GALLEY 66 File # 10Em

ECONOMIC BOTANY Wednesday Jan 14 2004 01:03 PM ebot 58_110 Mp_66 Allen Press • DTPro System GALLEY 66 File # 10em

CONSERVATION OF USEFUL PLANTS:AN EVALUATION OF LOCAL PRIORITIES FROM TWO INDIGENOUS COMMUNITIES IN EASTERN PANAMA1

SARAH PAULE DALLE AND CATHERINE POTVIN

Dalle, Sarah Paule (Department of Plant Science, Macdonald Campus of McGill University, 21111 Lakeshore Dr., Ste-Anne-de-Bellevue, QueÂbec, Canada H9X 3V9; e-mail: sar- [email protected]) and Catherine Potvin (Biology Department, McGill University, 1205 ave Dr. Pen®eld, MontreÂal, QueÂbec, Canada H3A 1B1). CONSERVATION OF USEFUL PLANTS:AN EVALUATION OF LOCAL PRIORITIES FROM TWO INDIGENOUS COMMUNITIES IN EASTERN PANAMA. Economic Botany 58(1):000±000, 2004. On both theoretical and practical grounds, respect for, and inclusion of, local decision-making processes is advocated in conservation, yet little is known about the conservation priorities on local territories. We employed interviews and ecological inventories in two villages in order to (1) evaluate the local perception of the conservation status of important plant resources; (2) compare patterns of plant use; and (3) compare perceived conservation status with population structure and abundance in the ®eld. One-third of the 35 species examined were perceived to be threatened or declining. These were predominantly used locally for construction or sold commercially, but were not necessarily rare in the ®eld. The destructiveness of harvest was the most consistent predictor of conservation status in both villages. Contrasting patterns were found in the two villages for the frequency of plant harvest and the relationship of this variable with conservation status. We suggest that local knowledge is an ef®cient means to rapidly assess the status of a large number of species, whereas population structure analysis provides an initial evaluation of the impact of harvest for selected species.

CONSERVACIOÂ NDEPLANTAS UÂ TILES:UNA EVALUACIOÂ NDEPRIORIDADES LOCALES EN DOS LOCAL- IDADES DEL ESTE DE PANAMAÂ . Tanto desde una perspectiva teoÂrica como praÂctica el respeto e inclusioÂn de los procesos locales de toma de decisiones es una forma en que la conservacioÂn puede ser promovida. Sin embargo, poco se sabe sobre las prioridades de conservacioÂn en territorios indõÂgenas. En el presente estudio se emplearon entrevistas e inventarios ecoloÂgicos en dos localidades indõÂgenas para (1) evaluar la percepcioÂn de los habitantes locales sobre el estado de conservacioÂn de recursos vegetales importantes; (2) comparar los patrones de uso de plantas; y (3) comparar la percepcioÂn del estado de conservacioÂn con la estructura de las poblaciones y la abundancia de las especies en el campo. Una tercera parte de las 35 especies estudiadas fueron percibidas como amenazadas o en proceso de declinacioÂn. Se trata de plantas utilizadas principalmente como materiales de construccioÂn o que aportan productos que son comercializados, pero que no son necesariamente escasas en el campo. En ambas localidades la variable que predijo el estado de conservacioÂn de forma maÂs consistente fue el grado de destruccioÂn de las plantas asociado a las praÂcticas de cosecha. Encontramos patrones con- trastantes entre las dos localidades con respecto a la frecuencia de cosecha y a la relacioÂn de esta variable con el estado de conservacioÂn de las plantas. Sugerimos que el uso de conoci- mientos locales es una forma e®ciente de evaluar con rapidez el estado de un gran nuÂmero de especies, en tanto que el anaÂlisis de la estructura de poblaciones aporta una evaluacioÂn inicial sobre el impacto de la cosecha para algunas especies de intereÂs. 1? Key Words: conservation, comparative ethnobotany, indigenous territories, local knowledge, Panama, population structure, rapid assessment, useful plants.

Efforts to conserve tropical forests are in- grounds, respect for, and inclusion of, local de- creasingly seeking the involvement of local cision-making processes is advocated (Norton communities. On both theoretical and practical 2001; Western et al. 1994), and it is increasingly recognized that cultural perception is an impor- 1 Received 29 June 2001; accepted 04 April 2003. tant component of conservation action. Further-

Economic Botany 58(1) pp. 000±000. 2004 ᭧ 2004 by The New York Botanical Garden Press, Bronx, NY 10458-5126 U.S.A. ECONOMIC BOTANY Wednesday Jan 14 2004 01:03 PM ebot 58_110 Mp_67 Allen Press • DTPro System GALLEY 67 File # 10em

2004] DALLE & POTVIN: CONSERVATION OF USEFUL PLANTS

more, conservationists are now seeking alliances knowledge in the two villages and identify with indigenous peoples, whose territories en- ecological and use characteristics associated compass a large proportion of the remaining for- with declining or threatened species; est in Latin America and elsewhere (Colchester (2) Contrast patterns of plant use in the two vil- 2000; Peres 1994; Schwartzman et al. 2000). lages, especially the frequency of harvest; One way in which local involvement may be and achieved is to focus conservation efforts on re- (3) Compare the perceived conservation status sources important to local economies (Ticktin et with population structure and abundance in al. 2002). A number of examples exist of local the ®eld. people adopting conservation actions (e.g., Our discussion then considers the utility of local Gadgil et al. 1993; Pinedo-Vasquez et al. 1990; knowledge and participation for plant conser- Ventocilla et al. 1995), and many of these are vation and makes recommendations for further motivated by perceived threats to local resource research. bases. Ethnobotany, by focusing on the knowledge, use, and management of plants by people, STUDY AREA can provide an understanding of the nature of The two focus villages of this study are situ- management problems and the potential for con- ated approximately 75 km apart. Ipeti is an Em- servation on local territories. Local knowledge bera village located along the Ipeti River near about the status of plant resources, for example, the foothills of Serrenia Maje (9Њ00Ј N, 78Њ05Ј has begun to be explored as a way to determine W), province of Panama, while Nurna is a Kuna conservation priorities (Potvin et al. 2002). Hel- village located near the headwaters of the Chu- lier and colleagues (1999) examined the use of cunaque River in the province of Darien (9Њ00Ј local knowledge for identifying declines in ani- N, 78Њ03Ј W). Both territories are covered by mal and plant species, as well as in forest cover forests classi®ed as ``humid tropical'' in the in highland Mexico, whereas vegetation change Holdridge life zone system (Instituto GeograÂ®co in an African savannah was assessed using local Nacional 1988) and range in elevation from ap- perceptions and population structure analysis by proximately 50 to 300 m above sea level. Av- Lykke (1998, 2000). Ethnobotanical research erage annual precipitation in the area is around can contribute to such an approach by charac- 2500 mm with annual temperature averages of terizing patterns of plant use and management 25ЊC (Instituto GeograÂ®co Nacional 1988). and investigating how these relate to plant de- There is a pronounced dry season from Decem- clines or conservation priorities. ber to April (Tosi 1971). In this paper we present results from a study Whereas both villages are found in ecologi- aimed at examining the conservation priorities cally similar regions, they differ in the degree of for plant resources of two villages in the bio- anthropogenic alteration of their surroundings. geographic region of Darien, Eastern Panama. Ipeti is a recent village organized as an indige- Darien is home to three indigenous groups who nous ``Tierras colectivas'' of 3198 ha. When the have cohabited on these lands for hundreds of Bayano hydroelectric reservoir was ¯ooded in years (the Kuna, the Embera, and the Wounaan) the mid-1970s, the Embera people living in the and is a global priority for forest conservation area were relocated to Ipeti and to the neigh- (Davis et al. 1997). Problems of resource dete- boring community of Piriati, both along the Pan- rioration and over-exploitation on indigenous american highway (Wali 1993). Travel to Ipeti territories in the region have been reported (Her- from Panama City takes ®ve hours by bus. The lihy 1986; Ventocilla et al. 1995). We employed largely forested indigenous territory is embed- a participatory approach with local people in one ded in a matrix of land used by colonists for Kuna and one Embera community to examine cattle ranching. There are 50 houses in Ipeti and the conservation status of traditional plant re- we estimate the population to be around 400 sources. The two villages studied differ substan- people. Most inhabitants of Ipeti speak Spanish tially in land-use history. Speci®cally, our goals with the exception of some older women; the were to: younger generation tends to use Spanish rather than Embera in their daily activities. The local (1) Evaluate the conservation status of impor- economy appears rather diversi®ed. Men and tant plant resources as determined by local women work for day wages in various colonist ECONOMIC BOTANY Wednesday Jan 14 2004 01:03 PM ebot 58_110 Mp_68 Allen Press • DTPro System GALLEY 68 File # 10em

