A Comparison of the Structure and Composition of Montane and Lowland Tropical Forest in the Serranía Pilón Lajas, Beni, Bolivia
Compiled by David N. Smith¹ (deceased) and Timothy J. Killeen²
Figures in the article are in PDF format and can only be viewed with Adobe Acrobat Reader (free download)
1. Herbario Nacional de Bolivia
2. Missouri Botanical Garden
Read the entire article in PDF format (ideal for printing).
A quantitative inventory is reported for two one hectare permanent plots established in the Serranía Pilón Lajas in Beni, Bolivia. The floristic diversity, as measured by tree species with diameter at breast height greater than 10 cm, was documented at 78 species/ha for the plot established on the piedmont at 270 m above sea level (Río Colorado) and at 146 species/ha for the plot established on the crest of the serranía at 900 m (Cumbre Pilón). The structure of the two plots fall within the range of values reported for mature humid tropical forest associations with 588 and 647 trees/ha and a total basal area of 26.4 and 30.6 m2/ha for Río Colorado and Cumbre Pilón respectively. The floristic composition between the two sites was moderately similar with 25 species in common, but differences in the abundance of individual species revealed greater differences. The Río Colorado plot was dominated by the palm Iriartea deltoidea and a widespread species in the Moraceae, Pseudolmedia laevis. The Cumbre Pilón was not dominated by any particular species, but the genus Inga was particularly diverse, as was the Rubiaceae, characteristics typical of the Andean montane forest formations. Both plots showed considerable floristic similarity with other inventories conducted in lowland tropical forest in adjacent regions of Bolivia. Sixty species are identified as being regionally common; these species were shown to be widely distributed taxa with distributions throughout western Amazonia.
Quantitative floristic inventories based on one hectare permanent plots have been used in recent years to characterize forest vegetation throughout the tropics. Typically, investigators are interested in documenting the structure and floristic composition of forest communities. Recently, permanent plots have been used to address changes that may be occurring in tropical ecosystems due to climate change (Phillips and Gentry, 1994) and their value as a tool in environmental monitoring programs is widely recognized. The first such study in Bolivia was conducted by Boom (1986) who sampled a "tierra firme" forest in northern Beni as part of an ethnobotanical study of the Chácobo indigenous group. Subsequent inventories in other parts of the Beni (de Aguila, 1995) and Santa Cruz (Vargas et al. 1994; Vargas, 1995) have contributed to the knowledge of the tropical forest formations in Bolivia. Several other studies are currently underway and the possibility for making biogeographical or phytosociological comparisons based on specimen-based, quantitative inventories is increasingly possible. This report provides data for two permanent plots established on the Serranía Pilón Lajas that was undertaken as part of a dendrological inventory of the region conducted between 1988 and 1992. The structure and composition is documented for a humid montane forest located on the crest of the serranía at 900 m and a humid lowland forest situated on the piedmont at 270 m above sea level (Figure 1).
The Serranía Pilón Lajas consists of a low mountain ridge that has a maximum elevation between 800 and 1200 m; the adjacent piedmont and the intermontane valley are between 300 and 500 m in elevation. Pilón Lajas is part of the geological region designated as the Faja Subandina (subandean strip) and is one of several parallel ridges that are oriented from the southeast to the northwest. These low mountain ranges are composed of Ordovician, Silurian, and Cretaceous sandstone, shale, and mudstone (Oblitas and Brockmann, 1978). The Quiquibey valley located to the southwest of the serranía is a syncline and its eastern escarpment is essentially a truncated anticline that is much more precipitous in comparison to the western slope. The soils on the piedmont are moderately deep, ranging from sandy loam to clay ultisols and entisols that are derived from Quaternary sediments with low to moderate levels of fertility; pH values range from 4.8 to 6.0. The upper soil horizons are generally well-drained, but many sites have a mottled B horizon indicating that these soils are saturated periodically. The soils on the serranía are extremely variable depending upon slope; in general, they are loamy clays that are slightly acidic to alkaline and range from entisols to oxisols (Merlot et al., 1994).
