respectively.

Williams (1995) subsequently described an additional species in Thenardia (T. chiapensis) , and included a few taxonomic notes on the genus.

Five additional names proposed in Thenardia have been trasferred to or synonymized under other genera (Forsteronia, and Rhynchanthera DC.) (see Excluded Taxa).

Generic Relationships

It is widely recognized that at present there is no modern intrafamilial classification system for the Apocynaceae (Fallen, 1983; Middelton, 1995). Treatments to date are not regarded by most authors as monophyletic. Schumann's treatment of the family in Engler and Prantl's Die natilrlichen Pflanzenfamilien (1895) is perhaps the most universally accepted for the family. Pichon (1948a, 1948b, 1950) and Leeuwenberg (1994) have provided subsequent treatments of the family. Pichon's work includes more detailed descriptions of the genera as well as provides some of the first observations on pollen for the family, however, the overall taxonomic arrangement varies only slightly from that of Schumann's. The treatment of Leeuwenberg is mainly a checklist of genera arranged by their suspected subfamilies and tribes. His treatment includes very little discussion of characters and the placement of the genera within the family.

Although there is no universal agreement as to the recognition and arrangement of tribes, subtribes and genera within the Apocynaceae, most authors (Schumman, 1895; Rosatti, 1989; Leeuwenberg, 1994) accept two subfamilies, although some authors have argued for three (Pichon, 1948a; Fallen, 1983). The two subfamilies can be distinguished by the following key:

Thenardia, as evidenced by its stamens fused to the style head, dextrorse aestivation and comose seeds, belongs in the Apocynoideae. The Apocynoideae has been divided into several different tribes by various authors (see below). Below is a brief history of the assignment of Thenardia within the family by various authors and a discussion of the genera most possibly related to Thenardia.

When Kunth (1819) erected Thenardia, a classification system of the Apocynaceae had not yet been established and he merely stated its resemblance to Echites. This is of little interest, however, because at that time Echites was nothing more than a grab-bag of miscellaneous Apocynoideae with over two hundred species. It was not until Woodson (1936) revised Echites that the genus ceased being artificial and became legitimately circumscribed, ultimately housing seven species.

De Candolle (1844) established the first classification system of the Apocynaceae, recognizing seven tribes in the family: Willughbeae, Carisseae, Plumereae, Parsonseae, Wrighteae, Alstoneae, and Echitideae. He placed Thenardia in the tribe Echitideae (= Apocyneae of Schumann) near Haemadictyon Lindl. (the species of which are at present divided between Lauhertia and Prestonia). The Echiteae as viewed by de Candolle, however, appears to represent an unnatural amalgam of artificially arranged taxa.

Bentham (1873) recognized three tribes (Carisseae, Plumerieae, Echitideae) in the Apocynaceae. In his work he placed Thenardia in the Echitideae subtribe Parsonsiineae (including, in this order; Prestonia, Thenardia, Forsteronia, Vallaris, Parsonsia, Lysonia (= Parsonsia), Pottsia, Isonema, Wrightia). The Parsonsieae of Bentham is an expanded version of de Candolle's Parsonsieae, which included only four genera (Vallaris, Parsonsia, Balfouria and Beaumontia; the latter two genera were placed elsewhere by Bentham). The additional members of Bentham's Parsonsieae came from Wrightieae of de Candolle and selected genera of his Echiteae (Prestonia, Thenardia, Forsteronia, Pottsia and Isonema). What explains the discrepancy between the genera recognized in the Parsonsieae by de Candolle and in that of Bentham is that both authors had different concepts for the taxon. De Candolle separated the Parsonsieae from the other tribes by its ovary of two united carpels, while Bentham defined the Parsonsieae by its exserted anthers and rotate corolla.

Schumann (1895) essentially followed the same treatment of Thenardia as Bentham, placing it in his subfamily Apocynoideae tribe Parsonsieae. His tribe Parsonsieae would encompass the same definition (exserted anthers) and genera as Bentham's, although Schumann s separation of genera differed. Whereas Bentham keyed Thenardia, Prestonia and Forsteronia from Parsonsia because of their restriction to the New World (vs. restricted to the Old World), Schumann keyed Thenardia, Prestonia and Forsteronia from Parsonsia based on their ovary of two free carpels (vs. ovary of two united carpels).