ECONOMIC BOTANY [VOL. 58 PETI AND I Fiber Leaves II. Ipeti (Embera) Name Principal use Part ÐÐ Ð Chunga Basketry Leaves trunk trunk Fruit, Latex Ð Ð Ð Leaf, fruit Yi Fruit, Trunk Ð Ð Ð dolls work wall timber, canoe I. Nurna (Kuna) Hand beater Branch Ð Ð Ð Edible, house Roof post Trunk Ð Ð Ð sapi sichit Name Principal use Part Life form voucher* Herbarium Bromeliaceae NSArecaceae UC 9a,c Herb Palm Oa Naba Ritual, edible Cyclanthaceae SD 238a Herb Nai-uar Basketry Petiole Nawala Basketry Petiole Fabaceae SD 200a,b Tree Baila-uar House post Trunk Ð Ð Ð Bignoniaceae SD 237a,b Liana Magep Ritual Leaf Ð Ð Ð Arecaceae SD 204a,d Palm Alar-uar Pot holder Stem Ð Ð Ð Boraginaceae SD 227aLecythidaceae SD 239a,b Tree Tree Ugar-uar Firewood Tupu-uar Edible Trunk Ð Fruit Ð Ð Ð Ð Ð Â King Meliaceae SD 236a,b Tree Kaoban Construction, L. f. Guttiferae SD 234a,b Tree Mutu Tool for bead- Andre (L. Ruiz & Pittier Fabaceae SD 223b Tree Ipup Firewood Trunk Ð Ð Ð L. H. (Kunth) Aubl. Tiliaceae SD 229a,b Tree Dubsip Rope ®ber Bark Ð Ð Ð (Bonpl.) (Ruiz & THNOBOTANICAL CHARACTERISTICS OF PLANTS SELECTED FOR STUDY BASED ON THEIR IMPORTANCE TO THE LOCAL CULTURE IN Species Family sp. Annonaceae SD 201a,c Tree Uichur- 1. E . ex Baker L. H. Bailey Pav. Harms) B. Verl. Bailey Pav.) Oken O. Berg URNA ABLE No idNo id Rutaceae Bombacaceae NS SD 233a,b Tree Tree Puarsip House post Sianele Edible, ritual Trunk Ð Ð Ð No idII. Plants in NurnaAechmea and magdalenae Ipeti Sterculiaceae SD 207a,b Tree Bunur- Symphonia globulifera Astrocaryum standleyanum Carludovica palmata Peltogyne purpurea Swietenia macrophylla Oxandra I. Plants only inApeiba Nurna tibourbou Arrabidaea chica Bactris coloradonis Cordia alliodora Gustavia superba Myroxylon balsamum T N ECONOMIC BOTANY Wednesday Jan 14 2004 01:03 PM ebot 58_110 Mp_69 Allen Press • DTPro System GALLEY 69 File # 10em

2004] DALLE & POTVIN: CONSERVATION OF USEFUL PLANTS Bark Trunk & Trunk Trunk Fruits Trunk no specimen collected. ϭ rope pot holder basketry ritual nial II. Ipeti (Embera) Name Principal use Part deposited at NYBG, NS ϭ Maquenque Construction, trunk Petiole, Fruit KiparaTrunk Body painting, Balsa Ladder, ceremo- deposited at SCZ, d ϭ . house walls ing post, ritual dolls I. Nurna (Kuna) Cosmetic, ritual deposited at FMNH, c ϭ ONTINUED Name Principal use Part 1. C ABLE T Life form deposited at PMA, b ϭ voucher* Herbarium Marantaceae UC 1a,c Herb Ð Ð Ð Bijao Cooking Leaves ArecaceaeArecaceae JH 2a,c UC 4a,c Palm Palm Soso Ila Thatch House walls Leaf Trunk Guagara Jira Thatch Floor Leaves Trunk Araceae SD 210a,b HemiepiphyteArecaceae Sargi UC Latching 5a,c Palm Roots Bejuco real Siler Basketry Basketry, Roots (H. H. Mart. Arecaceae NS Palm Ð Ð Ð Trupa Drink & oil Fruits (Mart.) L. Rubiaceae Tree Sabdur Aubl. Annonaceae UC 6a,c Tree Ð Ð Ð Malagueto Construction & Sw. Bombacaceae UC 2a,c Tree Ukur-uar Ladder, carry- (Willd. Hemsley Fabaceae SD 205a,b Tree Koibur Construction Trunk Cocobolo Black dye Trunk L. Meliaceae UC 8a,c Tree Ð Ð Ð Cedro amargo Canoe Trunk sp. Bombacaceae NS Tree Ð Ð Ð Cedro espino Canoe Trunk Species Family sp. Sapotaceae NS Tree Ð Ð Ð Nispero Construction Trunk sp. Bromeliaceae NS Herb Ð Ð Ð Pita sin espina Fiber Leaves sp. Rubiaceae NS Liana Ð Ð Ð Kidave Toothpaste ? sp. Arecaceae NS Palm Ð Ð Ð Uvita Construction & ex Link) Karst) Griseb. H. Wendl. H. Wendl. Kunth Karst * Collectors are Sarah Dalle (SD), Ultiminio Caisamo (UC) and Jane Hutton (JH); a No idNo idNo id Gramineae NS Bamboo NS NS Ð Shrub Ð Liana Ð Ð Ð Ð Ð Chiru Ð Ð Flute Tinta roja Motete Red Trunk dye Basketry Leaves Stem Bombacopsis Calathea latifolia Xylopia frutescens Cedrela odorata Manettia Manilkara Oenocarpus bataua Sabal mauritiiformis Socratea exorrhiza III. Plants in IpetiAechmea only Bactris Dalbergia retusa Genipa americana Heteropsis oblongifolia Oenocarpus mapora Ochroma lagopus ECONOMIC BOTANY Wednesday Jan 14 2004 01:03 PM ebot 58_110 Mp_70 Allen Press • DTPro System GALLEY 70 File # 10em

ECONOMIC BOTANY [VOL. 58

enterprises such as farms, logging companies, or EmberaÂ). Research in Nurna was coordinated private houses. All households also engage in with support from FundacioÂn Dobbo-Yala, a subsistence agriculture, ®shing, and hunting. In Kuna nongovernmental organization that has the past two decades, basket weaving (for wom- been working in the area since 1996 (see Lopez en) and, to a lesser extent, wood carving (for and Dalle 2001 for more details). men) have become an important source of sup- plementary income. The intensi®cation of bas- SPECIES SELECTION ketry raised the level of consciousness of the We worked with inhabitants in both villages community regarding resource availability, and to select species considered important to the lo- this concern facilitated the present study. cal culture. We did not focus on medicinal According to oral history, Nurna was estab- plants, as this topic was considered sensitive in lished approximately 70 to 100 years ago when both communities. In Ipeti, we conducted a one- a ®re destroyed a previous settlement located day workshop. Twenty men and 20 women met higher up in the San Blas mountains. Nurna is separately to establish a list of plants they con- part of the recently legalized (2000) Comarca sidered most important for Embera culture. The Wargandi. The village's traditional lands are ex- workshop yielded a list of some 38 species with tensive (approximately 10 000 ha). There is no 11 species overlapping between the respective road access to Nurna; during the time of the pre- lists of men and women. Twenty-two of these sent study (1998±1999), reaching Nurna from species were selected for our study. The criteria the city of Panama involved an eight-hour bus for inclusion were representation of male and ride and an eight-hour hike. Nurna is surrounded female choice, diversity of life form, uses, and by largely intact forests. Large predators, e.g., harvested plant parts. In Nurna, species were se- jaguar (Panthera onca) and harpy eagle (Harpia lected from an initial list of 70 species obtained harpyja), are still reported in the region of Nur- from free-listing interviews with two men and na by villagers (®eld observation). The popula- two women, as well as a group of ®ve young tion of 200 people is predominantly Kuna- men who volunteered as informants. Informants speaking. Only a few people (four younger men) were asked to list all species known in each of speak Spanish and can read and write. The local ®ve use categories (construction, weaving/utili- economy is based on swidden/fallow agriculture ty, edibles, ®rewood, ritual). This list was re- complemented by permaculture agroforestry viewed for accuracy with the ``Agriculture Com- along the riverbanks, hunting, and, to a lesser mittee,'' a group of four men who, among other degree, ®shing. Since approximately 1995, cash responsibilities to the community, were charged income has been obtained almost exclusively with coordinating with the research team. They through contracts established with Panamanian were asked to prioritize species in each category logging companies for the commercial extrac- according to their perceived importance. The ®- tion of valuable timber species, mainly Swieten- nal list of 23 plant species was selected from the ia macrophylla. Previously, cash income was ``priority'' species by maximizing diversity in much more limited and was derived from oc- uses, life form, and use by men/women. In both casional timber sales, as well as that of Ipecac villages, species were identi®ed using local ¯ora (Psychotria ipecacuanha) in the medicinal herb (Croat 1978; Gentry 1996; Henderson et al. markets in Panama City. At the onset of this 1995). Voucher specimens were collected and project, most inhabitants of Nurna showed rel- deposited at the the Herbario Nacional de Pan- atively little preoccupation for the status of their amaÂ (PMA) or sent to other herbaria for identi- resources, feeling that the surrounding forests ®cation. In a few cases we were unable to obtain were plentiful. a scienti®c identi®cation, largely because the species were never encountered in the ®eld. METHODS These species are referred to in this paper by In both villages the objectives and progress of their local names. The ®nal list of species from the research was presented and discussed with the two villages is shown in Table 1. the communities in the village congress. In Ipeti, work was coordinated directly with the local or- ECOLOGICAL INVENTORIES ganization OUDCIE (OrganizacioÂn para la uni- Ecological inventories were carried out to dad y el desarrollo de la comunidad de Ipeti- study the species abundance and distribution in ECONOMIC BOTANY Wednesday Jan 14 2004 01:03 PM ebot 58_110 Mp_71 Allen Press • DTPro System GALLEY 71 File # 10em

