Allozymic, Morphological, Phenological, Linguistic, Plant Use, and Nutritional Data on Wild and Cultivated Collections of Luffa aegyptiaca

Transcription

1 Allozymic, Morphological, Phenological, Linguisic, Plan Use, and Nuriional Daa on Wild and Culivaed Collecions of Luffa aegypiaca Mill. (Cucurbiaceae) from Nepal, Souhern China, and Norhern Laos Auhor(s): Kendrick L. Marr, Yong-Mei Xia, Nirmal K. Bhaarai Source: Economic Boany, 59(2): Published By: The New York Boanical Garden DOI: hp://dx.doi.org/ / (2005)059[0137:amplpu]2.0.co;2 URL: hp:// %5D2.0.CO%3B2 BioOne ( is a nonprofi, online aggregaion of core research in he biological, ecological, and environmenal sciences. BioOne provides a susainable online plaform for over 170 journals and books published by nonprofi socieies, associaions, museums, insiuions, and presses. Your use of his PDF, he BioOne Web sie, and all posed and associaed conen indicaes your accepance of BioOne s Terms of Use, available a Usage of BioOne conen is sricly limied o personal, educaional, and non-commercial use. Commercial inquiries or righs and permissions requess should be direced o he individual publisher as copyrigh holder. BioOne sees susainable scholarly publishing as an inherenly collaboraive enerprise connecing auhors, nonprofi publishers, academic insiuions, research libraries, and research funders in he common goal of maximizing access o criical research.

2 ALLOZYMIC, MORPHOLOGICAL, PHENOLOGICAL, LINGUISTIC, PLANT USE, AND NUTRITIONAL DATA ON WILD AND CULTIVATED COLLECTIONS OF LUFFA AEGYPTIACA MILL. (CUCURBITACEAE) FROM NEPAL, SOUTHERN CHINA, AND NORTHERN LAOS KENDRICK L. MARR, YONG-MEI XIA, AND NIRMAL K. BHATTARAI Marr, Kendrick L. (Museum of Anhropology, Universiy of Michigan, 4009 Museums Building, 1109 Geddes Avenue, Ann Arbor, MI , and Xishuangbanna Tropical Boanical Garden, Chinese Academy of Sciences, Menglun, Mengla, Yunnan, China ; Curren address: Deparmen of Naural Hisory, Royal Briish Columbia Museum, 675 Belleville Sree, Vicoria, BC, Canada V8W 9W2; Yong-Mei Xia (Xishuangbanna Tropical Boanical Garden, Chinese Academy of Sciences, Menglun, Mengla, Yunnan, China ), and Nirmal K. Bhaarai (Naional Herbarium Plan Lab, G.P.O. Box 938, Godhavari, Lalipur, Nepal). ALLOZYMIC, MORPHOLOGICAL, PHENOLOGICAL, LINGUIS- TIC, PLANT USE, AND NUTRITIONAL DATA ON WILD AND CULTIVATED COLLECTIONS OF LUFFA AE- GYPTIACA MILL. (CUCURBITACEAE) FROM NEPAL, SOUTHERN CHINA, AND NORTHERN LAOS. Economic Boany 59(2): , In order o explore he domesicaion and ehnoboany of Luffa aegypiaca, collecions were made from wo regions where wild plans and he domesicae occur: an area in and near souhern China, represened by Yunnan Province (China) and norhern Laos; and he Indian subconinen, represened by souheasern Nepal. The allozyme evidence was inconclusive wih respec o he region of domesicaion due o he small sample size of wild accessions from Nepal bu suggess a single place of domesicaion. Principal Componens Analysis of morphological characers revealed ha some accessions of he domesicae from Yunnan and Laos were more similar o he wild ype han were hose from Nepal. Compared o he wild ype, he domesicae had non-bier, larger, and indehiscen fruis, flowered earlier, and had hicker frui vascular bundles. The nuriional conen of radiional culivars was superior o ha of a single modern culivar ha was examined. Key Words: Luffa aegypiaca; Cucurbiaceae; domesicaion; ehnoboany; allozyme; morphology; phenology; nuriional conen; China; Laos; Nepal; wild. The genus Luffa (Cucurbiaceae) consiss of seven ropical species, four from he Old World 1 Received 2 January 2003; acceped 1 November and hree from he New World (Heiser and Schilling 1990). Two Old World species, L. aegypiaca Mill. (syn. L. cylindrica [L.] M.J. Roem.) and L. acuangula (L.) Roxb. have been domesicaed for heir edible immaure frui. They are imporan vegeables in ropical and Economic Boany 59(2) pp by The New York Boanical Garden Press, Bronx, NY U.S.A.

3 138 ECONOMIC BOTANY [VOL. 59 subropical counries hroughou he world, especially Asia and India. In English, boh species are known as loofah, and because he exocarp of L. aegypiaca is smooh, i is also described as smooh loofah. For use as a vegeable, he fruis are harvesed when sill ender, approximaely half he size of a maure frui. The exure is similar o ha of zucchini, bu is more mucilaginous. In India, hey are peeled, sliced, and fried wih spices as a ype of curry (Joshi 1995). In easern China, hey are sliced, fried, and served wih a sauce (Walers 1989). In Japan, hey are eaen fresh, or sliced and dried o be eaen laer (Porerfield 1943). A second imporan use is described by he English names: vegeable sponge, luffa sponge, and dishrag gourd (Burkill 1985; Jansen, Gildemacher, and Phuphahanaphong 1993). When he frui maures and dries, all ha remains of he mesocarp are he dried, angled vascular bundles. These form a durable, dense, siff, bu compressible fibrous marix. Worldwide, he uses for his fibrous marix include cleaning sponges, engine oil filers, seel helme linings, insulaion (from sound, shock, and emperaure), poholders, door and bah mas, insoles, sandals, and gloves (Jansen, Gildemacher, and Phuphahanaphong 1993), as well as suffing for pillows, maresses, and saddles (Sinno and Bloch 1943). Recenly, loofah sponges have been used in bioechnology as a marix o immobilize and suppor he growh of cells in liquid media (Ogbonna e al. 1994; Iqbal and Zafar 1993). There are addiional uses for L. aegypiaca. The shoo ips, young leaves, flower buds, and flowers are edible (Porerfield 1943; Keraudren-Aymonin 1975; Jansen, Gildemacher and Phuphahanaphong 1993), and he seeds, leaves, sem sap, and charred frui have medicinal uses (Jansen, Gildemacher, and Phuphahanaphong 1993; Neuwinger 1994; Schules 1990; Whisler 1990). In China, he ground seeds and whole plan are used for insec conrol (Yang and Tang 1988). In India, cale are allowed o breah he smoke of burning sponges o rea dyspepsia (Pal 1981), and he exocarp can be used as a recepacle for a lamp (Singh and Pandey 1998). Luffa aegypiaca is a perennial vine bearing brigh yellow flowers ha open during he day. Female flowers are soliary, while male flowers occur in racemes. Plans are self-compaible (Sinno and Bloch 1943). The leaves are palmaely five- o seven-veined and are slighly o deeply lobed. Luffa aegypiaca var. leiocarpa (Naud.) Heiser & Schilling is he wild variey, found from Myanmar o he Philippines and eas o Tahii and souh o Ausralia, and var. aegypiaca is he domesicae (Heiser and Schilling 1990). Fruis of he domesicae are generally no bier, whereas wild fruis are inensely bier, conaining cucurbiacins, bier eracyclic rierpenes ha characerize he Cucurbiaceae, and among he mos oxic naural subsances known (Neuwinger 1994). The original place of domesicaion of L. aegypiaca is unclear (Burkill 1935; DeCandolle 1959; Li 1970; Walers 1989). DeCandolle (1959) considered an African origin o be very unlikely and favored an Asian origin. Burkill (1935) suggesed eiher Africa or Asia, bu concluded ha i was impossible o esablish which place due o he aniquiy of is culivaion. Zeven and Zhukovsky (1975) also suggesed ropical Asia, possibly India, as he place of domesicaion. Heiser and Schilling (1990), however, ruled ou India because hey fel ha wild plans did no occur ha far wes. Par of he uncerainy over he place of domesicaion is due o differences in he repored range of var. leiocarpa. Alhough Heiser and Schilling (1990) concluded ha var. leiocarpa did no occur wes of Myanmar, DeCandolle (1959) repored wild plans in Malabar, Hindusan (i.e., India), Myanmar, and Sri Lanka, as well as Ausralia. Prain (1903) lised boh wild and culivaed L. aegypiaca in Bengal. Dalziel (1955) repored he presence in India of wo varieies, one edible and he oher bier and poisonous. Burkill (1985) saed ha in Wes Africa he wild fruis are bier and poisonous. I is unclear wheher all of he wild plans were ruly wild or were domesicaes, possibly escapes from culivaion, ha had revered o he bier form. Some Wes Africa indigenous names ranslae ino whie man s sponge, suggesing a relaively recen inroducion, a leas ino some areas (Burkill 1985). Asfaw and Tadesse (2001) lised boh wild and culivaed L. aegypiaca as edible species of Ehiopia, bu gave no vernacular names, which suggess ha he plan has no been in heir culure long, and hey did no menion wheher or no he wild plans have