Woodson never attempted a systematic treatment of the Apocynaceae as a whole. He did, however, tend to indicate relationships by constructing natural keys. In his key to the American genera of the Echitoideae (Woodson, 1933), he separated Thenardia along with Echites and Prestonia. He keyed Thenardia from Echites on its rotate corolla and exserted anthers (vs. salverform and included) and from Prestonia based on its lack of an annulate corolla moutH. Forsteronia was separated from the former genera based on its numerous colleters (“squamellae” of Woodson) alternate the calyx lobes (vs. solitary and opposite), the same difference used by Schumman.

Pichon (1950) was the first to recognize that Thenardia had an ovary of two united carpels. Because of this, he placed Thenardia, with Parsonsia and Artia, in his tribe Parsonseae subtribe Parsonsiinae, which was defined by its ovary of two united carpels. He keyed Thenardia from Parsonsia on its possession of developed intrapetiolar stipules (vs. reduced or absent) and corolla sizE. Artia was keyed from the former genera by its bent filaments. It seems evident that, due to their lack of fruiting specimens both de Candolle and Schumman refrained from placing Thenardia with Parsonsia. Indeed, if they were to have observed either the fruits or ovary of Thenardia they would have most likely classified it near Parsonsia.

Leeuwenberg (1994) placed Thenardia in his subtribe Parsonsiinae following this order: Delphyodon, Grissea, Parsonsia, Dewevrella, Thenardia, Artia, Pottsia, and Isonema. He neither provided a key to the genera nor discussed the differences between them. Leeuwenberg also defined the Parsonsiinae in a broader sense than had previous authors. His Parsonsiinae was defined as including lianas, with short corolla tubes, exserted anthers, and an ovary of two free or united carpels. The expanded definition of the Parsonsiinae is probably explained by the variability that exists within Parsonsia, a genus which possesses species with and without united carpels.

Relation to Parsonsia: Baillon (1890) regarded Thenardia as little distinct from Parsonsia, suggesting that the only real difference was that Thenardia was restricted to Mexico. He thus proposed that Thenardia be regarded as a section of Parsonsia (Baillon, 1891). He explains the presence of Thenardia in Mexico as being an extremely “ancient” introduction from “oceanic archipelagos where [Parsonsia] is present in high number.” An introduction of an Apocynaceae genus into the neotropics from Indonesia is not altogether improbable as the genus Alstonia shares members in both the Neotropics (Mexico and Central America) and the Paleotropics (Africa, Australia and Indonesia) (Gentry 1983a).

The species of Parsonsia most similar to Thenardia is P. rotata Maiden & Betche. The characters that P. rotata shares with Thenardia are watery sap, glabrous habit (except corolla throat), coiled filaments, leaves with obscure secondary veins, five free nectary scales, older stems with corky outgrowths and fruits of two fused follicles 15–20 cm long (J. B. Williams, 1996). Indeed, an illustration (ibid.) of the inflorescence of P. rotata even shows the inflorescence as compact and umbel-like.

Parsonsia mainly differs from Thenardia in its geography (Indonesia vs. Central America), lack of intra-petiolar stipules, valvate (occasionally dextrorose) aestivation, number of calycine squamella (0-many vs. 1), pubescent corolla throat (vs. glabrous) and stamens attached to the base of the corolla (vs. attached to the apex). In a broad scheme these differences seem inconsequential, however, many of the genera in the Apocynoideae are defined by fewer and seemingly less important characters (e.g., Mesechites differs from Mandevilla in only one character: branched vs. unbranched inflorescence).

Relation to Forsteronia: The only similarities between Thenardia and Forsteronia (besides their restriction to the Neotropics) is their possession of exserted anthers and fruits with fused follicles. Hansen (1985) suggests that the fused follicles have “little phylogenetic significance” (at least within Forsteronia) since they have evolved independently in several unrelated lineages. He likewise suggests that the moliniform structure of the follicles also represents convergence. Hansen mentions that Bentham placed Thenardia near Forsteronia i in the Parsonsieae, but does not suggest any close relation between two, supporting rather the relation of Forsteronia to Mandevilla on the basis of their shared colleters at the apex of the petiole along the upper surface of the leaf.

Forsteronia also differs from Thenardia in having many colleters alternate to the sepals (vs. solitary and opposite), domatia on the underside of the leaves and thyrsiform inflorescence (vs. subumbellate cyme) and a different style head shape. Of the possible candidates related to Thenardia, Forsteronia seems the most improbable.