2004] DALLE & POTVIN: CONSERVATION OF USEFUL PLANTS

both Nurna and Ipeti. Dalle et al. (2002) pro- the ®eld with S. Dalle; this team was accompa- vides a detailed description of the ®eld sampling nied by other men from the village. methodology. Brie¯y, the sampling consisted of a strati®ed random sampling design, in which a INTERVIEWS total of 50 transects in Ipeti and 52 transects in Interviews were carried out to characterize Nurna were laid out across an area of approxi- patterns of resource use and conservation status mately 3000 ha surrounding each village. In Ipe- in the two villages. Cultural differences between ti, this corresponded to the entire Tierras colec- Ipeti and Nurna forced us to conduct the eth- tivas whereas in Nurna it encompassed the re- nobotanical component in a different way in gions most frequently used for wild harvesting each community. In general, due to a number of of forest products, as well as some unharvested factors, including language and literacy, as well areas. The abundance of the species and habitat as experience with previous research and for- characteristics were recorded in three 12-m ra- eigners, access to people, especially women, 2 dius circular plots (452 m ) located at 250-m in- was greater in Ipeti than in Nurna. In Ipeti, eth- tervals along each transect. For trees and palms, nobotanical information was ®rst collected by abundance was recorded separately for each of means of a formal questionnaire focusing on ®ve ®ve size classes. Size classes (SC) for the trees key areas: knowledge transmission, plant phe- were de®ned as follows: SC1 Ͻ 0.5 m height, Ͼ Ͻ nology, location where found, harvest practices, SC2 0.5 m height and 10 cm dbh, SC3 10± and use. The questionnaires were implemented 20 cm dbh, SC4 20±40 cm dbh, and SC5 Ͼ 40 by two older women (Ͼ 50 years) and two youn- cm dbh. For the palms, size classes were SC1 Ͻ ger women (Ͻ 30 years) chosen by the com- 0.5 m height, SC2 Ͼ 0.5 m height, trunkless, munity. In each household, men were inter- SC3 trunk Ͻ 3 m height, SC 4 3±7 m height, viewed by one team while the second team in- SC5 Ͼ 7 m height. SC1 was recorded in a 3-m terviewed the women to minimize gender com- radius sub-plot. In the case of Socratea exorrhi- petition in answers. On any single day, a za, which has trunk development from the seed- questionnaire focusing on four to six plant spe- ling stage, SC2 was de®ned as palms 0.5 to 1 m cies was applied to 16 different community tall, and SC3 as 1 to 3 m tall. In this paper data members. A total of 236 interviews were carried on only the established size classes (2±4) are employed. In Nurna, data from three additional out between June and October 1998. More than transects (12 sampling sites), positioned in re- 90% of the 50 households in the village partic- gions of non-harvested Sabal mauritiiformis ipated. Participants ranged from 18 to 74 years populations, were included in the analyses of in age and had lived in Ipeti for varying lengths population structure. of time. The environmental variables recorded in each In January 1999, further interviews with focus plot were elevation, land-use, forest structure, groups were carried out in Ipeti in order to es- drainage, and soil color and texture. In Nurna, tablish consensus on six key questions: (1) How information was also recorded on harvest activ- much plant material is used for each species? (2) ities at each site. In Ipeti, ®eld work was initi- How often is the plant harvested? (3) Which part ated in June 1998 and ended in November of of the plant is most commonly used? (4) How that year, whereas in Nurna ®eld work was con- is it collected? (5) What is the gender difference ducted between September 1998 and April 1999. in work regarding these plants? (6) Which spe- In both villages the sampling order of the tran- cies should be targeted for conservation action? sects was randomized in order to control for sea- For questions one and two, the informants sonal and/or temporal differences. In Ipeti, the played a pile-sorting game (Bernard 1995) in village appointed a young man with a high which a series of cards representing the 22 spe- school professional degree in agronomy who cies were ordered according to the quantity and was trained during June 1998. Once suf®cient frequency they were perceived to be harvested. competence was gained in the ®eld techniques, For question six, informants were asked to put he teamed up with another man and they started together the species that were abundant, species the work independently. In Nurna, the village for which they foresaw a problem of supply, and congress nominated two men who were trained the species that were absent from the Tierras co- by S. Dalle. The two men alternated working in lectivas. The result recorded was the consensus ECONOMIC BOTANY Wednesday Jan 14 2004 01:03 PM ebot 58_110 Mp_72 Allen Press • DTPro System GALLEY 72 File # 10em

ECONOMIC BOTANY [VOL. 58

reached among all participants of one sex as with most experience with the species was used; they arranged the cards. when one gender claimed ignorance, the percep- In Nurna, structured interviews were used to tion of the other gender was employed. For Nur- obtain information on the most important ques- na, consensus was de®ned as a signi®cantly (P tions identi®ed from the Ipeti interviews. (1) Ͻ 0.10) non-random distribution in interview re- How frequently is the species used? (2) In what sponses, as assessed by chi-square tests. Abun- quantities is it collected? (3) Has the abundance dance was estimated from the ecological inven- of this species changed over your lifetime? In- tory as the total number of individuals recorded, formants were asked to identify each species and was transformed using the equation orally in the following categories: frequently/not ln(abundance ϩ 1), in order to achieve a normal frequently used, large quantity/small quantity, distribution and minimize outliers. Ecological declining/not declining. Interviews were carried guild, for the purposes of our study, was de®ned out in Kuna by three researchers from Funda- as pioneer, mid-successional/disturbance tolerant cioÂn Dobbo-Yala during a one-week period in or late successional. These classi®cations were October 1999. Because people (particularly derived from canonical ordinations of species women) in Nurna were more wary of partici- versus habitat from the ecological inventory pating in household surveys, the team worked (Dalle et al. 2002), as well as information in the with informants with whom they had closer re- literature and communication with other ecolo- lationships and who represented a range of ages. gists working in the region. A list of the codes One interview was carried out by S. Dalle with used in the data matrix is included in Appendix 1. one of the two men who worked in the ecolog- Stepwise discriminant analysis was used to ex- ical inventory. Fifteen male informants (30% of amine which variables from the use/ecology ma- the male population) were interviewed, half of trix best predicted the perception of conservation whom were over 40 years of age. status in both Ipeti and Nurna. The procedure was Participant observation was used to obtain ad- conducted using the program CANOCO for Win- ditional information on the plant part used and dows v.4 (ter Braak and Smilauer 1998), which methods of harvest in Nurna. Opportunities for provides Monte Carlo permutation tests for tests observation included witnessing harvesting and of signi®cance. The ethnobotanical characteristics agricultural activities during ®eld work for the considered were plant part used, frequency of ecological inventory, occasional outings with harvest, quantity harvested, mode of harvest (de- both men and women to collect ®rewood or structive or not), and degree of management (cul- work in agricultural ®elds, and living with two tivated and/or spared or not). The ecological char- different families. For both Ipeti and Nurna we acteristics included were life-form (woody or used participant observation and informal inter- not), guild, and ln(abundance). Relationships be- views to determine whether harvest for each tween the frequency of harvest and conservation species was non-destructive or not, and whether status with other variables in the use/ecology ma- species were managed through cultivation or trix were examined by means of Pearson corre- sparing. Harvest was considered non-destructive lation coef®cients (Systat v.9 for Windows, SPSS when the individual (e.g., ramets for clonal spe- Science, Chicago). cies) survived the harvest episode. To examine whether the frequency of harvest of species was related to characteristics of the STATISTICAL ANALYSIS habitats where they occur, the Fourth Corner Multivariate statistics were used to explore method (Legendre et al. 1997) was also used. the relationships of use and ecological variables This method allows correlations between habitat with their perceived conservation status. To pre- and species characteristics to be calculated, pare the data for analysis, a matrix of species based on the presence/absence of species at a versus use and ecological traits was developed. range of sampling stations. Here, the habitat For Ipeti, variables concerning conservation sta- characteristics of interest were successional tus and use characteristics were derived from the stage and the distance from the village of plots consensus reached in the card-ordering games. in the ecological inventory (used as an estimate For most analyses, men's and women's percep- of resource availability), whereas the species tions were pooled. In the case of disagreement characteristic examined was the frequency of between the sexes, the perception of the gender harvest. The signi®cance of the Pearson corre- ECONOMIC BOTANY Wednesday Jan 14 2004 01:03 PM ebot 58_110 Mp_73 Allen Press • DTPro System GALLEY 73 File # 10em

2004] DALLE & POTVIN: CONSERVATION OF USEFUL PLANTS

TABLE 2. SIMILARITY OF USE CHARACTERISTICS FOR THE TEN PLANTS COMMON TO IPETI AND NURNA. DIFFERENCES BETWEEN VILLAGES ARE CODED AS ``0'' AND AGREEMENTS AS ``1''.

Conservation Principal Harvest Plant part Abundance status use frequency Total Carludovica palmata 1 1 1 1 1 5 Genipa americana 1 1 1 1 1 5 Ochroma lagopus 1 1 1 1 1 5 Oenocarpus mapora 1 1 1 1 0 4 Sabal mauritiiformis 1 1 1 1 0 4 Socratea exorrhiza 1 0 1 1 0 3 Aechmea magdalenae 1 0 0 0 1 2 Dalbergia retusa 1 1 0 0 0 2 Heteropsis oblongifolia 1 0 1 0 0 2 Astrocaryum standleyanum 0 1 0 0 0 1 Total 97764

