Introduction

Madagascar is the primary center of diversity of Homalium Jacq. (Salicaceae) as it has been circumscribed since Warburg (1893), with six of the ten currently accepted sections present in the country and five of these endemic (Applequist, 2016a). The most recent full taxonomic treatment of the Malagasy species was published by Sleumer (1973). Recent revisionary studies (Applequist, 2016b, 2018a, 2018b, 2020; Wassel & Applequist, 2020) have substantially increased the number of species recognized in most sections, partly due to the availability of newer collections, and partly because Sleumer was inclined to lump distinct taxa or recognize them only at infraspecific levels if their floral anatomy was similar.

The last Malagasy section of Homalium not yet addressed by the author and her collaborators is Homalium sect. Antinisa (Tul.) Baill. ex Warb., which according to Perrier de la Bâthie (1940, 1946) and Sleumer (1973) only contained one species: H. involucratum (DC.) Hoffm. This vegetatively unremarkable taxon has an inflorescence morphology unique in the genus. The paired reniform bracts subtending each inflorescence node are strongly accrescent and become very large, especially at the inflorescence base. At anthesis the bracts spread to expose the flower(s) (Fig. 1), but later completely enclose them and overlap with the bracts from the two alternating nodes above. Homalium involucratum is occasionally confused with species of Tisonia Baill., also Salicaceae, which have huge accrescent sepals, but otherwise resembles no other species in the family.

Sleumer (1973) recognized three formae within Homalium involucratum. He considered most specimens to fall within H. involucratum f. involucratum, which is widespread in littoral to low-elevation eastern forests, while two other formae are geographically localized. Homalium involucratum f. lucidum (Scott Elliot) H. Perrier is confined to the extreme southeast of Madagascar; he distinguished it from f. involucratum by its more coriaceous, greenish-drying leaves with sometimes subcordate bases (vs. usually brownish-drying leaves with convex bases). Homalium involucratum f. hildebrandtii (Baill.) H. Perrier is native to extreme northern Madagascar, usually in mid-elevation forests; it is distinguished by its often larger, longer-petiolate, sometimes ovate-elliptic and acuminate leaves.

Both geographically restricted formae accepted by Sleumer (1973) were first published at higher ranks, Homalium involucratum f. lucidum as var. lucidum Baill. and f. hildebrandtii as H. hildebrandtii Baill., and both were reduced to the rank of forma by Perrier de la Bâthie (1940, 1946), both of which Sleumer (1973) maintained. Homalium involucratum has two other formae published by Perrier de la Bâthie (1940), f. viride H. Perrier and f. cloiselii H. Perrier, treated as synonyms of f. lucidum by Sleumer (1973). Homalium vatkeanum Hoffm. has been treated as an older species-level synonym for f. hildebrandtii. The synonymy of the latter two may be questionable, but their status cannot be addressed at this time as closer study of type material is necessary.

Fig. 1.

Homalium involucratum (DC.) Hoffm.: inflorescence at anthesis, with pairs of large bracts at each node spread open to expose the small greenish flowers.

[Andriambololonera et al. 133] [Picture: D. Rabehevitra]

I and most recent taxonomists consider it inappropriate to use the rank of forma for morphological variants that are well distinguished, often apparently fixed within populations, and with distinct ecological preferences. Available specimens were examined to support a proposal to recognize forma lucidum at a higher rank. This study was primarily based upon specimens available at MO, especially for the biometric data. Types and other specimens at P had been briefly observed during a visit to that institution several years previously, and specimen images were available through its website; images of types at other institutions were seen through JSTOR Global Plants website [ https://plants.jstor.org].

Critical examination of these specimens and images led to the conclusion that forma lucidum is clearly distinguished from the other two formae, which are more similar to one another and more overlapping in morphology, by multiple characters. In one locality where forma lucidum and forma involucratum are sympatric, they remain well distinguished. Recognition of the southeastern endemic taxon as a distinct species seems clearly to be merited. The epithet of var. lucidum Scott Elliot cannot be used at the species level because of the prior existence of Homalium lucidum Scott Elliot (a species in sect. Odontolobus Warb.). The epithet of the heterotypic synonym H. involucratum f. viride H. Perrier is therefore herein elevated to species rank. It is preferred over that of f. cloiselii because it is more informative, as usually greenish leaves are a notable characteristic of herbarium specimens.

Nomenclature and distribution of this species are provided below. Selected specimens listed favor collections present at both MO and P (which should usually also be at TAN or TEF). A preliminary assessment of its conservation status was made following IUCN (2012) criteria, with GeoCAT (Bachman & Moat, 2012) used to calculate Area of Occupancy and Extent of Occurrence.

Table 1.

Morphological distinctions between Homalium involucratum (DC.) Hoffm. and H. viride (H. Perrier) Appleq., with the most useful distinguishing features of the latter in boldface.