4 2005] MARR ET AL.: WILD AND CULTIVATED LUFFA 139 bier frui. Perhaps hese wild plans were escaped forms of he domesicae. There are no cerain idenificaions of Luffa remains from archaeological sies. Yen (1977) described archaeoboanical remains from Spiri Cave in Norhern Thailand, a sie ha was occupied approximaely 7,000 11,000 years BP. Throughou he assemblage were frui wall fragmens ha were enaively idenified as Luffa. There are early lieraure references o L. aegypiaca from China and India. I firs appears in Han (he dominan ehnic group of China) Chinese wriings in 600 A.D. (Porerfield 1943). However, many non-han ehnic groups who inhabied souhern China could have culivaed Luffa long before i was recorded in he Han lieraure. In Indian Rig Vedas from B.C. are recorded Sanskri names for L. aegypiaca ha ranslae as rale of he gods, beaing, seed vessel, or indisinc noise, all evidenly alluding o he sound made by he seeds when he maure dried frui is shaken (Decker-Walers 1998, 1999). I is no clear wheher i was he wild ype or he domesicae ha was described. Because archaeological remains of L. aegypiaca are lacking, we have underaken here o use indirec approaches o undersand he place and process of his species domesicaion. Our primary objecives were as follows: 1) o use evidence from allozyme and morphological variaion o gain insigh ino he original place of domesicaion; 2) o documen he effecs of human selecion on geneic diversiy as revealed by allozyme, morphological, and phenological variaion; 3) o compare he nuriional conen of differen accessions; and 4) o collec linguisic and plan use informaion. We sampled domesicaed and wild plans from wo possible regions of domesicaion of L. aegypiaca: a region including ropical souhwes Yunnan Province, China and adjacen areas of norhern Laos o represen souhern China/SE Asia; and ropical, cenral, and easern Nepal o represen he Indian subconinen (logisical issues prevened he collecion of samples from India). MATERIALS AND METHODS GERMPLASM AND DATA COLLECTION IN THE FIELD We visied villages of many differen ehnic groups (Table 1). In hese villages, many farmers coninue o raise old culivars of many crops. We recorded local names, how differen culivars were disinguished, and mehods of use. In Yunnan Province, China, we made collecions wihin an area approximaely o N, and o E, from Sepember 1997 o March 1998, and January o March We colleced in Laos in December 1997 and February 1999, wihin 60 km of Muang Xai, he capial ciy of Oudomaxai Province, a N and E. We colleced in Nepal in November 1998 and February 1999 in an area beween approximaely o N and o 88 E. In each village, we requesed seeds from several households. Villagers described he plans as old varieies or local varieies. We refer o hese as radiional culivars (Table 1). Packaged seeds of commercial culivars were widely available in markes and we refer o hese as modern culivars. We asked abou he presence of wild plans and colleced seeds of hese wherever hey were available. Each village and wild populaion was given a locaion number (Table 1). Unique accession numbers were assigned o he seeds from one household, and addiional numbers were used if i appeared ha he seeds from one household came from disincly differen frui. Thus from each locaion, one o several accessions were colleced. Villagers ofen sored he seeds in conainers such as he dried frui of Lagenaria siceraria (Molina) Sandley (bole gourd) mixed ogeher wih he seeds of many differen crops. In mos cases we removed seeds from a single Luffa frui. MORPHOLOGICAL ANALYSES Mos collecions were made during he dry season, when fruis were maure bu vines were dead. To undersand he morphology of he enire plan, seeds of 39 accessions from 31 collecion locaions were planed in a level, irrigaed field a he Xishuangbanna Tropical Boanical Garden (XTBG) (21 41 N and E) of he Chinese Academy of Sciences (CAS). I was locaed in he Xishuangbanna Dai Auonomous Prefecure of souhern Yunnan. The elevaion was 570 m and he climae monsoon ropical wih a mean annual rainfall of 1,557 mm and a mean annual emperaure of 21.4 C. These accessions represened he full range of geographical and morphological diversiy (a leas as un-