lation coef®cient between these variables was iables. Forward selection was used to select the tested using a Monte Carlo permutation test most important environmental variables for the (method 1, following Legendre et al. 1997) with CCA of population structure versus environ- 9999 permutations. ment. Finally, a partial CCA served to evaluate As discussed by Hall and Bawa (1993), com- the relationship between population structure parison of population structure between harvest- and harvest zone, controlling for the effect of ed and unharvested sites can provide an initial the environmental variables. In the case of one assessment of the impact of harvest on plant species (Sabal mauritiiformis) for which we populations. Here this analysis was carried out found a signi®cant effect in this last analysis, we using data on eight species from the ecological also calculated the amount of variance attribut- inventory in Nurna. To determine harvested and able to harvest zone, environmental variables, unharvested zones, maps of harvesting activity and a combination of these two, following the were examined by means of visualization in the method proposed by Borcard et al. (1992). GIS program Arcview, version 3.1 (ESRI, Cal- ifornia). Plots were coded as 1 (harvest occurs) RESULTS or 0 (no harvest known to occur) for each spe- RESOURCE USE cies, and a surface of harvest intensity was in- terpolated using a second-order Inverse Distance Thirty-®ve species from 20 families were se- Weighting (IDW) algorithm, based on the 12 lected for their cultural importance in the two nearest neighbors. These were compared with villages (Table 1). Of these, 10 were common to maps of harvest areas drawn with the two ®eld the two villages. These were most similar in assistants. Harvest zones were then determined terms of the plant part used, abundance, and for each individual species in relation to its dis- conservation status, and least similar in terms of tribution. To do this, the occurrence of each spe- the frequency of harvest and principal use (Table cies was overlaid on the corresponding harvest 2). Of the 10 species, the main use differed for intensity map and a ``harvest'' and ``control'' Astrocaryum standleyanum, Heteropsis oblon- zone de®ned. In the case of Sabal mauritiifor- gifolia, Aechmea spp., and Dalbergia retusa. mis, a third zone, ``past harvest,'' was de®ned The ®rst two species are used for basket weav- for plots near the village where the species oc- ing in Ipeti, but in Nurna serve primarily as curred but was no longer harvested. latching for house construction and for food, re- To compare population structures among har- spectively. However, historical or minor uses vest zones, a series of canonical correspondence converge in some cases. In Nurna, for example, analyses (CCA) were conducted on the popula- H. oblongifolia is occasionally used for basket- tion structure matrix (sites ϫ species abundance ry, while one man in Nurna reported that A. in each size class) to evaluate the relationship standleyanum was employed for weaving of with (a) harvest zone and (b) environmental var- hammocks in the past. Similarly, according to ECONOMIC BOTANY Wednesday Jan 14 2004 01:03 PM ebot 58_110 Mp_74 Allen Press • DTPro System GALLEY 74 File # 10em

ECONOMIC BOTANY [VOL. 58

oral history, `Òa'' (Aechmea magdalenae) was tance from the village. The negative sign on the used by the Kuna for thread (Salvador 1997), r-value indicates that early successional habitats corresponding to the Embera use. The species had more frequently harvested species, which is for which the main use converged were primar- consistent with the correlation analysis above. In ily used for construction (e.g., Sabal mauritii- Nurna, a signi®cant relationship was found only formis, Socratea exorrhiza), basketry (e.g., Car- for distance (r ϭϪ0.083, P Ͻ 0.0002), but not ludovica palmata, Oenocarpus mapora), and rit- for the stage of succession. Thus, closer but not uals (Ochroma lagopus, Genipa americana). necessarily early successional sites were more The frequency of harvest varied substantially likely to have frequently harvested species. among species in both Ipeti and Nurna. For Ipe- ti, the most frequently harvested species were LOCAL PERCEPTION OF Calathea latifolia and Carludovica palmate, CONSERVATION STATUS both on a weekly basis (Table 3). Ochroma la- In Ipeti, women's perception of the resources gopus, Dalbergia retusa, Genipa americana, As- was statistically correlated with that of men (Ta- trocaryum standleyanum, and ``Tinta roja'' were ble 4). In addition, both sexes' perception of reported to be harvested on a monthly basis. In species status was negatively correlated to the Nurna, seven species were considered to be frequency of harvest and abundance. Men's per- ``frequently'' harvested (at least several times a ception of status was more strongly associated year), whereas nine species were classi®ed as with ecological guild than the women's, al- infrequent (Table 3). The most frequently har- though the latter showed a similar trend (Table vested species were Carludovica palmata and 4). These results suggest that both sexes per- Sabal mauritiiformis. The chi-square analysis received plant status in a similar way. Overall, vealed signi®cant agreement (P Ͻ 0.10) among nine species were identi®ed as being absent respondents on the frequency of harvest of 16 from the Tierras Colectivas, and seven as pri- of the 23 species. The lack of consensus for the orities for conservation efforts (Table 6). In Nur- remaining seven species likely re¯ects differenc- na, statistically signi®cant consensus on peres in preference for, or experience with, partic- ceived changes in abundance was reached ular products (e.g., Genipa americana and Bac- among the 15 informants on a total of 11 spe- tris coloradonis are not harvested by everyone). cies: six species were considered declining, In subsequent analyses these species were con- while no decline was perceived for ®ve species sidered to be used with ``moderate'' frequency. (Table 6). To examine the relationship between harvest Stepwise discriminant analysis was performed frequency and availability, we evaluated the cor- to identify which ethnobotanical and ecological relation of harvest frequency with species abun- characteristics of species could best predict their dance, distance from the village, and species perceived status. For Ipeti, three variables were habitat. The frequency of harvest in Ipeti cor- found to be signi®cant in the forward selection related signi®cantly with species abundance as procedure: ln(abundance) (F ϭ 9.1, P ϭ 0.005), well as with guild (Table 4), suggesting that in non-destructive harvest (F ϭ 9.47, P ϭ 0.005), Ipeti the most frequently harvested species tend and harvest frequency (F ϭ 9.1, P ϭ 0.005). The to be abundant, early successional species. discriminant functions based on these three var- These relationships hold up when the seven spe- iables produced a signi®cant classi®cation (F ϭ cies absent in the ecological inventory are re- 11.8, P ϭ 0.001) and successfully classi®ed all moved from the analysis (data not shown). In but one of the species into the three groups: Nurna, however, no signi®cant correlation was `àbundant,'' ``danger,'' and `àbsent.'' The re- detected between frequency of harvest and ei- lationships among the three explanatory vari- ther of these variables (Table 5). Results from ables and the categories of perception are shown the fourth corner analysis revealed marked dif- in Fig. 1. Here, the ®rst axis (which represents ferences between the two villages in the rela- the ®rst discriminant function) serves to separate tionship of harvest frequency with habitat char- the `àbundant'' species from the `àbsent'' spe- acteristics. In the case of Ipeti, a signi®cant re- cies. In addition, species perceived to be `àbun- 4? lationship was found between frequency of har- dant'' tended to be frequently but non-destruc- vest and the successional stage (r ϭϪ0.15, P ϭ tively harvested. The group ``danger'' is posi- 0.0001), but no correlation existed with the dis- tioned in the positive portion of the second axis, ECONOMIC BOTANY Wednesday Jan 14 2004 01:03 PM ebot 58_110 Mp_75 Allen Press • DTPro System GALLEY 75 File # 10em

2004] DALLE & POTVIN: CONSERVATION OF USEFUL PLANTS 6 6 6 6 3 3 3 3 5 5 3 3 3 4 3 3 1 1 1 1 2 2 Last harvest*** II. Ipeti 0 0 0 0 1 1 1 0 0 0 3 2 2 2 1 1 4 4 3 3 3 3 Frequency of harvest** LEVEL OF CONSENSUS AMONG INFORMANTS : PECIES COMMON TO BOTH VILLAGES APPEAR IN URNA .S Manilkara sp. Aechmea sp. Aechmea magdalenae Bactris sp. Bombacopsis sp. Sabal mauritiiformis Socratea exorrhiza Heteropsis oblongifolia ``Chiru'' Manettia sp. Astrocaryum standleyanum Oenocarpus mapora Xylopia fructescens Oenocarpus bataua ``Motete'' Cedrela odorata Calathea latifolia Carludovica palmata Genipa americana ``Tinta roja'' Ochroma lagopus Dalbergia retusa .N 1 1 1 1 1 2 2 1 1 1 1 3 2 2 2 2 2 3 3 3 3 3 3 code* Species Frequency Ð Ð 0.008 0.008 0.021 0.005 0.001 0.439 0.317 0.071 0.071 0.046 0.012 0.046 0.225 0.317 0.439 0.763 0.467 0.001 0.001 0.003 0.008 -value P Ͻ Ͻ AS REPORTED IN INTERVIEWS , 2 y ago), 5 (sp. not present), 6 (don't know). Ͼ PETI I SORTING GAMES WITH FOCUS GROUPS Ð Ð - Chi- 7.118 7.118 5.333 8.067 0.6 1 3.267 3.267 4 6.25 4 1.471 1 0.6 0.091 0.529 9 7.118 square statistic 11.267 13.235 12.25 I. Nurna URNA AND N 17 17 12 15 15 15 16 15 15 16 16 16 17 16 15 11 17 16 17 17 16 16 17 RESULTS FROM PILE : 6.7 PETI 17.6 17.6 16.7 13.3 40.0 37.5 26.7 26.7 25.0 18.8 75.0 64.7 62.5 60.0 54.5 41.2 94.1 93.8 87.5 82.4 2 years apart), 2 (yearly), 3 (monthly), 4 (weekly). harvest N 100.0 100.0 frequent reporting Ͼ .I % Informants SQUARE TESTS - REQUENCY OF HARVEST FOR SPECIES IN Species . sp. 3. F ABLE *** Last harvest codes Ipeti: 1 (days or weeks ago), 2 (2±6 months ago), 3 (1±2 years ago), 4 ( * Codes for Nurna:** 1 Frequency (infrequent), of 2 harvest (moderate), Ipeti: 3 0 (frequent). rare or never, 1 ( Apeiba tibourbou Gustavia superba ``Bunur'' ``Sianele'' Symphonia globulifera ``Puarsip'' Dalbergia retusa Peltogyne purpurea Swietenia macrophylla Aechmea magdaleneae Astrocaryum standleyanum Oxandra Genipa americana Socratea exhorriza Cordia alliodora Bactris coloradonis Myroxylon balsamum ANALYZED WITH CHI T Carludovica palmata Sabal mauritiiformis Oenocarpus mapora Heteropsis oblongifolia Arrabidaea chica Ochroma lagopus BOLDFACE ECONOMIC BOTANY Wednesday Jan 14 2004 01:03 PM ebot 58_110 Mp_76 Allen Press • DTPro System GALLEY 76 File # 10em

ECONOMIC BOTANY [VOL. 58

TABLE 4. PEARSON CORRELATION COEFFICIENT FOR PERCEIVED STATUS AND FREQUENCY OF HARVEST WITH VARIOUS ECOLOGICAL AND USE INDICES,IPETI.Nϭ 15±22 SPECIES.1 STATISTICAL SIGNIFICANCE:* P Ͻ 0.05, ** P Ͻ 0.01, *** P Ͻ 0.001.

