48 AROIDEANA, Vol. 23 A reclassification of Sauromatum Schott and new species of Typhonium Schott (Araceae) w. L. A. Hetterscheid Chrysantenstraat 28 1214 BM Hilversum Netherlands P. C. Boyce Herbarium, Royal Botanic Gardens Kew, Richmond Surrey TW9 3AE United Kingdom

ABSTRACT situation in Biarum Schott). Nevertheless The generic status of Sauromatum is Wang & Li (1999) used the spathe base discussed and as a result it is merged into connation character combined with flow­ Typhonium, resulting in two new combi­ ering behaviour to indicate the generic sta­ nations. Additionally, two new species of tus of their new species Sauromatum gao­ Typhonium are described. ligongense Wang & Li.

INTRODUCTION SAUROMATUM AND 1YPHONIUM Recently a revision of the genus Ty­ Historical background phonium Schott was published (Sri boon­ Schott (1832) proposed Sauromatum as rna et ai., 1994) in which the genus Sau­ romatum was compared to Typhonium a new genus based on Arum guttatum and kept separate from it on one character Wallich (1831, non Salisbury 1796) and only, viz. the leaves developing from the Arum pedatum Link. It was classified in underground tuber only after the inflores­ his subtribe Euaroideae together with cence has developed, whereas in Typhon­ Arum L., Biarum and Typhonium. From ium species inflorescence and leaf the first two genera it was separated by (leaves) are said to develop simultaneous­ having staminodes only grouped below ly. This distinction is only partly true since the male flowers and not also above the Sauromatum (S. brevipes (HookJ.) N.E. male flowers (as in Arum and some Bia­ Br. behaves exactly like Typhonium s.s.) rum). Schott separated Sauromatum from while in Typhonium some species flower Typhonium by the connate spathe base, before the leaves develop. In Mayo et at. distant staminodes, the ovaries possessing (1997) the genera Sauromatum and Ty­ two rather than one ovule, the leaf ap­ phonium ("except Typhonium hirsutum", pearing after the inflorescence and being see p. 87, key) are kept separate solely on pedate rather than entire and the peduncle the basis of the spathe base being connate being very short. To Schott's species (S. (Sauromatum) or convolute (Typhon­ guttatum and S. pedatum) many were ium). The genus Lazarum A. Hay (Hay, added in subsequent years but mostly re­ 1992) was kept separate from Typhonium duced again to S. venosum (Dryand. ex Ai­ on the same basis and one additional, then ton) Kunth, with the exception of S. brev­ unique, character (septate spathe base) ipes. Sauromatum horsfieidii Miq. was but was later (Hay, 1997) merged with it, transferred to Typhonium by van Steenis thus underlining the relative taxonomic (1948). Recently, Wang & Li (1999) pub­ unimportance of the spathe-base closure lished Sauromatum gaoligongense, a spe­ in separating genera in this alliance (el the cies found in Yunnan. As a result three W. L. A. HETIERSCHEID, P. C. BOYCE, 2000 49 species remain in Sauromatum, viz. S. statement on p. 277 that Sauromatum venosum, S. brevipes and S. gaoligongense. horsfieldii Miq. "is not Sauromatum, in Since Schott's time the genus Typhon­ which the inflorescence always precedes ium has been expanded to a total of ca. the leaves, but a species of Typhonium, 40 species (Sriboonma et al., 1994) but nu­ . ... "It is remarkable therefore that S. brev­ merous new species from Thailand (Het­ ipes, which evidently follows the Typhon­ terscheid & Sookchaloem, in prep.) and ium flowering behaviour, is not included elsewhere await publication. at all in any discussion by the authors, al­ though it was originally described as a What are the differences? species of Typhonium (Hooker, 1893). Equally remarkable is the fact that no spe­ Brown (903) transferred Hooker's Ty­ cies of Sauromatum is involved in their phonium brevipes to Sauromatum stating phylogenetic analysis, whereas it seems that Hooker had overlooked the connate unavoidable that at least one such analysis spathe base "the distinctive character of should have been performed using Sau­ whereas the spathe base in Sauromatum", romatum as an outgroup. Instead, only Typhonium is convolute. Apparently Arum is used as an outgroup, which Brown did not consider the relative timing seems a debatable choice since at the of flowering and leaf development impor­ higher level of subfamily Aroideae this tant. Hooker (904) brought the latter evokes quite a number of unnecessary hy­ character to attention when discussing S. potheses of paralellism between Sauro­ brevipes in relation to Typhonium and matum and Typhonium. states that the simultaneous flowering and Typhonium shows quite a remarkable leaf development in this species "suggests pattern of variation in spathe and spadix the expediency of the two genera being morphology. The spadix of Sauromatum united". As a result Sauromatum and Ty­ venosum (Fig. 1) shows all the typical Ty­ phonium could at that time only be sep­ phonium characters and bears a remark­ arated by this single spathe base character. able resemblance to that of T giganteum Surprisingly no subsequent author consid­ (Fig. 2) and T hirsutum (Fig. 3). Sauro­ ered the taxonomic value of this character matum brevipes is morphologically very in both genera, although Murata & Mayo similar to T horsfieldii, as is S. gaoligon­ (991) suggested that this problem needed gense. to be looked into when Typhonium would Species of Typhonium usually flower be revised. This suggestion was brought alongside the developing leaves. In sev­ about by the finding of the remarkable Ty­ eral species however (T hirsutum, T phonium hirsutum (S.Y. Hu) Murata & brevipilosum, several undescribed species Mayo, which exhibits a partly connate from Thailand and Vietnam) the leaves are spathe base but otherwise shows all the still very immature when the plants flower. "classic" characters of Typhonium. Re­ In others [T mirabile (A. Hay) A. Hay, T cently Hay (997) merged the genus La­ praetermissum (A. Hay)] the inflorescenc­ zarum (Hay, 1992) with Typhonium, es develop before the leaves, or both thereby including yet another species with states occur in the same species (two new, a closed spathe-base. as yet undescribed species from Thailand, A revision of Typhonium has been un­ T roxburghii Schott (Fig. 4), T circinna­ dertaken recently by Sriboonma et al. tum). These observations indicate that (994) but despite Murata & Mayo's (1991) flowering before the leaves is not a suffi­ suggestion, Sauromatum was not includ­ cient character to separate genera from Ty­ ed in this revision nor extensively com­ phonium. mented upon. The authors consider that A few preliminary phylogenetic analy­ Sauromatum should remain separate from ses of a character matrix including all Ty­ Typhonium on the basis of the former's phonium and Sauromatum species (62 flowering behaviour judging from their species and 12 morphological characters) 50 AROIDEANA, Vol. 23