The climate of the area is humid to very humid with a mean annual precipitation of 2550 mm and a mean annual temperature of 25.9° C reported for Rurrenabaque, a small town located on the piedmont just north of the study area on the Beni River (Hanagarth, 1993). The rainy season extends from November to April, while the driest months of the year are July, August, and September. The area, like the rest of lowland Bolivia, experiences strong cold fronts during the austral winter and temperatures as low as 8° C have been recorded during the month of July. Although precise measurements do not exist, precipitation undoubtedly increases with increasing altitude while temperatures decrease.
The natural vegetation of the zone is humid tropical forest that changes in composition and structure according to elevation. The forest on the piedmont has been classified as a humid lowland forest, while the forest formations on the ridges has been designated as montane humid forest (Beck et al., 1993). These forest formations are composed of different plant associations that relate to changes in soil, humidity, and successional state. On the piedmont, there are examples of forest types with impeded drainage and approximately 25 km to the East are the inundated savannas of San Borja and Reyes which are part of the vast wetland complex of the Llanos de Moxos. The vegetation on the Serranía Pilón Lajas is a patchwork of different communities; landslides are common and a casual observer with the assistance of binoculars can distinguish a patchwork of plant communities in different successional stages. The fauna in the area is both interesting and diverse; although inventories are preliminary, numerous endemic and endangered species are known to exist in the area or in adjacent regions (Emmons, 1991; Merlot et al., 1994).
Land-use in the area is changing due to individuals practicing shifting agriculture. The areas most heavily affected by these activities are adjacent to a highway that was constructed between 1984 and 1991. In 1992 the Bolivian government declared much of the region as an indigenous territory, as well as recognizing the area as a Biosphere Reserve: (Figure 1). Currently, efforts are underway to organize a management framework for the protected area that includes participation of the indigenous groups, colonist communities, and conservation organizations (Merlot et al., 1994).
The studies were conducted in two one hectare permanent plots according to the guidelines outlined by Campbell (1989); field work was conducted in 1990 and 1991. In each plot, all trees with a diameter at 1.3 m (diameter breast height or "dbh") greater or equal to 10 cm were tagged, measured, identified, and catalogued. Trees with buttressed trunks or stilt roots were measured 20 cm above these structures. Each tree was marked with a consecutively numbered aluminum tree tag that was nailed 10 cm above the point of measurement. The total height of the tree was estimated visually; trees that straddled the boundaries of subplots were included if the midpoint was situated within the plot. Voucher collections were made for each species; individual trees that could be unequivocally identified in the field were not recollected. Identification of species was made by comparison of voucher specimens to fertile material collected throughout the course of the project. General collections have been distributed to LPB, MO, USZ, and BOLV, as well as a variety of other institutions; voucher specimens are deposited at LPB and MO.
The first plot was established on the piedmont in a forest remnant at the Colegio Técnico Agropecuario Río Colorado, a secondary school situated 35 km north of Yucumo on the highway to Rurrenebaque (Figure 1). The plot is a 100 x 100 m square composed of twenty five 20 x 20 m subplots; it is situated approximately 5 km to the northeast of the school campus facilities on a non-inundated terrace adjacent to the Río Colorado (14°05'S; 67°05'W; 270 m above sea level). According to local sources no previous agricultural or lumber extraction had occurred in the immediate vicinity of the study site in recent memory. The "Río Colorado" plot can be relocated by asking for assistance from the school administrators who have promised to maintain it as a long-term study sight; the boundaries of the plot are the four cardinal directions.