5 TABLE 1. COLLECTION LOCATION NUMBERS OF LUFFA AEGYPTIACA ACCESSIONS COLLECTED FROM CHINA (YUNNAN PROVINCE), NEPAL, AND LAOS. AC- CESSIONS FROM COLLECTIONS LOCATIONS INDICATED BY () WERE INCLUDED IN THE MORPHOMETRIC ANALYSIS. ALL SAMPLES EXCEPT THOSE IN [] WERE ANALYZED FOR ALLOZYME DIVERSITY. FOR EACH COLLECTION LOCATION, N DESIGNATES THE NUMBER OF ACCESSIONS THAT WERE COLLECTED, AND S DESIGNATES THE NUMBER OF INDIVIDUALS THAT WERE ASSAYED FOR THE ALLOZYME ANALYSIS. Locaion Elevaion number Village Prefecure/disric 1 (m) China (Yunnan Province) C1 (C2) C3 C4 (C5) C6 C7 C8 (C9) C10 C11 C12 (C13) (C14) C15 (C16) (C17) C18 (C19) (C20) (C21) (C22) (C23) C24 (C25) (C26) (C27) (C28) Guomu Mane Huiqing Shangyong Manzhang Huiwa Xiangduoya Sayyou Nanlaxiaozhai Xinxiuhua Daqingshu Mansa Daka Bakaxiaozhai Bayaxinzhai Lilachun Xuannawan Gedou Bangjiao Kongmudan Dawanian Xinmin Huidu Yiyuanshui Xibenxiang Jinghong Mk. Zhaguo Fenshuilin Dehong/Luxi Lincang/Cangyuan Xishuangbanna/Jinghong Xishuangbanna/Jinghong Xishuangbanna/Jinghong Xishuangbanna/Jinghong Simao/Ximeng Nujiang/Lushui Dehong/YingJiang Dehong/JieLe Dehong/JieLe Lincang/Zhengkang Lincang/Lincang Xishuangbanna/Jinghong Xishuangbanna/Jinghong , ,080 1, ,400 1,400 1, ,700 1, Type of accession 2 Ehnic group Local name m w w Dai Dai Dai Dai Dai Dai Aini Aini Aini Aini Aini Bulang Aini Jinuo Jinuo Lafu Lisu Jingpo Jingpo Jingpo Han Yao Yao Yao Han Han mamop buop buop buop makmop buop bweba bweba bweba bweba bweba loi, lol, beilol bwe(m)ba poblei poblei manoyxi gesidu bongsap bongsap bongsap, sapbuon si-gua lazai geray geray sigua rou-sigua Allozyme analysis N/S 4/14 8/21 5/12 2/6 5/6 5/5 4/11 5/11 3/8 2/6 2/3 5/11 3/8 3/11 2/3 6/9 7/8 3/5 7/7 6/9 5/5 3/6 3/8 2/7 4/6 1/1 5/12 8/ ECONOMIC BOTANY [VOL. 59

6 TABLE 1. CONTINUED. Locaion Elevaion number Village Prefecure/disric 1 (m) (C29) (C30) (C31) C32 [(C33)] Laos L1 L2 [(L3)] L4 (L5) (L6) [(L7)] Nepal (N1) (N2) N3 N4 (N5) N6 (N7) (N8) (N9) (N10) Lilachun Xiajili Mangmao Bayalaozhai Guofa Bannasay Laksip Banpuonsaa Baochuan Banhong Banhong Huikom Narayangadh Kalikaar Pahalaiya Bhaa Rambhauri Biranagar Budhabare Thanko Kienidha Cipadol Simao/Ximeng Lincang/Cangyuan Lincang/Lincang Xishuangbanna/Jinghong Xishuangbanna/Jinghong Oudomxai Oudomxai Oudomxai Oudomxai Oudomxai Oudomxai Oudomxai Narayani/Chiwan Lumbini/Tanahu Narayani/Bara Narayani/Bara Narayani/Parsi Kosi/Morang Mechi/Jhopa Bagmai/Kahmandu Mechi/Jhopa Bagmai/Kahmandu 1 Prefecure is equivalen o a Laoian Province and a Nepali Zone. 2 radiional culivar, m modern culivar, w wild populaion. 1,080 1,150 1,300 1,000 1,000 1,000 1,000 1,000 1,000 1, ,400 1,350 Type of accession 2 Ehnic group Local name w w w w Wa Han/Dai Aini Hmong Khmu Laolong Laoeng Laolong Laolong Nepali Nepali Nepali Nepali Nepali Nepali Nepali Nepali Nepali maimop sigua bweba daosua makbuop bierlamnar makbuop makbuop (ban)ghiraunla (ban)ghiraunla (ban)ghiraunla (ban)ghiraunla (ban)ghiraunla (ban)ghiraunla (ban)ghiraunla (ban)ghiraunla (ban)ghiraunla Allozyme analysis N/S 1/2 1/2 1/2 3/11 2/ 1/4 1/4 3/ 3/4 7/16 2/10 5/ 9/9 5/6 9/9 4/4 5/5 5/8 9/14 12/15 1/6 9/ ] MARR ET AL.: WILD AND CULTIVATED LUFFA 141

7 142 ECONOMIC BOTANY [VOL. 59 frui suggesed human selecion for hicker vascular bundle srands. In order o es his hypohesis, we measured, a 100, he srand widhs of a single represenaive vascular bundle aken from dried maure fruis of several accessions of he domesicae and wild plans. Fig. 1. Ouline of Luffa aegypiaca leaf. Numbers indicae dimensions ha were measured. Some of hese were used o calculae he following raios in Table 2: Raio 1 1/6; Raio 2 1/2; Raio 3 3/4; Raio 4 1/5. Measuremen 8 is he sinus deph along he leaf margin. Measuremen 9, he angle beween he vein and he peiole, was doubled for he PCA and he doubled value is used in Table 2. dersood from frui size and shape). They were grown from March 1998 unil November Voucher specimens were made from hese plans and deposied a he Universiy of Michigan (MICH) (see Appendix). Measuremens were aken of 20 quaniaive characers using hree o five leaves, flowers, and fruis per accession. Measuremens were averaged for each accession and hen summarized by culivar ype and counry of origin. To describe differences in leaf shape, we calculaed four raios (Fig. 1). Seed margins were winged o varying degrees, or no winged a all; hus seed lengh and widh measuremens alone do no capure he range of variaion. Because he wing is hin and conribues less mass han does seed endosperm, he characer seed mass was used o disinguish among accessions in which he seeds were he same lengh and widh bu differed in he widh of he wing. Principal Componens Analysis (PCA) was used o summarize he range and paern of morphological variaion based on a correlaion marix using Sysa (1997). The many uses of he fibrous marix of he ALLOZYME ANALYSIS We used cellulose-aceae gel elecrophoresis (Herber and Beaon 1993) for he allozyme analysis. Fresh young leaves were ground wih a small volume of sand in a ph 8.0, 0.1 M Tris- HCl buffer wih 10% PVPP, 5% sucrose, and bea-mercapoehanol (one Paseur pipee drop/ one ml buffer). A ph 8.5 Tris-Glycine (TG) elecrode buffer was used o resolve GDH (gluamae dehydrogenase), ADH (alcohol dehydrogenase, PGI (phosphoglucoisomerase), PGM (phosphoglucomuase), and AAT (asparae amino ransferase). A ph 7.0 Cirae-Morpholine (CM) buffer was used o resolve MDH (malae dehydrogenase), IDH (isocirae dehydrogenase), SKD (shikimae dehydrogenase), ACON (aconiase), LAP (leucine amino pepidase), ALD (aldolase), ME (malic enzyme), 6-PGD (6- phosphoglucodehydrogenase), G-3-PDH (glyceraldehyde-3-phosphae dehydrogenase [NADP]), MNR (menadione reducase), and G-6-PDH (glucose-6-phosphae dehydrogenase). TG gels were run for minues, CM gels for minues. According o availabiliy, he number of accessions assayed per collecion locaion varied (Table 1). For each collecion locaion or wild populaion, he number of alleles per locus (A), percen of loci polymorphic (P), mean expeced heerozygosiy (H exp ), and Nei s Geneic Ideniies (I) were calculaed using BIOSYS-1 (Swofford and Selander 1989). This program was used o produce a phenogram, based on I using he Unweighed Pair Group Mehod employing Arihmeic Averages (UPGMA). PHENOLOGICAL OBSERVATIONS On 17 March 1999, we planed seeds of five wild accessions from four populaions, 22 radiional culivar accessions from 17 collecion locaions, and a single new culivar a XTBG. The node on he main sem a which he firs flower appeared was recorded as was is gender and he gender of he flower(s) or inflorescences a each subsequen node.

