Rotenberg, James A


Established parks and nature reserves may be inadequate to preserve long-term biotic diversity, especially in tropical regions of Latin America where anthropogenic disturbance and land-conversion is an ongoing problem. Demand for economically productive land uses, such as cattle pastures or monoculture plantations, is one of the greatest threats to habitat and wildlife preservation. As a result, conservation biologists have turned to examining attributes of nonprotected lands to determine which land uses both support wildlife and make economic sense. I examined bird use of a plantation of Gmelina arborea (“white teak” or “melina” trees; hereafter “plantation”) in Guatemala to assess its suitability as bird habitat. Gmelina arborea is grown mainly for wood and paper pulp. Several compositionally different habitats were identified, based on the amount of natural vegetation grown within and among the plantation trees, forming a heterogeneous landscape. I detected 195 bird species from 45 families within this plantation-dominated landscape, and 144 species in plantation habitats combined. I observed a positive association between increased vegetative complexity and bird species richness; moreover, bird species richness attained levels statistically indistinguishable from those found in nearby forest fragments. Mixed plantation habitat containing 19-31% natural vegetative coverage supported bird communities that equaled or surpassed levels of species richness found by other researchers in shaded coffee (Coffea spp.) plantations. However, pure stands of G. arborea supported richness levels equal to those of grazed pasture; diversity levels associated with both these land uses were close to those reported for sun coffee plantations. Clearly, native vegetation played a significant role in enhancing bird species richness in the plantation habitat, and amounts of relative cover similar to or exceeding those in the plantation I studied should be considered in any management plan for G. arborea in Central America.

Received 4 April 2004, accepted 14 March 2006.

Key words: avian communities, bird habitat relationships, conservation, Gmelina arborea, Guatemala, Neotropics, plantation.

Rol Ecológico de Plantaciones de Árboles de Gmelina arborea en Guatemala: Una Evaluación de un Uso Alternative del Suelo para la Conservación de las Aves Tropicales

RESUMEN.-Los parques establecidos y las reservas naturales podrían ser inadecuadas para preserver a largo plazo la diversidad biológica, especialmente en las regiones tropicales de América Latina donde los disturbios antropogénicos y la conservación de la naturaleza son un problema frecuente. La demanda de tierras económicamente productivas, como los pastizales para ganado o las plantaciones de monocultivos, es una de las mayores amenazas para la preservación del hábitat y de la vida silvestre. Como resultado de esto, los biólogos de la conservación han vuelto a revisar los atributos de las áreas no protegidas para determinar cuáles usos de la tierra mantienen la vida silvestre y son viables económicamente. Examiné el uso por parte de las aves de plantaciones de árboles de Gmelina arborea (“teca blanca” o “melina”) en Guatemala para evaluar su calidad como hábitat para las aves. Las plantaciones de Gmelina arborea están destinadas principalmente a la producción de madera y pulpa para papel. Se identificaron varies ambientes que variaron en la composición según la cantidad de vegetación natural presente al interior y entre las plantaciones de árboles, formando un paisaje heterogéneo. Detecté 195 especies de aves pertenecientes a 45 familias en este paisaje dominado por las plantaciones, y 144 especies en la combinación de ambientes de G. arborea. Observé una asociación positiva entre un incremento en la complejidad de la vegetación y la riqueza de especies de aves; más aún, la riqueza de especies de aves alcanzó niveles estadísticamente indistinguibles de aquella encontrada en fragmentes de bosque vecinos. El ambiente combinado de G. arborea con una cobertura de vegetación natural de entre 19% y 31% sustentó comunidades de aves que igualaron o sobrepasaron los niveles de riqueza de especies encontrados por otros autores en plantaciones de café (Coffea spp.) a la sombra. Sin embargo, los rodales puros de G. arborea sustentaron niveles de riqueza iguales a los de los campos de pastoreo; los niveles de diversidad asociados con estos dos usos de la tierra fueron cercanos a aquellos presentados para las plantaciones de café al sol. Claramente, la vegetación nativa jugó un rol significative en incrementar la riqueza de especies de aves en las plantaciones de G. arborea. Por ende, la cantidad de cobertura relative similar o mayor que la de la plantación que estudié debe ser considerada en cualquier plan de manejo para esta especie de árbol en América Central.

