See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/225412942

Woody plants diversity, floristic composition and land

use history in the Amazonian rain forests of Madidi
National Park, Bolivia

Article  in  Biodiversity and Conservation · April 2008

DOI: 10.1007/s10531-008-9348-x

CITATIONS READS

33 252

1 author:

Manuel J. Macía
Universidad Autónoma de Madrid
167 PUBLICATIONS   1,975 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Amerindian and Afro-American Perceptions of Their Tradicional Knowlwlege in the Chocó Biodiversity Hotspot View
project

Madidi Project: A Modern project of floristic inventory and ecology in one of the most biodiverse regions of the world. View
project

All content following this page was uploaded by Manuel J. Macía on 22 May 2014.

The user has requested enhancement of the downloaded file.

Biodivers Conserv (2008) 17:2671–2690
DOI 10.1007/s10531-008-9348-x

O R I G I NA L P AP E R

Woody plants diversity, Xoristic composition and land

use history in the Amazonian rain forests of Madidi
National Park, Bolivia

Manuel J. Macía

Received: 13 August 2007 / Accepted: 25 January 2008 / Published online: 21 February 2008
© Springer Science+Business Media B.V. 2008

Abstract A Xoristic inventory of woody plants was carried out to analyse the relation-
ships between Xoristic similarity and geographical distance, and to compare the eVect of
land use history on the Xoristic composition between sites. Three lowland and two submon-
tane sites were studied in Madidi, Bolivia. In one site, there is evidence of an Inca ruin.
A total of 877 species and 12,822 individuals of woody plants with a diameter at breast
height ¸2.5 cm were recorded in 44 0.1–ha plots. Fisher’s Alpha index values were slightly
higher for the lowlands than for the submontane. Floristic similarity was higher within sites
than between sites as measured by both Sørensen and Steinhaus indexes. The fact that the
30 most important species per site (totalling 94 species) accounted for 61.7% of total
individuals, support the hypothesis that Amazonian plant communities are dominated by a
limited set of species, genera and families. On the other hand, 18 out of the 94 species were
reported in a single site, suggesting that some species are patchy in distribution and may be
environmentally determined. Both the oligarchy and environmental-determinism hypothe-
ses can be complementary in order to understand Xoristic patterns of this region. The Ruins
submontane site is Xoristically the most distinct, and past human disturbance is likely to be
the main reason. Since species diversity (ranging from 53 to 122 species per plot) and
density (ranging from 157 to 503 per plot) are highly variable in Madidi, to characterize the
diversity of a site, it is necessary to quantify an average of 10 0.1-ha plots in a relatively
small geographical area.

Keywords Beta diversity · Floristic patterns · Lianas · Past human disturbance ·

Plant communities · Sampling protocols · Southwest Amazonia · Tropical rain forest ·
Tropical trees

M. J. Macía (&)
Real Jardín Botánico, Consejo Superior de Investigaciones CientíWcas, Plaza de Murillo 2,
28014 Madrid, Spain
e-mail: [email protected]

1C
2672 Biodivers Conserv (2008) 17:2671–2690

Introduction

Floristic inventories in tropical rain forests practically never include all vascular plant hab-
its, because a complete inventory would be excessively time consuming. This is mainly due
to the high alpha diversity, the taxonomically poorly known Xora and the presence of some
life-forms, such as climbers and epiphytes, that are more diYcult to sample and quantify.
Most inventories have focused on trees, which traditionally have been considered as woody
self-supporting stems with a diameter at breast height (dbh) greater than or equal to 10 cm,
though other dbh cut-oV limits, mainly 2.5 cm, have been used as well (e.g. Gentry 1988;
Smith and Killeen 1998; Pitman et al. 2001; Romero-Saltos et al. 2001; Phillips et al.
2003a; ter Steege et al. 2003). Focusing on trees is logical, because they deWne the overall
physical forest structure, contribute the main part of forest biomass and represent a substan-
tial part of plant species diversity. However, other woody plants life-forms such as lianas,
contribute notably to species richness of a given region but only a minority of woody plants
studies include the inventory of trees and lianas within the same localities (Gentry and
Dodson 1987; Clinebell et al. 1995; Pérez-Salicrup et al. 2001; Duque et al. 2002).
Recent assessments of the two most widely used plot or transect based protocols of
Xoristic sampling (multiple 0.1-ha plots/transects including all woody plants ¸2.5 cm dbh,
vs. a 1-ha plot just for trees ¸10 cm dbh) conclude that the 0.1-ha method is the more
eYcient approach for studying Xoristic diversity and composition (Phillips and Miller
2002; Phillips et al. 2003b).
The Wrst vascular plants checklist of the larger Madidi region, deWned as the area which
includes three protected areas (the Madidi National Park, Apolobamba Biosphere Reserve,
and Área Natural de Manejo Integrado Pilón-Lajas) plus a region surrounding these areas,
contains 3,981 species from 23,515 collections (2,741 species for the Madidi National Park)
which characterize the region as the most plant-species rich of any forest in Bolivia (Parker
and Bailey 1991; Jørgensen et al. 2005a). The region includes 35 vegetation types in an ele-
vational gradient from Amazonian forests and savanna vegetation at ca 150 m, through
broad montane forests to subnival vegetation patches higher than 5,500 m (Navarro et al.
2004; Fuentes 2005). The focus of this study is the Amazonian rain forests (<1,100 m),
which contain the highest vascular plant alpha diversity (Jørgensen et al. 2005a).
The Madidi region has been reported as nearly without long-term human perturbation
(Parker and Bailey 1991), but there is evidence of Inca ruins dated to be more than 300-
years old, in the northern part of the National Park (Armentia 1897; Renard-Casevitz et al.
1988; Macía and Svenning 2005). To what extent such long-term human disturbances have
aVected current composition and structure of plant community of tropical rain forest, is an
unstudied issue in this region. Nevertheless, it has been reported that past anthropogenic
disturbance may have long-lasting eVects on the patterns of woody plant species composi-
tion (Balée and Campbell 1990; Thompson et al. 2002; Chazdon 2003; Heckenberger et al.
2003; cf. White and Hood 2004).
This study performs a comparative analysis of the woody plant diversity in the Amazo-
nian region of the Madidi National Park, Bolivia within the framework of a joint research
eVort by the Herbario Nacional de Bolivia, Missouri Botanical Garden and Real Jardín
Botánico de Madrid, established to inventory the poorly known vascular plant Xora and
vegetation of the Madidi region (Jørgensen et al. 2005b and papers cited therein). The
objectives of the present paper were: (i) to determine species richness, Xoristic composition
and structure of diVerent life-forms in Wve study sites in the lowlands and submontane
regions of the Madidi National Park; (ii) to analyse the relationships between Xoristic
similarity and geographical distance, as measured within sites and between sites; and

1C
Biodivers Conserv (2008) 17:2671–2690 2673

(iii) to compare the eVect of land use history on the Xoristic composition and forest struc-
ture of the Ruins site with respect to the other study sites.

Methods

Study area

Fieldwork was carried out in Madidi National Park, Amazonian Bolivia. The park protects a
total of 18,854 km2, ranging from 150 m to ca 6,000 m elevation, but we focused on the low-
land and submontane regions in the east-southeastern part of the Park, at elevations between
260 and 1,070 m (Fig. 1). The study area contains the last foothills of the eastern Andes and
often has a sharply dissected and steep topography which makes the area diYcult to access.
Rainfall varies considerably with elevation and topography. Mean annual precipitation is
2,335 mm and mean annual temperature 26.1°C (Navarro et al. 2004). The area experiences
a dry period of between 3 and 4 months (from May to August) a year, during which periodic
Patagonian winds (locally named sur or surazo) may occasionally drop the temperature to as
low as 4.5°C (Macía and Svenning 2005). The inventoried plots were located in valleys and
foothills landscapes where soil texture measured as the percentage of sand (>57% in all
sites), clay and silt was rather homogeneous between sites (Macía et al. 2007).
The vegetation in the study area is old-growth mature tropical rain forest, with more
than 90% of the landscape being non-inundated tierra Wrme forest and less than 10% being
sporadically Xooded forest by streams and rivers (Macía, personal observation). Perma-
nently-Xooded swamp forests were not found in the study area. Some high prized timber
tree species (e.g. Swietenia macrophylla, Cedrela odorata, Amburana cearensis) had been
logged in several areas with easy access near larger rivers before the area was declared a
National Park in 1995. A preliminary classiWcation of diVerent vegetation types for the
Madidi region is given in Fuentes (2005) and the Wrst annotated checklist of the vascular
plants for the region can be found in Jørgensen et al. (2005a).

