Agroforest Syst (2012) 86:1–16

DOI 10.1007/s10457-012-9503-y

Agroforestry species of the Bolivian Andes: an integrated

assessment of ecological, economic and socio-cultural plant
values
Regine Brandt • Heike Zimmermann •
Isabell Hensen • Juan Carlos Mariscal Castro •

Stephan Rist

Received: 20 September 2011 / Accepted: 23 February 2012 / Published online: 9 March 2012
Ó Springer Science+Business Media B.V. 2012

Abstract Agroforestry is a promising method for incorporating vegetation surveys, environmental studies,
enhancing land-use sustainability in the Bolivian Andes. and interviews on plant functions. Ecological, economic,
However, its benefits in terms of rural development are and socio-cultural values and the ecological apparency of
source: https://doi.org/10.7892/boris.17645 | downloaded: 11.11.2020

under-recognized due to gaps in understanding users’ plants were calculated, and detrended correspondence
perceptions while taking into consideration both local and principal component analyses helped to reveal the
and global environmental goals. Our study aimed to socio-ecological context of significant factors for plant
narrow these gaps by developing an analytical frame- distribution and uses. Results showed dominating seral
work for analyzing the site-specific socio-ecological woody species along an altitudinal gradient. Although
factors and interactions related to local woody species shrubs were more ecologically apparent than trees, trees
and assessing their ecological, economic, and socio- were perceived to be more valuable as the usefulness and
cultural plant values in order to identify the most prom- cultural importance of species increased with plant height
ising agroforestry species. The framework was then and timber availability. Phytosociological factors played
tested in an indigenous community at 2,760–3,830 m a.s.l., a minor but still significant role in perceived usefulness.
Schinus molle and Prosopis laevigata (\3,200 m a.s.l.),
Polylepis subtusalbida ([3,200 m a.s.l.), and Baccharis
dracunculifolia (both zones) were evaluated as most
Electronic supplementary material The online version of promising for agroforestry use. In conclusion, our
this article (doi:10.1007/s10457-012-9503-y) contains
supplementary material, which is available to authorized users. analytical framework proved to be a valuable tool for
context-specific agroforestry plant selection. Nonethe-
R. Brandt (&)  H. Zimmermann  I. Hensen less, economic, technical, and socio-cultural limitations
Institute of Biology/Geobotany and Botanical Garden,
of cultivating native agroforestry species were revealed
Martin-Luther-University Halle-Wittenberg,
Am Kirchtor 1, 06108 Halle (Saale), Germany as well. Agroforestry science and practice should,
e-mail: [email protected] therefore, focus on enhancing reproductive potentials of
existing woody vegetation, as well as problem-oriented
J. C. Mariscal Castro
horizontal dialogues between indigenous, expert, and
Agroecologı́a Universidad Cochabamba (AGRUCO),
Universidad Mayor de San Simón (UMSS), scientific actors.
Avenida Petrolera Km. 4 ‘, Cochabamba, Bolivia
Keywords Agroforestry  Bolivian Andes 
S. Rist
Local knowledge  Native woody species 
Centre for Development and Environment (CDE),
University of Bern, Hallerstrasse 10, Quantitative ethnobotany 
3012 Bern, Switzerland Socio-ecological plant values

123
2 Agroforest Syst (2012) 86:1–16

Introduction to their perceptions and behavior with respect to plant

use and management (Saunders et al. 2006). In this
Agroforestry is a traditional Andean land-use practice regard, Garibaldi and Turner (2004) suggest focusing
dating back to pre-colonial times (before 1530) when on species with outstanding socio-cultural values,
trees and shrubs were planted in agro-pastoral land- which they refer to as ‘‘cultural keystone species’’,
scapes for use as timber and firewood as well as for defined as ‘‘culturally salient species that shape in a
improving soil productivity and stabilizing hillsides major way the cultural identity of people, as reflected
and terraces (Chepstow-Lusty and Winfield 2000). in the fundamental roles these species have in diet,
Since the 1980s, many development projects have materials, medicine, and/or spiritual practices’’ (Gar-
proposed to revitalize Andean agroforestry due to its ibaldi and Turner 2004, no page number). Applying
potentials for enhancing land-use sustainability, and cultural keystone species may motivate land users to
for this purpose promoted the use of native woody support conservation and restoration initiatives,
species in traditional agrisilvicultural systems (trees including agroforestry. According to Turner (1988),
and crops) such as hedgerow intercropping, barrier the more widely a plant is used (intensity of use) and
hedges for soil erosion control, and shelterbelts, as the more use of this plant valued (quality of use) and
well as silvopastoral systems (trees on pastures) (e.g., directed towards this specific plant rather than towards
Reynel and Léon 1990). However, experiences have any other (exclusivity of use), the higher the plant’s
shown that native trees are rarely cultivated in the socio-cultural value. In contrast, negative plant attri-
farming systems of the Bolivian and Peruvian Andes butes such as toxicity reduce perceived socio-cultural
(e.g., Chepstow-Lusty and Winfield 2000; Ibisch plant values. In general, the degree to which a species
2002). Instead, monoculture plantations of fast-grow- is used and valued depends on this species’ biological
ing exotic trees are predominant in forestry and, plant attributes (traits) and on how well these traits and
consequently, in peasant livelihoods (Luzar 2007). their functions are known and recognized by the users.
Species of exotic Eucalyptus are the most popular, but This, in turn, is determined by the degree to which
they are heavily criticized due to their allelopathic these functions match with actors’ plant-related needs
effect on crops (Ahmed 1989) and understory vege- and priorities (Gausset 2004). Moreover, ecological
tation (Zhang and Fu 2009), depletion of soil nutrients plant attributes can influence users’ behavior accord-
and water (Kidanu et al. 2005), and potential inva- ing to the ‘‘ecological apparency hypothesis’’.
siveness (Richardson et al. 2004). Thus, Eucalyptus Adapted to ethnobotanical research by Phillips and
timber production tends to conflict with environmental Gentry (1993b), this hypothesis suggests that species
agroforestry goals related to the conservation of soils, which are frequent and dense and large in appearance
water, and biodiversity (Jose 2009; McNeely and are more visible and more readily available (i.e., more
Schroth 2006). Similarly, attempts to introduce agro- ecologically apparent), and therefore more likely to be
forestry aiming exclusively at producing environmen- used than less apparent plants. Rare or poorly acces-
tal services fail as well if peasants cannot derive any sible plants do not necessarily have low socio-cultural
socio-economic benefits (Alavalapati et al. 2004). In values, but compared to more apparent plants they
order to be feasible and attractive for land users, must provide higher-quality or more exclusive uses
agroforestry must provide favorable cost-benefit ratios that offset their low ecological apparency (Turner
within the given socio-ecological context (Gausset 1988).
2004), consider users’ socio-cultural perceptions, and From the point of view of sustainable agroforestry,
aim for compatibility with local livelihoods (Reed a basic problem related to the concept of cultural
2007). keystone species is that high socio-cultural plant
Considering agroforestry from an actor-oriented values are not necessarily compatible with high
perspective thus becomes a precondition for success- ecological plant values. The cultural keystone species
ful agroforestry. This requires evaluating context- approach is thus only useful for conservation purposes
specific subsistence, commercial, and socio-cultural if it is expanded to incorporate the assessment of
functions of agroforestry plants used (Madge 1995), potential positive and negative effects of plants on
by analyzing, among other things, peasants’ plant- soil, water, and biodiversity. For example, exotic
related knowledge and values, which in turn are linked species that bear the risk of becoming invasive need to

