Palms, 43(4),1999, pp. 194-199 Satakentia Revisited

• 1 JEAN-CHRISTOPHE PINTAUD Laboratoire de Botanique, [RD (ex ORSTOM), Nouméa, New Caledonia

2 HJROAKT SETOGUCH1 Makino Herbarium, Faculty ofScience, Tokyo Metropolitan University, lapan

Japanese palms are not numerous, essentially omote, in the southern Ryukyu Islands, close to eonfined to the southernmost arehipelagos of Taiwan. Three decades later, a bronze plate hon­ Bonin, Vol cano, and Ryukyu, but are neverthe­ ors the work of H. E. Moore, Jr., at the entrance less quite interesting and to sorne extent, still of a paved trailleading to the largest stand of Sa­ not weIl known. The most widespread of them is takentia at Yonehara on Ishigaki island. The site Livistona chinensis, represented by different va­ is declared as a National Treasure and is the is­ rieties in eaeh island group. However, apart from land's major tourist attraction. Each day, hun­ Livistona chinensis var. boninensis which is dreds ofJapanese tourists, brought by big buses, widespread and very abundant in the Bonin Is­ come to admire the famous palm and learn from lands and distinctive in the large, obpyriform, the preserve's guides, the story of Satakentia. glossy green fruits, the other varieties are unde­ The palm is also increasingly planted along scribed (a putative one in the Volcano Islands) or streets, roads, and in gardens aB around the is­ of doubtful if not unknown origin (Moore and land, of which it has become the emblematie Fosberg 1956), Since Livistona chinensis is one tree. of the most widely cultivated palms, it should Satakentia is typicaBy a Pacifie element, cu­ deserve further investigation. Another distinc­ riously endemic to these small islands that oth­ tive and elegant palm from this region is Clino­ erwise bear an essentially Asian flora, including stigma savoryanum, endemic to the humid up­ Arenga engleri, as far as palms are concerned. lands of the Bonin Islands. Clinostigma, one of Since the study of Moore, Satakentia was be­ the most widespread Pacific palm genera, reach­ lieved to be cIosely related to and scarcely dis­ es its northern limit there, weil north of the Tro­ tinct from Clinostigma (Moore 1969, Uhl and pic of Cancer. The most striking Japanese palm, Dransfield 1987). However, during the course of however, remains undoubtedly Satakentia liuki­ field studies on both genera in February, 1997, it uensis. This massive palm, described as a dis­ appeared to us that this was highly questionable, tinct and monotypic genus by Moore in 1969, is and that, on the contrary, these two genera were confined to two nearhy islands, Ishigaki and Iri- probably quite distant within the large and di­ verse subtribe Iguanurinae, tu which they be­ long. The palm that we collected on Ishigaki is­

1 PresenL address: Lahoratoire d'Ecologie Terrestre, 1:-3. land (Pintaud & Setoguchi 447, K) had an Ayenue du Colon"l Hoche. H.P. 440:3, 31405 Toulouse inflorescence enclosed in its bracts, showed a Cedex. France. flattened and bicarinate prophyll mueh shorter 1 Prest'nL addrt>ss: Facull.\- of InLegratt'd Human Studies. I)pparlnwnl of Nalural En" Îronrnen\ Sciences, Kyoto Univer­ than the protruding first peduncular bract (Fig. sity. "- yoto ü06--850 J. Japan. 1 A,B), this one complete. rostrate and including

--> 1: Four portrait pictures 1 A. H. Wel1 developed crownshaft of Satakentia liukiuen.