Ln Status Status Harvest Abun- (women) (men) frequency Quantity Trunk Guild dance Status (women) 1 Status (men) 0.821*** 1 Frequency Ϫ0.672** Ϫ0.716** 1 Quantity .055 Ϫ.013 .179 1 Trunk 0.101 0.268 Ϫ0.320 Ϫ.244 1 Guild 0.400 0.649** Ϫ0.444* Ϫ.352 0.324 1 Ln.Abundance Ϫ0.509* Ϫ0.803*** 0.493** 0.209 0.096 Ϫ0.510* 1

1 Pairwise deletion was used to accommodate missing values for men's perceived status (3 species) and quantity harvested (4 species).

indicating its association with destructively har- along the second axis. Consensus tended to be vested species. Both the groups ``danger'' and lacking, therefore, for species that were either `àbundant'' are associated with higher abun- early successional and destructively harvested dance in the ecological inventory than the `àb- (e.g., Cordia alliodora, Apeiba tibourbou) or lat- sent'' category; abundance, however, does not er successional species with non-destructive har- serve to distinguish between ``danger'' and vest and/or moderate harvest frequency (e.g., `àbundant'' groups. Oenocarpus mapora, ``Bunur''). It is interesting For Nurna, the forward selection procedure to note that frequency of harvest appears to cor- resulted in three variables being signi®cant with relate with different processes in the two com- 199 permutations: frequency of harvest (F ϭ munities. In Nurna, the most frequently har- 4.71, P ϭ 0.01), non-destructive harvest (F ϭ vested species were perceived to be in decline, 4.14, P ϭ 0.01), and guild (F ϭ 2.34, P ϭ 0.13). whereas frequent harvest was associated with The resulting classi®cation correctly classi®ed the `àbundant'' group in Ipeti. In both villages, 18 of the 23 species (78%) and was statistically however, destructive harvest appears to be as- signi®cant (F ϭ 4.2, P ϭ 0.003, 999 permuta- sociated with species perceived as in ``danger'' tions). The ®rst two variables (frequency and or ``declining,'' non-destructive harvest) served to separate the groups ``decline'' and ``no decline'' along the POPULATION STRUCTURE ANALYSIS ®rst axis, such that species perceived to be de- In Nurna, visualization of harvest zones (Fig. clining tended to be frequently as well as de- 2) as well as sketch maps revealed signi®cantly structively harvested (Fig. 1). The third group different patterns of harvesting among forest (``no consensus'') held an intermediate position products. Timber products harvested for con- with respect to this ®rst axis, and was negatively struction and ®rewood were reported to be har- associated with successionary stage (``guild'') vested close to the village, whereas harvesters

TABLE 5. PEARSON CORRELATION COEFFICIENT FOR PERCEIVED STATUS AND FREQUENCY OF HARVEST WITH VARIOUS ECOLOGICAL AND USE INDICES,NURNA.Nϭ 23 SPECIES.STATISTICAL SIGNIFICANCE:*P Ͻ 0.05, ** P Ͻ 0.01, *** P Ͻ 0.001.

Harvest Ln Status frequency Quantity Trunk Guild abundance Status 1 Frequency 0.462* 1 Quantity 0.234 0.539** 1 Trunk 0.313 Ϫ0.1 Ϫ0.469* 1 Guild 0.143 Ϫ0.238 Ϫ0.081 0.04 1 Ln.Abundance 0.055 0.107 0.191 Ϫ0.091 Ϫ0.526* 1 ECONOMIC BOTANY Wednesday Jan 14 2004 01:03 PM ebot 58_110 Mp_77 Allen Press • DTPro System GALLEY 77 File # 10em

2004] DALLE & POTVIN: CONSERVATION OF USEFUL PLANTS 9 0 1 0 0 0 0 0 0 32 56 13 92 55 70 644 504 272 553 466 361 2071 2 2 2 2 1 1 1 1 1 1 3 3 3 3 3 3 3 3 3 2 2 2 Status (pooled) Abundance II. Ipeti ? ? 2 2 2 2 1 1 1 1 1 1 3 3 3 3 3 3 3 2 2 2 (men) Status . ? 3 3 2 3 2 2 2 1 1 1 1 1 1 1 1 3 3 3 3 3 3 LEVEL OF CONSENSUS AMONG : Status** (women) TRANSECTS URNA 50 .N sp. sp. sp. HE NAMES OF SPECIES COMMON TO THE TWO COMMUNITIES .T Ochroma lagopus Bactirs Aechmea magdalenae Aechmea Manilkara Cedrela odorata Astrocaryum standleyanum Sabal mauritiiformis Socratea exorrhiza Bombacopsis quinata Xylopia frutescens ``Motete'' Oenocarpus mapora Genipa americana Calathea latifolia Carludovica palmata Oenocarpus bataua ``Chiru'' Heteropsis obongifolia Manettia reclinata ``Tinta roja'' Dalbergia retusa 2 1 1 2 0 0 12 33 61 71 33 865 102 204 337 172 455 338 199 373 205 328 2380 1 1 1 1 1 2 2 2 2 2 2 2 2 2 2 2 2 3 3 3 3 3 3 code* Abundance Species Status 0.071 0.071 0.02 0.033 0.005 0.197 0.439 0.796 0.796 0.796 0.782 0.439 0.439 0.763 0.593 0.564 0.405 0.001 0.001 0.001 0.005 0.033 0.109 -value P Ͻ Ͻ Ͻ CLASSIFICATIONS FROM FOCUS GROUPS I. Nurna : Chi- ratio square PETI 8.067 4.571 2.571 3.267 3.267 5.4 4.571 8.067 1.667 0.6 0.067 0.067 0.067 0.077 0.6 0.6 0.091 0.286 0.333 0.692 BUNDANCE IS THE TOTAL NUMBER OF INDIVIDUALS RECORDED IN .I .A 15 15 15 15 14 14 15 15 15 14 15 15 15 15 15 15 13 15 15 11 14 12 13 SQUARE TESTS 86.7 78.6 71.4 26.7 26.7 20 21.4 13.3 66.7 60 53.3 53.3 46.7 53.8 40 40 54.5 42.9 41.7 38.5 - decline N 100 100 100 perceiving % informants TO FACILITATE COMPARISON , OCAL PERCEPTION OF CONSERVATION STATUS AND ABUNDANCE FROM THE ECOLOGICAL INVENTORIES Species sp. 6. L ABLE * Conservation status for** Nurna: Conservation 1 status not for declining, Ipeti: 2 1 no abundant, consensus, 2 3 locally declining. endangered, 3 absent. ARE IN BOLDFACE T Heteropsis oblongifolia Sabal mauritiiformis Swietenia macrophylla Socratea exorrhiza ``Puarsip'' Oxandra INFORMANTS ANALYZED WITH CHI Aechmea magdalenae Gustavia superba Genipa americana Peltogyne purpurea Symphonia globulifera Oenocarpus mapora Myroxylon balsamum Carludovica palmata Dalbergia retusa Arrabidaea chica ``Sianele'' Apeiba tibourbou Astrocaryum standleyanum Bactris coloradonis Ochroma lagopus ``Bunur'' Cordia alliodora ECONOMIC BOTANY Wednesday Jan 14 2004 01:03 PM ebot 58_110 Mp_78 Allen Press • DTPro System GALLEY 78 File # 10em

ECONOMIC BOTANY [VOL. 58

Fig. 1. Correlation biplots representing discriminant analysis classi®cations of conservation status based on use and ecological variables. Species (circles) were assigned by the analysis to the closest group (stars). The names of mis-classi®ed species are followed by an asterisk. Quantitative predictors (e.g. frequency of harvest) are represented by vectors (arrows) while nominal precitors (e.g. non-destructive harvest) are represented by squares. The projection of species, groups, and nominal predictors at right angle onto the discriminant axes or the predictor vector approximates the value of that object on the axes/vector. The angle between the predictor vectors and the discriminant axes re¯ects their correlation (the smaller the angle, the greater the correlation). ECONOMIC BOTANY Wednesday Jan 14 2004 01:03 PM ebot 58_110 Mp_79 Allen Press • DTPro System GALLEY 79 File # 10em