Data collection

A total of 44 0.1-ha (50 £ 20 m) plots were inventoried at Wve diVerent sites along a north–
south geographical gradient (Fig. 1). The selection of the sites was based on satellite imagery,
available maps and accessibility. Three of them were located in the lowlands at 260–610 m
elevation in the vicinity of the Aguapolo, Tequeje and Yariapo rivers with 6–11 plots per
site, totalling 24 plots. The other two sites were in submontane forests at 735–1,070 m, in the
surroundings of Inca ruins (Fig. 2) and in the Serranía de Tumupasa, with 8 and 12 plots per
site respectively. All plots were placed in closed-canopy mature forests with no sign of any
recent disturbance, with the exception of the Ruins site albeit human disturbance was more
than 300 years ago (Armentia 1897; Renard-Casevitz et al. 1988). The forest structure at the
Ruins site was indistinguishable from any other mature forest in the region. Plots were
located to span the available habitat heterogeneity at each site, but each plot was placed to
avoid heterogeneity in forest physiognomy or soils and therefore installed in a single broad
forest type (tierra Wrme or Xoodplain) and excluding big canopy gaps. Plots within the same
site were at least 500 m apart. Further information on plot location, geographical coordinates
and elevations can be found appended to Macía and Svenning (2005).
All woody plants (trees, lianas and hemiepiphytes) with stems rooted independently
within a plot and with a dbh (measured at 1.3 m above ground for all life-forms) equal to or

1C
2674 Biodivers Conserv (2008) 17:2671–2690

68°W 67°40' W

R
oa ma
Ix

d
ia

to
Inca BOLIVIA

s
Ruins (8) Tumupasa
Río
14°S

Tequeje (6) Madidi National

Park

Tumupasa

Serranía de
Tumupasa (12)
San José de
Uchupiamonas
Río
Yariapo (11)

R rre
oa n a
R

d
u

to baq
Rí
oT
uic

eu
ih
14°30' S

Río
Aguapolo (7)
10 0 10 20 km

Fig. 1 Location of study area and the Wve study sites in Madidi National Park, Bolivian Amazonia. The number
of plots in each site is given in parentheses

greater than 2.5 cm were measured, inventoried and identiWed to species or to a Weld-tem-
porary name allocated to a morphological species concept (morphospecies). Multiple stems
were measured separately, but all stems rooting in the same place were counted as one indi-
vidual. Specimens (Macía et al. 3,876–7,051) were collected in 2001 and 2002 to voucher
each name (per site) and in all doubtful cases to identify a stem. All voucher specimens can
be found in the TROPICOS web page (http://www.mobot.mobot.org/W3T/Search/
vast.html). All specimens were Wrst sorted to species or morphospecies level. Then, the
sterile vouchers were identiWed by matching them with vouchers identiWed by specialists or

1C
Biodivers Conserv (2008) 17:2671–2690 2675

Fig. 2 Inca fort ruins at 950 m elevation in Madidi National Park, Amazonian Bolivia. In the surroundings
of the ruins, eight 0.1-ha plots were inventoried

professional botanists which are deposited at LPB, MA, MO, NY and USZ herbaria (acro-
nyms according to Holmgren et al. 1990). Duplicates of the specimens were distributed to
taxonomic specialists worldwide who agreed to identify them (see Acknowledgements). A
full set of duplicates is deposited at LPB and MA, and a nearly complete set at MO.
Unicates are kept at LPB. In the following analyses, specimens identiWed to subspeciWc
taxa were lumped under their species. All morphospecies were treated in the same way as
identiWed species.

Data analyses

For the inventoried sites, species diversity, Xoristic composition and similarity were mea-
sured with quantitative and qualitative indices. For each of these analyses, woody plants
were divided into six life-forms: (a) all woody plants (trees, lianas and hemiepiphytes
¸2.5 cm dbh); (b) all trees (¸2.5 cm dbh); (c) large trees (¸10 cm dbh); (d) small trees
(2.5 ¸ dbh < 10 cm); (e) all lianas (¸2.5 cm dbh); and (f) hemiepiphytes (¸2.5 cm dbh).
The hemiepiphytes were too rare for meaningful analyses and were therefore excluded (see
Table 1). Palms were included in the tree categories to facilitate analyses interpretation,
although they are not properly trees.
Species diversity values were broken down in diVerent life-forms, and expressed in
terms of species richness for each plot and averaged per site. The Fisher’s Alpha diversity
index was calculated to relate the total number of woody plants species and the total num-
ber of individuals (Fisher et al. 1943) for each plot and then averaged per site. Fisher’s
Alpha index is relatively insensitive to sample size and provides good estimates of the
overall diversity of tropical forests, even in the case of small forest samples (Condit et al.
1998).
To quantify and compare Xoristic composition between sites, the species Important
Value Index (IVI) was calculated as the sum of its relative density, its relative dominance
and its relative frequency (Curtis and McIntosh 1951). The frequency of a species for each
site is deWned as the number of 0.1-ha plots in which it is present, and the sum of all fre-
quencies as the total number of plots per site. Family Importance Value Index (FIVI) was
calculated as the sum of relative density, relative diversity and relative dominance for a
family (Mori et al. 1983).

1C
 2676

1C
Table 1 Total number of woody plant species and individuals per life-form at Wve study sites, totalling 44 0.1-ha plots inventoried in Madidi National Park, Bolivia

Attributes Lowlands Submontane

Aguapolo Tequeje Yariapo Ruins Tumupasa

Total woody plant species 285 (70–101) 312 (80–137) 335 (53–103) 313 (61–97) 416 (58–122)
Total woody plant families 62 (38–44) 72 (42–52) 66 (26–46) 66 (33–46) 77 (36–46)
Total woody plant individuals 1936 (228–349) 2075 (277–416) 2626 (157–295) 2811 (209–503) 3374 (208–471)
Trees ¸ 2.5 cm dbh species 220 (63–83) 249 (64–111) 258 (48–91) 256 (51–87) 337 (54–103)
Trees ¸ 2.5 cm dbh individuals 1715 (203–297) 1866 (246–363) 2343 (129–275) 2576 (191–475) 3125 (202–431)
Trees ¸ 10 cm dbh species 120 (26–38) 126 (27–46) 140 (18–41) 147 (28–43) 206 (26–56)
Trees ¸ 10 cm dbh individuals 423 (47–72) 485 (71–95) 749 (47–91) 743 (69–119) 852 (49–105)
Trees <10 cm dbh species 174 (49–63) 212 (51–89) 219 (39–67) 218 (33–69) 280 (40–89)
Trees <10 cm dbh individuals 1292 (142–229) 1381 (151–277) 1594 (78–195) 1833 (116–364) 2273 (135–337)
Lianas ¸ 2.5 cm dbh species 65 (8–19) 67 (7–25) 75 (2–16) 59 (4–21) 79 (4–19)
Lianas ¸ 2.5 cm dbh individuals 214 (12–52) 199 (8–61) 278 (2–62) 231 (16–93) 242 (4–39)
Hemiepiphytes ¸ 2.5 cm dbh species 5 (0–4) 5 (0–2) 5 (0–1) 4 (0–3) 5 (0–3)
Hemiepiphytes ¸ 2.5 cm dbh individuals 7 (0–4) 10 (0–5) 5 (0–1) 4 (0–3) 7 (0–3)
Mean Fisher’s Alpha index 43.7 (33.0–51.0) 51.1 (37.7–71.3) 41.7 (21.7–62.1) 35.1 (19.7–51.1) 42.1 (26.7–63.0)
Number of plots 7 6 11 8 12
Mean elevation (m a.s.l.) 330 (260–420) 480 (425–550) 530 (460–610) 915 (735–1045) 880 (800–1,070)

Mean plot Fisher’s Alpha index and elevation are also given per site. Range values between plots are shown in brackets
Biodivers Conserv (2008) 17:2671–2690
Biodivers Conserv (2008) 17:2671–2690 2677

To analyse the degree of Xoristic similarity within and between inventoried sites, two
similarity indices were calculated: the Sørensen index and the Steinhaus index. These two
indices are mathematically identical, but the Sørensen index uses only presence-absence
species data whereas the Steinhaus index also includes species abundance, measured as the
number of individuals per species (Legendre and Legendre 1998). It is convenient to com-
pare whether the results obtained with abundance data are diVerent from the results
obtained with presence-absence data. Similarity matrices were calculated separately for
diVerent life-forms: (a) all woody plants; (b) all trees; (c) large trees; (d) small trees; and (e)
all lianas.
Cluster analyses were computed to classify the Wve inventoried sites on the basis of their
Xoristic similarity, according to both the Sørensen and Steinhaus indices. An agglomerative
cluster which uses a proportional-link linkage algorithm was applied; the connectedness
level was set to 0.5 (i.e. midway between single and complete linkage). Clusters were per-
formed separately for the Wve life forms, but because clusters obtained for all trees and
small trees were very similar to the total woody plants cluster, they are not included here.
Structural composition for each site was analysed by comparing the distribution of all
trees and lianas according to dbh classes.
Sørensen and Steinhaus similarities and cluster analyses were computed in the
program Le Progiciel R (available online at http://www.bio.umontreal.ca/legendre/index-
English.html).
Owing to the sharply dissected topography in the study area, Xoods are sporadic and of
short duration compared to those further downriver in rolling landscapes, where forests
may be inundated for several weeks or months at a time. Therefore, the dichotomy of Xood-
plain and tierra Wrme forests is not important for deWning Xoristic patterns in the Madidi
region (see Macía and Svenning 2005; Macía et al. 2007 for details).