8 2005] MARR ET AL.: WILD AND CULTIVATED LUFFA 143 NUTRITIONAL ANALYSIS Fruis from eigh accessions of he living collecion a XTBG were harvesed when hey were approximaely half size. They were weighed, he skin removed, and he remaining issue weighed again. Proein, fiber, oal sugars, S, Mg, Mn, Cu, Fe, Zn, P, K and Ca were analyzed a he Yunnan Academy of Agriculural Sciences in Kunming, Yunnan. Toal sugar was analyzed using he Fehling Mehod, proein by he Semimicro Kjeldahl Mehod. Fiber was deermined afer processing wih dilue acid and alkali o eliminae saccharides and proeins, hen wih aceone o remove annin, pigmens, and remnan fas, hen heaed o 300 C and he ash weighed. Inorganic elemens were analyzed by ICP-AES (Induciviy Coupled Plasma Aomic Emission Specrum). RESULTS GERMPLASM AND DATA COLLECTION IN THE FIELD We colleced as much diversiy as possible, and acceped any samples we were offered. A oal of 221 accessions were colleced: 124 domesicae and 17 wild from China; 12 domesicae and seven wild from Laos; and 67 domesicae and one wild from Nepal (Table 1). Accession numbers are lised in he Appendix. Seeds of mos accessions were sored a 4 C in he seed bank of XTBG. The size, shape, and color of seeds (Fig. 2) and fruis (Fig. 3) varied grealy. Mos Yunnan accessions had black seed coas; however, some marke seeds were whie and a single whie seed was found in a wild frui from Yunnan (Fig. 2). Nepalese villagers disinguish beween whie (very pale green) and green (much darker green) fruis as well as whie and dark seeds. Luffa aegypiaca was grown in every village we visied, in home gardens or in upland fields devoed o swidden agriculure (in Yunnan and Laos) and planed wih rice, corn, or coon. Tree sumps in swidden fields were ofen lef for he vines o climb upon. Based on frui size and shape, i appeared ha a leas wo culivars were grown in mos villages. Young fruis and he dry maure frui sponges were common in markes. In larger ciies, vendors described heir frui as modern culivars. In smaller rural markes, boh modern and Fig. 2. Luffa aegypiaca seeds ha are radiional culivars colleced from villages unless oherwise noed. Top group are from Yunnan, middle group from Laos, and lower group from Nepal. Whie seed in op row and he one o is righ were of a modern culivar sold from an open bag by a seed vendor in a marke. Boom row of seeds from Yunnan are all from wild plans (he firs wo from he same frui). The las seed in he second row from Laos is from a wild plan. The las seed from he second row from Nepal is from a wild plan. radiional culivars were available. In Yunnan, modern culivars were commercially grown on bamboo rellises ha someimes covered several hundred square meers. Wild L. aegypiaca fruis were colleced from secondary foress near villages and were easily recognized by heir small, very bier frui (Fig. 3) and smaller seeds (Fig. 2). Some Yunnan villagers saed ha wild plans could also be found in dasenlin (lierally meaning big fores, probably referring o he primary fores). During January and February field rips in Yunnan and Laos, wild plans bore maure fruis, and mos vines had died. During our November field rip o Nepal, alhough we found several plans ha appeared o be wild, only one bore fruis wih maure seeds. Several seed accessions ha we were given by Nepali villagers were from puaively wild plans, bu hese urned ou o be he domesicae as heir seeds were he same size as he domesicae and, upon growing hem, we found ha heir frui was no bier. According o villagers, wild plans flowered laer han he

9 144 ECONOMIC BOTANY [VOL. 59 Fig. 3. Maure fruis of Luffa aegypiaca illusraing par of he range of variaion. All are radiional culivars unless noed. Counry of origin as follows: Top row, lef o righ: Nepal (bier ase), China, Nepal, Nepal, China, China, China (modern culivar); boom row, lef o righ: China, China, China, China, China, Nepal (wild). Scale is 15 cm long. domesicae and had rougher leaves ha ased more bier han he domesicae. Wild and domesicaed L. aegypiaca differed in a number of oher regards. Maure fruis of wild plans dehisced via a disal operculum ha separaed easily. This mechanism is necessary for a species ha proecs he developing as well as he maure seeds wih a bier mesocarp. By conras, he operculum of mos domesicaes only dehisced wih mechanical force, and he breakage was no necessarily along he abscission zone. The dry exocarp of wild fruis was visibly hinner han ha of mos domesicaes, whose hickness varied grealy among accessions. Wild fruis were olive green, wih dark sripes running lenghwise, while he domesicae ranged from pale o dark green. Pedicels of he domesicae were hicker han hose of wild fruis. In boh Yunnan and Nepal, we encounered wha appeared o be domesicaed L. aegypiaca bu wih bier fruis. These fruis were a leas wice as long as wild fruis. According o some Yunnan villagers, fruis from wo- o hree-year old vines are bier and look he same as nonbier fruis (only upon asing hem are hey known o be bier), bu plans grown from heir seeds will yield non-bier fruis. We planed seeds of some of he accessions above. Seeds colleced from locaion N5 produced vines ha bore bier frui. A family in one locaion, N7, complained ha hey could no eradicae plans wih bier frui ha grew ogeher wih plans ha yielded non-bier frui. We planed several seeds from one of heir bier frui. Some plans of his accession bore dark green bier frui and ohers bore ligh green nonbier frui. Each of he ehnic groups of Yunnan and Laos had heir own names for L. aegypiaca (Table 1). This is no surprising given ha among he non-han villages ha we visied in Yunnan alone, nine languages represening four major language groups (Ramsey 1987) were presen:

10 2005] MARR ET AL.: WILD AND CULTIVATED LUFFA 145 Tai language group (Dai ehnic group), Tibeo- Burman (Lisu, Aini, Lahu, Jinuo, and Jingpo), Miao-Yao (Yao), and Mon-Khmer (Wa, Bulang). We were able o learn he meanings of some of he names or modifiers ha are applied o L. aegypiaca. In Xishuangbanna Prefecure, he Lahu word is yixikefeixi, lierally a frui for washing a bowl. Dai people disinguish beween wild and domesicaed L. aegypiaca by naming he wild plan buopuen (uen bier) and he domesicae as buophom or buopban (hom fragran; ban swee). In he Jinuo language, wild L. aegypiaca is poblei aple (aple false) or poblei aku (aku bier). In he Yao language, domesicaed L. aegypiaca is biao geray (biao home) and wild is zha geray or geray im or gimdi geray (zha false, im bier, gimdi in he fores). Aini disinguish wild L. aegypiaca as bweba laha (laha wild) and fragran frui as bwebabaso (so fragran). A LaoTeng farmer in Laos gave hree names: bier lamnar mon for a shor, nearly round frui; bier lamnar wang for a longer frui, and bier kambier for an even more elongaed frui. In Nepal, ghiraunla, budune ghirauna, or ban ghiraunla (ban smooh) were he only names ha we encounered. Ghiraunla is also he name for L. acuangula, someimes as ire ghiraunla (ire like an arrow). The Mandarin word for L. aegypiaca is sigua. The modern culivar ha we colleced was named rou si-gua (rou flesh, evidenly alluding o he fleshy naure of his culivar). Si-gua was also used in Yunnan for wo oher Cucurbiaceae: Cucurbia ficifolia Bouche was fen sigua and Sechium edule (Jacq.) Sw. was yang sigua (yang foreign). In he Dai language, he word for S. edule also means foreign Luffa. In one village, C13, he vines of one plan bore clusers of frui a a single node. These developed from a raceme of perfec flowers borne in he place of he ypical raceme of male flowers. USAGE For use as a vegeable in Yunnan, he skin was removed and he frui was sliced and prepared as a soup or fried in oil. In Nepal, he frui was prepared in curries. Villagers described he fruis according o heir ase, size, shape, color, ime o ripening, and firmness of he fibrous marix. In Yunnan, domesicaes were caegorized as xiang (fragran or flavorful), no xiang, or bier. On several occasions, where villagers grew boh a long and a shor culivar, he shor culivar was idenified as being more xiang. In one Dai village, he culivar ha was more fragran was also said o have harder fibers. According o one Nepali villager, radiional culivars flowered over a longer period of ime, and he ase was preferred over modern culivars; however, modern culivars bore frui sooner. No all of he ehnic groups ha were familiar wih L. aegypiaca ae he frui. The Lisu, in he far wesern Yunnan Nujiang (Salween) River valley, were unaware of is use for food. Their name, gesidu, lierally means somehing used o wash he bowl. In his area, plans grew sponaneously, heir fruis were larger han he wild ype, and hey clearly were he domesicae. In addiion o he edible immaure frui, we encounered a number of oher uses. In one marke, a Jingpo woman sold fresh male flowers for eaing. In Yunnan, he Dai, Jingpo, Yao, and Aini peoples ea he young shoos of plans wih bier frui. I was unclear wheher or no hese plans were he wild ype. The Yao place he burned ash of wild or culivaed maure frui ono human or animal cus o preven infecion. One Dai woman old us ha a well-known use of male flowers, especially for women, is o boil hem ogeher wih an egg as a reamen for iron deficiency and dizziness. Han prepare a cough medicine from he liquid of cu sems. According o a Jingpo villager, wild L. aegypiaca was used o rea kidney disease and oher ho diseases, and he vine was placed a he enrance o a house as an offering o ghoss. In Laos, bier frui can be used o rea gall bladder problems. In one Nepal village, older people had used wild Luffa, previously more abundan, for an unsaed medicinal use. In anoher village, we were old ha he wild fruis are poisonous. Some Nepalese religious ceremonies require ha a goa be sacrificed. If a person is oo poor o afford a goa, hey can use four fruis from Luffa species o represen he legs of a goa and hus o serve as a subsiue sacrifice. In Yunnan, he Dai spli he maure frui fiber longiudinally and place i in he boom of he rice-cooking vessel ha is oherwise no sealed a he boom. This porous surface suppors he rice grains and creaes a space in he boom of he vessel ha is hen placed ino a wok o seam-cook he rice.

11 146 ECONOMIC BOTANY [VOL. 59 MORPHOLOGICAL STUDY The mean, sandard deviaion, and range of variaion for each characer used in he PCA are recorded in Table 2. While here was some overlap among accessions colleced from Yunnan, Laos, and Nepal, some differences exised as well. Some rends are highlighed here, including some characers ha are no lised in Table 2 because hey were no included in he PCA, followed by a summary of he PCA resuls. Fruis and seeds of wild plans were generally smaller and of a more uniform size han were hose of he domesicae (Table 2). A frui lengh o widh raio was useful o describe differences in frui shape. For he domesicae, his value was 1.1 o 12.0, wih he roundes frui from Laos and Yunnan, he narrowes from Nepal. For he wild ype his value was Vascular bundle srand hickness of maure fruis differed among accessions. Srands from wild fruis (n 9) were mm hick, hose of radiional culivars (n 39) were mm hick, and hose of a modern culivar were 0.33 mm hick. Leaf size and shape varied grealy among accessions. Mos domesicaes had larger leaves han wild plans. Some of he raios calculaed for leaf shape (Fig. 1) varied grealy among accessions (Table 2). Leaves wih low values for Raio 1 were rounder, leaves wih higher values were more elongaed. For Raio 2, he larger he value, he deeper were he lobes. In general, leaves of wild plans were more deeply lobed han were hose of he domesicae; however, a few domesicaed accessions also had deeply lobed leaves. Accessions wih more elongaed and/or deeper lobed leaves came from Nepal and Yunnan, whereas leaves of mos accessions from Laos were rounder wih shallower lobes. Floral pars of wild L. aegypiaca were generally smaller han were hose of he domesicae (Table 2). One characer of poenial use for which only casual observaions were made was he number of necar glands on he ouer surface of he calyx. These varied from zero o a leas 30 per calyx lobe, and heir number was fairly consisen among flowers wihin plans. Thiry-nine accessions were analyzed by PCA using 19 characers (Table 2). Seed lengh, widh, and mass conribued he mos o he variaion on he firs axis, which accouned for 28% of he oal variance. The second axis accouned for 12% of he variance wih Raio 1 (leaf blade lengh/widh), he angle of leaf lobe ips, and Raio 4 (leaf blade lengh/peiole lengh), conribuing he mos o he variaion. The hird axis accouned for 11% of he variaion, wih calyx lengh, peal lengh, and Raio 2 (he degree o which leaves are lobed) conribuing mos o he variaion. In he scaerplo (Fig. 4) from he PCA resuls, he single wild accession from Nepal (locaion N9) clusered close o one of he wild accessions from Laos (locaion L5), as well as wih domesicaes from Yunnan and Laos. Mos domesicaes from Yunnan and Laos clusered apar from hose of Nepal and closer o he wild accessions (Fig. 4). When more han one accession was measured from one locaion, i.e., C19, C20, N1 and N8, hey were ofen disinc from each oher (Fig. 4). Accessions of old culivars wih bier frui from locaions N5 and N7 clearly were disinc from he rue wild plans (Fig. 4). The modern culivar, C26, and accessions from locaion C20 clusered mos closely o Nepalese accessions, especially along he firs axis (Fig. 4). ALLOZYME RESULTS Tweny-nine loci of 16 enzymes were resolved: hree loci each for 6PGD, MDH, ADH, and ME; wo for PGM, PGI, ALD, AAT, and MNR; and one each for SKD, G6P, LAP, ACN, IDH, GDH, and G3PDH. Isozymes were labeled sequenially beginning wih 1 as he mos anodal locus. Allozymes were labeled sequenially beginning wih a as he mos anodal. In many populaions, no variaion was deeced a any locus. The only polymorphic loci were 6Pgd-2, Aa-2, Skd-1, Me-1, and Pgi-2, each of which had wo alleles in one o several populaions. The allele frequency daa marix is available from he firs auhor. Because few loci were polymorphic, measures of geneic diversiy were low. For he domesicae, he ranges were as follows: Yunnan (n 32), A , P %, H exp ; Laos (n 5), A , P %, H exp ; and Nepal (n 10), A , P %, H exp For he wild populaions, he ranges were: Yunnan (n 4), A , P %, H exp ; Laos (n 1), A 1.1, P 7.1%, and H exp 0.022; and Nepal (n 1), A 1.0, P 0.000%, and H exp