The second plot was located just below and to the northeast of the crest of the Serranía Pilón Lajas near the pass (or "cumbre") between the Quiquibey bridge and the community of Yucumo. The plot has an idiosyncratic design that reflects local topography, but essentially is L-shaped with one leg running downslope along a ridge and a second perpendicular leg that is situated on a terrace below a massive sandstone cliff (15°15'S; 67°00'W; 850-950 m). The soils ranged from superficial and rocky to locally deep in restricted spots; in general they were red and well drained. The "Cumbre Pilón" plot can be relocated by driving 24 km from Yucumo towards the Quiquibey bridge. The plot is situated approximately 1 km to the East of the highest point of the pass; the trail is situated to the left when facing towards the Quiquibey bridge and crosses a field of large boulders. The integrity of this plot as a long term study site has been compromised by subsequent tree felling by colonists; in addition, a large, natural mud-slide occurred during the 1992 rainy season and destroyed the vegetation in approximately 5 subplots.
Plot data was summarized according tothe stantard protocols (Cambpell, 1989): abundance or density is the number of trees registered in the plots, frequency is the number of subplots in which a single species has been recorded, and basal area is the sum of the cross-sectional areas of all the stems for each individual species as calculated from dbh values. These summary statistics are in turn used to calculate percentage values based on the totals for each individual parameter summed over all species. These percentage values are traditionally used for making comparison with other sites and are expressed as relative density, relative frequency and relative dominance. A single summary statistic or "importance value," is calculated by summing the three relative values. A family importance index is calculated in a similar fashion by summing the values for all species within a family; however, frequency data is excluded and a relative diversity value is employed that is based on the number of species for a specific family as a percentage of the total number of species.
The structure of the forest vegetation is represented in Figure 2. The number of individual plants in the different dbh size classes shows that both localities have the typical inverted J curve where abundance decreases with increasing diameter. Nonetheless, the two plots are markedly different with respect to the abundance of trees with different diameters. Although both plots were selected as being representative of "high forest," Río Colorado has more large trees when compared to Cumbre Pilón. This is demonstrated by a larger number of individuals with a dbh greater than 80 cm and, most noticeable, by the presence of a single very large tree (Hura crepitans) with a dbh of 162 cm at Río Colorado. Nonetheless, the Cumbre Pilón plot shows greater total basal area due to a higher total density of individual trees (see Appendix), as well as a greater number of trees in the mid-class size range between 30 and 60 cm in diameter (Figure 2).
The relationship between the area sampled and the number of species encountered is presented in Figure 3. The slope of the species-area curve approaches zero at approximately the 19th subplot for the Río Colorado plot, thus indicating that a one hectare sample was sufficient to characterize the structure and composition of this locality. Nonetheless, like most tropical forest sites continued sampling revealed additional taxa; common widespread species such as Swietenia macrophylla and Cariniana estrellensis were observed within 100 m of the sample plot. The greater species richness of the Cumbre Pilón plot is clearly demonstrated in Figure 3 and this trend was apparent at small sample sizes. The slope of the species-area curve does not approach zero and further sampling would be necessary for a reliable inventory of this forest type. The absolute values for abundance, frequency, and basal area for individual species for the two plots are presented in the Appendix.
The ten most important species at the Río Colorado plot are presented in Table 1. The two most important species are also the two most abundant species, Iriartea deltoidea and Pseudolmedia laevis; together these two species represent 28.4% of the total number of trees within the plot. In contrast, a total of 32 species or 41% of the total are represented by a single individual. The relationship between frequency and abundance is pronounced with a r = 0.89, while the correlation between abundance and dominance is r = 0.70. As such, the importance value essentially provides a measure of abundance that is weighted for species with very large trees. Evaluation of the plot data for individual species can, in some cases, permit the classification of specific species into a emergent, canopy, subcanopy, or understory trees. Hura crepitans, Spondias venosa, Sapium marmierii, Poulsenia armata, and Ficus killipii were emergents species. Most large trees were represented by only a few individuals, but Sapium marmierii and Poulsenia armata were represented in all class sizes. Palms were the most conspicuous elements of the canopy and subcanopy strata, as well as several species in the Annonaceae, Lauraceae, and Moraceae. Common understory trees were Lunania parviflora and Salacia macrantha; the most common understory plants were shrubs in the Rubiaceae and Melastomataceae that did not meet the measurement criteria of 10 cm dbh. The most important family in the Río Colorado plot was the Moraceae, which was the most diverse of all the families, as well as the most dominant (Table 2). Other diverse families were the Palmae, Lauraceae, and Annonaceae. The Euphorbiaceae was ranked as important due largely to three individuals of Hura crepitans with large diameters. This species has been characterized as a pioneer species that requires high light intensity for canopy recruitment (T. Gullison, pers. comm.). Since H. crepitans was not recorded in the small diameter size-classes, this provides supporting evidence for the supposition that this study site was a mature forest community with little previous disturbance.
