On the taxonomic identity of Arisaema pierreanum Engl . ( Araceae ) in Vietnam

Trang 52 On the taxonomic identity of Arisaema pierreanum Engl. (Araceae) in Vietnam  Van Hong Thien Industrial University of Ho Chi Minh City Institute of Tropical Biology, Vietnam Academy of Science and Technology  Nguyen Phi Nga University of Science, VNU-HCM  Luu Hong Truong Southern Institute of Ecology, Vietnam Academy of Science and Technology (Received on June 5th 2015, accepted on 26th September 2016 ) ABSTRACT The taxonomic identity of the rare Arisaema pierreanum Engl. has been disagreed by botanists, mostly due to the lack of diagnostic characteristics of its female inflorescence. In this paper, a detailed description of the species is provided based on our new living collections from the type location, Ba Den Mountain, by which we reconfirm its specific status and eliminate a former question about its possible conspecific form of A. roxburghii Kunth and A. harmandii Engl.


INTRODUCTION
Arisaema pierreanum Engl.was first described in 1920 by Engler based on Pierre's collection from Ba Den Mountain, southern Vietnam in 1869 [2].No more collections of this species were known until a recent report by Nguyen et al. (2014), who revisited the mountain to collect the plant in 2013 [9].During this duration, the taxonomic status of the species was disagreed by other authors.Gagnepain (1942) [4], Murata (1984) [7] accepted it as a good species.However, in their recent revision of the genus, Gusman and Gusman (2006) reduced it to be a synonym of A. roxburghii Kunth because: (1) Nicolson (1999) determined that A. harmandii Engl.was synonym of A. roxburghii, and (2) they supposed that the types of A. pierreanum and A. harmandii were collected from the same place, i.e.Ba Den Mountain, NW of Ho Chi Minh City in southern Vietnam, and thus both taxa were conspecific [6].Nguyen et al. (2014) disagreed with Gusman and Gusman (2006) by accepting A. pierreanum as a good species based on its male spadix with long cylindrical, light conical fertile portion and stamens not in groups and densely arranged [9].However, only the character "stamens not in groups" appears to be of diagnostic importance to distinguish A. pierreanum from A. roxburghii as the others vary in the later species as stated in Gusman & Gusman (2006), especially in its populations observed in Malaysia.A clear delimitation of A. pierreanum, therefore, needs supplemental information, especially on the female inflorescences that have been missed in both collections by Pierre in 1869 and Nguyen et al. in 2013.During our re-survey of the Araceae in southern Vietnam, we successfully sampled both living male and female inflorescences of A. pierreanum from the type location, Ba Den Mountain, in June 2015 (Thien93, SGN, PHH).
Our examination of the collected living materials indicates its distinct characteristics and therefore we re-describe it here with the updated information based on our collection.

Methods
Plant specimens were collected and processed following the Kew Botanic Garden protocols (Bridson & Forman, 1999).

Discussion
All collections of Arisaema pierreanum obviously support its placement in the section Fimbriata [6] (recently recommended by Murata (2013) to be moved to the sect.Attenuata [8]) as this species has: deciduous habit, sub-globose, tuberous, subterranean stem; rooting around the central shoot only; 1-2 bilaterally symmetrical leaves; trifoliolate leaf blade; and glabrous, subulate and slender spadix appendage.In this section, A. barbatum, A. maxwellii, A. condaoense, A. pachystachyum, A. roxburghii (syn. A. harmandii) and A. sizemoreae were reported to have yellow spathes like that of A. pierreanum [6].However, unlike A. pierreanum, A. barbatum and A. pachystachyum have yellow green rather than true yellow spathes and neuters or projections on the spadix appendix.A. sizemoreae has very yellow spathe but it is different from A. pierreanum in having spadix appendix with a brush of yellow-green upright fleshy bristles.In fact, A. pierreanum appears to be the most similar to A. maxwellii and A. condaoense, but it is distinguishable from the two later in having a combination of always auriculate spathe limb, long and glabrous male and female spadix appendices, non-grouping stamens and yellow synandria.These characters also make A. pierreanum different from A. roxburghii (syn. A. harmandii).Furthermore, although the limb of the later is sometimes auriculate, it is actually translucent -a character which is not seen in A. pierreanum.Based on the descriptions twice by Engler (1898Engler ( & 1920) [1, 2]) [1, 2], especially the figure 29 in page 164 (Engler, 1920), and examination of the specimens mentioned above and our collections, we do not doubt that A. harmandii is a synonym of A. roxburghii [6] and thus it is not a synonym of A. pierreanum.In fact, we have collected A. roxburghii in the coastal Kien Luong Karst Hills (Kien Giang Province) and Cambodia-bordered That Son Mountains (An Giang Province), both in the Mekong Delta; these collections match Harmand's (1875Harmand's ( -1877) ) in all characters and in locality-i.e. the "mouths of the Mekong River" [1,2].Meanwhile, Ba Den Mountain, where A. pierreanum has been collected by Pierre in 1869, Nguyen et al. in 2013 andby us in 2015, is a famous religion and tourist location in Tay Ninh Province.It is located in the southeastern Vietnam but isolated from the Mekong Delta.These facts do not support the idea that the types of A. harmandii and A. pierreanum were collected from the same location [6].

CONCLUSION
With updated information on its both male and female inflorescences collected from the type location, A. pierreanum is obviously distinct from its close species including A. roxburghii, and therefore it should be accepted as a good species.