12 TABLE 2. COLLECTED FROM CHINA, LAOS, AND NEPAL. (SEE FIG. 1FOR ILLUSTRATION OF LEAF RATIO CALCULATIONS.) Organ SUMMARY STATISTICS (MEAN S.D.; RANGE) FOR CHARACTERS USED FOR A PRINCIPAL COMPONENTS ANALYSIS OF LUFFA AEGYPTIACA ACCESSIONS Characer China modern (n 1) Leaf Lengh (mm) Angle a lobe ip ( ) Marin sinus deph (mm) 1.3 Angle beween veins 124 Raio Raio Raio Raio Frui Frui lengh (cm) 29 Frui widh (cm) 8 Seed lengh (mm) 16 Seed widh (mm) 9 Seed mass (mg) 111 Female flower Ovary lengh (mm) 37 Ovary widh (mm) 7 Corolla lengh (mm) 45 Sepal lengh (mm) 15 Sepal widh (mm) 7 Pedicel lengh (mm) 66 Domesicaed China/Laos radiional (n 23) (87 145) (81 141) ( ) (55 138) ( ) ( ) ( ) ( ) (9 50) 7 3 (4 15) ( ) ( ) (61 118) (19 53) (5 10) 42 7 (27 54) (10 17) (4 10) (13 140) Nepal radiional (n 9) ( ) (77 130) ( ) (75 117) ( ) ( ) ( ) ( ) (16 76) 6 1 (4 7) (12 15) ( ) (67 167) (26 83) (6 10) 50 6 (38 59) (12 16) (5 8) (45 133) China/Laos (n 5) Wild (87 119) (37 114) ( ) (70 99) ( ) ( ) ( ) ( ) 9 2 (6 11) 4 1 (4 5) (9 11) ( ) (38 59) (16 27) (5 7) 40 8 (34 53) (9 17) (4 6) (21 36) Nepal (n 1) ] MARR ET AL.: WILD AND CULTIVATED LUFFA 147

13 148 ECONOMIC BOTANY [VOL. 59 Fig. 4. Scaerplo of Luffa aegypiaca accessions from a Principal Componens Analysis of morphological characers. Each poin is of a single accession, labeled by is collecion locaion number. Numbers followed by * are wild plans. See Table 1 for collecion locaions. Some accessions had a 6PGD phenoype ha was difficul o inerpre. Mos plans had hree bands ha were inerpreed as homodimers of hree homozygous loci. However, in all 13 accessions ha were homozygous for 6Pgd-2a, a fourh band migraed furher on he gel han 6Pgd-1a. The four-band phenoype was no observed in wild plans. Wihin accessions, all individuals had he same phenoype for 6PGD. No individuals were heerozygous, even in villages where boh phenoypes (he hree band and he four band) were deeced. Those accessions wih 6Pgd-2a came from some villages in Xishuangbanna Prefecure, Yunnan, and one village from Laos, all wihin he Mekong River drainage. For hese accessions (hose wih four bands), here does no appear o be a correlaion wih morphological characers; for example, he frui lengh o widh raio varied from As we could no propose a genoype o accoun for his phenoype, he mos conservaive inerpreaion was o no score his fourh band. The allelic complemens of domesicaed and wild L. aegypiaca differed very lile. The very rare Pgi-2a, (homozygous in wo individuals of a single accession from C20) and he less rare 6Pgd-2a occurred only in he domesicae. The Fig. 5. Phenogram based on Nei s Geneic Ideniies for allozyme analysis of Luffa aegypiaca. Numbers represen collecion locaions (see Table 1). Number preceded by * are wild populaions. Copheneic correlaion coefficien rare Aa-2a was deeced in a few accessions only from wo wild populaions, C28 and L5. The single wild accession from Nepal differed from all bu one accession in being fixed for Me- 1b. A single accession of he domesicae from N7, a locaion a few km from he wild populaion, also conained Me-1b. All oher wild or domesicae accessions from China, Nepal, and Laos were homozygous for Me-1a. The allelic complemens of wild L. aegypiaca from Nepal and China/Laos differed by he occurrence of Me-1b in he former and Aa-2a in he laer.

14 2005] MARR ET AL.: WILD AND CULTIVATED LUFFA 149 Fig. 6. Node a which firs flower appeared for accessions of Luffa aegypiaca. Each bar represens a single accession, labeled by is collecion locaion number. See Table 1 for collecion locaions. There was very lile allozyme divergence among mos collecion locaions (Fig. 5). Collecion locaions of he domesicae from Yunnan, Laos, and Nepal cluser ogeher wih wild populaions from Yunnan and Laos. Two radiional culivar collecion locaions (all individuals fixed for 6Pgd-2a) clusered apar from he remaining populaions. The nex mos disinc populaion was he wild populaion from Nepal, N9. PHENOLOGY OBSERVATIONS Wild accessions began flowering laer han did hose of he domesicae (Fig. 6). Several radiional culivars from Nepal flowered earlier han hose from Laos and Yunnan. The modern culivar was among he earlies o flower. Female flowers appeared 2 15 nodes laer han did male flowers. Thereafer, a each node, here was usually a raceme of male flowers and, on a longer pedicel, a single female flower. NUTRITIONAL ANALYSIS For each parameer, he nuriional conen varied o some exen among he eigh accessions (Table 3). The modern culivar (C26) generally ranked he lowes among accessions in all parameers excep for sugar concenraion. All accessions had higher values for percen proein, P, Ca, and Fe han hose repored by Jansen, Gildemacher, and Phuphahanaphong (1993). DISCUSSION Based on morphological characers, he area of souhern China/Laos is favored over Nepal as he place of domesicaion of L. aegypiaca. This assumes ha hose accessions ha have changed he leas from he wild ype are probably more primiive, and herefore he region in which hey occur is likely o be he earlies place of domesicaion. Many accessions of he domesicae from Yunnan and Laos were more similar o he wild ype han were domesicaed accessions from Nepal. Because ropical Nepal probably does no represen he range of variaion presen in he enire Indian subconinen, analysis of samples from India would be very useful, and in fac he quesion canno be answered wihou such samples. The lack of allozyme divergence among collecion locaions of he domesicae from China, Laos, and Nepal suggess a single place of domesicaion. However, hese resuls do no clearly indicae he cener of domesicaion. The main limiaion is he low sample size of wild populaions, especially from he Indian subconinen.