Results

Species diversity and density

A total of 12,822 individuals representing 877 species (359 genera and 94 families) of
woody plants ¸2.5 cm dbh were found in 44 inventoried 0.1-ha plots. Mean total numbers
of woody plant species and individuals were slightly smaller in the lowland sites than in the
submontane sites (Table 1). The lowland Yariapo and the submontane Tumupasa sites had
both the highest species richness and density for all studied life-forms and size classes.
Free-standing trees represented 90.7% of the total individuals and 78.3% of the total
species, whereas values for lianas were 9.1% and 24.6% and hemiepiphytes 0.2% and 1.9%,
respectively (Table 1). The total species percentages sum to more than 100 because some
liana species in their juvenile stages were found growing as free-standing trees or rarely vice
versa. In the lowland Aguapolo site, the lowest tree species richness was reported and in the
submontane Ruins site the lowest liana species richness was reported. Hemiepiphytes were
poorly represented in all sites, with similar number of species and individuals.
The Wve most common tree species in terms of number of individuals were Rinorea
viridifolia (770), Iriartea deltoidea (720), Rinorea guianensis (274), Styloceras brokawii
(267) and Pseudolmedia laevis (224); they accounted for 19.4% of tree individuals. The
Wve most common liana species, Roentgenia bracteomana (69), Byttneria pescapriifolia
(44), Petrea maynensis (35), Combretum laxum (29) and Dalbergia frutescens (26),
accounted for 17.4% of liana individuals.

1C
2678 Biodivers Conserv (2008) 17:2671–2690

Table 2 Comparison of the 10 most important families (in bold) at Wve study sites, totalling 44 0.1-ha plots
in Madidi National Park, Bolivia

Family Lowlands Submontane

Aguapolo Tequeje Yariapo Ruins Tumupasa

Annonaceae 9.79 13.63 5.27 7.25 6.60

Arecaceae 21.93 29.80 31.81 11.15 19.43
Bignoniaceae* 11.84 7.23 9.61 9.15 8.50
Buxaceae – 1.98 11.52 1.45 3.71
Cecropiaceae 1.50 2.42 4.92 8.16 1.99
Chrysobalanaceae 14.60 4.84 3.32 6.87 5.92
Euphorbiaceae* 2.88 12.87 18.13 10.08 8.62
Flacourtiaceae 6.05 11.37 7.99 4.86 9.18
Lauraceae 9.59 7.59 13.22 9.42 9.27
Leguminosae* 21.28 17.95 22.27 18.56 18.60
Melastomataceae 2.48 4.17 2.90 18.36 3.65
Meliaceae 11.73 20.79 14.32 5.43 19.37
Moraceae 17.83 20.34 19.13 20.94 18.67
Myrtaceae 11.12 4.84 3.24 7.54 7.08
Piperaceae 2.87 1.23 9.93 4.53 7.41
Rubiaceae* 8.90 10.93 9.75 32.19 16.96
Violaceae 25.36 13.17 6.94 10.60 25.69

Numbers indicate the Family Importance Value Index (FIVI), which was obtained as the sum of their relative
density, relative dominance and relative diversity. Families with the symbol * include liana species

Mean plot Fisher’s Alpha index values were slightly higher for the lowlands than for the
submontane region (Table 1). The highest diversity value was recorded in the lowland
Tequeje site, whereas the lowest value was recorded in the submontane Ruins site.

Family and species composition

According to the FIVI, the most important families were Arecaceae (comprising 11
species), Lauraceae (47), Leguminosae (100) and Moraceae (31) (Table 2). The families
Bignoniaceae (43 species, mainly lianas), Euphorbiaceae (24), Meliaceae (20), Rubiaceae
(55) and Violaceae (5) also had high FIVI values in four study sites. Overall, Leguminosae,
Rubiaceae and Lauraceae were the most species rich tree families whereas the richest for
lianas were Bignoniaceae (40 species), Leguminosae (31) and Malpighiaceae (30). The ten
most important families per site (totalling 17 families; Table 2) accounted for 66.4% of the
individuals and 56.2% of the species. In the lowland Yariapo site, Buxaceae (only repre-
sented by the small tree Styloceras brokawii) was among the most important families, but
in Aguapolo not a single stem of this species was found. The families Chrysobalanaceae
and Myrtaceae were among the most important only in Aguapolo, and Piperaceae only in
Yariapo. Cecropiaceae and Melastomataceae only ranked among the most important families
at the submontane Ruins site.
According to the IVI, the 30 most important species per site (totalling 94 species)
accounted for 61.7% of the total individuals, and both Iriartea deltoidea and Rinorea viridi-
folia were the sole species that reached high values in all Wve sites (Table 3). In the lowlands,
nine species (Celtis schippii, Leonia crassa, Lunania parviXora, Meliosma herbertii, Otoba
parvifolia, Pseudolmedia laevis, Socratea exorrhiza, Sorocea briquetii and Tetragastris
altissima) were recorded as the most important on all three inventoried sites, and all of them

1C
Biodivers Conserv (2008) 17:2671–2690 2679

reached ¸10 cm dbh (Table 3). However, in the submontane forests only Piper obliquum,
Pseudolmedia laevigata and Quararibea wittii were important species shared in both sites.
Based on IVI, a total of 18 species were very important at a single site, especially in the
submontane Ruins which was characterized by its speciWc suite of 9 species (Amaioua
guianensis, Aspidosperma marcgravianum, Connarus perrottetii, Helicostylis tomentosa,
Hippotis sp., Miconia pyrifolia, M. splendens, Pourouma guianensis, and Sacoglottis
mattogrossensis; Table 3). Of the other nine species, Wve of them were found only at
Aguapolo (Brosimum alicastrum, Cheiloclinium cognatum, Hirtella racemosa, Pachira
sp., and Trichilia aV. pleeana), two in Tumupasa (Condaminea elegans and Rinorea guian-
ensis) and one each in Tequeje (Oxandra aV. acuminata) and Yariapo (Mollinedia ovata).
Ten out of the 94 species were small trees that never reached 10 cm dbh, and three other
species were lianas (Byttneria pescapriifolia, Petrea maynensis and Roentgenia bracteo-
mana), which were reported only in the lowlands.

Floristic similarity between sites

Floristic similarity was clearly higher within sites than between sites for all life-forms and
size classes as measured with both Sørensen and Steinhaus indexes (Tables 4 and 5, respec-
tively). The coeYcients obtained with the Sørensen index were generally higher than the
coeYcients obtained with the Steinhaus index, but overall patterns were similar.
The submontane Ruins site showed the lowest Xoristic similarity values in comparison
with the other four sites for all life-forms and using both indexes (Tables 4 and 5). In
general terms, the resemblance coeYcients were similar between the lowland and submon-
tane Tumupasa sites for all life-forms with the exception of lianas, which were slightly
higher among the lowlands with both indexes. The lowland Yariapo and Tequeje sites had
the highest coeYcients between sites albeit in the case of lianas, values were much more
similar.
The Xoristic similarity clusters based on the Sørensen and Steinhaus indexes were simi-
lar for all the life-forms (Fig. 3). The Ruins site was separated from the other sites in all the
plant group dendrograms. The Yariapo, Tequeje and Tumupasa sites were closer together
whereas Aguapolo site was slightly more diVerent than the other lowland sites.

Community structure

The distribution of tree diameter classes showed an inverse J-shape curve with high similarity
among all the study sites (Fig. 4). Within the tree category, small trees (<10 cm dbh) repre-
sented 72% of the individuals whereas large trees (¸10 cm dbh) comprised 28% of the
individuals. Structurally, lianas showed more variability in the mean number of individuals
per site than trees. The lowlands region had higher number of liana stems than the
submontane region, and Aguapolo was the most abundant liana site. The community structure
of the Ruins site was very similar to the other inventoried sites (Fig. 4).