RECENT CONSERVATION STUDIES have identified alternative land-use practices that may sustain avian forest biodiversity (Donald 2004). These hitherto little-studied land uses, sometimes referred to as “countryside habitats” (Daily et al. 2001), may support a substantial fraction of the original avifauna. For example, shaded coffee (Coffea spp.) plantations have been found to support large numbers of Neotropical migratory bird species, while at the same time having economic benefits (Aguilar-Ortiz 1982, Perfecto et al. 1996, Greenberg et al. 1997, Petit et al. 1999, Sherry 2000). The vegetative complexity created by growing individual coffee plants under overlying canopy trees attracts bird species to shaded coffee plantations. This reflects the well-known positive relationship between bird species richness and increased vegetative complexity (MacArthur and MacArthur 1961). Identifying shaded coffee plantations as bird réfugia in otherwise deforested landscapes is significant, because most large-scale agricultural operations, such as cattle pastures and monoculture plantations, do not support high levels of biodiversity compared with the species-rich forests they replaced (Smith 1974, Disney and Stokes 1976, Carlson 1986, Cruz 1987, Evans 1992, Saab and Petit 1992, Sawyer 1993, Cardoso da Suva et al. 1996). Thus, identifying which land use provides the greatest ecological value for tropical bird communities and identifying which groups of species within that community use these habitats is essential for effective conservation.

In eastern Guatemala, Costa Rica, and Mexico, the past decade has seen the establishment of plantations of Gmelina arborea (Verbenaceae), a rapidly growing deciduous tree native to India that is frequently planted for paper pulp and is commonly called “white teak,” “yemane,” or “melina,” depending on where it is being grown (Lamb 1968). Changing land practices guided by new economic trends show that it may be financially advantageous to grow trees rather than cattle (Reforestadora Simpson pers. comm.). It is estimated that there are between 700,000 and 1 million ha of G. arborea under cultivation in the tropics and subtropics, with at least another 100,000 ha expected to be added by 2020, specifically in Central America and parts of Southeast Asia (Dvorak 2004). Although plantations of G. arborea exist elsewhere, there has not been an extensive study of the avian community they support. Therefore, the question arises: does a plantation of G. arborea (“hereafter “plantation”) support a significant subset of the original forest birds that once lived in Guatemala’s tropical lowland areas, now largely deforested, or are these plantations another agricultural practice that is a poor substitute for native forest bird habitat?

To answer this question, I examined bird use of a Guatemalan plantation, and of adjacent native forest fragments and pastures, to assess the plantation’s suitability as habitat for forest bird species. Because the effects of such plantations on the species richness and composition of Neotropical bird assemblages were unknown, I focused on the following questions. (1) How does bird species distribution vary over the forest-plantation-pasture landscape, and are those patterns associated with vegetative differences among plantation habitats? (2) What types of species compose the bird community? And (3) how effective is a plantation of G. arborea as a “foster ecosystem” for forest birds as compared with other anthropogenically created habitats, such as shaded coffee and other agro-ecosystems?


Study site.-I worked in a landscape dominated by a 7,000-ha plantation and native forest fragments in the tropical lowlands of eastern Guatemala. The plantation is ~5 km south of the town of Río Dulce (Fronteras), in the Department of Izabal (Fig. 1). Average yearly rainfall is 2,014 mm (data from the plantation’s owners, Reforestadora Simpson). The plantation was established in 1988 on a previously deforested site that had been used as cattle pasture for >20 years.

The plantation was managed specifically for paper-pulp production under the following regime: (1) trees were planted in a standardized 3 × 3 m grid plan; (2) during the first six months, tree establishment was aided by physical clearing of weeds and herbicide application, after which no other site maintenance was practiced; and (3) whenever possible, pre-existing native habitat features such as stream corridors or forest fragments were retained inside the plantation. The regime resulted in trees growing to heights of 8-15 m in 2-4 years, with canopies typically closing at the end of the second or during the third year. Many stands eventually reached 15-20 m by their fifth or sixth year, some with thick understories.