15 150 ECONOMIC BOTANY [VOL. 59 TABLE 3. NUTRITIONAL CONTENT OF LUFFA AEGYPTIACA. ALL VALUES ARE ACCORDING TO DRY WEIGHT. FRUIT FROM A SINGLE ACCESSION WERE ANALYZED FROM EACH COLLECTION LOCATION. SEE TABLE 1 FOR COLLECTION LOCATIONS. Cu mg/kg Mn mg/kg Zn mg/kg Fe mg/kg Mg % S P K Ca % % % % Sugar % Fiber % Proein % Collecion locaion [7] 18[7] 65[6] 90[7] 0.25[8] 0.17[8] 2.34[8] 0.56[7] 0.22[7] 32.9[1] 6.7[8] 16.5[7]* Tradiional Culivars C5 C16 C23 C25 N5 N8 N10 Modern Culivar C Jansen e al. (1993) * Numbers in [] are he ranking for his collecion ou of he eigh collecions ha were analyzed. Even in China and Laos, all of he wild populaions ha we sampled occur only in a single waershed, ha of he Mekong River. Neverheless, hese resuls esablish a baseline of geneic markers upon which o build. A poenially key geneic marker is he Me-1 enzyme locus. If furher collecions of wild plans from he Indian subconinen lack Me-1a, his would be srong evidence in suppor of a souhern China/SE Asian origin, because Me-1a was fixed in wild populaions from China and Laos and was also fixed in all of he domesicae, excep for he single accession from Nepal. The collecion of he wild plan from locaion N9 is imporan because i documens he presence of wild L. aegypiaca on he Indian subconinen. Fixaion for he Me-1b allele indicaes ha i is cerainly disinc from he domesicae, and is morphological similariy o wild plans from Laos confirm ha i is ruly a wild ype. Human selecion for earlier flowering is an obvious advanage for any crop. The fac ha among he old culivars of Yunnan and Laos are accessions ha iniiae flowering laer, some nearly as lae as wild plans, is addiional evidence ha he domesicae here has diverged less from he wild condiion han have accessions from Nepal, hough one of he old accessions from Nepal also iniiaed flowering relaively lae. Earlier flowering in Nepal may be due o he higher laiude, which may resul in a shorer growing season. Phenological variaion exiss in many crops. For example, a sudy of 544 radiional rice culivars in Laos found he number of days o flowering ranged from (Roder e al. 1996). Our single observaion of perfec flowers may be he firs such repor for L. aegypiaca. Roy and Saran (1990) horoughly reviewed sex expression in he Cucurbiaceae and repored ha L. aegypiaca is sricly monoecious. Alhough we analyzed he nuriional qualiy of only one modern culivar, i is ineresing o noe ha his culivar was uniformly inferior for all nuriens wih he excepion of he one consiuen ha may have he greaes influence on ase, sugar. These resuls are among he many reasons why i is imperaive ha radiional culivars be preserved. There is some evidence for bi-direcional selecion for vascular bundle hickness from our limied comparison of wild frui o differen ra-

16 2005] MARR ET AL.: WILD AND CULTIVATED LUFFA 151 diional culivars. Palaabiliy of he immaure frui may have been enhanced by selecion for hinner vascular bundle srands, whereas oher uses, for example as a scrubbing implemen, would have favored hicker srands and hence, a ougher fibrous marix. Humans have also seleced for greaer fiber densiy in he leaves of Agave fourcroydes Lem. ha is also culivaed for is fibers (Colunga-Garcia Marin, Esrada- Loera, and May-Pa 1996). Domesicaion ypically resuls in he loss of seed dispersal (Heiser 1988). I is obvious ha his is a rai ha humans would selec for in crops such as grains and legumes, bu i is less clear why i would be seleced for in L. aegypiaca since he seed is no he issue ha is harvesed direcly; neverheless, many accessions had indehiscen frui. Maure wild fruis reain a few seeds in he fibrous marix; however, mos fall ou hrough he operculum. Early in heir domesicaion, fruis ha reained heir seeds would have been easily disinguished from hose ha did no, and heir selecion would have given humans conrol over propagaion of he nex generaion. Because few, if any, species provide he useful properies of he maure L. aegypiaca frui fiber as well as he use of he immaure frui for food, humans may have found i o be an imporan species o culivae. The use of he maure fiber o seam rice is closely linked o he preparaion and consumpion of gluinous rice, which is ypically seamed, no boiled (Golomb 1976), and may have conribued o he selecion of larger and non-bier fruis. Gluinous rice is culivaed only in SE Asia among Tai speaking groups, or groups ha have had conac wih hem (Golomb 1976). Seam cooking requires a layer a he boom of he cooking vessel ha will allow seam o pass upwards hrough he rice grains, while a he same ime prevening hem from slipping ou of he boom. The maure L. aegypiaca frui fiber mees his need very well. This pracice may have favored he selecion of a non-bier muan; oherwise, he flavor of he rice would be desroyed by he bierness impared by he waer-soluble cucurbiacins. Among he names used in souhwes Yunnan (Table 1), here is lile evidence of borrowing among ehnic groups. This does no a all indicae independen domesicaion by each ehnic group, bu i does poin o a long period of knowledge of he plan. The lack of similariy beween Indian and Souheas Asian names also suggess independen knowledge of he species (Decker-Walers 1999). There is one possible excepion o his paern. Among he early Dravidian names in India for L. aegypiaca, L. acuangula, and Trichosanhes cucumerina var. anguina (L.) Haines are pir, pira, and bira (Decker- Walers 1998). These names may be relaed o a Laoian name for L. aegypiaca ha is also based on bier (Table 1). The morphological and phenological diversiy ha villagers are conserving in siu may be due o inrogression from wild plans. Inrogression from he wild ype may be he cause of bierness in some accessions of he domesicae, raher han reversion o he wild characer sae. Allozymes did no allow us o disinguish beween hese wo possibiliies because of he overall lack of unique alleles in wild versus domesicaed populaions. Inrogression wih wild plans is almos cerainly he reason for he presence of he Me-1b allele in he single domesicaed accession from Nepal (locaion N7) given he close proximiy of wild plans. In summary, he domesicaion of Luffa aegypiaca has resuled in several changes ha Heiser (1988) describes as a common resul of human selecion: loss of bierness, larger frui, changes in frui color, larger seeds, and loss of dispersal. Oher noable changes from he wild ype include earlier flowering and, in many accessions, hicker vascular bundles. The low allozyme diversiy in domesicaed L. aegypiaca is also a common feaure of mos crop plans (Doebley 1989). Analysis of boh wild and domesicaed accessions from addiional pars of his species range would conribue grealy o undersanding is place of domesicaion. ACKNOWLEDGMENTS For heir suppor and encouragemen during his research, KLM hanks Prof. Xu ZaiFu, Prof. Chen Jin, Prof. Yin ShouHua, Ms. Li LiMing, Fang ChunYan, Yang Wei, Deng XiaoBao, and Xiao Li and Xiao Xia (for managing he living collecion), all from he Xishuangbanna Tropical Boanical Garden. Prof. Long ChunLin and Su YongGe, Kunming Insiue of Boany (CAS), also provided much encouragemen during his invesigaion. Thanks o Mr. KhamPhao Bundala, of he Oudamxy Rural Developmen Office in Laos, for his assisance as a ranslaor during fieldwork in Laos. We are graeful for he assisance of he Wardens of Chiwan Naional Park and Parsa Wildlife Refuge in Nepal for logisical assisance. This research was suppored by a gran (INT ) from he U.S. Naional Science Foundaion o Dr. Richard I. Ford, Museum of Anhropology, The Universiy of Michigan, and by a gran from he Conservaion, Food, and Healh Foundaion. KLM hanks Dr. Richard Ford and Ms. Susan F. Campbell for encouraging and faciliaing his research. Bruce Bohm, Deena Decker-Walers, and hree anonymous reviewers made useful suggesions on an earlier draf of his paper. Wihou he