a second, also complete and similar peduncular bract, and a third and even fourth, incomplete but prominent additional ones (Fig. 2 B). This condition was different from lhat described by Moore (1969) who stated that the inflorescence • of Salakenlia had a complete, l'ostrate prophyll enclosing a similar first peduncular bract, like Clinostigma. He noted however, the large addi­ tional, incomplete, peduncular bracts. The rea­ son for this discrepancy is obvious when looking at Moore's picture showing the inflorescence and its bracts (Fig. 2 A). Compared with our Figure 2 B, it is clear that the prophyll is missing in Moore's collection (Moore el al. 9382). In fact, the inflorescence bud collected by Moore was at an advanced stage of development, arter the falling of the prophyll, since the first peduncular bract had already split abaxially, exposing the second one, on the tree collected by Moore (Fig. 1 C). What Moore called the pl"Ophyll corre­ sponds evidently with the first peduncular bract of our collection and therefore Moore was misled in interpreting the inflorescence structure. Moreover, the inflorescence from Moore's col­ lection deposited at BH shows three annular scars, the first corresponding with the over­ looked prophyl1 and the fol1owi ng ones to the two complete peduncular bracts. Another collection 2 A. Dissecled inOorescence (Moore et al. 9382) showing complele first and second peduncular bracls and incomplele made on the mature tree of the Tsukuba Botani­ but ver)' prominent Ihird one, the prophyll is missing (Fig. 3, cal Garden's glasshouse (Pinlaud & Higuchi lower, in Moore (969). 2 B. Dissecled innorescence (Pinlaud 448, K) showed the same inflorescence structure & Higuchi 448) showing from top to bottom prophyll, first as in our first collection on lshigaki, proving that and second complete peduncular braets and the third one, mu~h redu~ed. it is constant and diagnostic. Only the size of the incomplete additional peduncular bracts is vari­ able, as noted by Moore (Fig. 2 A,B). This inflo­ rescence structure is very rare in Iguanurinae, (Dowe and Uhl 1989). Therefore, the similarity since it is found in only one other palm, Car­ in inflorescence structure between Salakenlia poxylon macrospermum (Dowe and Uhl 1989, and Carpoxylon should be regarded as a conver­ Dowe et al. 1997), endemic to Vanuatu. It is also gence rather than as a sign of a close affinity. very different from that of Clinosligma which re­ ally has a prophyll enclosing a si.milar first pe­ Ecology of Satakentia duncular bract, followed by 2-3 inconspicuous, Moore said about the Yonehara grove in Ishi­ ridge-like bracts. There are a number of other gaki that "these lrees appear la be essentially differences between the two genera, which are the same age and have probably grown from summarized in Table 1. 50, with this new infor­ seedlings left when mature palms were cut for mation, what are the affinities of Salakenlia and the cabbage or edible bud during World War Il'' Cfinosligma? A cladistic analysis of the whole while the trees he saw on Iriomote "were larger Iguanurinae based on morphological characters than those at Yonehara and very impressive, (Pintaud, 1999) suggests that Satakenlia is a being in an undislurbed habitat away from evi­ rather unspecialized and isolated member of the dence of human activity." In faet, Moore's state­ subtribe while Clinosligma belongs to a more ment abou t the Yonehara stand is incorrect. derived group including Carpoxylon which has There is evidence that this grove is also natural already been recognizec! as close to Clinosligma and not the result ofhuman disturbance. Moore's l'INTALIl ANIl SETO(;UCHI, SATAKENTIA HEVISITI<:1l 197

Table 1. Synopsis oj'difl"erences belwecn Satakentia and CI i nosligma.

Sfllalœnl;a Clillfl:itigma

slt'rHler, (~:\p()st'd 10 ('"rm a thi('k, [ohu:-;l, not exposed or varioll:'ily \,pry promint'Ilt ro()I-I)()s~ ("':\p()~p(L oflen forrning stilh LipJlp 01 ,llPalh very prominent, chartaeeou:-; in('olbpicuous

ILlml'nla Oll mic!rilJ Of] "haxial fTH-'difixed, Iwislccl, hrown IIlt'difixed. fiaI. elo,;ely app...''''''[ 10 ,-..IJrrill'l' or l'illrl,ll' 1~1(' lamina, rnemhran()us-lran~llIs('f'nl Pt-'dunl·lt-' of inf1on':O;(TrI(,f' ~It-'nd(~r, f~l()ngalt' \('n .,dlor!. inflakd IndllrTh'lll or inf!oft-':--.n'tH'e \"f~r~ shortlj hui denspl) hrown i nflort,,,:..;('pn('t-' glahrous -1 epi dolt'-torrlen 1OSt' l'rophyll mllr:h ,horler 1han lirsl sirnilar 10 and erH'losing fîrsl p(~dun('ular pr:duncular hral'! [lracl Fir,;1 pedunr:u]ar hrar:1 thick. wood, L!Jin. ht-'rbaceous Sp,'ond pedunr:ular hral'! complete. rostrale. endosÎng tht-' very "mail. mueh reduced, orten 10 a ]01'1

Înf1ore .... c.. 'If"t' in IJud riclge Third and forlh peduneular hn\l'I,; ill('ornpidf' but vt'ry prominenl, inc()nspi('uous or lacking (·harté.lceous Kar:hillap ratlwr stoul very slender Triad hrar:l,; founded acuminale Second bracleole sepal-like nol,;epal-likp ThinI hra"lpole rounded deltate Staminale bud rwarly syrnmeLrical markedlyasymmelrieal Staminale sepals rounded lanceolaLe Slaminaie pela]s boat-shaped deltoid, nearly Hal Pislillode as ]onji as slamens, swollen. eapilate very shorL, hroadly conical, trifid Slaminodes :3 (5-) 6 Perianth residue on fruit prorninenl ,mail