2004] DALLE & POTVIN: CONSERVATION OF USEFUL PLANTS

opportunistically harvested during hunting ex- cursions or other outings, which might also ex- plain the distribution of harvest activities for this species. Finally, a number of other products are harvested in more localized regions, including Socratea exorrhiza, Oenocarpus mapora, and Gustavia superba. In general, harvest ranges depend on a variety of factors, including the distribution and quality of the resource, the quantities required, the dif®culty of transport or han- dling, and the harvesting strategy (e.g., oppor- tunistic or deliberate, individual or collective). This variation in the spatial patterns of harvest meant that we could not identify a single ``harvest'' zone in Nurna. Instead harvest zones were determined according to product type. In Ipeti, on the other hand, it was not possible to de®ne harvest zones at all due to the differing system of land tenure, in which people do not have much communal forest available to them. In- stead, people in Ipeti usually harvest forest products speci®cally on their own ``parcelas.'' Any unharvested lands, therefore, would in theory exist within individual ``parcelas'' as opposed to outlying communal lands as in Nurna. Population structure analysis, therefore, was conducted only for Nurna, and only on the eight species that were suf®ciently abundant and for which information on harvest activities was available. The objective here was to evaluate whether population structure varied among harvest zones. CCAs of population structure versus harvest zone revealed signi®cant relationships for Sabal mauritiiformis, Myroxylon balsamum, Apeiba tibourbou, and Gustavia superba (Table 7). However, once the relationships between species and environment were accounted for, only Sabal mauritiiformis retained a signi®cant effect of harvest intensity. These results are consistent with the hypothesis that declining species Fig. 2. Harvest intensity maps for Nurna illustrat- would exhibit signi®cantly different population ing a range in harvest patterns: a) localized harvestÐ structures among harvesting zones for six of the Socratea exorrhiza, b) regional harvestÐconstruction eight species: Sabal mauritiiformis and the ®ve materials, and c) widespread harvestÐHeteropsis oblongifolia. The village is represented by a square, at species considered ``not declining'' or for which the centre of the maps. Darker tones represent a higher no consensus was reached. The hypothesis is not density of plots in the ecological inventory for which supported in the case of the other two species harvesting for the product was reported to occur. perceived to be declining (Socratea exorrhiza and Oxandra spp.). In the case of Sabal mauritiiformis, variance actually range quite far to harvest species such partitioning revealed that 10.7% of the variation as Sabal mauritiiformis and Heteropsis oblon- in the population structure data was explained gifolia, required in large quantities for house by harvest zones alone, 18.4% by environment construction. Heteropsis oblongifolia is often only, and 10.7% by covariation of environmen- ECONOMIC BOTANY Wednesday Jan 14 2004 01:03 PM ebot 58_110 Mp_80 Allen Press • DTPro System GALLEY 80 File # 10em

ECONOMIC BOTANY [VOL. 58 , . status of conservation Local perception decline decline decline no consensus no consensus no consensus no decline no decline ARE SHOWN IN BOLDFACE -value 0.001 0.495 0.498 0.11 0.417 0.645 0.925 0.409 P 0.05 Ͻ Fig. 3. Biplot from the CCA of Sabal mauritiiformis population structure vs. harvest zones, controlling 5.516 0.854 0.724 2.172 0.913 0.57 0.189 0.899 for environmental variation. Size classes are repre- VALUES - sented by circles, harvest zones by squares. The pro- P Population structure vs.

. jection of either size class or harvest zones at right

3.8 0.8 6.6 1.9 1.6 0.2 1.2 angle onto the axes approximates the value of that ob- harvest controlling for environment 15

explained F-ratio ject on the axis. HARVEST ZONES AND ENVIRONMENTAL VARIABLES % variation .

tal variables and harvest zones together, for a

PERMUTATIONS total of 39.8% of variation in population struc- 0.002 0.001 0.136 0.003 0.052 0.002 0.064 0.001 ture explained. In the biplot of Sabal mauritii- 999 formis population structure versus harvest intensity, controlling for environment (Fig. 3) the ®rst

2.63 3.393 1.979 2.797 1.848 3.186 3.041 3.918 axis is dominated by a gradient of distance from the village, with SC2 being abundant in the harvest and control zones, having apparently been

2.2 4.7 reduced near the village. In addition, the size Population structure vs. environmental variables 29.1 30.7 31.1 16.1 30.7 23.3 PERFORMED WITH explained F-ratio P-value

, class 4 is more closely associated with the con- % variation trol zone. For this species, the size classes that

3 1 5 5 5 2 6 are most frequently harvested are SC2 and SC3 10 selected variables (non-destructively) and to a lesser degree SC4 No. enviro. (destructively).

DISCUSSION 0.001 0.528 0.144 0.086 0.038 0.019 0.571 0.008 COMPARISON OF CONSERVATION STATUS AND PATTERNS OF PLANT

ARLO PERMUTATION TESTS In the present study, six of the 23 species in C

9.749 0.77 1.924 2.342 2.788 3.208 0.7 6.425 Nurna were perceived to have declined in abundance within living memory, whereas in Ipeti ONTE seven of the 22 study species were considered M Population structure vs. harvest

3 2.1 6.3 5.1 7.4 0.6 7.4 as conservation priorities. Thus villagers per- 21.3 explained F-ratio P-value

% variation ceived a third of the species examined as potentially threatened. These species are predominantly used for construction in the local econo- 75 27 89 37 54 42 84 sites 127 No. of my (e.g., houses, canoes), whereas three species provide products sold on regional or national markets (such as Swietenia macrophylla in Nur- ESULTS FROM CANONICAL CORRESPONDENCE ANALYSES OF POPULATION STRUCTURE VS na and Astrocaryum standleyanum and Dalber- IGNIFICANCE TESTS ARE gia retusa in Ipeti). In general, the pressure re- 7. R .S sulting from commercialization of forest prod-

Species ucts is well recognized, but less attention has URNA ABLE T Sabal Socratea Oxandra Oenocarpus Myroxylon Apeiba Astrocaryum Gustavia N focused on subsistence products (Pinedo-Vas- ECONOMIC BOTANY Wednesday Jan 14 2004 01:03 PM ebot 58_110 Mp_81 Allen Press • DTPro System GALLEY 81 File # 10em

2004] DALLE & POTVIN: CONSERVATION OF USEFUL PLANTS

quez et al. 1992), despite the fact that subsis- 2000). In both Ipeti and Nurna we found that tence products represent the majority of species accessibility is indeed related to the intensity of harvested from the forest (Pinedo-Vasquez et al. use, yet in different ways. In Ipeti, frequently 1990; Prance et al. 1987). Species used for con- used species were abundant early successional struction were also reported declining for a sa- species, but these did not occur closer to the vannah in Senegal (Lykke 2000) and in one village. On the other hand, in Nurna the opposite montane village in Chiapas, Mexico, studied by was found, with frequency of use being related Hellier et al. (1999). In the latter study, ®rewood to distance only. and fenceposts predominated among declining These results appear to stem from the con- use categories for a second montane village. trasting ecological and cultural contexts of the In both villages, a combination of ecological two villages. Perhaps the most evident are the (abundance, guild) and use (frequency, mode of contrasting ecological contexts: Ipeti has a lower harvest) variables was necessary to predict per- percentage of forest cover (44.6% of plots in the ceived conservation status. The interaction of ecological inventory vs. 76% in Nurna), higher use and ecology is seen especially in Nurna, population density, and is situated in a highly where the ability of the early successional spe- fragmented landscape as compared to Nurna. cies to tolerate disturbance seemed to buffer the This suggests that reduced or degraded forest impact of use. However, our results do not sug- cover may have led inhabitants in Ipeti to rely gest any dichotomy between timber and non- more heavily on early successional species than timber products in terms of conservation status in Nurna. In contrast, in Nurna where mature in either village. This is indicated by the lack of forest is still abundant, villagers appear to rely any relationship with the plant part harvested. on species whose distributions make them more Instead, the destructiveness of harvest emerges accessible rather than on early successional hab- as the most consistent predictor of conservation itats. The lack of major land-use change as yet status between the two villages. Although it is in Nurna (Lopez and Dalle 2001) would appear often assumed that harvest of NTFPs is non-de- to lead to a situation where harvest pressure is structive, as discussed by Peters (1996), NTFPs still the major conservation concern, as opposed are indeed sometimes destructively harvested, as to land-use change apparently being the major is the case here for Sabal in Nurna and Astro- pressure on plant populations in Ipeti. It cannot caryum in Ipeti. Demographic studies have be discarded, however, that cultural and histori- found that survival of adult life stages is often cal differences in plant use may be another pos- crucial to maintaining palm and other long-lived sible explanation. In Ipeti, where a much higher plant populations (Joyal 1996; Pinard 1993). level of cultural and economic change as com- The intensity of use of a species, as noted by pared to Nurna has been experienced, several of Peters (1996), is important for understanding hu- the plants studied have fallen into disuse and man impacts on resources. From an ecological many younger people no longer know how to standpoint, the more frequently a resource is use them (e.g., Aechmea spp., ``Chiru,'' Manet- harvested, the greater the impact should be. We tia spp.). These species are all found in late suc- found that the frequency of harvest was related cessional habitat and could at least in part ex- to species perceived as declining in Nurna, but plain the trend for frequent use of early succes- with those species in Ipeti that were considered sional species in Ipeti. ``abundant.'' We thus examined the hypothesis In examining the 10 species common to both that harvest frequency is related to resource case studies, we found that cultural differences availability. It has been suggested that for me- appear to exist in the use of plant resources be- dicinal purposes, people tend to rely more on tween the two villages, particularly in terms of species from early successional/disurbed than the frequency of harvest. Some of these differ- late successional habitats (Frei et al. 2000; Kohn ences can be attributed to different uses (e.g., 1992; Stepp and Moerman 2001; Voeks 1996). Astrocaryum standleyanum, Heteropsis oblon- Species in early successional habitats are gifolia, Dalbergia retusa). On the other hand, thought to be more available due to the high the more frequent use of Sabal mauritiiformis, densities at which they occur (Stepp and Moer- Socratea exorrhiza, and Oenocarpus mapora man 2001) or because such habitats are often (used in house construction) in Nurna may be found closer to human settlements (Frei et al. due to the more widespread use of non-tradi- ECONOMIC BOTANY Wednesday Jan 14 2004 01:03 PM ebot 58_110 Mp_82 Allen Press • DTPro System GALLEY 82 File # 10em