The Cumbre Pilón plot has four species that are ranked as of approximately equal importance (Table 3); together they represent 17.5% of the total number of trees found in that plot. In contrast to the Río Colorado plot, the occurrence of species represented by a single individual was even greater (56%). The four species with a largest basal area (Poulsenia armata, Tetragastris mucronata, Dendropanax arboreum, and Sapium glandulosum) are characterized by having numerous stems that also have relatively large diameters. The tallest trees in the plot (i.e., > 35 m tall) were Poulsenia armata, Tetragastris mucronata, Terminalia amazonica, Terminalia oblonga, Myroxylon balsamum, Pterocarpus amazonum, Inga nobilis, Swartzia cf. jorori, Clarisia racemosum, Clarisia biflora, Helicostylis sp., Otoba parviflora, Virola peruvianum, Pouteria krukovii, and Ampelocera ruizii. The subcanopy and understory was dominated by Hasseltia floribunda and two species of Allophylus. The most important family in the plot was the Leguminosae with a total of 25 species and large values for abundance and basal area. The most diverse genus was Inga with a total of 11 species. Other diverse families were the Moraceae, Rubiaceae, Sapotaceae, Myrtaceae, Euphorbiaceae, Lauraceae, and Meliaceae, while the Burseraceae, Flacourtiaceae, Sapindaceae and Araliaceae had high importance values due to the relative abundance of one or two species (Table 2).
Table 4 shows the total values for the two sites and compares them with selected other localities in the Neotropics. Both forest communities fall within the range of tropical forest associations with respect to tree density and total basal area. The piedmont plot was most notable for the abundance and diversity of palm species, a phenomenon that is not infrequent in lowland forest communities; in contrast, palms are typically less common at higher elevations (Gentry, 1992). Atypical, is the low diversity and abundance of the Leguminosae at Río Colorado, a family that is usually ranked as the most diverse in lowland forest inventories (Gentry, 1988a, 1992). The Moraceae was important on both plots, as were the Annonaceae, Lauraceae, Euphorbiaceae, and Meliaceae. The Rubiaceae was particularly diverse at the Cumbre Pilón plot; several genera were present that are considered to be characteristic of montane forests, particularly Cinchona, Ladenbergia, Macrocnemum, Coussarea, and Alibertia. The Serranía Pilón Lajas is blanketed in clouds during the rainy season; nonetheless, the forest lacks taxa such as Weinmannia, Podocarpus, Juglans, Oreopanax, and Clethra that are common at higher elevations in cloud forest formations in other parts of Bolivia (Vargas, 1995; Beck & Smith, 1995). An undescribed species in the genus Berrya Roxb. occurred at both sites, but was more common on the piedmont; this species is apparently endemic to premontane forests in this part of Bolivia and adjacent Peru (A. Gentry, pers. com.). Berrya was previously thought to be a genus restricted to Asia, but it is closely related and possibly not distinct from the Caribbean and African taxon Carpodiptera Burret. The description of this new species awaits the collection of sufficient fertile material to ensure its appropriate classification.