123
Agroforest Syst (2012) 86:1–16 3

be evaluated in terms of their negative effects on following research questions: (1) What is the floristic
biodiversity (Nuñez and Simberloff 2005). However, composition of local woody species, and what are the
knowledge gaps regarding the prediction of exotic determining environmental factors? (2) Which spe-
invaders’ likeliness to threaten native biodiversity and cies’ functions are important to local livelihoods, and
ecosystem functions (Kolar and Lodge 2001) call for which factors control species’ uses? (3) Which species
using native agroforestry species in order to prevent provide the highest ecological, economic, socio-cul-
invasions. Moreover, reforestation with native woody tural, and integrated values for agroforestry under
species can contribute significantly to restoring and the local conditions? Examination of these questions
conserving biodiversity (Jose 2009), which is of high was followed by a discussion on the implications for
significance in the Bolivian Andes given that natural applied community-based agroforestry.
woodlands remain only in the form of relicts (Hensen
2002; Navarro and Maldonado 2002). Besides sup-
porting biodiversity conservation goals, agroforestry Study site
species should also possess soil-enriching and soil-
stabilizing capacities (Jose 2009) that could help to The pilot study for testing our analytical framework
mitigate widespread and critical soil degradation took place in the Quechua-speaking indigenous com-
problems in the Bolivian Andes (Zimmerer 1993). munity Tres Cruces (17°280 –17°300 S, 66°270 –66°290
Against this background, potential agroforestry W, *850 ha, 2,760–3,830 m a.s.l.), situated in the
species need to meet context-specific criteria simulta- municipality of Tapacarı́, Cochabamba, Bolivia. This
neously regarding ecological, economic, and socio- semiarid region receives an average of 600 mm of
cultural sustainability. Although the need for assessing annual precipitation, with [80% of the rainfall
agroforestry species from an integrated perspective occurring between November and March (Ramadas
has been recognized (e.g., Jose 2011), previous No. 401-17, 1971–2003, Bolivian National Meteorol-
research focuses either on evaluating ecological plant ogy and Hydrology Service, SENAMHI). Annual
attributes such as growth and survival (e.g., Butterfield mean temperature is about 11°C (Honorable Alcaldı́a
1995; Mahboubi et al. 1997) or on describing plant Municipal de Tapacarı́ 2003). According to Navarro
uses (e.g., Langenberger et al. 2009). Quantitative and Maldonado (2002), the study area extends over the
techniques were developed and applied for determin- biogeographic Peruvian Puna Province (in transition
ing use values (e.g., Bennett and Prance 2000; Phillips to Bolivian-Tucuman Province). Native woody veg-
and Gentry 1993a), cultural values (e.g., Garibaldi and etation consists of hedges and shrublands at field
Turner 2004; Tardı́o and Pardo-de-Santayana 2008) margins, on waysides and stony terrace walls, in
and the ecological apparency of plants (e.g., Lucena ravines, and on fallow land.
et al. 2007). Assessing conservation status, albeit The peasant population mainly depends on small-
considering habitat rather than single species (Her- scale subsistence farming based on the production of
nando et al. 2010), is another example for a quanti- tubers such as Solanum tuberosum, cereals (e.g., Zea
tative evaluation using ecological criteria. McDonald mays, Triticum sativum, Chenopodium quinoa),
et al. (2003) did an integrated evaluation of plants’ legumes, vegetables, and fruits. The rearing of live-
usefulness and ecological attributes, but without using stock (ovine, caprine, bovine) complements agricul-
quantitative techniques. A quantitative tool suitable tural activities. More than 60% of the communal
for an integrated and context-specific assessment of territory is affected by hydric soil erosion, which
agroforestry species is thus not yet available. endangers agro-pastoral productivity (Aguilar and
The aims of the present study were to develop an Bracamonte 2002). This led to the establishment of
analytical framework for analyzing the site-specific soil conservation measures (e.g., terraces, dams)
socio-ecological factors and interactions related to complemented with agroforestry systems of native
local woody species, and to test it by calculating and exotic trees and shrubs (e.g., Buddleja coriacea,
ecological, economic, and socio-cultural plant values Spartium junceum) such as hedgerow intercropping
as a basis for identifying the most appropriate and barrier hedges, boundary plantings, and shelter-
agroforestry species in an indigenous peasant com- belts in the course of a participatory rural development
munity of the Bolivian Andes. We addressed the project (1999–2002). Moreover, family fruit orchards

123
4 Agroforest Syst (2012) 86:1–16

with peach (Prunus persica) and exotic tree planta- values to identify the significant factors controlling
tions (Eucalyptus globulus, Pinus sp., Cupressus sp.; local floristic compositions and plant uses (Fig. 1).
[3,600 m a.s.l.; *15 ha) for timber and firewood
production were cultivated. Vegetation surveys and environmental studies

In the first phase of our study, we determined the

Methods distribution, dominance, height, and floristic composi-
tion of woody species growing in the community of Tres
We constructed an analytical framework in order to Cruces. In addition, we analyzed environmental vari-
analyze the site-specific socio-ecological factors and ables (see below) of the study plots. Vegetation surveys
interactions related to local woody species. For this were conducted in 42 plots ranging from 2,816 to
purpose, we carried out vegetation surveys and 3,774 m a.s.l. between November 2005 and January
environmental studies, as well as ethnobotanical data 2006, during the rainy season. Sampling was spatially
sampling and classification. The results provided the organized according to three agro-ecological production
basis for calculating the context-specific ecological, zones as defined by local peasants: ura, situated below
economic, and socio-cultural plant values in order to 3,200 m a.s.l.; chawpi, situated at 3,200–3,500 m a.s.l.;
identify the most appropriate agroforestry species for and pata, situated above 3,500 m a.s.l. In each zone, we
the study area. We concluded our analyses by randomly selected 14 plots with woody vegetation. As
combining environmental site variables and plant we focused on naturally growing species, exotic tree

Fig. 1 Analytical framework of woody plant-related socio- for floristic composition of woody plants; principal component
ecological interactions for identifying promising agroforestry analysis (PCA) and linear regression analysis for determining
species, based on ecological and ethnobotanical indices applied significant species traits for usefulness of woody plants;
in our study (in italics): EV ecological plant value, RI relative Spearman rank analysis for calculating correlations between
importance, ICI identified cultural influence. Statistical methods ecological apparency (Importance Value [IV]), usefulness (RI)
applied in our study (in italics): detrended correspondence and cultural influence (ICI)
analysis (DCA) for analyzing significant environmental factors

123
Agroforest Syst (2012) 86:1–16 5

plantations were not considered. In order to comply with with ten semi-structured individual interviews about
criteria of structural and floristic homogeneity of vernacular names and uses of woody plants with local
vegetation units, we defined different plot sizes for peasants and three group interviews (n = 7–12 par-
shrublands (30–100 m2) and hedges on terrace walls or ticipants). Fresh plant specimens were used to opti-
waysides (20–60 m2). Geographic position, altitude, mize plant recognition. In addition, we conducted
and exposition of plots were measured using a handheld ‘‘freelisting interviews’’ (Quinlan 2005) with local
GPS device (GARMIN Geko 201); inclination was families between December 2007 and January 2008
estimated. A mixed topsoil sample of six subsamples about actual uses and the cultural importance of plants.
10 cm in depth was taken for each plot, except for plots Freelisting is a well-established ethnographic method
on stony terrace walls (n = 3). We identified woody defined as ‘‘listing things in a domain in whatever
species (trees, shrubs, subshrubs, perennial herbs, and order they come to mind’’ (Quinlan 2005, p 1), and
vines with woody stems) with a height C0.5 m. For all thus allows for identifying the informants’ most
species surveyed, maximum height per plot was mea- salient items. It was thus considered a suitable method
sured and dominance was estimated in terms of for discovering plants that were frequently used by the
percentage accuracy. For uncertain taxa, we collected informants, but not detected in our vegetation survey
plant vouchers and identified them by comparison with plots because they grew rarely or not at all in the study
plant material from the herbarium of Cochabamba area. The freelisting interviews were accompanied by
(BOLV) and in consultation with specialists. Air-dried participatory mapping of the communal territory and
soil samples were crushed and sieved (2-mm). We participatory observation of families’ plant uses (e.g.,
measured soil pH in calcium chloride (20 g soil, 50 ml construction materials, medicinal plants in home
0.1 mol/l KCl) and determined soil electrical conduc- gardens). The languages employed were either Span-
tivity (EC) in water (20 g soil, 50 ml H2O). Soil texture ish or Quechua (support of interpreter). The interviews
was estimated by finger testing and by identifying were taped, translated (Quechua to Spanish), and
visible features of moist soil samples (Boden 1994). transcribed. Vernacular plant names were related to
Total carbon and nitrogen contents were analyzed using scientific names in accordance to IPNI (2011), and
a CN-analyzer (Vario EL, ELEMENTAR). some unknown plant terms were identified in the
herbarium of Cochabamba (BOLV).
Ethnobotanical data sampling and classification We then registered the plant uses of the given
woody species. Based on local taxonomy, we regarded
Furthermore, we analyzed and classified plant-related the varieties of Cestrum parqui with different pheno-
economic and socio-cultural variables, such as the use types (flower colors) and chemical-pharmacological
and valuation of plants, as a basis for comprehending properties (toxicity) as distinctive species. On the
the knowledge of local people with regard to 60 woody basis of local plant use categorization and considering
species that were either surveyed in the study plots the plant use classification of Cook (1995), we
(n = 46 species of 58) or additionally mentioned as grouped all plant uses into eleven use categories: (1)
useful plants by local informants (n = 14 species). construction, (2) tools, (3) fuel, (4) food (incl.
We chose a stepwise methodological procedure to beverage), (5) fodder, (6) medicine (incl. spiritual
gather local plant knowledge of a representative group healing), (7) environmental use (e.g., soil manage-
of the community of Tres Cruces. In December 2005, ment), (8) field tools (e.g., fence, shelter), (9) domestic
data sampling started with two ‘‘walks-in-the-woods’’ use (e.g., kitchenware, furniture), (10) social/spiritual
(Phillips and Gentry 1993a) next to our vegetation use (rituals, ornaments) and (11) commercialization.
survey plots. Each 3-h walk was guided by a local Within each use category, different use types per plant
peasant who explained vernacular names and uses of were registered, resulting in open lists (e.g., food-
local woody plants growing in their natural habitats. specific use types of Clinopodium bolivianum: herbal
As participants utilize numerous attributes for recog- tea, spice). A classification developed by Arrázola
nizing plants, the walk-in-the-woods approach is et al. (2002) supported categorizing of medicinal use
reliable but time-consuming, and it only allows for a types.
small number of participants (Thomas et al. 2007). For Data on perceived negative plant attributes, such as
these reasons, we complemented our data sampling toxicity or weediness, were registered as well, and