Relation to Echites: Echites is divided (Woodson, 1936) into two subgenera, E. subG. Echites and E. subG. Pseudechites, that are separated by their corolla lengths and sap (30–80 mm and milky vs. 12–25 mm and watery, respectively). Thenardia is more similar to E. subG. Pseudechites (represented by the species E. agglutinata Jacq. [Prestonia agglutinata (Jacq.) Woodson of some authors], E. tuxtlensis Standley, E. turhinata Woodson, and E. woodsoniana Monach.).

Thenardia and Echites subG. Pseudechites share the following characters: watery sap, one colleter opposite each sepal, intra-petiolar stipules, leaves with obscure secondary veins, and an extensive growth form from woody terrestrial stems (the above observations were taken from a population of E. woodsoniana by the author), as well as a neotropical distribution. Echites subG. Pseudechites mainly differs from Thenardia in having a racemous inflorescence, included anthers and fruits of two free follicles.

In conclusion, I refrain from placing Thenardia within a subtribe due in part to the artificial nature of the tribes and subtribes as circumscribed to date and in part to the considerable convergence that exists within the Apocynoideae. I am inclined, however, to support a relation between Thenardia and Echites subG. Pseudechites (as described above). Ultimately, this is a hypothesis and I believe that until a thorough cladistic analysis as constructed by molecular evidence is formulated the relationship of many of the genera in the Apocynoideae will remain speculative.

Morphology

Leaves. The leaves of Thenardia are opposite and entire, as in all neotropical lianas of the ApocynoideaE. Thenardia lacks colleters at the apex of the petiole and domatia on the underside of the leaves, typical of Forsteronia, a genus which Thenardia is often confused. The leaves have obscure secondary veins, which are alternate in formation.

Although the Apocynaceae is typically regarded as lacking stipules, Thenardia, and all other Mexican Apocynoideae, possess intra-petiolar stipules. These stipules are not to be confused with “colleters”, they are in fact multi-cellular foliaceous structures arising from the leaf axil. Two of the species are pubescent along the petioles and midribs of the leaves beneath, while the third species (T. chiapensis) has wholly glabrous leaves. The leaves of T. floribunda are often two to three times longer and broader than the other two species. Field studies have shown that the above mentioned characters are uniform and consistent within a population.

Inflorescence. The inflorescence in Thenardia is a pedunculate di-to trichotomously branched umbelloid axillary cyme. The inflorescences have a congested growth form which often makes them appear to be umbels. The umbel-like appearance of the inflorescence is a distinctive trait that helps to readily distinguish the genus from other neotropical viney Apocynoideae. The distinctiveness of the inflorescence lead Sessé and Mociño to name the species T. umbellata (see T. floribunda below).

The number of flowers per inflorescence ranges from 6–15 in T. chiapensis and T. galeottiana and between 20–30 for T. floribunda. The more congested inflorescence of the latter species helps to distinguish it from the former two species.

Flowers. The calyx in Thenardia is regular, pentamerous, and with a solitary colleter opposite to the adaxial surface of each sepal. The sepals are narrowly triangular and 1–2 mm long in T. chiapensis and T. floribunda. In T. galeottiana the sepals are quite distinctive being narrowly elliptic, curled, and 4–6 mm long. The shape and dimensions of the sepals for each species corresponds to the shape and dimensions of the bracts subtending the pedicel.

The corolla is rotate, wholly glabrous and with wholly exserted stamens. The exserted stamens are a unique feature of Thenardia, helping to distinguish it from other Mexican Apocynoideae genera (ex. Laubertia and Prestonia) which have only their anther tips exserted past the corolla orifice. The filaments are attached to the apex of the corolla throat as opposed to Parsonsia which has its filaments attached to the bottom of the corolla tube. The filaments are coiled in T. floribunda. Coiled filaments were also observed in one of the specimens (Martinez 124) of T. galeottiana, while the filaments of T. chiapensis are apparently always straight. The aestivation of Thenardia is dextrorse as are all other members of the Mexican Apocynoideae.

Fruit. The fruit of Thenardia is composed of two fused, herbaceous, pendulous, dehiscent follicles developing from two united ovaries. The follicles are moliniform and can reach a length of 40 cm (T. floribunda).

The seeds are 6–20 per follicle, fusiform 0.8–2.0 cm long, glabrous, and each with a micropylar coma. The surface is essentially smooth, lacking any surface sculpturing which could prove taxonomically useful. The overall structure of the seeds are essentially similar to the seeds of other genera of the Apocynoideae.