picture (Fig. 3 B), shows a part of the Yonehara ticed. The factors initiating these major periods grove with numerous Satakentia emerging above of regeneration are not certain. It is possible that lhe forest canopy, just as it is today (Fig. 3 A). the establishment follows large tree fall gaps or Palms which germinated during Word War II are landslides caused by the frequent typhoons quite unlikely to reach such a size after 20 years which affect the island. When an open area ap­ and the fact that the trees present today are not pears, the Satakentia palms might be able to col­ taller indicates that they were already mature onize it rapidly due to their efficient reproduc­ when the photograph was taken in 1966, and that tive system, In normal conditions. there is the population is stable. Moreover, Satakenlia insufficient regeneration to allow the future re­ presents a gregarious syndrome similar to that placement of the existing trees, the numerous described by Pintaud and Hodel (1998) for Ken­ seedlings that germinate each year are killed by liopsis species growing in natural conditions in the dead leaves of the adults within the stands New Caledonia: a dense population of mature and are also usually unable to establish them­ trees of similar size (and probably age) with liule selves in the dark understory of the adjacent rain juvenile establishment beneath due to continu­ forest. ous fall of dead leaves and synchronous phenolo­ gy of all individuals with production of massive Conservation Status amounts of small fruits with readily germinating The extent of Satakenlia in Ishigaki is very seeds. The gregarious behaviour of Satakentia is Iimited. It occurs only at the Yonehara site, with also evident in clearly undisturbed conditions a main area of gregarious patches and scattered, on lriomote, as illustrated by Moore (Fig. 3 C). isolated trees not far away, in rain forest on sand­ The establishment of each stand of Salakentia stone hill slopes. Remnant small patches of has occurred over a very lirnited time span, and forests on the fiat lowlands including sorne Sa­ lhere are stands of different ages as Moore no- takentia and isolated rernnant trees indicate lhat 198 PALMS [VOl.. 43

3 A. The Sal.aken.tia grove al Yonehara as il is Ioda)'. 3 B. The same grove pholographed by Moore in 1966 (Fil'. 2 in Moore 1969). .3 C. A gregariOlls Sland of Salakenliu on 1riomole, pholographed b)' Moore (Fig. 9 in Moore J969). 1999] PINTAUD AND SETOGUCHI: SATAKENTIA REVISITED 199 the species was formerly more widespread in Tokyo Metropolitan University for facilities to these areas which are nowadays mostly convert­ work in Japan for JCP, to Natalie Uhl at the L. H. ed to sugar cane fields or otherwise very dis­ Bailey Hortorium, Comell University, where turbed and devoid of natural vegetation. The Moore's collections were studied, to John Drans­ forests on the island are only conserved on hill field for suggesting the title and to the redaction slopes which are not suitable for cultivation. The of Principes for reprinting the pictures from remaining population at Yonehara is protected Moore's article on Satakentia. as a National Treasure but the boundary of the preserve includes only the main grove and LITERATURE CITED should be extended in order to include aIl the forested area where Satakentia is likely to occur DOWE,1. L., J. BENZIE AND E. BALLMENT. 1997. Ecology and and regenerate. This is particularly important genetics of Carpoxylon macrospermum H. Wendl. & for a gregarious species showing probably the Drude (Arecaceae), an endangered palm from Vanuatu. Biological Conservation 79: 205-216. most active regeneration after unpredictable DOWE, J. L. AND N. W. UHL. 1989. Carpoxylon macrosper­ natural events which can affect any part of its ac­ mum. Principes 33(2): 68-73. tuaI or potential habitat. mCN. 1994. mCN Red List Categories. Gland, Switzerland. On Iriomote, Satakentia occurs in isolated, MOORE, H. E., JR. 1969. Satakentia. A new genus of Pal­ mae-Arecoideae. Principes 13(1): 3-12. mostly gregarious populations in undisturbed ---, AND F. R. FOSBERG. 1956. The palms of Micronesia forest, the island biota being much less alterated and the Bonin Islands. Gentes Herbarium 8(6): than in Ishigaki, and declared as a National 423-477. Park. In conclusion, we can assess the conserva­ PINTAUD, 1.- C. 1999. Phylogénie Biogéographie et Ecologie tion status of this palm as Low Risk but Conser­ des Palmiers en Nouvelle Calédonie. PhD. Thesis, Uni­ versity of Toulouse. 376 pp. vation Dependant (LRcd) according to the new --- AND D. R. HODEL. 1998. A revision of Kentiopsis, a mCN Red List categories (1994). genus endemic to New Caledonia. Principes 42(1): 32-33,41-53. Acknowledgments UHL, N. W. AND J. DRANSFIELD. 1987. Genera Palmarum, a classification of palms based on the work of H. E. We are grateful to Dr. Masanobu Higuchi for Moore, Jr. L. H. Bailey hortorium and The International assistance in Tsukuba Botanical Garden, to the Palm Society. Allen Press, Lawrence, Kansas, USA.