123
6 Agroforest Syst (2012) 86:1–16

were considered as additional information in the final on soil and biodiversity in agroforestry systems. For
discussion of the evaluation of most suitable agrofor- this purpose, we developed and applied the ‘‘Ecolog-
estry plants. ical Plant Value’’ (EV) approach by using data from
secondary sources (Fjeldså and Kessler 1996; ISSG
Ecological, economic, and socio-cultural plant 2011; Killeen et al. 1993; Mahboubi et al. 1997; MBG
values 2011; Reynel and Léon 1990; Torrico et al. 1994).
Being (1) native to and (2) characteristic of natural
Based on our data recorded in vegetation studies, we woodlands, as well as the potentials for contributing to
calculated the ecological apparency for each of the (3) soil fertility and (4) erosion and wind control were
woody species surveyed (n = 58 species) by using the classified as species attributes with positive effects on
index ‘‘Importance Value’’ (IV), which is frequently biodiversity and soil conservation. In contrast, being a
applied in phytosociological (e.g., Boom 1986) and (5) exotic and probably invasive species or (6)
ethnobotanical studies (e.g., Lucena et al. 2007). IV is allelopathic or soil-degrading species were evaluated
usually measured in terms of relative frequency (RtF), as attributes with potential negative effects. Per
relative dominance (RtDo), and relative density (RtD) species, EV was calculated and categorized according
(e.g., Boom 1986; Lucena et al. 2007). As we to Table 1.
considered different life-forms, RtD was not regarded In order to assess social plant values, ethnobotany
as meaningful. Instead, relative height (RtH) contrib- provides suitable indices for collecting relevant qual-
utes to ecological apparency (Phillips and Gentry itative information and translating it into quantifiable
1993b). Thus, we calculated IV by using the modified data (Phillips and Gentry 1993a). For each of the
formula IV = RtDo ? RtF ? RtH, in which RtF was woody plants considered in ethnobotanical data sam-
the plant occurrence number divided by the total pling (n = 60 species), we calculated the economic
number of occurrences. RtDo was calculated as (subsistence and commercial) value by using the
species basal area divided by the total basal area; Relative Importance (RI) index (Bennett and Prance
RtH was calculated as species average maximum 2000, modified by Albuquerque et al. 2006). RI
height divided by the total number of average measures plant use numbers independently of the level
maximum heights for all species. IV values were of agreement among informants (‘‘informant consen-
between 0 (not ecologically apparent) and 300 (most sus’’, e.g., Phillips and Gentry 1993a), and can
ecologically apparent). therefore be applied to data sampled by using different
After that, we calculated ecological values for each ethnobotanical methods, as was done in our study. As
of the woody plants considered in the ethnobotanical a first step, we identified the sum of use categories and
data sampling (n = 60 species) in order to predict use types of each species. RI was then calculated based
their potential positive and negative ecological effects on the formula RI = NUC ? NT, in which NUC was

Table 1 Calculation of ecological plant value (EV) (n = 60 species)

Index Elements Rating of Index Index value Definition
elements group

EV Beneficial attributes: 0–1 Ecological plant value:

(1) Native; (2) characteristic of woodlands; 1 4 Highly beneficial
(3) contribution to soil fertility;
(4) contribution to erosion and wind control
Risky attributes: 2 2–3 Moderately beneficial
3 0–1 Low/no beneficial effects
(5) Invasive; (6) allelopathic or soil degrading 4 \0 Risky
EV is composed of six elements: four beneficial and two risky plant attributes. Each attribute was rated from 0 (no) to 1 (yes). Based
on benefits per species added up and risks subtracted, EV ranges from -2 (risky) to 4 (highly beneficial). On this basis, species were
categorized into four EV groups

123
Agroforest Syst (2012) 86:1–16 7

the number of use categories of a species divided by species, we applied detrended correspondence analy-
the total number of use categories of the most useful sis (DCA) (R package ‘‘vegan’’, function ‘‘decorana’’)
species and NT was the number of use types attributed on all plots except for two with deficient samples and
to a species divided by the total number of use types of those on stony terrace walls, since soil data for the
the most useful species. RI values ranged from 0 (not latter were not available (n = 37 plots of 42; n = 54
useful) to 2 (most useful) (Albuquerque et al. 2006). species).
As RI is based on plant use numbers and is Furthermore, we linked environmental, human, and
independent of informant consensus, the index does plant-related data. To analyze which phytosociologi-
not consider socio-cultural functions of plants (Albu- cal factors and species traits were correlated with RI,
querque et al. 2006). For this reason, we additionally we applied principal component analysis (PCA) (R-
calculated socio-cultural values for each of the woody package ‘‘vegan’’, function ‘‘prcomp’’) with previ-
plants considered in ethnobotanical data sampling ously logarithmized data (RtDo, RtF, RtH, RI) for 46
(n = 60 species) by using the ‘‘Identified Cultural woody species. PCA was complemented by linear
Influence’’ (ICI) index (Garibaldi and Turner 2004). regression analyses to estimate the extent to which RI
Based on our own ethnobotanical data complemented was determined by correlating factors. Spearman rank
by data from Aguilar (1997) and Ponce (2003), ICI correlations were used to examine whether RI was
was calculated considering species’ (1) intensity of linked to IV (n = 46 species) or ICI (n = 60 species),
use, (2) multiplicity of use, (3) use in language or as respectively. Underlying species numbers differed
indicator, (4) role in narratives and ceremonies or because not all species considered in ethnobotanical
symbolism, (5) persistence of use, (6) irreplaceability, data sampling were surveyed in vegetation plots, and
and (7) use as a trade item (see Table 2). The higher vice versa.
the ICI rating (maximum 35) of a species, the more
likely it was to represent a local cultural keystone
species. Results
Taking account of ecological, economic, and socio-
cultural plant values, these indicators provided the Socio-ecological features of local woody
basis for an integrated assessment of most promising vegetation
agroforestry species.
In our study plots, we surveyed 58 woody species
Statistical analyses belonging to 15 families (Leguminosae was consid-
ered as one family). Asteraceae was the most domi-
Statistical analyses were performed using the R nant family, representing 31% of taxa. Solanum (six
software (R Foundation 2010, version 2.11.1). In species, Solanaceae) and Baccharis (five species,
order to identify significant environmental site vari- Asteraceae) were the most diverse genera. Shrubs
ables that determine floristic compositions of woody (47%) were the most dominant life-form. Most species

Table 2 Calculation of Identified Cultural Influence (ICI) (n = 60 species)

Index Elements Rating of Index Index value Definition
elements group

ICI (1) Intensity of use; (2) multiplicity of use; Identified cultural influence:
(3) use in language, as indicator; 0–5 1 28–35 Very high
(4) role in narratives and ceremonies,
2 21–27 High
symbolism; (5) persistence of use;
(6) irreplaceability; (7) trade item 3 14–20 Moderate
4 7–13 Low
5 0–6 No
ICI is composed of seven elements, each of which is rated from 0 (no) to 5 (yes, very high). Added up, ICI ranges from 0 (no) to 35
(very high) per species. On this basis, species were categorized into five ICI groups