Pollen. The pollen of Thenardia is spherical, triporate, and with a smooth perforate surface (Fig. 1). The diameter of the grains ranges from 30–60 µm (Table 1). Thenardia chiapensis and T. galeottiana have much smaller grains than T. floribunda, a difference correlated with the latter species also having much larger anthers than the former two (6–7 mm vs. 2.5–5.0 mm). There is not a difference in the diameter of the pollen apertures between the three species. All species have round apertures ranging from 3.5–4.0 (5.0) µm.

Fig.1.

Microphotograph of T. chiapensis pollen. Microbar =100 µm.

The structure of pollen grains in the Apocynoideae is essentially uniform. The majority of species have spherical, 3–4-porate (occasionally 5 to polypantopo-rate) and with smooth perforate surfaces (Erdtman, 1952; Nilsson, 1990; Sampson and Anusarnsunthorn, 1990; Roubik and Moreno, 1991; Nilsson et al., 1993). Huang (1989) has shown that the number, arrangement and shape of the pores are occasionally useful diagnostic characters for recognizing genera. The latter characters can be observed using either light or scanning electron microscopy (SEM). The same author also shows that species of a genus can in many cases be distinguished by their exine stratification using freeze/fracture methods with a transmission electron microscope (TEM). The present study and others (e.g., Roubik and Moreno, 1991) show that species of a genus can occasionally be identified by their pollen diameter.

Table 1.

Pollen and pore size in Thenardia species.

A comparison of SEM photos of pollen in Thenardia (Fig. 6) with SEM photos of Forsteronia (Hansen, 1985; Nilsson et al., 1993) and Parsonsia (Sampson and Anusarnsunthorn, 1990) show the three genera to be little different from one another in pollen characters. Because of the uniform and convergent nature of pollen grains in the Apocynoideae, their study in Thenardia provides at present no further insight into its relationship with other genera.

Chromosome Numbers. There have been no reported chromosome counts of Thenardia, and none were done during the present study. A cursory review of chromosome counts of the Apocynaceae up to 1991 (the last published account of the Index to Plant Chromosome Numbers; Goldblatt and Johnson, 1994) shows that at present the only chromosome counts for genera closely related to Thenardia are for Echites (x=6), Parsonsia (x=9), Pentalinon (x=6) and Prestonia (x=9). A chromosome count of Thenardia could help to resolve its placement within the family. However, because studies (Van der Laan and Arends, 1985) suggest that aneuploidy reduction from the family's basic number of x=11 might have occurred multiple times in the family, a count of either x=6 or x=9 for Thenardia could represent a case of convergence rather than a relationship to either of the above mentioned genera. It is of interest, however, that the chromosome count of Prestonia is from P. acutifolia (Benth.) K. Schum. a species that could be considered a member of Echites subG. Pseudechites based on its included anthers and reduced sepals (see Gentry (1983b) and J. K. Williams (1996) regarding the differences between Prestonia and Echites). The count of x=9 for both Prestonia acutifolia and Parsonsia suggests a possible chromosomal relation between the two and with the morphological similarities between Thenardia and Echites subG. Pseudechites acknowledged, a possible relation between Thenardia and Parsonsia can possibly be inferred. This relationship is highly speculative, however, with so few chromosome counts in the Apocynaceae to compare, (the count of x=6 for Echites was taken from a species in E. subG. Echites).

Methods

The morphological study is based on field observations and the examination of approximately 80 herbarium specimens from the following institutions: F, GH, LL, MA, MEXU, MO, NY, TEX, US.

The pollen of the three species was collected from dried herbarium specimens, TEX/LL. The pollen was not rehydrated, acetolyzed or critical point dried. Observations of Parsonsia pollen by Sampson and Anusarnsunthorn (1990) suggest that the skipping of these procedures produces no significant change in either the diameter of the pollen grains or the pores. Pollen grains were mounted onto a SEM stub with double-sided adhesive tape placed in a dessicator for 24 hours and coated with gold using a LADD Sputter Coater. Observations of the pollen were made with a Phillips 515 Scanning Electron Microscope (Cell Research Center, University of Texas at Austin) and photographed using Polaroid type 55 positive/negative film.

Taxonomic Treatment

Fig. 2.

Distribution map of Thenardia; circles = T. floribunda, squares = T. galeottiana, triangles = T. chiapensis.

Fig. 3.

Photograph of T. floribunda, by J. K. Williams.

Fig. 4.

Reproduction of plate of T. floribunda from Kunth (1819).

Fig. 5.

Holotype of T. chiapensis (TEX).

Fig. 6.

Isotype of T. galeottiana (US).