Discussion

Diversity and Xoristic composition of Madidi lowlands and submontane forests

Species diversity and density are highly variable in the Madidi Amazonian rain forests,
with total woody plant species ranging from 53 to 122 and the total number of individuals

1C
2680 Biodivers Conserv (2008) 17:2671–2690

Table 3 Comparison of the 30 most important species (in bold) at Wve study sites, totalling 44 0.1-ha plots
in Madidi National Park, Bolivia

Family ScientiWc name Lowlands Submontane

Aguapolo Tequeje Yariapo Ruins Tumupasa

Anac. Tapirira guianensis Aubl. 1.08 1.54 1.48 1.61 2.01

Anno. Oxandra aV. acuminata Diels 7.34
Unonopsis Xoribunda Diels 5.75 3.31 1.77 0.81 0.76
Apoc. Aspidosperma marcgravianum Woodson 5.18
Aral. Dendropanax arboreus (L.) Decne. & Planch. 0.92 2.35 1.95 0.20 0.86
Arec. Astrocaryum murumuru Mart. 5.29 0.31 2.66 1.28
Euterpe precatoria Mart. 6.39 3.31 2.13 1.51
Iriartea deltoidea Ruiz & Pav. 6.22 21.16 20.82 8.53 18.07
Socratea exorrhiza (Mart.) H. Wendl. 2.40 5.01 4.01 1.87 0.49
Bign. **Roentgenia bracteomana 2.95 1.76 2.00
(K. Schum. ex Sprague) Urb.
Bomb. Pachira (MJM 6897) 2.68
Quararibea wittii K. Schum. & Ulbr. 1.59 4.80 2.39 5.73
Burs. Protium rhynchophyllum (Rusby) D.C. Daly 3.90 2.85 1.78 0.60
Protium spruceanum (Benth.) Engl. 6.25 0.29
Tetragastris altissima (Aubl.) Swart 3.06 2.73 2.85 1.81 4.56
Buxa. *Styloceras brokawii A.H. Gentry & 2.14 12.28 1.29 4.45
R.B. Foster
Capp. Capparis prisca J.F. Macbr. 0.38 1.84
Cecr. Pourouma guianensis Aubl. 7.51
Cela. Gymnosporia urbaniana (Loes.) Liesner 1.95 0.63 1.86
Chry. *Hirtella racemosa Lam. 12.18
Licania kunthiana Hook. f. 1.96 3.72 0.68
Comb. Terminalia oblonga (Ruiz & Pav.) 0.43 2.05 2.29 0.69
Steud. Cambess.
Conn. Connarus perrottetii (DC.) Planch. 3.82
Cyat. Cyathia delgadii Sternb. 2.74 0.49
Dich. Tapura juruana (Ule) Rizzini 2.75 0.98 0.54 0.65 3.21
Elae. Sloanea eichleri K. Schum 2.34 1.10
Sloanea fragrans Rusby 0.25 0.31 2.35
Euph. Alchornea glandulosa Poepp. 1.50 2.45 0.29
*Aparisthmium cordatum (Juss.) Baill. 4.26 0.28
Drypetes (MJM 4967) 6.73 0.29 1.69
Hieronyma alchorneoides Allemão 0.27 2.23 2.76 0.14
Mabea anadena Pax & K. HoVm. 0.75 3.29
Flac. Hasseltia Xoribunda Kunth 1.54 4.36 2.46 1.36 4.82
Lunania parviXora Spruce ex Benth. 2.68 4.88 4.65 2.94
Hipp. *Cheiloclinium cognatum (Miers) A.C. Sm. 4.22
Humi. Sacoglottis mattogrossensis Malme 2.64
Laci. Lacistema aggregatum (P.J. Bergius) Rusby 1.66 0.23 0.17 2.32 0.16
Laur. Ocotea obovata (Ruiz & Pav.) Mez 0.58 0.47 2.62 0.99
Lecy. Eschweilera (MJM 7018) 4.04 0.29
Legu. Inga thibaudiana DC. 0.25 2.56 0.14
Swartzia myrtifolia Sm. 2.11 0.21
Mela. Miconia pyrifolia Naudin 4.17
Miconia splendens (Sw.) Griseb. 2.55
Meli. Guarea guidonia (L.) Sleumer 3.68 1.03 0.36 0.27 2.36
Guarea kunthiana A. Juss. 0.73 0.22 1.80 0.41 3.28
Guarea macrophylla Vahl 5.70 2.75 0.86
Guarea pterorhachis Harms 4.68 3.71 0.58 4.11
Guarea purusana C. DC. 2.19 0.71

1C
Biodivers Conserv (2008) 17:2671–2690 2681

Table 3 continued

Family ScientiWc name Lowlands Submontane

Aguapolo Tequeje Yariapo Ruins Tumupasa

Guarea (MJM 6832) 0.50 2.41 0.90

Trichilia inaequilatera T.D. Penn. 1.24 0.78 0.64 0.20 1.83
Trichilia pallida Sw. 1.15 2.09 0.51 0.83
Trichilia pleeana (A. Juss.) C. DC. 0.76 3.69 1.76 0.52 3.13
Trichilia (MJM 4662) 0.79 0.77 0.46 2.68
Trichilia aV. pleeana (A. Juss) C. DC. 2.50
Moni. *Mollinedia ovata Ruiz & Pav. 4.87
Mora. Brosimum alicastrum Sw. 4.03
Clarisia biXora Ruiz & Pav. 0.51 1.44 2.85 0.46
Clarisia racemosa Ruiz & Pav. 1.17 0.24 1.03 0.40 1.96
Helicostylis tomentosa (Poepp. & Endl.) Rusby 5.67
Poulsenia armata (Miq.) Standl. 0.38 1.02 2.86 1.80
Pseudolmedia laevigata Trécul 1.22 10.86 3.31
Pseudolmedia laevis (Ruiz & Pav.) J.F. Macbr. 6.51 6.83 6.91 0.21 4.77
Sorocea briquetii J.F. Macbr. 4.05 2.97 2.64 0.20 0.32
Trophis caucana (Pittier) C.C. Berg 0.49 2.08
Myri. Otoba parvifolia (Markgr.) A.H. Gentry 4.17 2.65 5.34 0.21 3.83
Virola calophylla (Spruce) Warb. 2.45 0.76
Virola sebifera Aubl. 0.82 2.94 1.33 0.49 0.22
Myrs. Stylogyne cauliXora (Mart. & Miq.) Mez 3.34 3.16 0.51
Myrt. Myrcia paivae O. Berg 0.24 2.80
Myrciaria Xoribunda (H. West ex Willd.) 3.14 0.34 2.49 0.19
O. Berg
Nyct. Neea (MJM 5577) 2.16 2.23 0.90
Pipe. *Piper heterophyllum 1.82 3.39
*Piper obliquum Ruiz & Pav. 0.41 1.01 2.42 2.58 2.01
*Piper pseudo-arboreum Yunck. 0.17 3.15
Poly. Triplaris setosa Rusby 3.58 0.56 1.61 0.95 1.05
Rubi. Amaioua guianensis Aubl. 13.77
Bathysa peruviana K. Krause 8.30 0.48
Condaminea elegans Delprete 1.85
Hippotis (MJM 6235) 4.65
Macrocnemum roseum (Ruiz & Pav.) Wedd. 1.00 1.87 2.33
Sabi. Meliosma herbertii Rolfe 2.69 3.46 2.33 0.21 1.66
Sapi. Allophylus petiolulatus Radlk. 0.23 0.91 0.22 3.30
Sipa. *Siparuna biWda (Poepp. & Endl.) A. DC. 0.25 6.51 3.31 0.48 0.28
*Siparuna guianensis Aubl. 6.22 0.20 0.73
Sola. Solanum (MJM 6036) 0.23 0.17 2.73
Ster. **Byttneria pescarpriifolia Britton 0.23 2.52
Theobroma cacao L. 1.54 1.07 1.82 0.43 2.60
Tili. Pentaplaris davidsmithii Dorr & C. Bayer 8.11 0.83 2.26 1.63
Ulma. Celtis schippii Standl. 3.98 2.16 4.36 0.20 1.45
Verb. **Petrea maynensis Huber 3.27 0.59 0.34
Viol. Leonia crassa L.B. Sm. & A. Fernández 3.15 2.44 3.71 0.72
Leonia glycycarpa Ruiz & Pav. 0.20 5.02
Rinorea guianensis Aubl. 15.85
Rinorea viridifolia Rusby 23.65 11.37 3.42 6.02 10.76

Numbers indicate species Important Value Index (IVI) which were obtained as the sum of their relative den-
sity, relative dominance and relative frequency, respectively. The species with * symbol are understorey trees
(species whose stems did not reach 10 cm dbh in the inventory), and those with ** are lianas

1C
2682 Biodivers Conserv (2008) 17:2671–2690

Table 4 Pair-wise comparison of Xoristic similarity as measured by the Sørensen index (for presence/
absence data) between three lowland and two submontane study sites as recorded at 44 0.1-ha plots in Madidi
National Park, Bolivia

