Nectandra globosa Mez, Jahrb. Bot. Gart. Berlin 5: 415. 1889. (non Laurus globosa Aublet).
Nectandra globosa var. barbeyana Mez, .Jahrb. Bot. Gart. Berlin 5: 416. 1889 (syntype Blanchet 1863, isotype, fl. NY.)
Nectandra pallida Miquel, Stirp. Surinam. 196, pl. 59A. 1736. (holotype, Kappler 1827, fl. U.)
Nectandra vaga Meissner in DC. Prodr. 15>: 153. 1864. (syntypes, SchomburgTc 896, Saqot 811, As.NY.)
Nectandra leucantha Miquel, Stirp. Surinam. 196. pl. 58. 1851.
Type. Arboream floribus albis fructque onustam, ad ripam fluminibus prope Bergendaal, Suriname, m. Majo, Focke (holotype, fl., fr.).
Vernacular Names. Canau (Cardona) ; Canau-yek, Tabau-yek (Steyermark).
This widespread and variable species is constant in bearing branchlets that for the most part maintain minute, appressed, fulvous pubescence through the flowering stage, becoming sparingly pubescent in fruit; and inflorescences with large flowers which haA'e triangular, usually papillose anthers with a fleshy connective, the cells in an arclike formation in the lower portion. The inflorescences are usually shorter than the leaves and may be few- to many-flowered, often more or less corymbose. The oblong, ovoid, or subglobose ovary is usually equal in length to the style, and the fruit is ellipsoid. Few fruiting specimens are available.
The material which I have seen from British Guiana, representing the typical N. pisi Miquel vegetatively, is consistent in having fulvous pubescence on branchlets and inflorescences. The flowers are densely pubescent without, the stamens with densely papillose, fleshy anthers; those of series I and II are broadly or triangularly ovate with rounded or slightly obtusely subacuminate tips; those of series III are short and obovate or obovate-truncate; and the ovary is usually subglobose, occasionally ovoid, and usually equals the style. These characteristics accord very well with Schomburgk 896 and Sagot 811, syntypes of N. vaga /? vulgaris and y major of Meissner, which I have seen. Blanchet 1544bis, and 1863, holotype of N. globosa var. barbeyana Mez; both represent N. pisi.
I have not examined Miquel's type of N. pisi, based on Focke's specimen, but the illustration is not in conflict with the inclusion of the numbers cited by Mez, excluding his reference to Aublet's Laurus globosa for which no specimen is mentioned.
The Macarena specimens cited show leaves with 8 or 9 pairs of veins, glabrescent, reddish-brown, angular branchlets, and slightly smaller flowers with the structure typical of the species. Material from Vaupes (Cuatrecasas 7458) and some numbers from northern Colombia not here cited, also show leaves with this number of veins, but the stamens of series I and II consistently have long triangular-ovate anthers, while those of series III are broader and less attenuate at the apex.
Nectandra leucantha of Bentham, cited by Mez under N. globosa, based on Spruce 3677, which is at hand, differs only slightly from the Guiana materials examined. The leaves tend to be more rounded at the base with much more regular venation, the latter calling to mind N. caucana of Colombia, a species with yellowish-pubescent branchlets and leaves with 12 or more pairs of veins. The inflorescences of Spruce 3677 and the collections of N. pisi from Amazonas, Venezuela, above cited, are usually very full-flowered and often more corymbose, but a similar internal structure of the flowers prevails. These variations appear, from the rather sparse material available, to be more or less consistent within the sandstone area of this part of Venezuela.
Nectandra ambigua Meissner (in D C . Prodr. 15: 158. 1864), based on Schomburgk 29 (isotype, fl. US), is very near the specimens representing N. pisi, except for the opposite, 5 or 6 nerved leaves; the anthers, however, are long-ovate or triangular-ovate, and the ovary, if globose, nearly equals the style.
Nectandra lucida Nees, the type of which was collected by Martius near the Japura River in upper Amazonia, I have not seen, but Schomburgk 568, cited by Mez, is at hand. This is similar to N. ambigua in flower as well as leaf-shape and venation, but the branchlets and inflorescences tend to be more glabrescent, and the leaves are alternate instead of opposite. Several collections by Krukoff from the Rio Acre and Rio Embira regions of Brazil have been referred to N. lucida, but the veins of these numbers are consistently 8 or 9 pairs.
The Nectandra globosa based on Aublet's Laurus globosa (PI. Guiane 1: 364. 1775) supposedly has globose fruit. I have seen no specimens belonging to the pisi-coimplex, collected in the area we are considering, with this type of fruit. It is of interest to note that except for minutely pubescent red-brown, angular branchlets, the collection of Wurdack d- Monachino 39722 (fl. NY) from Serrania Imataca, Territorio Delta Amacuro, is very like the description of N. pallida Miquel, the type of which I have not seen.
Wurdack & Guppy 158 (fl. NY), from Hato La Vergarena, varies from the Bolivar and Guiana collections in having more regularly lanceolate-elliptic leaves, and fewer, stricter, racemose inflorescences not branching at the tip as do the other specimens from this general region of Venezuela.
From the above comments, it is obvious that the species is not a stable one, to say the least, and undoubtedly needs much more critical study. It is obvious, also, that a definitive treatment must be delayed until such time as more abundant fruiting material is available from all of the regions now represented in our herbaria by only flowering material.