Gentry (1995) has shown that montane forest up to 1500 m are similar to lowland forest both in terms of diversity and composition. The data presented here supports that hypothesis in part. A comparison of the species composition of these two plots with recently conducted inventories in humid forest on the Andean piedmont in Santa Cruz (Vargas, 1995) and the adjacent alluvial plain of the Beni (de Aguila, 1995) reveals that 60 species occur at least at two of the four study sites (Table 5). Although the number of shared taxa between the montane sight and the lowland sites are approximately the same, the calculation of an index weighted for unequal sample sizes shows that the lowland plots are more similar to each other than they are to the montane forest plot is (Table 6). The lower values are due to the greater diversity of the montane forest community, which incorporates most of the lowland taxa, as well as a variety of montane forest taxa that are absent from the lowland sites. The differences between the montane and lowland plots are even more pronounced when the abundance values of individual species are used for calculating the similarity index. This reflects the fact that the most abundant species in the montane plots are different from the three lowland plots, while the most abundant species in three lowland plots tend to be the same.
A comparison with inventories in other Amazonian countries reveals some conspicuous trends. Pseudolmedia laevis is ranked within the top twenty important species at Alto Ivon (Boom, 1986), Jatun Sacha (Neill & Palacios, manuscript), Añangu (Balslev et al., 1987), and is present, if not abundant, at Rio Xingu (Campbell et al., 1986). In Bolivia, P. laevis generally is considered to be the most abundant species in lowland humid forests in the Pando, Beni, and northern Santa Cruz (Boom, 1986; Saldias et al., 1994; Vargas et al., 1994; Arroyo et al., 1994; de Aguila, 1995). Likewise, several palms are widely distributed and abundant; Iriartea deltoidea, Socratea exorrhiza, Euterpe precatoria, and Oenocarpus bataua have been reported for several of the sites in Peru and Ecuador (Balslev et al., 1987; Kahn & Mejia, 1990). In Bolivia, Iriartea deltoidea is more common in the forests at the base of the Andes and the adjacent alluvial plain of the Beni, while being absent from the Brazilian Shield region of northern Santa Cruz. Socratea exorrhiza and Euterpe precatoria are more widely distributed being found commonly throughout the Bolivian orient north of 15° S on the Brazilian Shield and as far south as 17°45' S on piedmont and foothills of the Andes. A comparison of the common species listed in Table 6 with published checklists from Ecuador (Neill & Palacios, 1989) and Peru (Bracko & Zaruchi, 1993) reveal that 53 of these 60 species are widely distributed taxa that occur throughout western Amazonia. In contrast, there is little floristic overlap between the forests situated at the base of the Andes and the lowland humid forests in northern Bolivia (Table 6) or on the Brazilian Shield in eastern Bolivia (Killeen, unpublished data). The importance of ongoing botanical inventory in the region is highlighted by comparing the list of trees from this study with the recently published Guía de Arboles de Bolivia (Killeen et al., 1993). Twenty nine species not DIVd in this treatment of Bolivia trees are reported, or 14.3% of the total of 203; undoubtedly, the list would be greater if the undetermined specimens were identified.
The moderate levels of tree species diversity recorded for lowland humid forest communities in Bolivia (Table 7) and the abundance of species that are widely distributed can be interpreted in a variety of ways. Such characteristics would support the assumption that these forests have only recently become established in a biogeographical time frame and would conform to the hypothesis that Amazonian forest vegetation has expanded southward during the Holocene (Ratter, 1992; Prado & Gibbs, 1993). Likewise, there is widespread evidence that much of lowland Bolivia has been subject to recurrent catastrophic disturbance caused by the fluvial dynamics of both the major and secondary river systems (Hanagarth, 1993; Gullison et al., 1995). Such disturbance has resulted in the formation of a diverse assemblage of habitats, particularly wetland plant communities; however, it has not been conducive to the development of rich forest communities typically found on well-drained soils throughout the Neotropics. Finally, the data from these plot studies support the generally accepted hypothesis that tree species richness is correlated with latitude and precipitation (Gentry, 1988a).