Life-form Sites Lowlands Submontane

Aguapolo Tequeje Yariapo Ruins Tumupasa

Total woody plants Aguapolo 0.38 § 0.12

Tequeje 0.23 § 0.06 0.39 § 0.06
Yariapo 0.23 § 0.10 0.28 § 0.10 0.32 § 0.12
Ruins 0.11 § 0.05 0.11 § 0.05 0.09 § 0.06 0.31 § 0.14
Tumupasa 0.20 § 0.06 0.23 § 0.05 0.23 § 0.08 0.12 § 0.09 0.32 § 0.09
Trees ¸ 2.5 cm dbh Aguapolo 0.41 § 0.14
Tequeje 0.25 § 0.07 0.42 § 0.06
Yariapo 0.26 § 0.12 0.30 § 0.11 0.36 § 0.13
Ruins 0.13 § 0.06 0.12 § 0.06 0.10 § 0.07 0.33 § 0.15
Tumupasa 0.21 § 0.07 0.25 § 0.05 0.25 § 0.09 0.13 § 0.10 0.35 § 0.10
Trees ¸ 10 cm dbh Aguapolo 0.35 § 0.16
Tequeje 0.21 § 0.10 0.27 § 0.11
Yariapo 0.22 § 0.13 0.23 § 0.10 0.28 § 0.12
Ruins 0.07 § 0.04 0.07 § 0.05 0.07 § 0.05 0.26 § 0.14
Tumupasa 0.17 § 0.08 0.19 § 0.07 0.18 § 0.10 0.08 § 0.07 0.24 § 0.09
Trees <10 cm dbh Aguapolo 0.38 § 0.14
Tequeje 0.24 § 0.06 0.39 § 0.06
Yariapo 0.23 § 0.11 0.28 § 0.11 0.32 § 0.13
Ruins 0.12 § 0.06 0.12 § 0.06 0.09 § 0.08 0.30 § 0.14
Tumupasa 0.19 § 0.07 0.23 § 0.05 0.23 § 0.09 0.12 § 0.09 0.33 § 0.11
Lianas ¸ 2.5 cm dbh Aguapolo 0.20 § 0.11
Tequeje 0.10 § 0.07 0.19 § 0.11
Yariapo 0.10 § 0.08 0.11 § 0.11 0.13 § 0.12
Ruins 0.04 § 0.06 0.03 § 0.04 0.02 § 0.04 0.11 § 0.12
Tumupasa 0.08 § 0.07 0.08 § 0.07 0.09 § 0.08 0.04 § 0.06 0.10 § 0.09

Mean § standard deviation Sørensen coeYcients are presented for diVerent life-forms and size classes

ranging from 157 to 503 among the 0.1-ha plots. This clearly indicates that maximizing
understanding of the diversity of a site requires inventory of several 0.1-ha plots to be
installed in diVerent sites with evidence of variation in environmental variables (soil, Xood-
ing and topography) and historical features, both of which strongly inXuence alpha diver-
sity (Gentry 1988; Duivenvoorden 1995; Svenning 1999; Tuomisto et al. 2003; Vormisto
et al. 2004b; Fine et al. 2005).
The most species-diverse 0.1-ha plots were also the plots with the highest density. For
trees, this is a common pattern also shown in many other past studies (Romero-Saltos et al.
2001; Duque et al. 2002; Phillips and Miller 2002) but in the case of lianas it is unclear.
This could be due to sampling error caused by the smaller number of liana individuals
inventoried per plot. In any case, liana density and diversity were relatively high (9.1% of
total individuals, representing 24.6% of total species); this conWrms earlier observations
that Madidi region has high concentration of lianas compared to other Neotropical sites
(Foster 1991; Foster and Gentry 1991; Burnham 2004). This could be due to the high den-
sity of large trees that seem to be greater than normal in Amazonia (Macía and Svenning
2005).
The coeYcients obtained in the Fisher’s Alpha diversity index indicate that there is no
clear diVerences in species richness between the lowland and submontane sites, although
there is high variability in plot-diversity coeYcients, which in some cases is greater than

1C
Biodivers Conserv (2008) 17:2671–2690 2683

Table 5 Pair-wise comparison of Xoristic similarity as measured by the Steinhaus index (for abundance data)
between three lowland and two submontane study sites as recorded in 44 0.1-ha plots in the Madidi National
Park, Bolivia

Life-form Sites Lowlands Submontane

Aguapolo Tequeje Yariapo Ruins Tumupasa

Total woody plants Aguapolo 0.38 § 0.16

Tequeje 0.20 § 0.08 0.33 § 0.10
Yariapo 0.16 § 0.09 0.23 § 0.11 0.24 § 0.12
Ruins 0.09 § 0.08 0.09 § 0.08 0.07 § 0.07 0.23 § 0.13
Tumupasa 0.16 § 0.07 0.21 § 0.06 0.19 § 0.09 0.10 § 0.09 0.28 § 0.11
Trees ¸ 2.5 cm dbh Aguapolo 0.41 § 0.17
Tequeje 0.22 § 0.09 0.35 § 0.11
Yariapo 0.17 § 0.10 0.24 § 0.11 0.26 § 0.13
Ruins 0.10 § 0.09 0.10 § 0.08 0.07 § 0.08 0.24 § 0.13
Tumupasa 0.17 § 0.08 0.22 § 0.07 0.20 § 0.09 0.10 § 0.10 0.29 § 0.12
Trees ¸ 10 cm dbh Aguapolo 0.30 § 0.15
Tequeje 0.18 § 0.09 0.27 § 0.13
Yariapo 0.18 § 0.12 0.23 § 0.13 0.25 § 0.15
Ruins 0.06 § 0.05 0.09 § 0.08 0.09 § 0.09 0.24 § 0.13
Tumupasa 0.15 § 0.08 0.19 § 0.08 0.19 § 0.11 0.10 § 0.09 0.25 § 0.11
Trees <10 cm dbh Aguapolo 0.33 § 0.15
Tequeje 0.18 § 0.09 0.31 § 0.09
Yariapo 0.13 § 0.07 0.20 § 0.09 0.22 § 0.12
Ruins 0.08 § 0.10 0.08 § 0.09 0.05 § 0.06 0.20 § 0.13
Tumupasa 0.14 § 0.07 0.18 § 0.07 0.16 § 0.08 0.09 § 0.10 0.27 § 0.12
Lianas ¸ 2.5 cm dbh Aguapolo 0.19 § 0.11
Tequeje 0.08 § 0.07 0.13 § 0.08
Yariapo 0.08 § 0.06 0.08 § 0.09 0.09 § 0.10
Ruins 0.02 § 0.04 0.01 § 0.02 0.02 § 0.04 0.06 § 0.07
Tumupasa 0.05 § 0.05 0.06 § 0.06 0.06 § 0.06 0.03 § 0.05 0.07 § 0.07

Mean § standard deviation Steinhaus coeYcients are presented for diVerent life-forms and size classes

100% within sites. The average diversity value for the Tumupasa submontane site was sim-
ilar to the average values for the lowland sites, although the Tequeje lowland site clearly
showed the highest value and the submontane Ruins site showed the lowest.
The results obtained here conWrm past observations that dominant families represented
in tropical forests are highly predictable (e.g. Gentry 1988; Richards 1996; Terborgh and
Andresen 1998; ter Steege et al. 2000). Families found with the highest FIVI are basically
among the same most important families reported in many other Amazonian Xoristic stud-
ies and particularly in Bolivia, when using the same sampling protocol (Foster and Gentry
1991; Kessler and Helme 1999; Phillips and Miller 2002; Araujo-Murakami 2005a, b, c).
In some families, the FIVI was high because they were among the most species-rich (e.g.
Leguminosae, Rubiaceae and Lauraceae), while other families scored high because they
include extremely abundant species (e.g. Iriartea deltoidea in the Arecaceae, Styloceras
brokawii in the Buxaceae or Rinorea viridifolia and R. guianensis in the Violaceae). For the
lianas, the most important families were the most species-rich (Bignoniaceae, Leguminosae
and Malpighiaceae).
The fact that the 30 most important species per site (totalling 94 species) accounted for
61.7% of total individuals and more speciWcally, the Wve most abundant tree and liana spe-
cies accounted for 19.4% and 17.4% of all individuals respectively, support the hypothesis
that Amazonian plant communities are dominated by a limited set of species, genera and
families (Pitman et al. 2001; Burnham 2004; Vormisto et al. 2004a; Macía and Svenning

1C
2684 Biodivers Conserv (2008) 17:2671–2690

Total woody-plants > 2.5 cm Trees > 10 cm Lianas > 2.5 cm

Yar Yar Yar
Sørensen index

Tum Teq Tum

Te q Tum Agu

Agu Agu Teq

Rui Rui Rui

Yar Yar Ya r
Steinhaus index

Te q Teq Tum

Tum Tum Teq

Agu Agu Agu

Rui Rui Rui

Similarity
0 0.5 1

Fig. 3 Floristic similarity clusters for diVerent life-forms at Wve inventoried sites, totalling 44 0.1-ha plots in
the lowlands and submontane areas of the Madidi National Park, according to both Sørensen index (species
presence-absence data) and Steinhaus index (species abundance data). The clusters for all trees
(dbh ¸ 2.5 cm) and small trees (2.5 ¸ dbh < 10 cm) are not included in this Wgure because they were very
similar to the total woody plant cluster. The site abbreviations are as follows: Agu. for Aguapolo; Rui. for
Ruins; Teq. for Tequeje; Tum. for Tumupasa; and Yar. for Yariapo