I identified six habitat types within the forest-plantation landscape: forest, Gmelina, Gmlina-forest (Gm-forest), Gmelina-stream (Gm-stream), Gmelina-fence (Gm-fence), and pasture. Forest consisted of fragments of 1-100 ha comprising mostly secondary forest, with patches of larger, emergent trees characteristic of primary forest. Plantation habitat was primarily monotypic, though interspersed with small numbers of native species. The next three habitat types were “mixed” Gmelina, consisting of a stand of G. arborea trees mixed with varying amounts of natural vegetation, either a forest fragment (Gm-forest), a stream corridor with remnant forest-like habitat (Gm-stream), or a living fence (Gm-fence). Gm-forest fragments ranged from 1 ha to 100 ha, whereas stream corridor widths ranged from 3 m to 100 m. A living fence is created when ranchers use fence stakes cut from local trees and those stakes resprout. Pasture consisted of adjacent and nearby pasture habitats that likely matched the cattle pastures the plantation replaced. These pastures were typical of the area with mostly open grassland and scattered trees.

Bird surveys.-I established 135 survey points, stratified by the habitat types described above. Because my goal was to characterize variation in the avian community living within this large, relatively undescribed landscape, I used an extensive (rather than intensive) pointcount sampling strategy, with many points scattered over a broad area (Ralph et al. 1993). Points were constrained to be at least 200-250 m apart, and ≥75 m from another habitat type. We thus avoided counting the same individuals at adjacent points and minimized the effects from nearby habitats. In the “mixed” habitats, points were placed along the edge of the two vegetation types to ensure sampling of approximately equal portions within a 50-m-radius circle. I sampled birds using a 50-m-radius, 10-min point count, immediately followed by a 10-min area search (Hutto et al. 1986). I recorded bird species presence, detected by sight or sound, of those birds that were actively using the habitat (i.e., not flyovers). The 10-min area search was used to confirm bird calls, songs, and detection distances, as well as to detect more cryptic species. In addition, I noted all species seen within the surveyed habitat type during the count period, even if outside the 50-m radius. However, these detections were not used in statistical analyses.

Counts were conducted six days per week, on nearly a weekly basis, during all seasons in 1998 and in the winter, spring, and fall of 1999. In 1998, all points were sampled at least once until the beginning of the fall migration (the first day fall migrants were detected), at which time I began to resample the entire set. Because of logistical constraints, only a subset of the 135 points was surveyed during each of the remaining seasons, resulting in a range of one to four survey visits per point. Ultimately, 76% of the points were surveyed at least once each year, and ~50% were surveyed two or more times in either year. To account for the disparity in the number of point-count visits, I included the number of times a particular point was surveyed as a covariate in later analyses. Point counts then yielded a comparative index of species richness within each habitat type. Point-count surveys were conducted on mornings without rain or strong winds, between sunrise and 1030 hours.

Habitat sampling. -I used the James and Shugart (1970) method of vegetation sampling to quantify point-scale structural and floristic habitat variables. Using a 0.04-ha circle at each bird survey point, I measured 11 structural characteristics (Table 1) and identified and counted the number of trees ≥4 cm diameter at breast height (DBH). All trees were identified to morphospecies in the field, and herbarium samples were taken to attribute species (or at least family or genus) to each morphospecies when possible. To quantify the amount of vegetation coverage across the forest-plantationpasture landscape, a geographic information system (GIS) database was constructed by hand digitizing a 1:11,000 scale map of the plantation provided by Reforestadora Simpson into ARC/INFO (ESRI, Redlands, California). The map (Fig. 1) was of sufficient detail to define polygon coverages for the major vegetation elements, including naturally forested vegetation, swamp, pasture, and G. arborea and to demarcate each of the six habitat types defined above. Map features were ground-truthed for accuracy, and the amounts of coverage of each of the six habitats were calculated using the GIS database. These data were then used to determine proportional coverage of native vegetation and G. aborea in each habitat type.

Statistical analysis. -I tested variation in each vegetation variable and tree species richness (number of tree species per unit area) across habitat types with one-way analysis of variance (ANOVA) followed by Tukey’s multiplecomparison procedure. Because of the large number of high correlations among the structural variables, I used principal component analysis (PCA) to identify the major independent patterns of covariation in structural attributes. Principal component analysis constructs new, synthetic variables that are independent, linear combinations of the originals. I retained all components with eigenvalues ≥1 for further analyses (Gauch 1982). Each component was interpreted by examining its factor loadings, and I considered raw variables with factor loadings > |0.5| as contributing to the ecological interpretation of a particular component.