123
8 Agroforest Syst (2012) 86:1–16

were naturally growing native species (91%), of which

32% grew exclusively in the agro-ecological produc-
tion zone ura, 13% grew in chawpi, 28% occurred
exclusively in pata, and 27% were distributed across
two or all agro-ecological production zones. Four
percent were cultivated native species (e.g., Buddleja
coriacea) not locally grown, but originating from
Bolivia, and 5% were exotic. Baccharis dracunculi-
folia, Cestrum parqui, and Lepechinia graveolens
were among the most dominant and frequent woody
species and had the highest IV. Exotic trees such as
Eucalyptus globulus had the highest RtH, but the
native Polylepis subtusalbida and Schinus molle
reached high RtH values as well (Table A.5 in
Appendix Supplementary material).
Results from the DCA showed a high eigenvalue
(0.82) for axis 1, which indicated good species Fig. 2 Detrended correspondence analysis (DCA) of woody
dispersion along the axis and explained 49.7% of species (n = 54) and site variables of 37 plots (only signif-
icantly correlated variables and those with weak tendencies
total variance. Axis 2 showed an eigenvalue of 0.39 towards axes were plotted). Vectors: Altit: altitude [m a.s.l.],
and explained 23.2% of variance. Altitude (r = CN: total soil carbon [%]/nitrogen [%], EC: soil EC [ls], Incl:
-0.999, p B 0.001) and soil pHKCl (r = 0.998, inclination [°], pH: soil acidity (pHKCl). Data points of
p B 0.001) correlated most with axis 1. Hence, they multifunctional species with RI [ 0.5 labelled by abbreviated
names (algr: Aloysia gratissima, amar: Ambrosia arborescens,
were the most determining environmental factors for babu: Baccharis buxifolia, badr: Baccharis dracunculifolia,
species distribution. C/N ratios correlated significantly bapa: Baccharis papillosa, bape: Baccharis pentlandii, beco:
(p B 0.01) more with axis 1 (r = -0.776) than with Berberis commutata, buco: Buddleja coriacea, caen: Calceo-
axis 2 (r = 0.630). Soil EC correlated with axis 2 laria engleriana, cepa: Cestrum parqui, clbo: Clinopodium
bolivianum, eugl: Eucalyptus globulus, gyps: Gynoxys psilo-
(r = 0.936, p B 0.05). Inclination showed a weak phylla, kasa: Kaunia saltensis, legr: Lepechinia graveolens,
tendency towards axis 2 (r = 0.846, p = 0.058), and miov: Minthostachys ovata, posu: Polylepis subtusalbida, prcu:
soil texture and exposition did not correlate signifi- Proustia cuneifolia, prla: Prosopis laevigata, prpe: Prunus
cantly (p [ 0.05) with axis 1 or axis 2 (Fig. 2). persica, saha: Salvia haenkei, sape: Sambucus peruviana, satr:
Salpichroa tristis, scmo: Schinus molle, seay: Senna aymara,
Peasants in the study site assessed most plants as secl: Senecio clivicolus)
useful (95.0%, n = 57 species of 60). Table 3 shows a
rather uniform distribution of species citations among
use categories, except for the less cited social/spiritual narratives and beliefs (e.g., Schinus molle: death after
use and commercialization. The most cited uses were planting).
for medicine, fodder, fuel, and environmental uses. In addition to the numerous beneficial functions of
Fodder and fuel were also most diverse in species. plants, peasants also mentioned negative plant attri-
Within the use categories, 68 use types were regis- butes, such as leaf toxicity to livestock (Cestrum
tered. The most diverse use categories in terms of parqui, Kageneckia lanceolata), and attributes such as
numbers of use types were domestic uses (n = 16) and thorns and spines (e.g.,, Prosopis laevigata), which
medicine (n = 11). Fifty percent of use types were were perceived controversially, as they were benefi-
based on timber use, 45% on the use of leaves, flowers, cial in fences, but could also cause injuries and
and fruits, and 5% on tree crown use (Table A.6 in decreased the plant’s palatability (Table A.5 in
Appendix Supplementary material). Besides their Appendix Supplementary material).
material uses, plants had cultural roles, such as in
naming places (e.g., Prosopis laevigata: ‘‘thaqo Plant values and factors controlling usefulness
loma’’), as phenological indicators (e.g., Prosopis
laevigata: fructification indicates corn productivity), Polylepis subtusalbida was ecologically most benefi-
in rituals (e.g., Schinus molle: spiritual healing), or in cial, because this endemic and vulnerable tree is

123
Agroforest Syst (2012) 86:1–16 9

Table 3 Numbers of relative citation (# citation) of use categories [%], and species attributed (# species) to use categories [%]
(n = 60 species)
Use category all com con dom env fie fod food fuel med soc tool

# Citation [%] 100.0 1.7 8.0 9.9 11.1 10.8 12.7 6.4 11.6 15.7 1.9 10.2
# Species [%] 95.0 10.0 36.7 43.3 55.0 40.0 70.0 25.0 66.7 50.0 10.0 25.0
Use categories: all all uses, com commercialization, con construction material, dom domestic use, env environmental use, fie field
use, fod fodder, food, fuel, med medicine, soc social/spiritual use, tool

characteristic of high-Andean woodlands and has Prosopis laevigata: tools, Eucalyptus globulus: con-
beneficial soil-related attributes. Alnus acuminata, struction) and intensity of use (e.g., Baccharis dra-
Schinus molle, and Prosopis laevigata were also cunculifolia: field uses). Schinus molle and Prosopis
ecologically valuable. In contrast, Eucalyptus globu- laevigata had unique cultural roles, and timber of
lus was assessed as most unsuitable for use in Prosopis laevigata was even bartered with neighbor-
agroforestry systems due to its high absorption rate ing peasant communities. However, only exotic tree
of soil nutrients and moisture, its allelopathic effects timber (e.g., Eucalyptus globulus) provided monetary
on crops, and its potential to invading the natural value. Despite their lower socio-cultural values,
vegetation (Tables 4, A.5 in Appendix Supplementary Berberis commutata, Buddleja coriacea, Polylepis
material). subtusalbida and Sambucus peruviana also provided
Based on the high numbers of use categories and multiple uses (Table 4). In general, culturally influ-
use types, Schinus molle, Prunus persica, Eucalyptus ential species also demonstrated high numbers of
globulus, Prosopis laevigata, and Baccharis dracun- use categories and use types, and consequently,
culifolia had the highest RI values, and except for high RI values (Table A.7a–b in Appendix Supple-
Prunus persica also the highest ICI values. They were mentary material). This resulted in a highly significant
not easily replaceable due to their multifunctionality strong correlation between RI and ICI (R = 0.87,
(e.g., Schinus molle: 31 use types), use quality (e.g., p \ 0.001).

Table 4 Woody species with highest ecological (EV), economic (RI), and socio-cultural (ICI) values
# Species EV # Species RI # Species ICI

1 Alnus acuminata 4 1 Schinus molle 2.00 1 Prosopis laevigata 24

2 Escallonia resinosa 4 2 Prunus persica 1.42 2 Schinus molle 23
3 Polylepis subtusalbida 4 3 Eucalyptus globulus 1.41 3 Eucalyptus globulus 18
4 Prosopis laevigata 4 4 Prosopis laevigata 1.41 4 Baccharis dracunculifolia 17
5 Schinus molle 4 5 Baccharis dracunculifolia 1.38 5 Clinopodium bolivianum 13
6 Acacia visco 3 6 Berberis commutata 1.19 6 Lepechinia graveolens 13
7 Baccharis dracunculifolia 3 7 Buddleja coriacea 1.15 7 Polylepis subtusalbida 13
8 Berberis commutata 3 8 Polylepis subtusalbida 1.15 8 Prunus persica 13
9 Buddleja coriacea 3 9 Sambucus peruviana 1.15 9 Kageneckia lanceolata 11
10 Dodonaea viscosa 3 10 Clinopodium bolivianum 1.05 10 Sambucus peruviana 10
11 Gynoxys psilophylla 3 11 Pinus sp. 1.02 11 Calceolaria engleriana 9
12 Kageneckia lanceolata 3 12 Gynoxys psilophylla 0.96 12 Cestrum parqui (amarillo) 9
13 Kenthrothamnus weddelianus 3 13 Lepechinia graveolens 0.95 13 Kenthrothamnus weddellianus 9
14 Plazia daphnoides 3 14 Baccharis papillosa 0.93 14 Pinus sp. 9
15 Proustia cuneifolia 3 15 Kaunia saltensis 0.93 15 Salvia haenkei 9
16 Salix humboldtiana 3 16 Proustia cuneifolia 0.93 16 Spartium junceum 9
17 Senna aymara 3 17 Senna weddelliana 0.93 17 Passiflora spp. 8
The most promising agroforestry species were identified based on high integrated plant values (bold)