2005; cf. Tuomisto et al. 2003). These observations also add support to several studies that
report the canopy palm Iriartea deltoidea is frequently one of the most abundant species in
both 0.1-ha and 1-ha plot inventories in Western Amazonia (Clark et al. 1999; Pitman et al.
2001; Vormisto et al. 2004a). Similarly, Rinorea species (here R. viridifolia and R. guian-
ensis) are remarkably dominant in Bolivian lowland forests (Foster and Gentry 1991; Flo-
res et al. 2002; Araujo-Murakami et al. 2005a, b) as well as in other Western Amazonia
rain forests (Romoleroux et al. 1997; Romero-Saltos et al. 2001; Phillips and Miller 2002).
Future studies are needed to clarify the ecological mechanisms creating such large-scale
community structure, and particularly for Rinorea species.
A total of 18 dominant species were reported in a single site in this study (Table 3)
which could be explained by several factors including deterministic characteristics, such as
soil or other environmental site variables that create deWned forest mosaics (Gentry 1988;
Tuomisto et al. 1995, 2003) and to a lesser degree by random eVects (Hubbell 2001). Both
hypotheses have been reported as important in explaining diVerences in species composi-
tion among sites (Duque et al. 2002; Phillips et al. 2003; Tuomisto et al. 2003; Vormisto
et al. 2004a; Macía et al. 2007). Further studies which quantitatively measure the environ-
mental heterogeneity for the dominant species would be of great interest to understand
patterns of patchy distribution of Amazonian plants.
Two of these points of view (oligarchic dominance and environmental-determinism
hypotheses) have usually been considered as opposite lines of thinking to explain plant
distribution in Western Amazonia (Tuomisto et al. 2003). However, both views are not
mutually exclusive and may in fact be complementary and necessary in order to understand
Xoristic patterns in the Amazonian rain forests of the Madidi region.

Long-term eVects of human disturbance

The Ruins submontane site is notably diVerent Xoristically from all inventoried sites for the
following evidences. First, mean plot Fisher’s Alpha index was the lowest in the study area,

1C
Biodivers Conserv (2008) 17:2671–2690 2685

Fig. 4 Relative distribution of Trees Lianas

80 9
diameter classes between trees
and lianas ¸2.5 cm dbh respec- 70 8
Aguapolo
tively, for the Wve inventoried 60 7
sites in the lowlands and submon- 50 6
tane areas of Madidi National 40 5
Park, Bolivia. Bars indicate stan- 4
30
dard deviation 3
20 2
10 1
0 0
70
Tequeje
60 7
50 6
5
40
4
30
3
20 2
10 1
0 0
Number of individuals (%)

70
Yariapo 7
60
6
50
5
40
4
30 3
20 2
10 1
0 0
70
Ruins
60
50 6
5
40
4
30
3
20 2
10 1
0 0
70
Tumupasa
60
50
5
40
4
30 3
20 2
10 1
0 0
50-59.9
10-19.9
20-29.9

40-49.9

>9.5
8.5-9.49
30-39.9

3.5-4.49
4.5-5.49

7.5-8.49
2.5-9.9

>60

5.5-6.49
2.5-3.49

6.5-7.49

DBH class (cm) DBH class (cm)

and also the minimum value for all plots was recorded here. Second, the families Cecropia-
ceae and Melastomataceae showed very high FIVI values compared with other sites. These
two families are characteristic of gaps in rain forests and are predominant in second
growth (Gentry 1993). Third, according to the IVI values it contains the highest number of

1C
2686 Biodivers Conserv (2008) 17:2671–2690

exclusive dominant species (9) which were found only at this site. Fourth, the highest
dissimilarity, i.e. lowest similarity values between sites as measured with both species
presence/absence data (Sørensen index) and species abundance data (Steinhaus index), was
found in the pair-wise comparison involving the Ruins site for all life-forms. Fifth, Xoristic
similarity clusters computed with both Sørensen and Steinhaus indexes showed the Ruins
site as the most dissimilar for all plant groups. However, its forest structure was similar to
the other studied sites and indistinguishable from other mature forests, although lianas
showed more structural variability than trees.
Some studies have reported that past human disturbance may have long-lasting eVects
on plant community (Balée and Campbell 1990; Thompson et al. 2002; Chazdon 2003; van
Gemerden et al. 2003) although others did not Wnd any current anthropogenic eVect (White
and Hood 2004). The present study adds to the Wrst series of observations and infer past
human disturbance is likely to be the main reason of Xoristic diVerences in the Ruins site.
It is widely accepted that disturbance promotes liana abundance in secondary and frag-
mented forests (Putz 1984; Schnitzer and Carson 2000; Schnitzer 2005), but in cases of
very old disturbance, liana regeneration is lower (DeWalt et al. 2000). In the case of the
Ruins site, where disturbance occurred more than 300 years ago, liana density was similar
to other inventoried sites, but its diversity was the lowest.

Comparison of woody plants in 0.1-ha inventories in Bolivia

Using the 0.1-ha plot/transect protocol, the Amazonian rain forests of Bolivia (<1000 m
elevation) proved to be its most diverse forests (Table 6; Jørgensen et al. 2005a). However,
there are some dry forest plots in the Department of Pando, which challenge these lowland
rain forests as having the highest diversity (Phillips and Miller 2002).
The present study supports the observation that liana abundance is higher in seasonal
regions (Schnitzer 2005) since Bolivian dry forests overall have a higher number of liana
individuals compared to other wet habitats (Table 6).
The Gentry-transect sampling protocol (2 £ 500 m) produced higher values of woody
plant diversity and density as compared to a single 0.1-ha plot in all habitats (Table 1 vs.
Table 6). The plot sampling protocol (20 £ 50 m and 10 £ 100 m) produced similar diver-
sity and density data. These diversity and density diVerences between transects and plots
may be due to several factors. It is clear that a narrow transect protocol will study a more
heterogeneous area than a plot protocol, and therefore the number of species should be
higher, but it is unclear why plant density would be higher. Perhaps, the Gentry-transects
did not strictly follow a compass-direction line and changed periodically to incorporate
some new woody plants to the inventory. In any case, this transect vs. plot comparison indi-
cates that to gain a more comprehensive overall idea of woody plant richness patterns in a
given region, it is necessary to inventory several plots/transects at an average of 10 0.1-ha
plots (here between 6 and 12 plots) and at least 500 m apart in a relatively small geograph-
ical area (here between 20 and 40 km2 per site), as plot diversity and density data can vary
more than 100% within sites (Tables 1, 6). Since the two most common 0.1-ha plot proto-
cols produced similar data, their application to inventory woody plants make no qualitative
diVerence. However, in the interest of rigor it is recommended to be consistent and use a
single sampling method within a site or region.
It has been suggested that for more rigorous comparisons of diversity at diVerent sites,
one should include samples with approximately equal numbers of individuals (ignoring the
area), using the same sampling protocol and dbh limit, and preferably with 2,000+ individ-
uals inventoried (Condit et al. 1998). In order to make more precise Xoristic comparisons of

1C
 Table 6 Comparison of diversity and density of woody plants (trees and lianas) in 0.1-ha plot/transect inventories (inv.) carried out in Bolivia

Vegetation type Elevation Department Woody plants ¸ 2.5 cm Trees ¸ 2.5 cm Lianas ¸ 2.5 cm Alpha Plot/Transect No. Reference
(m) Fisher dimension (m2) Inv.
Spp. Indiv. Spp. Indiv. Spp. Indiv.

Lowland forest 260–420 La Paz 285 1,936 220 1715 65 214 92.2 1000 (50 £ 20 m) 7 This study
Lowland forest 280 La Paz 204 434 151 341 53 93 150.2 1000 (500 £ 2 m) 1 Foster and Gentry (1991)
Lowland forest 300–360 La Paz 310 2,680 274 2369 36 311 90.6 1000 (100 £ 10 m) 12 Araujo-Murakami et al. (2005a)
Lowland forest 300–600 La Paz 294 2,776 262 2532 32 244 83.1 1000 (100 £ 10 m) 13 Araujo-Murakami et al. (2005b)
Ridge top lowland foothills 360–380 La Paz 175 483 131 398 44 85 98.6 1000 (500 £ 2 m) 1 Foster and Gentry (1991)
Lowland forest 425–550 La Paz 312 2,075 249 1866 67 199 101.9 1000 (50 £ 20 m) 6 This study
Lowland forest 460–610 La Paz 335 2,626 258 2343 75 278 101.9 1000 (50 £ 20 m) 11 This study
Submontane forest 735–1045 La Paz 313 2,811 256 2576 59 231 90.2 1000 (50 £ 20 m) 8 This study
Biodivers Conserv (2008) 17:2671–2690