The research was conducted in collaboration with the Herbario Nacional de Bolivia with funds provided by the United States Agency for International Development (grant LAC 0605-G-SS-8027-00) and the American Philosophical Society, Michaux Grants for Forest Botany.
de Aguila C., M. 1996. Determinación de la diversidad botánica de árboles en la región del Bosque Chimanes en el bosque húmedo de llanura del Beni, Bolivia. Tesis de Grado, Facultad de Agronomía, Universidad Técnica del Beni, Trinidad, Bolivia.
Arroyo P., L., T. Ruíz de Centurión, M. Saldias P. & T. J. Killeen. 1994. Estructura y composición de una isla de bosque y un bosque de galería en el Parque Nacional "Noel Kempff Mercado". VI Congreso Latinoamericano de Botánica, Mar de Plata. Libro de Resúmenes. Pp. 719.
Balslev, H., J. Luteyn, B. Ollgaard & L. B. Holm-Nielsen. 1988. Composition and structure of adjacent unflooded and floodplain forest in amazonian Ecuador. Opera Botanica 92: 37–57.
Beck, S. G., T. J. Killeen & E. García E. 1993. Vegetación de Bolivia. Pp. 6–25. In: T. J. Killeen, S. G. Beck & E. García E. (eds.) Guía de Arboles de Bolivia. Herbario Nacional de Bolivia & Missouri Botanical Garden, La Paz, Bolivia.
Beck, S. G. & D. N. Smith. (in press) El bosque montano húmedo de Apa Apa, Sud Yungas, Bolivia; un ejemplo de la diversidad florística Andina. In: M. Libermann. (ed.) II Simposio Internacional, Desarrollo Sostenible de Ecosistemas de Montaña: Manejo de Areas Frágiles en Los Andes, La Paz, Bolivia.
Boom, B. 1986. A forest inventory in Amazonian Bolivia. Biotropica 18: 287–294.
Bracko, L. & J. L. Zarucchi. 1993. Catalogue of the flowering plants and gymnosperms of Peru. Monographs in Systematic Botany from the Missouri Botanical Garden, v. 45.
Campbell, D. G. 1989. Quantitative inventory of tropical forests. Pp. 524–533. In: D. G. Campbell & H. D. Hammond (eds.) Floristic Inventory of Tropical Countries. New York Botanical Garden, New York.
Campbell, D. G., D. C. Daly, G. T. Prance & U. N. Maciel. 1986. Quantitative ecological inventory of terra firme and v&aaacute;rzea forest on the Rio Xingú, Brazilian Amazon. Brittonia 38: 369-393.
Dallmeier, F., R. B. Foster, C. B. Romano, R. Rice & M. Kabel. 1991. User's guide to the Beni Biosphere Reserve Biodiversity Plots 01 ad 02, Bolivia. Smithsonian Institution / Man and the Biosphere Biological Diversity Program – Beni Biological Station, Academia de Ciencias. La Paz, Bolivia.
Emmons, L. 1991. The mammals of Alto Madidi. In: T. Parker & B. Bailey. (eds.) A Biological Assessment of the Alto Madidi Region and Adjacent Areas of Northwest Bolivia. Conservation International, RAP Working Papers, v.1, Washington, DC.
Gentry, A. H. 1988a. Changes in plant community diversity and floristic composition on environmental and geographical gradients. Ann. Missouri Bot. Gard. 75: 1–34.
Gentry, A. H. 1988b. Tree species richness of upper amazonian forests. Proc. Natl. Acad. Sci. USA 85: 156–159.
Gentry, A. H. 1992. Diversity and floristic composition of Andean forests of Peru and adjacent countries: implications for their conservation. Memorias del Museo de Historia natural, U.N.M.S.M. (Lima) 21: 11–29.
