Morphology of the Lecythis poiteaui clade.
Scott Alan Mori, Caroline Carollo, Nathan P. Smith
Lecythis Loefl.
From Huang et al. (in review)
Lecythis poiteaui clade (72% BS; Figs. 1A, 5)—This clade is found from central to eastern Amazonia and disjunct in the coastal forests of eastern Brazil (Huang, 2010). It is sister to the L. ollaria clade but the five species recovered in it lack the single coil (Fig. 5C, 5D, 5F, 5K; character 26) of that clade. None of the coded morphological characters provide synapomorphies for the L. poiteaui clade, but members of this clade possess a long, oblique or geniculate (Fig. 5G) instead of an erect style (Fig. 4G); roundish (Figs. 5J, 5P) instead of longer than broad seeds (e.g., Fig. 8D, 8E); dendritic seed venation (Fig. 8M–P); and absent (Fig. 5J) or vestigial (Fig. 5P) versus a more developed (Figs. 3G, 4F, 9A–C) basal aril (characters 47, 48). The species of this clade generally have the androecial hood closed (= closed androecium, character 33, Figs. 5D, 5F, 5K) and petals tightly pressed against the androecium, presumably to stop entry into the flowers by non-pollinators and yellow at the apex of the opening into the androecial hood, a color that usually directs bees to a pollinator reward.
The exceptions are Lecythis barnebyi S. A. Mori (Fig. 5A) and L. poiteaui. These two species are nocturnal, and bats have been observed taking nectar from their flowers and, thus, are presumed to be the pollinators (Mori & Prance, 1990). These two species also possess similar cuticular papillae on the abaxial leaf blade surface (see Fig. 96 in Mori & Prance, 1990; character 5), a massive number of stamens (character 34), open androecia (Figs. 6A, 6B; character 33), petals not pressed against the androecium (Fig. 5A) and the presence of at least some anthers (or possibly antherodes) on the hood (character 32). Mori (1990b) placed L. brancoensis (R. Knuth) S. A. Mori, along with the two other bat-pollinated species, in Lecythis sect. Poiteaui, and this relationship was supported by Huang et al. (2011). In contrast, this study places L. brancoensis in the Lecythis chartacea clade. Thus, if L. brancoensis is found to be bat-pollinated as suggested by Mori (1990b), and our phylogeny correctly reflects evolutionary relationships, bat pollination in the New World Lecythidaceae has evolved twice.