123
10 Agroforest Syst (2012) 86:1–16

Economically and culturally influential woody availability. In comparison, phytosociological factors

plants were also among the most apparent, reflected such as RtF and RtDo played a minor, but still
by highly significant correlations between ICI and IV significant, role in plant valuation.
(R = 0.76, p \ 0.001) as well as between RI and IV In summary, the species with the highest integrated
(R = 0.71, p \ 0.001). Culturally most influential values, assessed as being the most promising agrofor-
plants occurred exclusively in the agro-ecological estry plants, were Schinus molle, Prosopis laevigata,
production zone ura, but moderately culturally influ- Polylepis subtusalbida, and Baccharis dracunculifolia
ential plants were also surveyed in chawpi and pata. (Table 4).
Trees and shrubs showed moderate and high cultural
influence, whereas subshrubs and perennial herbs had
lower socio-cultural values on average (Table A.7b in Discussion
Appendix Supplementary material). This corre-
sponded with the results of PCA for determining Woody vegetation shaped by altitude, agro-
whether phytosociological factors (RtF, RtDo) or pastoral land use, and subsistence plant use
species traits (RtH, life-form) correlated with useful-
ness and economical value (RI) and, consequently, Floristic compositions reflected around five times
with cultural influence. The analyses resulted in 37.4% higher RtF and RtDo values for (sub-) shrubs than for
of variance for the first axis (highest rotation: trees. This was in line with the low frequency and
RI = 0.56, RtH = 0.51), and 30.0% for the second dominance determined for Polylepis subtusalbida
axis (highest rotation: RtF = -0.58, RtDo = -0.53). trees characteristic of high-Andean woodlands, which
High correlations were shown between RtH and the were mostly replaced by seral stages of grass- and
life-form ‘‘tree’’, as well as between RtF and RtDo shrublands due to human-induced processes such as
(Fig. 3). Post hoc linear regression analyses resulted in grazing, burning, and timber use (Hensen 2002;
a higher relationship of RI with RtH (R2 = 0.53) than Navarro et al. 2005; Thomas et al. 2010a). This is
with RtDo (R2 = 0.34) (adjusted values, p \ 0.001). why Baccharis papillosa and Clinopodium bolivia-
Hence, economic and socio-cultural value of woody num were frequently present in our study plots of the
plants increased with rising plant height and timber agro-ecological production zone pata, as they are
characteristic species of grasslands situated above
3,500 m a.s.l. (Navarro and Maldonado 2002), along
with Gynoxys psilophylla, which is representative for
field edges and fallows (Hensen 2002). A mosaic of
degraded forest relicts and seral grass- and shrublands
also characterized the vegetation of our study area in
chawpi and ura that refers to the upper part of Andean
dry valleys. Schinus molle and Prosopis laevigata
were the only characteristic woodland taxa surveyed,
while the most dominant species were Kentrothamnus
weddellianus and Baccharis dracunculifolia, which
are characteristic of seral shrublands situated below
3,500 m a.s.l. (Navarro and Maldonado 2002). The
dominant and frequently growing woody species
surveyed in our study area represent plant types
adapted to local traditional agro-pastoral land use, as
they are characterized by specific survival strategies,
Fig. 3 Principal component analysis (PCA) of usefulness (RI), such as quickly developing extensive roots (e.g.,
phytosociological data (RtF, RtDo) and traits (RtH, life-form) of Baccharis sp.) or production of plant toxins (e.g.,
local woody species (n = 46). Life-form: tree (T), shrub (Sh),
Cestrum parqui) (Ruthsatz and Fisel 1984). We
subshrubs/herbs (Su/H). Data points of multifunctional species
with RI [ 0.5 labelled by abbreviated names (see Caption of conclude that cultivating woody plants that are well-
Fig. 2) suited to agro-pastoral land use may advance the

123
Agroforest Syst (2012) 86:1–16 11

successful establishment and maintenance of Andean timber markets (Chepstow-Lusty and Winfield 2000).
agroforestry systems—provided that they are not Due to this absence of promising monetary returns, we
perceived negatively in ecological (e.g., poisonous) concluded that time- and labor-intensive establish-
or socio-cultural terms (e.g., Schinus molle: belief in ment and management of native agroforestry trees
death after planting, see above)—and should therefore would hardly be adopted and refined by local land
be considered in all activities. In particular the users, especially in view of continuously declining soil
potential use of native shrubs is worth considering productivity and the unfavorable macroeconomic
more closely. conditions that force Andean peasants to increasingly
Apart from human-induced factors, altitude was spend labor on non-farming activities in order to earn
shown to be a most significant factor for floristic additional income (Zimmerer 1993). The lack of
composition as well, not only in our study area but also markets for agroforestry timber and non-timber prod-
in an adjacent region (Thomas et al. 2010a). In ucts is not only a Bolivian but a common problem
addition, soil pHKCl (acidity) inversely and highly (Langenberger et al. 2009), and successful commer-
significantly correlated with altitude. This was possi- cialization also depends on peasants’ capacity to
bly an effect of decreasing temperature (e.g., Alexan- participate in local and regional trading activities. For
der et al. 1993) and thus decreasing decomposition all these reasons, we conclude that achieving eco-
rates and higher leaching intensities based on lower nomic sustainability of Andean agroforestry systems
evapotranspiration rates at higher altitudes. C/N using native species is a key problem, and gaining
showed rising ratios with increasing altitude, possibly commercial benefits from native species is quite
due to decelerating organic C turnover. However, impossible for peasants without long-term external
plant litter quality and quantity also played crucial support. In our view, therefore, under the current
roles in regulating C/N ratios (Dahlgren et al. 1997). conditions it is more feasible to implement less time-
Our data therefore reflected that climatic factors and labor-consuming types of agroforestry systems
strongly controlled soil properties and vegetation that aim mainly at meeting multifunctional subsis-
types, which in turn were also shaped by mutual tence needs such as fuelwood and fodder plants,
soil-vegetation interactions. With regard to Andean described in more detail below. Such agroforestry
agroforestry, our results emphasize that selecting systems should be given particular consideration.
species suited to altitude-dependent climatic and Approximately 70% of local woody plants supplied
edaphic conditions is a crucial factor of successful fuel and fodder, respectively. These use categories
management. were thus highly unspecific in terms of the species
Regarding local plant-related livelihood needs and concerned, but the species utilized were not equally
uses, as much as 95% of plant use types registered in valued. For instance, Baccharis dracunculifolia was
our study area focused on meeting domestic (e.g., fuel, favored for use as fuel based on high heating values,
construction) and agro-pastoral subsistence needs and Lepechinia graveolens was appreciated as a
(e.g., fodder, tools), which is characteristic of small- highly palatable fodder plant. Apart from the quality
holder farming systems that usually focus on multi- of use, peasants also cited plant availability as crucial
functional production (Wiersum 2004). In contrast, to preference, which was in line with the ecological
only 5% of plant use types represented commercial apparency hypothesis (Phillips and Gentry 1993b),
uses. Especially Eucalyptus had promising commer- and was also confirmed by correlating RI and IV in our
cial values due to its dominance in Andean timber study. Due to the absence of external energy sources,
markets (Chepstow-Lusty and Winfield 2000). Along peasants depend heavily on using fuel plants in order
with high subsistence values (e.g., construction, fuel, to sustain their livelihoods. At the same time,
medicine), these commercial values accounted for fuelwood harvesting has been identified as a driving
Eucalyptus’ high local economic and socio-cultural factor for continuing Andean woodland destruction
values. By contrast, in spite of the promising charac- (Thomas et al. 2010b). Similarly, while livestock
teristics of some species such as the fast-growing rearing is an integral part of local agro-pastoral land
Alnus acuminata, the timber of native trees does not use, livestock grazing can reduce natural fodder plant
provide monetary values (Ibisch 2002). The main regeneration (e.g., Polylepis sp., see Hensen 2002) and
reason for this is the dominant role of exotic trees in may result in soil surface compaction and reduced