Fig. 1. Typhonium venosum: spadix base Fig. 2. Typhonium giganteum: spadix (spathe cut lengthwise). Photo: B. v.d. base (spathe cut lengthwise). Zwaan.

have learned from Dr. V. Soukup (USA) consistently show a group of species well­ that a chemical difference exists between nested in Typhonium. This group consists T (S.) venosum and other species of Ty­ of T (5.) venosum, T (S) brevipes , T (S.) phonium. His chemical analyses of acid gaoligongense, T hirsutum, T brevipilos­ content of seeds of Typhonium species um, T horsfieldii and T giganteum. Char­ show that T (S.) venosum curiously lacks acters used in this anaylses were: a certain acid that is present in all other 1. shape of underground parts species of Typhonium that have been 2. shape of leafblade screened. However, since this character 3. presence/ absence of indumentum seems to be an autapomorphy for T ven­ 4. closure of spathe base osum, it is uninformative in phylogenetic 5. shape of spathe limb matters at this level of comparison be­ 6. shape of appendix tween Sauromatum and Typhonium. 7. types of staminodes On the basis of what has been said 8. shape of lower staminodes above, Sauromatum should not retain its 9. morphological change from lower to generi c status but should be merged with upper staminodes (when different) Typhonium. This requires the fo ll owing 10. shape of upper staminodes in com­ nomenclatural changes: parison to lower ones 11. spathe base presence during fruiting Typhonium brevipes Hook. f. , FI. Brit. Ind. 6: 511 (1893)-Sauromatum 12. temporal development of leaf in rela­ brevipes (Hook. f.) N.E . Brown, Gard. tion to inflorescence Chron. III , 34(2): 93 (903); Hook. f. , In several personal communica tions we Bot. Mag. t. 7940 (904); Engler, W. L. A. HETIEHSCHElD, P. C. BOYCE, 2000 51

Fig. 3. Typhonium hirsutum: spadix base Fig. 4. Typhon-ium roxburghii: fl owering (spathe cut lengthwise). before leaf development.