Submontane forest 800–1070 La Paz 416 3,374 337 3125 79 242 124.8 1000 (50 £ 20 m) 12 This study
Montane rain forest 1240–1385 Cochabamba – – 83 328 – – 35.8 1000 (50 £ 20 m) 3 Macía and Fuertes (unpubl.)
Lower montane forest 1500–1550 La Paz 115 268 89 229 17 27 76.3 500 (250 £ 2 m) 1 Phillips and Miller (2002)
Lower montane forest 1520–1560 La Paz 159 514 141 470 12 28 78.8 1000 (500 £ 2 m) 1 Phillips and Miller (2002)
Ridge top montane 1560 Cochabamba – – 44 437 – – 12.2 1000 (100 £ 10 m) 1 Macía and Fuertes (unpubl.)
rain forest
Montane rain forest 2380–2450 La Paz 98 550 78 418 16 90 34.7 1000 (500 £ 2 m) 1 Phillips and Miller (2002)
Montane forest: ceja 3225–3503 La Paz – – 66 3360 – – 11.6 1000 (100 £ 10 m) 10 Araujo-Murakami et al. (2005c)
de monte
Moist forest 100 Santa Cruz 116 445 70 288 46 157 51.0 1000 (500 £ 2 m) 1 Phillips and Miller (2002)
Moist forest 280 Santa Cruz 111 320 74 218 37 102 60.2 1000 (500 £ 2 m) 1 Phillips and Miller (2002)
Moist forest 375 Santa Cruz 63 180 37 105 26 75 34.5 1000 (500 £ 2 m) 1 Phillips and Miller (2002)
Dry forest 100 Pando 169 331 130 262 36 66 138.3 1000 (500 £ 2 m) 1 Phillips and Miller (2002)
Dry forest 160 Pando 149 358 116 295 32 62 95.8 1000 (500 £ 2 m) 1 Phillips and Miller (2002)
Dry forest 300 Santa Cruz 83 395 53 275 30 120 32.0 1000 (500 £ 2 m) 1 Phillips and Miller (2002)
Dry forest 700–800 Santa Cruz 47 179 45 169 2 10 20.8 600 (300 £ 2 m) 1 Phillips and Miller (2002)
Dry forest 818–1015 La Paz 171 4,709 – 4291 – 418 34.8 1000 (100 £ 10 m) 13 Fuentes et al. (2004)
Dry forest 1000 La Paz 79 465 50 331 29 134 27.3 1000 (500 £ 2 m) 1 Foster and Gentry (1991)
Dry forest 1020–1200 La Paz 80 339 59 278 – – 33.0 1000 (500 £ 2 m) 1 Kessler and Helme (1999)
Chaco upland thorn forest 250 Santa Cruz 32 360 27 351 5 9 8.5 1000 (500 £ 2 m) 1 Phillips and Miller (2002)
Chaco upland thorn forest 350 Santa Cruz 41 420 29 348 12 72 11.2 1000 (500 £ 2 m) 1 Phillips and Miller (2002)
Chaco upland thorn forest 350 Santa Cruz 51 366 43 249 8 117 16.1 1000 (500 £ 2 m) 1 Phillips and Miller (2002)

Fisher’s Alpha diversity index was computed on total woody plants when data were available for trees and lianas, otherwise just for trees
2687

1C
2688 Biodivers Conserv (2008) 17:2671–2690

diVerent tropical forests sites and assess their diversity in future forest inventories, the
results of the present study support using this approach in the 0.1-ha plot protocol (measuring
all woody plants ¸2.5 cm dbh) as well as considerating this minimum number of approx-
imately 2000 individuals per site.

Acknowledgements I am grateful to J. Quisbert for collaboration during data collection and herbarium
processing, and all the Tacana people from several communities who assisted during Weldwork. Thanks to
S.G. Beck and the staV at Herbario Nacional de Bolivia for working space and facilities, and to M.T. Tellería
for support and encouragement. I acknowledge Ministerio de Desarrollo Sostenible y Medio Ambiente and
Dirección General de Biodiversidad of Bolivia for providing the necessary permits for conducting the Weld-
work in Madidi National Park, and the following specialists for kindly providing identiWcations of specimens:
P. Acevedo, W.S. Alverson, W.R. Anderson, G. Aymard, S.G. Beck, C.C. Berg, P. Berry, T. Borsh, A.E. Brant,
J.M. Cardiel, T. Croat, D. Daly, L.J. Dorr, R. Duno, H.J. Esser, R. Foster, A. Fuentes, B.K. Holst, O. Huber,
P.M. Jørgensen, L. Kelly, M. Lehnart, R. Liesner, L. Lohmann, J. Miller, J.D. Mitchell, M. Nee, R. Ortiz, J.J.
Pipoly III, G.T. Prance, J. Pruski, S.S. Renner, L. Sánchez, R. Seidel, C. Stace, B. Ståhl, P.F. Stevens, C.M.
Taylor, W. Thomas, H. van der WerV, and B. Wallnöfer. Finally, thanks to A. Ibáñez and P.M. Jørgensen for
their helpful comments on an earlier draft. This work was made possible by support from Consejería de
Educación, Comunidad de Madrid, and Consejo Superior de Investigaciones CientíWcas (programa I3P).

References

Araujo-Murakami A, Bascopé F, Cardona-Peña V, De la Quintana D, Fuentes A, Jørgensen P, Maldonado C,

Miranda T, Paniagua-Zambrana N, Seidel R (2005a) Composición Xorística y estructura del bosque
amazónico preandino en el sector del Arroyo Negro, Parque Nacional Madidi, Bolivia. Ecol Boliv
40(3):281–303
Araujo-Murakami A, Cardona-Peña V, De la Quintana D, Fuentes A, Jørgensen P, Maldonado C, Miranda T,
Paniagua-Zambrana N, Seidel R (2005b) Estructura y diversidad de plantas leñosas en un bosque
amazónico preandino en el sector del río Quendeque, Parque Nacional Madidi, Bolivia. Ecol Boliv
40(3):304–324
Araujo-Murakami A, Jørgensen P, Maldonado C, Paniagua-Zambrana N (2005c) Composición Xorística y
estructura del bosque de ceja de monte en Yungas, sector de Tambo Quemado-Pelechuco, Bolivia. Ecol
Boliv 40(3):325–338
Armentia N (1897) Límites de Bolivia con el Perú por la parte de Caupolican. OWcina nacional de inmigración,
estadística y propaganda geográWca, La Paz
Balée W, Campbell DG (1990) Evidence for the successional status of liana forest (Xingu River Basin,
Amazonian Brazil). Biotropica 22:36–47
Burnham RJ (2004) Alpha and beta diversity of lianas in Yasuní, Ecuador. For Ecol Manage 190:43–55
Chazdon RL (2003) Tropical forest recovery: legacies of human impact and natural disturbances. Perspect
Plant Ecol Evol Syst 6:51–71
Clark DB, Palmer MW, Clark DA (1999) Edaphic factors and the landscape-scale distributions of tropical
rain forest trees. Ecology 80:2662–2675
Clinebell RR II, Phillips O, Gentry AH, Stark N, Zuuring H (1995) Prediction of neotropical tree and liana
species richness from soil and climatic data. Biodivers Conserv 4:56–90
Condit R, Foster RB, Hubbell SP, Sukumar R, Leigh EG, Manokaran N, Loo de Lao S, LaFrankie JV, Ashton
PS (1998) Assessing forest diversity on small plots: calibration using species-individual curves from 50-
ha plots. In: Dallmeier F, Comiskey JA (eds) Forest biodiversity research, monitoring and modeling.
MAB series, vol 20. UNESCO, Paris, pp 247–268
Curtis JT, McIntosh RP (1951) An upland forest continuum in the prairie-forest border region of Wisconsin.
Ecology 32:476–496
DeWalt SJ, Schnitzer SA, Denslow JS (2000) Density and diversity of lianas along a chronosequence in a
central Panamanian lowland forest. J Trop Ecol 16:1–19
Duque AJ, Sánchez M, Cavelier J, Duivenvoorden JF (2002) DiVerent Xoristic patterns of woody understorey
and canopy plants in Colombian Amazonia. J Trop Ecol 18:499–525
Duivenvoorden JF (1995) Tree species composition and rain forest-environment relationships in the middle
Caquetá area, Colombia, NW Amazonia. Vegetatio 120:91–113
Fine PVA, Daly DC, Villa G, Mesones I, Cameron KM (2005) The contribution of edaphic heterogeneity to
the evolution and diversity of Burseraceae trees in the western Amazon. Evolution 59:1464–1478