123
12 Agroforest Syst (2012) 86:1–16

water infiltration (Ruthsatz and Fisel 1984). Fuelwood people to manage natural vegetation in a more
and livestock grazing are thus important plant uses but sustainable way. Moreover, recognizing traditional
conflict with environmental agroforestry goals. Agro- medicinal knowledge may improve rural health care
forestry systems should therefore focus on providing (Mathez-Stiefel and Vandebroek 2012).
additional fuelwood and fodder resources in order to
reduce the pressure on natural woodlands. Further- Promising species and implications
more, we suggest using agroforestry species adapted for agroforestry
to livestock grazing (e.g., re-shooting ability), and
motivating peasants to apply community-based norms The importance of Polylepis trees and woodlands for
and control mechanisms in order to regulate fuelwood preserving biological diversity and their unique envi-
harvesting and livestock grazing and thereby balance ronmental services, such as absorbing and storing air
use and conservation of fuel and fodder plants. moisture at high altitudes where no other trees can
Not all plant uses are per se plant-destructive. For grow, is widely recognized (Fjeldså and Kessler 1996;
instance, peasants perceived more than half of local Gareca et al. 2010). Alnus acuminata—not surveyed in
woody species as beneficial to soil productivity; this our study plots—is another high-Andean tree which is
was especially the case for Lepechinia graveolens, well-known as being ecologically valuable. It hosts
which grows along field margins, as well as Polylepis nitrogen-fixing actinomycetes (Frankia) in its root
subtusalbida and Schinus molle, which can even be nodules and therefore improves soil-fertility (Russo
found in cultivated plots. The organic matter of these 1990). In our study area, Schinus molle and Prosopis
species is used as organic fertilizer, in addition to laevigata contributed to environmental services in the
livestock dung. Furthermore, immaterial values (aes- agro-ecological production zone ura, while Baccharis
thetical, emotional, and spiritual) of the vegetation as a dracunculifolia was the most widely distributed
whole were locally recognized. In accordance with pioneer and soil-enriching species on fallow plots in
Andean beliefs that life is a dynamic interplay of all zones. However, rather than specific species,
natural-material, social, and spiritual aspects (Rist and peasants recognized woody species as whole as
Dahdouh-Guebas 2006), woody plants such as Schinus essential landscape elements which—with the excep-
molle were used in rituals, such as q’owa (giving to tion of Eucalyptus globulus—improve climatic, soil,
eat) and ch’alla (giving to drink), in order to maintain and water conditions and, accordingly, ensure suc-
positive relationships with Pachamama (Mother cessful agro-pastoral production. Summarized, the EV
Earth) and thereby ensure (re)production and well- index developed in our study provides useful data for
being. Spiritual beliefs are rooted in the principle of evaluating species’ ecological effects. However, it is
reciprocity (Mathez-Stiefel et al. 2007); accordingly, based on secondary information which is still incom-
psychosomatic disorders are believed to be caused by plete. Further experimental ecological studies and data
disturbed social or spiritual interrelations. Spiritual of promising native and exotic agroforestry plants of
healing rituals thus aim at restoring these imbalances the Bolivian Andes are therefore needed to improve
(Vandebroek et al. 2004). According to local beliefs, these insights.
mostly native medicinal woody plants (e.g., Schinus Eucalyptus globulus, Schinus molle, Prosopis lae-
molle, Lepechinia graveolens) possessed beneficial vigata, and Baccharis dracunculifolia were among the
spirits for curing, but cultivated exotic plants (e.g., species with the highest economic and socio-cultural
Eucalyptus globulus, Sambucus peruviana) were values due to their beneficial traits (e.g., timber) and
assimilated in spiritual healing as well. Half of the high ecological apparency. In contrast, Berberis
local woody species were used as medicinal plants, commutata and Buddleja coriacea also provided high
with most medicinal uses (*90%) directed towards economic values, but exhibited low socio-cultural
treating somatic diseases such as pains, infections, and values. Buddleja coriacea is a recognized native
gastrointestinal and dermatological disorders. Possi- agroforestry tree (e.g., Chepstow-Lusty and Winfield
bly, emphasizing non-destructive plant uses (e.g., 2000) and has been promoted by rural development
environmental, medicine) and recognizing immaterial initiatives. Local knowledge regarding the use of
plant values (e.g., spiritual) in community-based Buddleja is thus a result of external knowledge
agroforestry science and practice may motivate local transfer. However, the short time period since

123
Agroforest Syst (2012) 86:1–16 13

implementation (*10 years) possibly induced low species in accordance with peasants’ needs and
active species uses and cultural assimilation. In environmental adaptation.
contrast, knowledge regarding the use of locally In summary, using our analytical framework helped
grown Berberis commutata is locally rooted. Low to identify the most promising agroforestry species.
abundance and accessibility, unfavorable spines, and However, as shown by the example described further
out-dated use types (e.g., coloration of wool) were below, relying exclusively on quantitative assessment
among the factors responsible for low use intensities as expressed in the analytical framework discussed
of Berberis. RI and ICI values correlated significantly here bears the risk of any model, in the sense that it can
based on usefulness as a shared crucial plant attribute. represent part, but never the full range of socio-
Nevertheless, in our study Buddleja and Berberis ecological complexities involved in issues of sustain-
demonstrated that using RI and ICI was complemen- able development. A careful re-contextualization of
tary, because plant valuation considered use numbers quantitative data achieved using an analytical frame-
combined with use quality, use intensity, and socio- work with complementary qualitative data is therefore
cultural values of plants. recommended. The need for such re-contextualization
Schinus molle, Prosopis laevigata, Polylepis sub- is evident in the case of the use of Polylepis in agro-
tusalbida, and Baccharis dracunculifolia were iden- pastoral systems, which is controversial despite this
tified as the most promising agroforestry species based species’ high integrated values. Indeed, agricultural
on high integrated plant values within the given socio- production benefits from Polylepis based on the fertile
ecological context. In comparison, Escallonia resin- organic matter produced by its foliage or on its ability
osa and Alnus acuminata provide high ecological to create more favorable microclimatic conditions
values and potential usefulness as well (e.g., Chep- (e.g., Fjeldså 2002). On the other hand, however,
stow-Lusty and Winfield 2000; Torrico et al. 1994), Hensen (2002) reported that Polylepis can also have
but only received low economic and socio-cultural negative effects on crops due to light or root compe-
values as they were sparsely distributed in our study tition. Moreover, livestock grazing is in conflict with
area. Their importance would perhaps increase if they the natural regeneration of Polylepis.
were promoted, but—as already shown by Buddleja In our study area, Polylepis trees were rare and
and Berberis—this is highly dependent on how much grew in places distant from inhabited and cultivated
their plant uses effectively contribute to livelihoods. areas. Their high ecological plant values from a
On the contrary, both Eucalyptus globulus and Prunus scientific perspective (e.g., Fjeldså 2002) were there-
persica received high economic and socio-cultural fore hardly recognized by peasants, who additionally
values but low ecological values as they were exotic reported that previous efforts of Polylepis cultivation
species and, in the case of Eucalyptus, soil- and had not succeeded. However, peasants also stated that
biodiversity-degrading. The negative attributes of Polylepis had recently regenerated in uncultivated
plants were not quantified in our study, but must also ravines and slopes due to reduced harvesting pressure
be considered in plant assessment; this becomes since the establishment of exotic trees. From this
apparent in the case of Kageneckia lanceolata, which perspective, even Eucalyptus plantations may con-
is a promising agroforestry tree but locally known to tribute to preserving natural woodlands, as they help to
possess leaves that induce poisonous effects to live- reduce the exploitation of native species. However,
stock. Overall, plant indices provide useful estima- great care must be taken as to how and where they are
tions of plant values (Turner 1988). Integrated in our planted in order to prevent soil degradation (Fjeldså
analytical framework, the plant indices were therefore 2002). Moreover, if native species are less used and
shown to be suited not only to assess and rank replaced by cultivated trees, this also decreases their
promising agroforestry species, but also to detect perceived value (Lawrence et al. 2005), which in turn
ecological and biological plant attributes (e.g., life- may affect local tree conservation. In our study area,
form, dominance) that determined the usefulness of only access to useful woody plants was restricted by
plants in the peasants’ perception (Lawrence et al. community-based norms and rules, which are impor-
2005). This in turn leads to a better understanding of tant tools for regulating native tree timber uses
environment-human-plant relationships, which is the (Mariscal and Rist 1999). By using Polylepis as an
basis for identifying most promising plant traits and example, our study also showed that the decline in