Kew by William Malcolm in 1774 pflanzenr. rY.23F (Heft 73): 127 (BM, holot). (1920). Syntypes: Darjeeling, 7.500 ft. , Arum fugax Sa lisb., Prodr. Stirp. Clarke 26708 (KI); Jo re Pokri, 7600 ft, Chap. Allerton: 260 (1796), nom. il­ Cammie s.n. (KI). leg. Typhonium gaoligongense (Wang & H. Arum pedatum Willd., Enum. Pl, Li) Hett. & P.e. Boyce comb. nov. Suppl 54 (1814)- Sauromatumpe­ Basionym: Sauromatum gaoligon­ datum (Willd.) Schott in Schott & genseWang & H. Li. , Acta Bot. Yunn. , Endl, Melet. Bot.: 17 (1832)- Alo­ suppl 11: 61 (1999) Type: China, casia pedata (Willd.) Raf. , Fl. Tel­ Yunnan prov. , Boashan Xian, Li Heng lur. 3: 64 (1837)- S. guttatum & C. Ruckert 11309A (KUN). (Wall) Schott var. pedatum (Will d.) Typhonium venosum (Dryand. ex Aiton) Engl , Pflanzenr. ry.23F CHeft 73): Hert. & P.e. Boyce comb. nov. Bas­ 125 (1920). ionym: Arum venosum Dryand. ex Ai­ Arum pedatum Fisch. Ex Spreng. , ton, Hort. Kew. 3: 315 (1789); Will d, Syst. Yeg. 3: 769 (1826), non. Willd . Spec. 4: (1805) 497- Desmesia ven­ (1814), nom. illeg. osum (Dryand. ex Aiton) Raf. , Fl Tel­ Arum clavatum. Desf., Tabl Ecole lur. 3: 63 (1837)- Sauromatum ven­ Bot., ed.3: 385 (1829). osum (Dryand. ex Aiton) Kunth, Arum guttatum Wall, Plant. Asiat. Enum. Pl 3: 28 (1841)- S. guttatum Rar. 2: 10, t. 115 (1831)- Sauro­ (Wall.) Schott var. venosum (Ait.) matum guttatum (Wall.) Schott in Engl, pflanzenr. ry.23F (Heft 73) Schott & Endl. , Melet. Bot.: 17 (1920) 125-Type: Plant of unknown (1832), nom. illeg. based on Amm origin introduced into cultivation at guttatum Wall non Salisb. 52 ARO IDEANA, Vol. 23