1C
Biodivers Conserv (2008) 17:2671–2690 2689

Fisher RA, Corbet AS, Williams CB (1943) The relation between the number of species and the number of
individuals in a random sample of an animal population. J Anim Ecol 12:42–58
Flores JG, Batte-Batte C, Dapara J (2002) Caracterización de la vegetación del río Undumo y su importancia
para la conservación de la fauna silvestre. Ecol Boliv 37:23–48
Foster R (1991) Plant communities of Alto Madidi, Bajo Tuichi, and the foothill ridges. In: Parker III TA,
Bailey B (eds) A biological assessment of the Alto Madidi Region. RAP Working Papers 1. Conservation
International, Washington DC, pp 15–19
Foster R, Gentry A (1991) Plant diversity. In: Parker III TA, Bailey B (eds) A biological assessment of the
Alto Madidi Region. RAP Working Papers 1. Conservation International, Washington DC, pp 20–21
Fuentes A, Araujo A, Cabrera H, Canqui F, Cayola L, Maldonado C, Paniagua N (2004) Estructura, composición
y variabilidad del bosque subandino xérico en un sector del valle del río Tuichi, ANMI Madidi, La Paz
(Bolivia). Rev Boliv Ecol Conserv 15:41–62
Fuentes A (2005) Una introducción a la vegetación de la región de Madidi. Ecol Boliv 40(3):1–31
Gentry AH (1988) Changes in plant community diversity and Xoristic composition on environmental and
geographical gradients. Ann Missouri Bot Gard 75:1–34
Gentry AH (1993) A Weld guide to the families and genera of woody plants of northwest South America
(Colombia, Ecuador, Peru) with supplementary notes on herbaceous taxa. Conservation International,
Washington DC
Gentry AH, Dodson C (1987) Contribution of nontrees to species richness of a tropical rain forest. Biotropica
19:149–156
Heckenberger MJ, Kuikuro A, Kuikuro UT, Russell JC, Schmidt M, Fausto C, Franchetto B (2003) Amazonia
1492: pristine forest or cultural parkland? Science 301:1710–1714
Holmgren PK, Holmgren NH, Barnett LC (1990) Index Herbariorum, part I: the herbaria of the world, Regnum
Vegetabile 120, 8th edn. The New York Botanical Garden, New York
Hubbell SP (2001) The uniWed neutral theory of biodiversity and biogeography. Princeton University Press,
Princeton
Jørgensen PM, Macía MJ, Fuentes A, Beck SG, Kessler M, Paniagua N, Seidel R, Maldonado C, Araujo-
Murakami A, Cayola L, Consiglio T, Killeen TJ, Cabrera WH, Bascopé F, De la Quintana D, Miranda
T, Canqui F, Cardona-Peña V (2005a) Lista anotada de las plantas vasculares registradas en la región de
Madidi. Ecol Boliv 40(3):70–169
Jørgensen PM, Macía MJ, Killeen TJ, Beck SG (eds) (2005b) Estudios botánicos de la región de Madidi. Ecol
Boliv 40(3):1–452
Kessler M, Helme N (1999) Floristic diversity and phytogeography of the central Tuichi valley, an isolated
dry forest locality in the Bolivian Andes. Candollea 54:341–366
Legendre P, Legendre L (1998) Numerical ecology. Elsevier, Amsterdam
Macía MJ, Svenning JC (2005) Oligarchic dominance in western Amazonian plant communities. J Trop Ecol
21:613–626
Macía MJ, Ruokolainen K, Tuomisto H, Quisbert J, Cala V (2007) Congruence between Xoristic patterns of
trees and lianas in a southwest Amazonian rain forest. Ecography 30:561–577
Mori SA, Boom BM, De Carvalho AM, Dos Santos TS (1983) Ecological importance of Myrtaceae in an eastern
Brazilian wet forest. Biotropica 15:68–70
Navarro G, Ferreira W, Antezana C, Arrázola S, Vargas R (2004) Bio-corredor Amboró Madidi, zoniWcación
ecológica. Editorial FAN, Santa Cruz
Parker III TA, Bailey B (eds) (1991) A biological assessment of the Alto Madidi region and adjacent areas of
northwest Bolivia. RAP Working Papers 1. Conservation International, Washington, DC
Pérez-Salicrup DR, Sork VL, Putz FE (2001) Lianas and trees in a liana forest of Amazonian Bolivia.
Biotropica 33:34–47
Phillips OL, Miller JS (2002) Global patterns of plant diversity: Alwyn H. Gentry’s forest transect data set.
Monogr Syst Bot Missouri Bot Gard 89:1–319
Phillips OL, Núñez P, Lorenzo A, Peña A, Chuspe ME, Galiano W, Yli-Halla M, Rose S (2003a) Habitat
association among Amazonian tree species: a landscape-scale approach. J Ecol 91:757–775
Phillips OL, Vásquez R, Núñez P, Lorenzo A, Chuspe ME, Galiano W, Peña A, Timaná M, Yli-Halla M, Rose
S (2003b) EYcient plot-based Xoristic assessment of tropical forest. J Trop Ecol 19:629–645
Pitman NCA, Terborgh JW, Silman MR, Núnez P, Neill DA, Cerón CE, Palacios WA, Aulestia M (2001)
Dominance and distribution of tree species in upper Amazonian terra Wrme forests. Ecology 82:2101–2117
Putz FE (1984) The natural history of lianas on Barro Colorado Island, Panama. Ecology 65:1713–1724
Renard-Casevitz FM, Saignes T, Taylor AC (1988) Al este de los Andes: relaciones entre las sociedades
amazónicas y andinas entre los siglos XV y XVII. Ediciones Abya-Yala, Quito
Richards PW (1996) The tropical rain forest: an ecological study, 2nd edn. Cambridge University Press,
Cambridge

1C
2690 Biodivers Conserv (2008) 17:2671–2690

Romero-Saltos H, Valencia R, Macía MJ (2001) Patrones de diversidad, distribución y rareza de plantas

leñosas en el Parque Nacional Yasuní y la Reserva Étnica Huaorani, Amazonía ecuatoriana. In:
Duivenvoorden JF, Balslev H, Cavelier J, Grandez C, Tuomisto H, Valencia R (eds) Evaluación de
recursos vegetales no maderables en la Amazonía noroccidental. IBED, Universiteit van Amsterdam, pp
131–162
Romoleroux K, Foster R, Valencia R, Condit R, Balslev H, Losos E (1997) Árboles y arbustos (dap ¸ 1 cm)
encontrados en dos hectáreas de un bosque de la Amazonía ecuatoriana. In: Valencia R, Balslev H (eds)
Estudios sobre diversidad y ecología de plantas. PontiWcia Universidad Católica del Ecuador, Quito, pp
189–215
Schnitzer SA (2005) A mechanistic explanation for global patterns of liana abundance and distribution. Am
Nat 166:262–276
Schnitzer SA, Carson WP (2000) Have we missed the forest because of the trees? Trends Ecol Evol 15:375–376
Smith DN, Killeen TJ (1998) A comparison of the structure and composition of montane and lowland tropical
forest in the serranía Pilón Lajas, Beni, Bolivia. In: Dallmeier F, Comiskey JA (eds) Forest biodiversity in
North, Central and South America, and the Caribbean. MAB series, vol 21. UNESCO, Paris, pp 681–700
Svenning JC (1999) Microhabitat specialization in a species-rich palm community in Amazonian Ecuador. J
Ecol 87:55–65
ter Steege H, Sabatier D, Castellanos H, Van Andel T, Duivenvoorden J, De Oliveira AA, Ek R, Lilwah R,
Maas P, Mori S (2000) An analysis of the Xoristic composition and diversity of Amazonian forests
including those of the Guiana Shield. J Trop Ecol 16:801–828
ter Steege H, Pitman N, Sabatier D, Castellanos H, van der Hout P, Daly DC, Silveira M, Phillips O, Vazquez
R, van Andel T, Duivenvoorden J, Adalardo de Oliveira A, Ek R, Lilwah R, Thomas R, van Essen J,
Baider C, Maas P, Mori S, Terborgh J, Núñez P, Mogollón H, Morawetz W (2003) A spatial model of
tree
-diversity and tree density for the Amazon. Biodivers Conserv 12:2255–2277
Terborgh J, Andresen E (1998) The composition of Amazonian forests: patterns at local and regional scales.
J Trop Ecol 14:645–664
Thompson J, Brokaw N, Zimmerman JK, Waide RB, Everham III EM, Lodge DJ, Taylor CM, García-Mont-
iel D, Fluet M (2002) Land use history, environment, and tree composition in a tropical forest. Ecol Appl
12:1344–1363
Tuomisto H, Ruokolainen K, Kalliola R, Linna A, Danjoy W, Rodriguez Z (1995) Dissecting amazonian
biodiversity. Science 269:63–66
Tuomisto H, Ruokolainen K, Yli-Halla M (2003) Dispersal, environment, and Xoristic variation of western
Amazonian forests. Science 299:241–244
van Gemerden BS, OlV H, Parren MPE, Bongers F (2003) The pristine rain forest? Remnants of historical
human impacts on current tree species composition and diversity. J Biogeogr 30:1381–1390
Vormisto J, Svenning JC, Hall P, Balslev H (2004a) Diversity and dominance in palm (Arecaceae) commu-
nities in terra Wrme forests in the western Amazon basin. J Ecol 92:577–588
Vormisto J, Tuomisto H, Oksanen J (2004b) Palm distribution patterns in Amazonian rainforests: what is the
role of topographic variation? J Veg Sci 15:485–494
White DA, Hood CS (2004) Vegetation patterns and environmental gradients in tropical dry forests of the
northern Yucatan Peninsula. J Veg Sci 15:151–160

1C
View publication stats