123
14 Agroforest Syst (2012) 86:1–16

timber harvest due to alternative fuel sources may also Kaunia saltensis) and subshrubs (e.g., Minthostachys
support woody species’ natural succession, especially ovata, Clinopodium bolivianum)—contribute to meet-
when associated with communal regulations of live- ing subsistence needs, offer environmental benefits,
stock grazing and fire control. Furthermore, natural and foster emotional and spiritual well-being; some
succession of valuable species can be stimulated by even have commercial potentials (e.g., Alnus acumi-
enhancing their reproductive potential through spe- nata). Agroforestry should therefore pay attention to
cific management practices (e.g., weeding, mulching) the management of all adapted interstitial species and
or by dispersing their seeds and seedlings (Wiersum vegetation types that grow along with agro-pastoral
2004). In our study area, controlled plant regeneration systems (Backes 2001) rather than focusing only on
of Schinus molle was already being practised by local the cultivation of selected woody species. Further-
peasants because they particularly appreciate this more, our work shows that it is necessary to break
species for storing fodder in its crown and for its conceptual barriers, not only between social and
providing shade. Nonetheless, it was only rarely natural sciences but also between the local knowledge
cultivated. One reason for this was that local knowl- of peasants and the external knowledge of scientists
edge of native plant propagation was generally low in and development experts. This requires an intimate
the study area; an additional reason probably was that understanding of the realities of peasants’ livelihoods
spiritual beliefs hold the planting of Schinus molle and recognition of their ecological knowledge, as well
responsible for causing death. In view of this example, as their integration as equal partners into horizontal
we suggest that making use of the successional dialogues based on mutual learning between science-
processes in agroforestry is an effective alternative based and local knowledge and practices. Further-
to tree cultivation. more, such a transdisciplinary approach allows for
recognizing and making more systematic use of
previously unknown local potentials, such as local
Conclusion norms and plant management techniques, as well as
the often-ignored cultural and spiritual plant values,
This paper presents an analytical framework for for improving community-based agroforestry science
integrated socio-ecological assessment of agroforestry and practice.
species and describes how this framework was
empirically tested, helping to identify most promising Acknowledgments We greatly appreciate the financial
support received from the Andrea-von-Braun-Stiftung in
agroforestry species within the locally existing range
Munich, Germany. Our most sincere thanks go to the syndicate
of woody species in an indigenous peasant community assembly of Tres Cruces and the local leaders of the sub-central
of the Andean highlands of Bolivia. Emphasis was Waka Playa for authorizing our research, and to all local peasants
placed on integrating ecological, socio-economic, and who facilitated the study with confidence and ready cooperation.
We are also thankful for the technical assistance of the BioAndes
cultural aspects as well as on considering scientific and conservation and development program, funded by the Swiss
indigenous knowledge alike. Application of the ana- Agency for Development and Cooperation (SDC). Moreover, we
lytical model to evaluate 60 woody species occurring would like to acknowledge the scientific support of the Swiss
in the study area showed that Schinus molle, Prosopis National Centre of Competence in Research (NCCR) North–
South: Research Partnerships for Mitigating Syndromes of
laevigata, Polylepis subtusalbida, and Baccharis
Global Change, a research program co-funded by the Swiss
dracunculifolia can be considered as the most prom- National Science Foundation (SNSF), SDC, and the participating
ising agroforestry species in this area. However, institutions. Furthermore, we are thankful to the Bolivian
interview data and participatory observations also National Meteorology and Hydrology Service (SENAMHI),
Cochabamba, for providing climatic data, as well as to Magaly
indicated that economic, technical, practical, and
Mercado and the working team of the herbarium of Cochabamba
socio-cultural obstacles impede the cultivation of (BOLV) for their support in plant identification. We are very
these native species in agroforestry systems. As grateful to our colleagues Sarah-Lan Mathez-Stiefel (CDE),
complementary or alternative practices to cultivation, Rolando Sánchez, Sonia Medrano, Roger Juárez, Deicy Mejı́a
(AGRUCO), Michael Beckmann, Christine Voigt, Catharina
enhancing the reproductive potentials of natural
Landschulz, Ronny Warzecha, and Heidi Hirsch (MLU) for
woody vegetation and selected valuable species cooperating in field studies, interview translation, soil analyses,
should therefore be considered. Numerous species— and graphic presentation. Finally, we would like to acknowledge
especially shrubs (e.g., Lepechinia graveolens, Marlène Thibault and Danny McCluskey for proofreading the

123
Agroforest Syst (2012) 86:1–16 15

manuscripts, and the anonymous reviewers who have Garibaldi A, Turner N (2004) Cultural keystone species: implica-
contributed to improving this paper. tions for ecological conservation and restoration. Ecol Soc
9(3):1. http://www.ecologyandsociety.org/vol9/iss3/art1/
Gausset Q (2004) Ranking local tree needs and priorities
through an interdisciplinary action research approach.
J Transdiscipl Environ Stud 3(1):1–17
References Hensen I (2002) Impacts of anthropogenic activity on the veg-
etation of Polylepis woodlands in the region of Coc-
Aguilar LC (1997) Predicción del tiempo y su influencia en la habamba, Bolivia. Ecotropica 8:183–203
organización de la producción en la comunidad de Tres Hernando A, Tejera R, Velázquez J, Núñez MV (2010) Quan-
Cruces, provincia Tapacarı́. Tesis de licenciatura. AG- titatively defining the conservation status of Natura 2000
RUCO, Universidad Mayor de San Simón, Cochabamba forest habitats and improving management options for
Aguilar LC, Bracamonte R (2002) Diagnóstico participativo y enhancing biodiversity. Biodivers Conserv 19:2221–2233
plan de ordenamiento predial de la comunidad Tres Cruces. Honorable Alcaldı́a Municipal de Tapacarı́ (2003) Ajuste del
AGRUCO, Municipio Tapacarı́, Cochabamba plan de desarrollo municipal Tapacari (PDM) 2003–2007.
Ahmed P (1989) Eucalyptus in agroforestry: its effects on Programa de Inversión Rural Participativo (PDCR II),
agricultural production and economics. Agrofor Syst AGRUCO, Cochabamba
8:31–38 Ibisch PL (2002) Evaluation of a rural development project in
Alavalapati JRR, Shrestha RK, Stainback GA, Matta JR (2004) southwest Cochabamba, Bolivia, and its agroforestry
Agroforestry development: an environmental economic activities involving Polylepis besseri and other native spe-
perspective. Agrofor Syst 61:299–310 cies—a decade of lessons learned. Ecotropica 8:205–218
Albuquerque UP, Lucena RFP, Monteiro JM, Florentino ATN, IPNI (2011) The International Plant Names Index. Royal
Almeida C (2006) Evaluating two quantitative ethnobo- Botanic Gardens Kew, Harvard University Herbaria,
tanical techniques. Ethnobot Res Appl 4:51–60 Australian National Herbarium. http://www.ipni.org.
Alexander EB, Mallory JI, Colwell WL (1993) Soil-elevation Accessed 17 July 2011
relationships on a volcanic plateau in the southern Cascade ISSG (2011) Global Invasive Species Database. Invasive Species
Range, northern California, USA. Catena 20:113–128 Specialist Group. http://www.issg.org/database/welcome.
Arrázola S, Atahuachi M, Saravia E, Lopez A (2002) Diversidad Accessed 3 June 2011
florı́stica medicinal y potencial etnofarmacológico de las Jose S (2009) Agroforestry for ecosystem services and envi-
plantas de los valles secos de Cochabamba, Bolivia. Rev ronmental benefits: an overview. Agrofor Syst 76:1–10
Boliv Ecol Conserv Ambient 12:53–85 Jose S (2011) Managing native and non-native plants in agro-
Backes MM (2001) The role of indigenous trees for the conser- forestry systems. Agrofor Syst 83:101–105
vation of biocultural diversity in traditional agroforestry Kidanu S, Mamo T, Stroosnijder L (2005) Biomass production
land use systems: the Bungoma case study. In situ conser- of Eucalyptus boundary plantations and their effect on crop
vation of indigenous tree species. Agrofor Syst 52:119–132 productivity on Ethiopian highland vertisols. Agrofor Syst
Bennett BC, Prance GT (2000) Introduced plants in the indig- 63:281–290
enous pharmacopoeia of northern South America. Econ Killeen TJ, Garcı́a E, Beck SG (1993) Guı́a de árboles de
Bot 54(1):90–102 Bolivia. Herbario Nacional de Bolivia, Missouri Botanical
Boden AG (1994) Bodenkundliche Kartieranleitung. E. Schwe- Garden, La Paz
izerbart’sche Verlagsbuchhandlung, Stuttgart Kolar CS, Lodge DM (2001) Progress in invasion biology:
Boom BM (1986) A forest inventory in Amazonian Bolivia. predicting invaders. Trends Ecol Evol 16:199–204
Biotropica 18(4):287–294 Langenberger G, Prigge V, Martin K, Belonias B, Sauerborn J
Butterfield RP (1995) Promoting biodiversity: advances in (2009) Ethnobotanical knowledge of Philippine lowland
evaluating native species for reforestation. For Ecol Manag farmers and its application in agroforestry. Agrofor Syst
75:111–121 76:173–194
Chepstow-Lusty A, Winfield M (2000) Inca agroforestry: les- Lawrence A, Phillips OL, Ismodes AR, Lopez M, Rose S, Wood
sons from the past. Ambio 29(6):322–328 D, Farfan AJ (2005) Local values for harvested forest
Cook FEM (1995) Economic botany data collection standard. plants in Madre de Dios, Peru: towards a more contextua-
Royal Botanic Gardens Kew, Richmond lised interpretation of quantitative ethnobotanical data.
Dahlgren RA, Boettinger JL, Huntington GL, Amundson RG Biodivers Conserv 14:45–79
(1997) Soil development along an elevational transect in the Lucena RFP, Araújo E, Albuquerque UP (2007) Does the local
western Sierra Nevada, California. Geoderma 78:207–236 availability of woody Caatinga plants (northeastern Brazil)
Fjeldså J (2002) Polylepis forests-vestiges of a vanishing eco- explain their use value? Econ Bot 61(4):347–361
system in the Andes. Ecotropica 8:111–123 Luzar J (2007) The political ecology of a ‘‘forest transition’’:
Fjeldså J, Kessler M (1996) Conserving the biological diversity Eucalyptus forestry in the southern Peruvian Andes. Eth-
of Polylepis woodlands of the highland of Peru and Bolivia. nobot Res Appl 5:85–93
A contribution to sustainable natural resource management Madge C (1995) Ethnography and agroforestry research: a case
in the Andes. NORDECO, Copenhagen study from the Gambia. Agrofor Syst 32:127–146
Gareca EE, Hermy M, Fjeldså J, Honnay O (2010) Polylepis Mahboubi P, Gordon AM, Stoskopf N, Voroney RP (1997)
woodland remnants as biodiversity islands in the Bolivian Agroforestry in the Bolivian Altiplano: evaluation of tree
high Andes. Biodivers Conserv 19:3327–3346

