International Aroid Society
Dracontium Pollen

The pollen grains of Dracontium are exclusively monosulcate. The monosulcate aperture type has been generally regarded as the most primitive for angiosperms as well as monocots (Muller, 1970; Walker & Doyle, 1975; Zavada, 1983) and Araceae in particular (Grayum, 1992). Monosulcate pollen may be described as heteropolar, boat-shaped and bilaterally symmetrical, and therefore, it may be assumed these conditions, which are all intimately correlated with the monosulcate aperture type (Walker & Doyle, 1975), represent the most primitive stages of polarity, pollen shape and pollen grain symmetry in Araceae (Grayum, 1992, p. 123). Grayum (1992) also suggested that boat-shaped, elliptic grains are more primitive than boat-shaped, elongated ones; both types are found in Dracontium.

Dracontium pollen grains, as in all other aroid genera except Xanthosoma and Chlorospatha (Grayum, 1992), are shed in monads (Fig. 74- 82, Table 3). Monads are considered basic for angiosperms as a whole.

The size of Dracontium pollen grains ranges from 17.9 x 31.8 (D. amazonense) to 31.3 x 48 um (D. changuango). Grayum (1992, p. 128), using measurements of only the longest axis, suggested that the primitive pollen size of Araceae lies in the range of 22-34 um and there is a pervasive trend in the direction of larger pollen. This implies that Dracontium has both primitive and derived pollen types. However, pollen size may vary according to the quality of mineral nutrition, local climate, and other environmental factors (Muller, 1979). Different pollen sizes were found in different populations of the same species (D. spruceanum).

Nonapertural exine sculpturing of Dracontium pollen may be reticulate (D. changuango, Fig. 83), subreticulate (D. amazonense, Fig. 84), foveolate (D. spruceanum, Fig. 85), punctate (D. angustispathum, Fig. 86). Reticulate exine is widely regarded as primitive for monocots (Walker & Doyle, 1975), since the earliest known fossil pollen, Liliacidites, has this type of exine sculpturing (Doyle, 1973; Wolfe et al. 1975). However, Thanikaimoni (1970) and Harley (1990) suggested that the psilate exine type (similar to punctate exine type) is primitive in palms. In Araceae, Grayum (1992, p. 130) also characterized punctate exine as nearly as primitive as psilate type; he observed that foveolate and reticulate exine types were equally common and could not decide which of the two "subtypes" was more primitive. In any case, the exine type of Dracontium is among the most primitive Araceae.

In Dracontium, a species may exhibit more than one exine type or subtype. This indicates that the evolution of exine characters has occured in the genus. A cladistic analysis involving all other characters will be required to establish whether thee is a transformational sequence from punctate to foveolate then reticulate exine in Dracontium. Then, the evolutionary scheme of exine sculpturing in Araceae can be reconsidered. Grayum (1992, 130-132, fig. 158) described a reduction in size of the brochi toward the 'ends' in Dracontium prancei, but the significance of this is unknown.

Exine stratification has rarely been studied in Araceae (Grayum, 1992) and was unknown in Dracontium until the present study. Pollen exine stratification from three populations of D. spruceanum have been investigated using TEM. The ektexine of Dracontium pollen consists of a tectum and an infratectal granular layer (Fig. 87). This granular type of infrastructure, much like the granular wall structure found in the Nymphaeaceae, Magnoliaceae, and Annonaceae (Zavada, 1983), is found commonly in gymnosperms (Xi & Wang, 1989). Endexine is very rare in monocots, and when present is different from the endexine observed in dicots (Zavada, 1983). It has been sought in only a few genera of Araceae (Grayum, 1992). The endexine of Dracontium pollen is lamellate, with 4-6 layers. As far as I am aware, lamellated endexine has never been reported for angiosperms. It exists in most gymnosperms except some species of Podocarpus (Xi & Wang, 1989). I suggest, therefore, that Dracontium is among the most primitive of aroids.