Fig. 5. Typhonium brevipilosum: habit. Fig. 6. Typhonium brevipilosum: spadix base (spathe cut lengthw ise). Arum sessiliflontm Roxb ., Fl. Ind., ed .1832, 3: 507 (1 832)- Saurama­ nubicum Scho tt va r. angolense tum sessiliflorum (Roxb.) Kunth, (N .E. Br.) Engl. , Pflanzenr. IY.23F Enum. PI. 3: 28 (1841) - S. gutta­ (Heft 73): 126 (1920). tum (Wall .) Schott var. sessiliflorum j aimenostia Jernandop oana Guinea (Roxb.) Eng!. , Pfl anzenr. IV.23F & Gomez Mo r. In E. Guinea, Ensa­ (Heft 73): 125 (1920) yo Geobot. Guin. Continent. Es­ Sau romatum a byssin icum Schott, pan .: 248 (1946). Syn. Aroid.: 25 (1 856). Sauramatw n n u bicum Schott, Syn. NEW SPECIES OF TYPHON/UM Aroid.: 25 (1856). Sauram a/u m puncta tum K. Koch, Typhonium brevipilosum Hett. & Size­ Wochenschr. Gartnerei Pfl anzenk. more sp. nov. a Typhonio hirsuto 1: 263 (1858) spatha laevi nee hirsuta, fo lio perfecte Sauramatu m simlense Schott, Oes­ pedato foliolis plus quam quinque ha­ terr. Bot. Z. 8: 349 (1858)- S. gut­ benti , appendice minus robusta dif­ tatum (Wall.) Schott va r. simlense ferr o Typus: Indo nesia, Sumatera, (Scho tt) Eng!. , Pfl anzenr. IV.23F West Sumatera, near Aeksah, Hetter­ (Heft 73): 125 (1920). scheid H AR.097-T (orig. coli . Size­ Sauromatw n pulchrum Miq. , Ann. more s.n.) fl owered in cult. in Leiden Mus. Bot. Lu gduno-Batavi 1: 221 Bot. Garden, 29 November 1999 (hol­ (1 864)- 5. guttatum (Wall.) Schott otypus L, spirit eoll.). Figs. 5, 6. va r. pulchrum (Miq.) Engl. , pflan­ Tuber depressed globose, to ca. 6 em in zenr. N.23F (Heft 73): 125 (1920) diam ., to ca. 4 em high, develo ping several Sauramatum cmgolense N.E . Br.in Ol­ seasonal offsets, these broadly and shortly iver, Fl. Tro p. Afr. 8: 142 (190 1)-S. fu siform and covered with strong, nearly W. 1. A. HETTERSCHEID, P. C. BOYCE, 2000 53 woody scales. Leaf solitary or paired; pet­ verruculate, white. Stamens 1.5-2 mm iole to 50 cm long, 1 cm in diam., pale long, pale yellow; thecae fusiform. Stami­ green with or without indistinct greyish nodes straight or slightly curved, fusiform spots, surface more or less densely set conical, to ca. 1 cm long, 3 mm in diam., with short (ca. 0.5 mm) hairs; leaf-blade top sometimes laterally compressed, pur­ deeply pedatisect; leaflets elliptic to ellip­ plish, acute or subacute. tic lanceolate, to 30 cm long, to 10 cm in' Notes-this species is closely related to T. diam., acute, bright pale green, upper and from N. Thailand and Yunnan. lower surface covered with short hairs. In­ hirsutum The leaf of T. bears a re­ florescence solitary, developing alongside brevipilosum young, immature leaf (the latter maturing markable resemblance to that of T. (Sau­ except for the hairs after flowering), short pedunculate, base romatum) venosum, covered by three rather large, visible cat­ being absent in the latter and a different aphylls; cataphylls broadly triangular­ colour pattern on the petiole. A consider­ able geographic gap exists between the al­ ovate, the largest 10 cm long, 4 cm in ledged sister-species T. and diam. at the base, outside dark greyish brevipilosum brown, covered with acute, translucent, T. hirsutum. sometimes acicular hairs. Peduncle very Typbonium circinnatum Hett. & ]. short, subterraneous, ca. 0.5 cm long, ca. Mood sp. nov. ab omnibus speciebus 0.8 cm in diam., pale green, covered with ceteris generis Typhonii folio super­ hairs as on the cataphylls. Spathe elongate ficie superiore clare cinereo-virenti, elliptic, 20-23 cm long, 9-10 cm in diam., limbo spathae valde circinatim resi­ base and limb separated by a strong con­ lienti, staminodiis pergrandibus irre­ striction, base ovate, ca. 4 cm long, ca. 2.5 gulariter conicus vel cubiformis dis­ cm in diam., outside pale green, smooth, tinguibilis. Typus Vietnam, Dak Lak, inside grooved, glossy pale purple, up­ Dray Sup area, secondary forest, wards near the constriction turning pale 1990, Hetterscheid H.AR.258-T (orig. green; limb 19 cm long, oblique, acute, call. Adams s.n.) flowered in cult. in outside smooth, glossy pale green with a Leiden Bot. Garden, (holotypus L, purple flush, inside smooth, dull purplish spirit call.). Figs. 7, 8, 9. brown, near the base pale green. Spadix shorter than spathe, ca. 18 cm long; female Underground part a short tuberous rhi­ part conical, ca. 0.7 cm long, ca. 1.1 cm in zome, ca. 2 cm long, 1.5 cm in diam. Pet­ diam., flowers congested; sterile part be­ iole to 6 cm long, 0.4 cm in diam., smooth, tween female and male part 4 cm long, the pale green; lamina elliptic ovate, to 15 cm lower 0.5-1 cm with staminodes, the re­ long, to 10 cm in diam., base cordate, mar­ mainder naked, white, longitudinally gins overlapping, top acute, upper surface grooved; male part cylindrical, ca. 2 cm with impressed venation, bright greyish long, ca. 0.8 cm in diam., base and top blue-green. Inflorescence appearing be­ oblique, flowers congested; appendix very fore or simulaneous with the leaves; pe­ shortly stipitate, elongate cylindric-conical, duncle hidden in the soil, 1-2 cm long, ca. ca. 11 cm long, ca. 1 cm in diam. at the 0.3 mm in diam., pale green; spathe 10- base, top obtusish, base truncated, surface 23 cm long, base and limb separated by a shallowly, irregularly furrowed, pale red­ contriction, base broadly ovate, 1.3-2.5 cm dish brown, producing a strong, unpleas­ long, 1-1.5 cm in diam., outside glossy ant smell at female anthesis and warming pale green, inside as outside but dull, limb up. Ovaries elongate, cylindric, 3 mm lanceolate, acute, base widened to ca. 1.5 long, 1.4 mm in diam., subangulate, white cm, upper part tapering to the top, acute, with a faint purple flush near the top, uni­ very strongly circinnately recoiled over the locular, with one or two basal ovules; stig­ entire length, outside glossy bronze col­ ma sessile, large, depressed-hemispheric, oured, inside lower part dull purple, upper 1-1.5 mm diam., ca. 1 mm high, densely part silvery bronze. Spadix as long as the 54 AROIDEANA, Vol. 23

Fig. 7. Typhonium circinnatum: habit.

