Pseudodracontium N.E.Br.

The genus Pseudodracontium was established by N.E. Brown in 1882. He introduced one new species, P. anomalum N.E. Br., and transferred Amorphophallus lacourii Lind. & Andre´ to Pseudodracontium. The first mentioned species was chosen as the lectotype species by Nicolson (1967). Subsequent years saw the addition of several new names in Pseudodracontium. These were dealt with taxonomically in a revision of the genus by Serebryanyi (1995) which left a total of 6 accepted species, 4 of which were newly introduced in that revision. After the revision only one new name was published in relation to Pseudodracontium, namely Amorphophallus glaucophyllus Hett. & Serebryanyi (Hetterscheid, 2006): Because it was already then strongly suspected by the authors that Pseudodracontium would soon be transferred to Amorphophallus, this new species, clearly a member of "Pseudodracontium", was published as a species of Amorphophallus. Pseudodracontium has always been thought to be very closely related to Amorphophallus and both genera solely make up the tribe Thomsonieae. Thomsonieae is now regarded to be the sister group of Caladieae (incl. Zomicarpeae; Cabrera et al., 2008; Cusimano et al., 2011).

PSEUDODRACONTIUM AND AMORPHOPHALLUS
Hetterscheid(1994) already suggested that Pseudodracontium may well be a part of Amorphophallus instead of its sister genus. This was based on morphological observations which resulted in a small suite of apomorphies for Amorphophallus + Pseudodracontium, whereas without Pseudodracontium there was no single character to be found for a monophyletic Amorphophallus (Hetterscheid, 1994; Serebryanyi, 1995). The monophyly of Pseudodracontium itself was never disputed and strongly supported by a suite of morphological (near-) autapomorphies (Serebryanyi, 1995; Van der Ham & Van Heuven, 2001; Van der Ham et al., 1998, 2000, 2005). The most obvious ones are: unilocular thecae, slender filaments (fully free or more or less connate to form a slender column,Fig. 1b),appendix with recognizable staminodial structure (Fig. 1c),pollen with "polar caps" (smooth or irregularly structured polar areas, fig. 1d).
We have always been puzzled by the retention of seemingly juvenile leaf architecture in many adult plants of Pseudodracontium taxa (fig 1a). In almost all taxa the central main segment of the leaf is considerably less complex than the lateral ones, and there is a general morphocline from more complex to rather simply divided leaf laminas (or the reverse). In some specimens the leaf remains fully juvenile with only three, very large, leaflets in adult plants, as is seen in seedlings, also in many other Amorphophallus species. In nearly all seedlings ever observed by us (several hundred) the first seedling leaf lacks the central segment (leaflet) entirely. Thepolarcaps of the pollen grains appear to be a retained non-adult phase of pollen formation. Looking at the inflorescence, we observe a few characters thatmay be seen as "plesiomorphic" for Amorphophallus at large, like the appendix wall with fully developed staminodes and the long, free or partly fused filaments. We hypothesize that the morphogenic pathways of several phenotypic characters in Pseudodracontium is deregulated. It may well be that the chaotic molecular and morphological patterns display an evolutionary phase Pseudodracontium is going through whereby (past and ongoing) hybridisation (see also paragraph 3.1) may well have been the onset of the deregulation. Heterochrony seems to bepart of this deregulation, leading to adult plants with partly juvenile morphologies.
Pseudodracontium as a group within Amorphophallus is considered to be closely related to A. longituberosus (Engl.) Engl. & Gehrm. and its immediate allied species A. albispathus Hett., A. coudercii (Bogn.) Bogn. and A. tenuispadix Hett. This group turned up again and again in morphology-based attempts to unravel the phylogeny of Amorphophallus. In these analyses Pseudodracontium species were always associated with the aforementioned group as a sister group. Even in chemical analyses of the scent of species in this alliance (Kite & Hetterscheid, 1997; Kite et al., 1998), traces of a compound (4-methoxyphenetyl alcohol, or "anise oil") unique to it, were found in P. lacourii (Lind. & Andre´) N.E. Br. and P. fallax Serebryanyi. This seemed to strengthen the suggested relationship. The trouble with all full-morphology-based phylogeny reconstructions of Amorphophallus was that no statistically relevant support for this grouping could be found, not even applying the Implied Weighing procedure of the TNT program of Goloboff et al. (2004; Hovenkamp & Hetterscheid, 2008). The species group of A. longituberosus reappeared in all analyses but never associated with Pseudodracontium.
Molecular analyses of Amorphophallus + Pseudodracontium published to date (Grob et al., 2002, 2004; Sedayu et al., 2010) strongly support the sister-group relationship of Pseudodracontium to the A. longituberosus group within Amorphophallus and thus confirms the earlier morphology-based hypotheses. More recent molecular analysis (Randomized Axelerated Maximum Likelihood [RAxML]) by the second author using the markers ITS1, FLint2, rbcL and matK and applied to ca. 130 species of Amorphophallus (incl. Pseudodracontium) again strongly supports the aforementioned grouping of Pseudodracontium in Amorphophallus (Claudel et al., in prep., Hetterscheid & Claudel, in prep.) but expands it with a small strongly supported clade of 3 (possibly only 2) species (A. saraburiensis Gagn., A. scutatus Hett. & T.C. Chapman and A. tenuistylis Hett.). This group of the Amorphophallus species mentioned + Pseudodracontium is part of a larger and strongly supported clade of the "Continental Asia-II"-clade of Sedayu et al. (2010).
In the light of all evidence discussed above it is here decided to reduce the genus Pseudodracontium to the synonymy of Amorphophallus. The necessary new combinations and one new name are presented below. The tribe Thomsonieae will become monotypic.