123
16 Agroforest Syst (2012) 86:1–16

species and greenhouse growth of wheat on soils treated forestales útiles para el productor agropecuario. Tomo II: Las
with tree leaves. Agrofor Syst 37:59–77 especies. Proyecto FAO Holanda/DGFF, Lima
Mariscal JC, Rist S (1999) Tipos de relaciones bosque-comun- Richardson DM, Binggeli P, Schroth G (2004) lnvasive agrofor-
idad y normas tradicionales de uso y acceso a la vegetación estry trees: problems and solutions. In: Schroth G, Da
boscosa. El caso de las comunidades de Chorojo y Chullpa Fonseca GAB, Harvey CA, Gascon C, Vasconcelos HL, Izac
K’asa de las provincias Quillacollo y Tapacarı́ en el de- AMN (eds) Agroforestry and biodiversity conservation in
partamento de Cochabamba. AGRUCO, PROBONA, tropical landscapes. Island Press, Washington, pp 371–396
Cochabamba Rist S, Dahdouh-Guebas F (2006) Ethnosciences—a step
Mathez-Stiefel SL, Vandebroek I (2012) Distribution and towards the integration of scientific and indigenous forms
transmission of medicinal plant knowledge in the Andean of knowledge in the management of natural resources for
highlands: a case study from Peru and Bolivia. Evid Based the future. Environ Dev Sustain 8:467–493
Complement Altern Med 2012:1–18 Russo RO (1990) Evaluating Alnus acuminata as a component
Mathez-Stiefel SL, Boillat S, Rist S (2007) Promoting the in agroforestry systems. Agrofor Syst 10:241–252
diversity of worldviews: an ontological approach to bio- Ruthsatz B, Fisel U (1984) The utilization of natural resources
cultural diversity. In: Haverkort B, Rist S (eds) Endoge- by a small community on the highlands of Bolivia and its
nous development and bio-cultural diversity: the interplay effects on vegetation cover and site conditions. Erdwiss
of worldviews, globalization and locality. COMPAS, CDE, Forsch 18:211–234
Leusden, pp 67–81 Saunders CD, Brook AT, Myers OE (2006) Using psychology to
MBG (2011) Tropicos. Missouri Botanical Garden. http:// save biodiversity and human well-being. Conserv Biol
www.tropicos.org. Accessed 17 July 2011 20(3):702–705
McDonald MA, Hofny-Collins A, Healey JR, Goodland TCR Tardı́o J, Pardo-de-Santayana M (2008) Cultural importance
(2003) Evaluation of trees indigenous to the montane forest indices: a comparative analysis based on the useful wild
of the Blue Mountains, Jamaica for reforestation and plants of southern Cantabria (northern Spain). Econ Bot
agroforestry. For Ecol Manag 175:379–401 62(1):24–39
McNeely JA, Schroth G (2006) Agroforestry and biodiversity Thomas E, Vandebroek I, Van Damme P (2007) What works in
conservation—traditional practices, present dynamics, and the field? A comparison of different interviewing methods
lessons for the future. Biodivers Conserv 15:549–554 in ethnobotany with special reference to the use of photo-
Navarro G, Maldonado M (2002) Geografı́a ecológica de graphs. Econ Bot 61(4):376–384
Bolivia: vegetación y ambientes acuáticos. Centro de Thomas E, Van Damme P, Goetghebeur P (2010a) Some factors
Ecologı́a Difusión Simón I. Patiño, Santa Cruz determining species diversity of prepuna and puna vege-
Navarro G, Molina JA, De la Barra N (2005) Classification of tations in a Bolivian Andes region. Plant Ecol Evol
the high-Andean Polylepis forests in Bolivia. Plant Ecol 143(1):31–42
176:113–130 Thomas E, Douterlungne D, Vandebroek I, Heens F, Goetg-
Nuñez MA, Simberloff D (2005) Invasive species and the cul- hebeur P, Van Damme P (2010b) Human impact on wild
tural keystone species concept. Ecol Soc 10(1): r4. http:// firewood species in the rural Andes community of Apil-
www.ecologyandsociety.org/vol10/iss1/resp4/ lapampa, Bolivia. Environ Monit Assess 178:333–347
Phillips O, Gentry AH (1993a) The useful plants of Tambopata, Torrico G, Peca C, Beck SG, Garcı́a E (1994) Leñosas útiles de
Peru: I. Statistical hypotheses tests with a new quantitative Potosı́. Proyecto FAO/Holanda/CDF ‘‘Desarrollo forestal
technique. Econ Bot 47(1):15–32 comunal en el Altiplano boliviano’’, Potosı́
Phillips O, Gentry AH (1993b) The useful plants of Tambopata, Turner NJ (1988) ‘‘The importance of a rose’’: evaluating the
Peru: II. Additional hypothesis testing in quantitative eth- cultural significance of plants in Thompson and Lillooet
nobotany. Econ Bot 47(1):33–43 Interior Salish. Am Anthropol 90:272–290
Ponce D (2003) Previsión del clima y recreación del conocimiento Vandebroek I, Calewaert JB, De Jonckheere S, Sanca S, Semo
indı́gena como estrategia para la conservación de la diversi- L, Van Damme P, Van Puyvelde L, De Kimpe N (2004)
dad cultivada en los Andes bolivianos: el caso de la comun- Use of medicinal plants and pharmaceuticals by indigenous
idad de Chorojo, provincia Quillacollo, departamento communities in the Bolivian Andes and Amazon. Bull
Cochabamba. Tesis de maestrı́a. AGRUCO, Universidad World Health Organ 82(4):243–250
Mayor de San Simón, Cochabamba Wiersum KF (2004) Forest gardens as an ‘intermediate’ land-
Quinlan M (2005) Considerations for collecting freelists in the use system in the nature-culture continuum: characteristics
field: examples from ethnobotany. Field Methods 17(3):1–16 and future potential. Agrofor Syst 61:123–134
R Foundation (2010) The R project for statistical computing. Zhang C, Fu S (2009) Allelopathic effects of Eucalyptus and the
http://www.r-project.org. Accessed 24 Sept 2010 establishment of mixed stands of Eucalyptus and native
Reed MS (2007) Participatory technology development for species. For Ecol Manag 258(7):1391–1396
agroforestry extension: an innovation-decision approach. Zimmerer KS (1993) Soil erosion and labor shortages in the
Afr J Agric Res 2(8):334–341 Andes with special reference to Bolivia, 1953–91: impli-
Reynel C, Léon J (1990) Árboles y arbustos andinos para agro- cations for ‘‘conservation-with-development’’. World Dev
foresterı́a y conservación de suelos. Tomo I: Especies 21(10):1659–1675

123