ECONOMIC BOTANY [VOL. 58

tional building materials in Ipeti as compared to vide important ecosystem services, such as nu- Nurna. However, another reason may be that in trient cycling (O'Hara 1998), and play crucial Nurna house construction and maintenance is roles in many plant-animal interactions (Redford carried out collectively, whereas in Ipeti it is or- 1996). Furthermore, conservation measures ganized by household. In Nurna, then, a man aimed at locally important resources can also will harvest materials such as Sabal mauritiifor- lead to the protection of a threatened habitat mis, Heteropsis oblongifolia, Socratea exorrhi- (Ticktin et al. 2002). za, and Oenocarpus mapora every time a house In addition to stimulating local participation, is built or repaired in the village (two to three the involvement of local people in conservation times a year during our research), whereas in can be a valuable source of knowledge (Gadgil Ipeti harvest events will be linked to each fam- et al. 1993). Here we compared the local per- ily's needs (every 8±10 years). This difference ceptions of species declines with population was not accounted for in our interviews. Further structure analysis, a rapid-assessment method studies on the comparative ethnobotany and his- that has been used for evaluating the effect of torical ecology of the Embera and Kuna are nec- harvest pressure (Hall and Bawa 1993; Murali essary to elucidate the dynamics of resource use et al. 1996). We expected that species declining in these communities. due to harvest pressure would exhibit signi®- cantly different population structures among LOCAL PERCEPTION AS A TOOL harvesting zones, whereas non-declining species FOR CONSERVATION would not. This was corroborated for one of the At the onset of this study we chose to focus three ``declining'' species (Sabal mauritiiformis) on useful plants, since we assumed that local and all ®ve of the ``non-declining'' species. people would be most likely to have interest in In the case of S. mauritiiformis, the lower conserving species of importance to them. This density of the most frequently harvested size was certainly the case in Ipeti, where the com- classes of S. mauritiiformis near the village is munity quite readily expressed interest in initi- consistent with a hypothesis of over-harvesting ating efforts to restore and/or enrich populations for this species. However, for Oxandra spp. and of useful species. In Nurna, the participatory na- Socratea exorrhiza (perceived as declining but ture of the research stimulated discussion in the no signi®cant difference in population structure), village regarding the state of their resources, and it is possible that population structure is not a by the end the community decided to continue good indicator for their status. In a previous pa- mapping parts of their territory, and to experi- per, we found that the occurrence of these two ment with planting of Sabal. To what extent can species increased with distance from Nurna, in- such initiatives contribute to conservation? dependently of environmental variation, and that In the conservation literature, considerable at- all size classes tend to co-occur (Dalle et al. tention has been focused on rare species and/or 2002). These species may therefore be in¯u- small populations, since these are thought to be enced by land-use or other changes that affect more susceptible to extinction (Groombridge the establishment or maintenance of entire pop- 1992). In our study, nine species were reported ulations rather than in¯uencing speci®c size to be absent from the territory. These species classes, as is expected with harvesting. could be the focus of restoration efforts with the Our experience indicates that local knowledge local community. However, apart from these may have several advantages for rapid-assess- ``absent species,'' it is important to note that ment of plant resources. First, it is possible to conservation status did not correlate with abun- collect data on large numbers of species. In the dance. Thus the species prioritized for cultural case of population structure, sampling in the purposes were not necessarily rare and may not ®eld is much more labor intensive, particularly be threatened on a regional level. From a local for less abundant species that may require ad- perspective, however, some important resources ditional sampling (e.g., Murali et al. 1996). In are required in high abundance (e.g., thatch) and addition, non-harvested sites may not be avail- may be prioritized for this reason. From a sci- able, as in the case of Ipeti. A second advantage enti®c viewpoint, conservation efforts aimed at of local knowledge is that it does not assume a such species can nonetheless have an ecological priori any speci®c cause of the decline itself. On basis. For example, abundant species often pro- the other hand, a hypothesis for the cause of ECONOMIC BOTANY Wednesday Jan 14 2004 01:03 PM ebot 58_110 Mp_83 Allen Press • DTPro System GALLEY 83 File # 10em

2004] DALLE & POTVIN: CONSERVATION OF USEFUL PLANTS

decline is necessary in order to determine what certain plants are used more than others, and should be the ``control'' for population structure how plant use responds to changes in the land- analysis (e.g., harvested vs. unharvested or be- scape. We have also noted differences in Nurna tween different land-uses). Controls are neces- and Ipeti in the use of the landscape, with early sary because trends in populations cannot be de- successional habitats versus proximate habitats duced from the population structure alone (Con- having more frequently harvested species. This dit et al. 1998, but see Lykke 1998). Based on result suggests that a better description of the these advantages, we suggest that local knowl- structure of the landscape, in terms of the abun- edge may be best used as an initial assessment dance and spatial distribution of habitat types, of a large number of species. Population struc- could potentially increase our understanding of ture analysis may then be used as an initial test how and why plants are used. of the hypothesis of harvest pressure for selected Finally, the use of local knowledge as a tool species in regions where unharvested (or other for rapid assessment requires a closer examina- control) sites exist. To this end, the method used tion of how perceptions regarding trends in bio- here to map harvest activities and factor out en- diversity are actually constructed, and how cog- vironmental variation can be employed. nitive, socioeconomic, or cultural factors may affect such perceptions. In Ipeti, women and DIRECTIONS FOR FUTURE RESEARCH men seem to have similar perceptions of status, The two case studies presented here illustrate which was the case in Lykke's (2000) study. how assessments of multiple species can be con- However, Hellier et al. (1999) suggest that the ducted to improve our understanding of conser- perceived importance of a given product, mem- vation priorities on indigenous territories. Our ory, and experience with the resource may in¯u- results suggest that certain characteristics, such ence informants' ability to report species de- as the destructiveness of harvest, may be com- clines. Further studies of the use of local knowl- mon to many declining species, whereas others edge, where possible, should be carried out in such as species abundance or the intensity of areas where long-term ecological data (e.g., harvest may depend on the type of resource use abundance over time) are available. This could or the speci®c ecological and cultural context. then be used to assess local knowledge. Further studies comparing a range of different ecological, cultural, and socio-economic situa- ACKNOWLEDGMENTS tions are needed to better evaluate the consisten- We would like to thank the many people who contributed their time, cy of these patterns and to elucidate causal re- knowledge, and efforts to this project. Lidia Barrigon, Inocencio Caisa- mo, Omaira Casama, Aurelio Chiari, Domingo DõÂaz, Andrebe Gonzalez, lationships. Similar studies could also examine Jane Hutton, Heraclio LoÂpez, Iguaigdigili LoÂpez, Lubertina Mosquera, the conservation status of species within speci®c Bornaje Pacheco, Nelson Perez, Melania Sarco, and Yanet Sierra provided invaluable assistance in the ®eld. We are equally indebted to the nu- use classes (e.g., construction materials), per- merous people in both Ipeti and Nurna who in various ways participated mitting a more detailed comparison of plant use in the research. S. D. would like to thank FundacioÂn Dobbo-Yala for their patterns in relation to conservation status and guidance and support. Institutional support in Panama was also provided by the Smithsonian Tropical Research Institute. Julie VelaÂsquez Runk and ecological characteristics. Richard Condit provided information regarding the habitat of some spe- A better understanding of harvesting patterns cies. We are grateful to Pierre Legendre for his assistance and expertise is imperative for ethnobotany to address issues in the multivariate analyses. We thank Tamara Ticktin, Kendra Mc- Sweeney, and an anonymous reviewer for useful comments on earlier of sustainable plant use. From our study it ap- versions of this manuscript. This research was funded by a grant from pears that a number of cultural, ecological, and the International Development Research Centre (Ottawa, Canada). S. D. gratefully acknowledges a postgraduate fellowship from the Natural Sci- historical factors may in¯uence which species ence and Engineering Research Council of Canada. are more frequently used. These will need to be teased out through more comparative studies. In LITERATURE CITED human ecology, optimal foraging theory has been used to examine hunting practices (e.g., Balee, W., ed. 1998. Advances in historical ecology. Winterhalder and Lu 1997). In this study, Columbia University Press, New York. Bernard, H. R. 1995. Research methods in anthro- switching between alternative resources was pology: Qualitative and quantitative approaches. 2? modeled according to the abundance and quality Altamira Press, Walnut Creek, CA. of the resource. Such a framework, in combi- Borcard, D., P. Legendre, and P. Drapeau. 1992. nation with historical approaches (e.g., Balee Partialling out the spatial component of ecological 1998), may be of value in understanding why variation. Ecology 73(3):1045±1055. ECONOMIC BOTANY Wednesday Jan 14 2004 01:03 PM ebot 58_110 Mp_84 Allen Press • DTPro System GALLEY 84 File # 10em