Fig. 8. Typhonium circinnatum: inflo­ Fig. 9. Typhonium circinnatum: spadix rescence. base (spathe cut lengthwise). W. L. A. HETTERSCHEID, 1'. C. BOYCE, 2000 55 spathe, sessile; female zone cylindric, 2-3 and John Mood (USA) for bringing mate­ mm long, ca. 3 mm in diam., consisting of rial of the two newly described species to only 2 "rows" of flowers, these congested; his attention. Both authors wish to thank sterile part 1.5-2.5 cm long, base with a Dr. Victor Soukup (USA) for sharing with few distant or more congested staminodes, us his data on the chemical composition remaining part naked, smooth; male zone of Typhonium seeds and Bert van der cylindric, ca. 3 rom long, ca. 3 mm in Zwaan for providing the photgraph of Ty­ diam., flowers congested; appendix fili­ phonium venosum. All photos by the first form, slightly attenuate at the base, top author, except where otherwise indicated. acute, ca. 8-20 cm long, 2 mm in diam., smooth, creamy yellow. Ovaries elongate, LITERATURE CITED 1.4 rom long, 0.8 rom in diam., unilocular, Brown, N. E. 1903. Sauromatum brevipes uni-ovulate, top tapering to the stigma, N.E. Br. (n. sp.). Card. ehron. 34(2); whitish green; stigma sessile, depressed, 93-94. ca. 0.2 mm high, ca. 0.5 mm in diam., ru­ Hay, A. 1992. A new Australian genus of gulose. Male flowers butterfly-shaped in Araceae, with notes on generic limits cross-section, truncate, pinkish orange, and biogeography of the Areae. j. connective not raised. Staminodes clavate, Linn. Soc., Bot. 109: 427-434. lower part thin, stalk-like, sometimes di­ ---.1997. Two new species and a new chotomous, to ca. 1 mm long, ca. 0.3 mm combination in Australian Typhonium in diam., dark purple, upper part hugely (Araceae tribe Areae). Edinburgh j. swollen, conical, angulate-ovate or cubic, Bot. 54(3); 329-336. base truncated, top truncate or subacute, Hetterscheid, W. L. A. & D. Sookchaloem 2-3 mm long, ca. 2 mm in diam., surface (in prep.). New species of Typhonium shallowly rugulose, greyish brown. Berries in Thailand and a revised key to the elliptiC, 7 mm long, 4 mm in diam., one­ Thai species. Thai Forest Bull. seeded; seeds orbicular or slightly cordate, Hooker, J. D. 1893. Aroideae in J. D. laterally flattened, 3 mm long, 2.5-3 mm Hooker, The Flora of British India. in diam., blackish. vol. 6: 490-556. Reeve & Co., London. ---. 1904. Sauromatum brevipes. Bot. Note-a most remarkable species because Mag. t. 7940. of the highly ornamental greyish blue­ Mayo, S. J., J. Bogner & P. C. Boyce. 1997. green upper leaf surface, the uniquely cir­ The Cenera of Araceae, xii + 370 pp. cinnately recoiled spathe-limb and the Royal Botanic Gardens, Kew. large staminodes. The flowering behaviour Murata, J. & s. J. Mayo, 1991. A new com­ (with or without leaves) is an intermediate bination in Typhonium (Araceae). situation between "typical" Typhonium Kew Bull. 46(1): 129-131. Cleaf and inflorescence simultaneous) and Schott, H. W., 1832. Aroideae in H. W. what is found in e.g. T. (Sauromatum) Schott & S. Endlicher, Meletemata Bo­ venosum [flowering before the leaf (see tanica: 15-22.·Vienna. above»)' Sriboonma, D., J. Murata & K. Iwatsuki, 1994. A revision of Typhonium (Ara­ ACKNOWLEDGMENTS ceae). j. Fac. Sci. Univ. Tokyo, Sect.] The first author wishes to express his Bot. 15(4); 255-313. thanks to the Keeper and staff of the Her­ Steenis, C. G. G. J. v., 1948: Typhonium barium, Royal Botanic Gardens, Kew for horsfieldii (Miq.) Steen. Bull. Jard. permision to visit and to borrow material Bot. Buitenz., III, 17: 403. essential to this revision. Thanks also to Dr Wang, Z.-L. & H. Li, 1999. Sauromatum Alan Radcliffe-Smith, Kew, for the Latin gaoligongense (Araceae), a new spe­ translation of the diagnoses. Furthermore cies from Gaoligong mountains. Acta he wishes to thank Mary Sizemore (USA) Bot. Yunnan., suppl. 11: 61-64.