17 152 ECONOMIC BOTANY [VOL. 59 willingness of farmers from many villages o share heir seeds, knowledge, and hospialiy, his sudy would no have been possible. LITERATURE CITED Asfaw, Z., and M. Tadesse Prospecs for susainable use and developmen of wild food plans in Ehiopia. Economic Boany 55: Burkill, H. M The useful plans of Wes Tropical Africa. Ediion 2. Royal Boanical Gardens, Kew. The Whiefriars Press Limied, London, England. Burkill, I. H A dicionary of he economic producs of he Malay Peninsula. Crown Agens for he Colonies, London, England. Colunga-Garcia Marin, P., E. Esrada-Loera, and F. May-Pa Paerns of morphological variaion, diversiy, and domesicaion of wild and culivaed populaions of Agave in Yucaan, Mexico. American Journal of Boany 83: Dalziel, J. M The useful plans of Wes Tropical Africa. Crown Agens for Oversea Governmens and Adminisraions, London, England. DeCandolle, A Origin of culivaed plans. Hafner Publishing Co., New York. Reprin of he second ediion published in Decker-Walers, D. S Cucurbis, Sanskri, and he Indo-Aryas. Economic Boany 53: Sanskri, Modern Indo-Aryan, and Dravidian names for Cucurbis. Occasional Papers of he Cucurbi Nework No. 1. The Cucurbi Nework, Miami, Florida. Doebley, J Isozymic evidence and he evoluion of crop plans. Pages in D. E. Solis and P. S. Solis, eds., Isozymes in plan biology. Dioscorides Press, Porland, OR. Golomb, L The origin, spread and persisence of gluinous rice as a saple crop in mainland Souheas Asia. Journal of Souheas Asia Sudies 7:1 15. Heiser, C. B Aspecs of unconscious selecion and he evoluion of domesicaed plans. Euphyica 37: Heiser, C. B., and E. E. Schilling The genus Luffa: A problem in phyogeography. Pages in D. M. Baes, R. W. Robinson, and C. Jeffrey, eds., Biology and uilizaion of he Cucurbiaceae. Cornell Universiy Press, Ihaca, NY. Herber, P. D. N., and M. J. Beaon Mehodologies for allozyme analysis using cellulose aceae elecrophoresis: A pracical handbook. Helena Laboraories, Beaumon, TX. Iqbal, M., and S. I. Zafar Vegeable sponge: A new immobilizaion medium for plan cells. Bioechnology Techniques 7: Jansen, G. J., B. H. Gildemacher, and L. Phuphahanaphong Luffa P. Miller. Pages in J. S. Siemonsma and K. Piluek, eds., Plan resources of Souh-Eas Asia. No. 8. Vegeables. Pudoc Scienific Publishers, Wageningen, The Neherlands. Joshi, P Ehnoboany of he primiive ribes in Rajashan. Rupa Offse Priners, Jaipur India. Keraudren-Aymonin, M Cucurbiacées. Pages in A. Aubréville and J.-F. Leroy, eds., Flore du Cambodge Laos e du Viê-Nam. Muséum Naional D Hisoire Naurelle, Paris, France. Li, H. L The origin of culivaed plans of Souheas Asia. Economic Boany 24:3 19. Neuwinger, H. D African ehnoboany: Poisons and drugs. Chapman and Hall, London, England. Ogbonna, J. C., Y. C. Li, Y. K. Liu, and H. Tanaka Loofa (Luffa cylindrica) sponge as a carrier for microbial cell immobilizaion. Journal of Fermenaion and Bioengineering 78: Pal, D. C Plans used in reamen of cale and birds among ribals of Easern India. Pages in S. K. Jain, ed., Glimpses of Indian ehnoboany. Oxford and IBH Publishing Co., New Delhi, India. Porerfield, W. M Luffas as hey are used by he Chinese. Journal of he New York Boanical Garden 44: Prain, D Bengal plans: A lis of he phanerogams, ferns and fern-allies indigenous o, or commonly culivaed in, he Lower Provinces and Chiagong, wih definiions of he naural orders and genera, and keys o he genera and species. Wes, Newman and Co., London, England. Ramsey, S. R The languages of China. Princeon Universiy Press, Princeon, NJ. Roder, W., B. Keoboulapha, K. Vannalah, and B. Phouaravanh Gluinous rice and is imporance for hill farmers in Laos. Economic Boany 50: Roy, R. P., and S. Saran Sex expression in he Cucurbiaceae. Pages in D. M. Baes, R. W. Robinson, and C. Jeffrey, eds., Biology and uilizaion of he Cucurbiaceae. Cornell Universiy Press, Ihaca, NY. Schules, R. E Biodynamic cucurbis in he New World ropics. Pages in D. M. Baes, R. W. Robinson, and C. Jeffrey, eds., Biology and uilizaion of he Cucurbiaceae. Cornell Universiy Press, Ihaca, NY. Singh, V., and R. P. Pandey Ehnoboany of Rajashan, India. Scienific Publishers (India), Jodhpur. Sinno, E. W., and R. Bloch Luffa sponges, a new crop for he Americas. Journal of he New York Boanical Garden 44: Swofford, D. L., and R. B. Selander BIOSYS- 1: A compuer program for he analysis of allelic variaion in populaion geneics and biochemical sysemaics (Release 1.7). Naural Hisory Survey, Champaign, IL. Sysa Version for windows. SPSS Inc. 7/