ECONOMIC BOTANY [VOL. 58

Colchester, M. 2000. Self-determination or environ- medicinal plants from primary and secondary mental determinism for indigenous peoples in trop- growth by the Runa of Eastern lowland Ecuador. ical forest conservation. Conservation Biology Journal of Ethnobiology 12(1):141±152. 14(5):1365±1367. Legendre, P., R. Galzin, and M. L. Harmelin-Vivi- Condit, R., R. Sukumar, S. P. Hubbell, and R. B. en. 1997. Relating behavior to habitat: Solutions to Foster. 1998. Predicting population trends from the fourth-corner problem. Ecology 78(2):547±562. size distributions: A direct test in a tropical tree Lopez, H., and S. Dalle. 2001. El desarrollo sostenible community. American Naturalist 152:495±509. en la reserva bioloÂgica y comarca Kuna de War- Croat, T. B. 1978. Flora of Barro Colorado Island. gandõÂ, DarieÂn. Pages 220±227 in Stanley Heckadon Stanford University Press, Stanford, CA. Moreno, ed.,Panama: Puente BioloÂgico. Instituto Dalle, S., H. Lopez, D. Diaz, P. Legendre, and C. Smithsonian de Investigaciones Tropicales, Pana- Potvin. 2002. Spatial distribution and habitats of ma. useful plants: An initial assessment for conserva- Lykke, A. M. 1998. Assessment of species composition on an indigenous territory, Panama. Biodiver- tion change in savanna vegetation by means of sity and Conservation 11(4):637±667. woody plants' size class distributions and local in- Davis, S. D., V. H. Heywood, and A. C. Hamilton, formation. Biodiversity and Conservation 7:1261± eds. 1997. Centres of plant diversity: a guide and 1275. strategy for their conservation, 3. World Wide Fund 2000. Local perceptions of vegetation change for Nature (WWF) and ICUNÐWorld Conserva- and priorities for conservation of woody-savanna tion Union, Cambridge, UK. vegetation in Senegal. Journal of Environmental Frei, B., O. Sticher, and M. Heinrich. 2000. Zapotec Management 59:107±120. and Mixe use of tropical habitats for securing me- Murali, K. S., Uma Shankar, R. Uma Shaanker, K. dicinal plants in Mexico. Economic Botany 54(1): N. Ganeshaiah, and K. S. Bawa. 1996. Extraction 73±81. of non-timber forest products in the forests of Bi- Gadgil, M., F. Berkes, and C. Folke. 1993. Indige- ligiri Rangan Hills, India. 2. Impact of NTFP ex- nous knowledge for biodiversity conservation. Am- traction on regeneration, population structure, and bio 22(2±3):151±156. species composition. Economic Botany 50:252± Gentry, A. H. 1996. A ®eld guide to the families and 269. genera of woody plants of northwest South Amer- Norton, B. G. 2001. Conservation biology and envi- ica (Colombia, Ecuador, Peru). The University of ronmental values: Can there be a universal earth Chicago Press, Chicago. ethic? Pages 71±102 in C. Potvin, M. Kraenzel, and Groombridge, B. 1992. Global biodiversity: Status of G. Seutin, eds., Protecting biological diversity: the Earth's living resources. Chapman and Hall, Roles and responsibilities. McGill-Queen's Univer- London. sity Press, Montreal. Hall, P., and K. Bawa. 1993. Methods to assess the O'Hara, J. L. 1998. Monitoring nontimber forest impact of extraction of non-timber tropical forest product harvest for ecological sustainability: A products on plant populations. Economic Botany case study of Huano (Sabal mauritiiformis) in the 47(3):234±247. RõÂo Bravo Conservation and Management Area, Hellier, A., A. C. Newton, and S. Ochoa Gaona. Belize. Pages 195±208 in R. B. Primack, D. B. 1999. Use of indigenous knowledge for rapidly as- Bray, H. A. Galletti, and I. Ponciano, eds., Timber, sessing trends in biodiversity: A case study from tourists, and temples: Conservation and develop- Chiapas, Mexico. Biodiversity and Conservation 8: ment in the Maya Forest of Belize, Guatemala, and 869±889. Mexico. Island Press, Washington, DC. Henderson, A., G. Galeano, and R. Bernal. 1995. Peres, C. A. 1994. Indigenous reserves and nature Field guide to the palms of the Americas. Princeton conservation in Amazonian forests. Conservation University Press, Princeton, NJ. Biology 8(2):586±588. Herlihy, P. H. 1986. A cultural geography of the Em- Peters, C. M. 1996. Observations on the sustainable bera and Wounan (Choco) Indians of Darien, Pan- exploitation of non-timber tropical forest products: ama, with emphasis on recent village formation and An ecologist's perspective. Pages 19±39 in M. Ruiz economic diversi®cation. Doctoral dissertation, PeÂrez and J. E. M. Arnold, eds., Current issues in Louisiana State University, New Orleans. non-timber forest products research. Centre for In- Instituto GeograÂ®co Nacional. 1988. Atlas Nacional ternational Forestry Research (CIFOR). 3? de la Republica de PanamaÂ. Instituto GeograÂ®co Pinard, M. 1993. Impacts of stem harvesting on pop- Nacional ``Tommy Guardia,'' PanamaÂ. ulations of Iriartea deltoidea (Palmae) in an ex- Joyal, E. 1996. The palm has its time: An ethnoecol- tractive reserve in Acre, Brazil. Biotropica 25(1): ogy of Sabal uresana in Sonora, Mexico. Econom- 2±14. ic Botany 50(4):446±462. Pinedo-Vasquez, M., D. Zarin, and P. Jipp. 1992. Kohn, E. O. 1992. Some observations on the use of Community forest and lake reserves in the Peru- ECONOMIC BOTANY Wednesday Jan 14 2004 01:03 PM ebot 58_110 Mp_85 Allen Press • DTPro System GALLEY 85 File # 10em

2004] DALLE & POTVIN: CONSERVATION OF USEFUL PLANTS

vian Amazon: A local alternative for sustainable Stepp, J. R., and D. E. Moerman. 2001. The impor- use of tropical forests. Pages 79±86 in D. C. Nep- tance of weeds in ethnopharmacology. Journal of stad and S. Schwartzman, eds., Non-timber prod- Ethnopharmacology 75(1):25±31. ucts from tropical forests: evaluation of a conser- ter Braak, C. J. F., and P. Smilauer. 1998. CANO- vation and development strategy. Advances in Eco- CO release 4 reference manual and user's guide to nomic Botany. New York Botanical Garden, Bronx, Canoco for WindowsÐSoftware for canonical NY. community ordination. Microcomputer Power, Ith- , Zarin, P. Jipp, and J. Chota-Inuma. 1990. aca, NY. Use-values of tree species in a communal forest Ticktin, T., G. De la PenÄa, C. Ilsley, S. Dalle, and reserve in Northeast Peru. Conservation Biology T. Johns. 2002. Participatory ethnoecological re- 4(4):405±416. search for conservation: Lessons from case studies Potvin, C., R. J.-P., G. Patenaude, and J. Hutton. in Mesoamerica. Pages 575±584 in J. R. Stepp, F. 2002. The role of indigenous peoples in conser- S. Wyndham, and R. K. Zarger, eds., Ethnobiology vation actions: a case study of cultural differences and biocultural diversity: Proceedings of the sev- and conservation priorities. Pages 159±176 in P. Le enth international congress of ethnobiology. Uni- Prestre, ed., Governing global biodiversity: The versity of Georgia Press, Athens, GA. evolution and implementation of the Convention of Tosi, J. A. 1971. Zonas de vida: Una base ecoloÂgica Biological Diversity. Global Environmental Gov- para investigaciones silvicolas e inventariacioÂn for- ernance Series. Ashgate Publishing Company, Al- estal en la Republica de PanamaÂ. FO: SF/PAN 6 dershot, England. No. 2 Food and Agriculture Organization of the Prance, G. T., W. BaleÂe, B. M. Boom, and R. L. United Nations, Rome. Carneiro. 1987. Quantitative ethnobotany and the Ventocilla, J., V. NuÂnÄez, F. Herrera, H. Herrera, case for conservation in Amazonia. Conservation and M. Chapin. 1995. Los indõÂgenas kunas y la Biology 1(4):296±310. conservacioÂn ambiental. MesoameÂrica 29:95±124. Redford, K. H. 1996. Not seeing animals for the trees: Voeks, R. A. 1996. Tropical forest healers and habitat The many values of wild animals in forest ecosys- preference. Economic Botany 50(4):381±400. tems. Pages 41±57 in M. Ruiz-Perez and J. E. M. Wali, A. 1993. The transformation of a frontier: State Arnold, eds., Current issues in non-timber forest and regional relationships in Panama, 1972±1990. products research. Center for International Forestry Human Organization 52(2):115±129. Research, Bogor, Indonesia. Western, D., M. Wright, and S. C. Strum, eds. 1994. Salvador, M. L., ed. 1997. The art of being Kuna. Natural connections: Perspectives in community- UCLA Fowler Museum of Cultural History, Los based conservation. Island Press, Washington, DC. Angeles. Winterhalder, B., and F. Lu. 1997. A forager-re- Schwartzman, S., A. Moreira, and D. Nepstad. source population ecology model and implications 2000. Rethinking tropical forest conservation: Per- for indigenous conservation. Conservation Biology ils in parks. Conservation Biology 14(5):1351± 11(6):1354±1364. 1357.

APPENDIX 1. CODING USED FOR THE MATRIX OF ECOLOGICAL AND USE VARIABLES.

Variable Type Codes Perceived status Ordinal/Qualitative Ipeti: 1 (abundant), 2 (danger), 3 (absent) Nurna: 1 (no decline), 2 (no consensus), 3 (decline) Plant part used Qualitative Leaves, Trunk or Fruits Quantity harvested Ordinal Ipeti: 1 (little), 2 (medium), 3 (large) Nurna: 1 (small), 2 (no consensus), 3 (large) Frequency of harvest Ordinal Ipeti: 0 (never, I don't know, or species does not exist here), 1 (more than 2 years apart), 2 (annually), 3 (monthly), 4 (diurnally or weekly) Nurna: 1 (infrequent harvest), 2 (no consensus), 3 (consensus on frequency) Mode of harvest Binary 0 (destructive), 1 (non-destructive) Ecological guild Ordinal 1 (pioneer) 2 (mid-successional/disturbance tolerant), 3 (late successional) Management Binary 0 (no management), 1 (cultivated and/or spared) Life-form Binary 0 (not woody), 1 (woody)