Gentry, A. H. 1995. Patterns of diversity and floristic composition in Neotropical montane forests. In S. P. Churchill, H. Blaslev, E. Forero & J. L. Luteyn (eds). Biodiversity and Conservation of Neotropical Montane Forests. The New York Botanical Garden, New York.
Gullison, R. E., S. N. Panfil, J. J. Strouse & S. P. Hubbell. 1995. Ecology and management of mahogany (Swietenia macrophylla King) in the Chimanes Forest, Beni, Bolivia. Botanical J. Linnean Soc. (in press).
Hanagarth, W. 1993. Acerca de la Geoecología de las Sabanas del Beni e el Noreste de Bolivia. Instituto de Ecología, La Paz.
Jardim, A., R. Quevedo, S., M. Saldias P., L. Arroyo P. & T. J. Killeen. 1994. Estructura y composición de los bosques secos en Parque Nacional "Noel Kempff Mercado". VI Congreso Latinoamericano de Botánica, Mar de Plata. Libro de Resúmenes.
Kahn, F. & K. Mejia. 1990. Palm communities in wetland forest ecosystems of Peruvian Amazonia. Forest Ecology and Management 33: 169–170.
Killeen, T. J. , S. G. Beck & E. García E. (eds.) Guía de Arboles de Bolivia. Herbario Nacional de Bolivia & Missouri Botanical Garden, La Paz, Bolivia.
Magurran, A. E. 1988. Ecological Diversity and its Measurement. Princeton University Press, Princeton, New Jersey.
Merlot, L., Brugioni, I. & L. F. Terceros (eds.) 1994. Implementación de la Reserva de Biósfera Territorio Indígena Pilón Lajas. Veterinarios Sin Fronteras, La Paz, Bolivia.
Neill, D. A. & W. A. Palacios. (manuscript). Quantitative inventory of one hectare of tropical wet forest on the upper Río Napo, Amazonian Ecuador.
Neill, D. A. & W. A. Palacios. 1989. Arboles de la Amazonía Ecuatoriana, Lista preliminar de especies. Dirección Nacional Forestal, Quito, Ecuador.
Oblitas G., J. & C. E. Brockmann H. 1978. Mapa Geológico de Bolivia. GEOBOL (Servicio Geológico de Bolivia) and YPBF (Yacimientos Petroliferos Fiscales de Bolivia), La Paz, Bolivia.
Prado, D. E. & P. E. Gibbs. 1993. Patterns of species distributions in the dry seasonal forests of South America. Ann. Missouri Bot. Gard. 80: 902–927.
Phillips, O. L. & A. H. Gentry. 1994. Increasing turnover through time in tropical forests. Science 263: 954–958.
Ratter. J. A. 1992. Transitions between cerrado and forest vegetation in Brazil. Pp. 417–429. In: P. A. Furley, J. Procter & J. A. Ratter (eds.) Nature and Dynamics of Forest - Savanna Boundaries. Chapman & Hall, London.
Saldias P., M. 1991. Inventario de árboles en el bosque alto del Jardín Botánico de Santa Cruz, Bolivia. Ecologia en Bolivia. 17: 31–46.
Saldias P., M., R. Guillen V. B. Mostacedo C. & T. J. Killeen. 1994. Composición y estructura de un bosque alto húmedo del Parque Nacional "Noel Kempff Mercado". VI Congreso Latinoamericano de Botánica, Mar de Plata. Libro de Resúmenes, Pp. 774.
Vargas C., I. 1995. Estructura y Composición de Cuatro Sitios Boscosas en el Parque Nacional Amboro. Tesis de Grado, Universidad Autónomo Gabriel René Moreno, Santa Cruz, Bolivia.
Vargas C., I, T. R. de Centurión & M. Saldias P. 1994. Parcela permanente de investigación en la Reserva de Vida Silvestre Ríos Blanco y Negro, Santa Cruz, Bolivia. Revista de la Sociedad de Estudios Botánicos (Santa Cruz) 1: 9–32.