I estimated bird species richness (BSR) at a survey point by calculating the average number of species detected per visit. I then used ANOVA followed by Tukey’s multiple comparison procedure to test for differences in mean BSR among habitats. The point-specific estimate of BSR should not be biased by the number of visits each point received, though that estimate will be more precise for those points that were surveyed more often. However, to control for the disparity in the number of point-count visits, I included the number of times a particular point was surveyed as a covariate in the analysis. I assessed the association between BSR and the two PCA habitat variables using simple correlations. To assess seasonal use of the plantation by migrants, I calculated the proportion of migrant species out of all species detected per month.

Avian habitat affinities.-Habitat affinities of birds were determined from two classification methods (Stiles 1980, Canterbury et al. 2000) that take into account individual forest bird species’ tolerance to structural changes in habitat. I based my classifications on independently identified habitat associations from the literature (Howell and Webb 1995), rather than on where I found species, to avoid circularity of inference. I identified six habitat-use classes (Appendix): “forest specialists” (species intolerant of forest disturbance and found in forests only), “edge-tolerant forest species” (species found in forest, but using one or more disturbed forest habitats, including forest edge, gallery forest, and secondary forest), “edge specialists” (nonforest species favoring disturbed edges), “open-semiopen species” (found only in open, nonforested habitats), “generalists,” and “other.” The generalist assemblage includes species that use a wide variety of vegetation types. To distinguish generalists from other categories, I defined “generalist” as a species using both open and any other type of closed forested or forest-like habitat. The “other” class included non-upland species identified during the surveys. I calculated the percentage of occurrence of each habitat-use class across the entire landscape and by plantation-landscape habitat type.

Comparisons to other agro-ecosystem bird communities.-To evaluate how the plantation compared with other land uses, I compared my results with previous studies of bird communities in several types of tropical agro-ecosystems. These agro-ecosystems included three types of shade coffee, sun coffee, a rustic cardamom (Elettaria cardamomum) plantation, a pine (Pinus spp.) plantation, and three types of agroforests (Thiollay 1995, Greenberg et al. 1997, Petit et al. 1999). Because each study compared bird species richness in the agro-ecosystem to that of local forests, I did the same, by calculating the ratio of mean BSR of a particular land use within the plantation-forest-pasture landscape to mean BSR of the local forest from the present study. Those studies differ from mine in elevation, latitude, and, in one case, hemisphere (Asia); however, by preserving the relationship of each land use to its local forest habitat, the ratio takes into account these differences and provides a context in which to set the plantation.

All statistical analyses were performed with SPSS, version 10.0 (SPSS, Chicago, Illinois). The alpha level for each test was α = 0.05, and all results are reported as means ± SE.


Habitat.-The six study habitats varied in both vegetation structure and floristics as characterized by the 11 structural vegetation measures and tree species richness (Table 1). The mixed plantation habitats showed low variability, whereas forest vegetation had the greatest canopy cover and DBH, and pasture habitat had the lowest values for 7 of the 11 variables. Vegetation variables that may have attracted forest bird species to plantation habitat (i.e., those contributing most to vegetative complexity) were reduced in Gmelina, but were enhanced by other vegetation in the mixed plantation habitats. For example, understory height in Gmelina was about 1-2 m lower than in either mixed plantation or forest. Mean shrub abundance was 40-50% lower in Gmelina than in mixed plantation habitats, and 64% lower than in forest. Mean tree species richness was greatest in the edge habitats of Gm-stream and Gm-forest, and lowest in pasture and Gmelina.

The proportion of natural vegetation present within the mixed plantation habitats varied from a high of 31% in Gm-forest to a low of 9% in pure Gmelina stands. The Gm-stream and Gm-fence habitats had 19% and 12% forest coverage, respectively.

The PCA of habitat structure yielded two new, synthetic variables (eigenvalues >1). Factor loadings indicated that the first principal component (PC1) represented a gradient of increasing “forest-like structure” associated with increasing tree height, canopy cover, canopy size, understory height, shrub abundance, and plant and dry-leaf litter groundcover, and decreasing grass cover. Thus, forest points had high positive scores on this axis, whereas pastures had high negative scores; plantation habitats fell in between. The second principal component (PC2) described “tree attributes,” contrasting increasing tree diameter (DBH) with decreasing tree abundance. Pastures and forests had similar high scores on this axis, as trees in both habitats tended to be larger and more sparsely distributed than the thinner, more densely packed G. arborea.

Avian community patterns.-I detected 195 species from 45 families within the forest-plantation-pasture landscape during the present study, including those noted outside the 50-m radius of a point count. Of these, 170 were upland species. I observed 144 species in the four plantation habitats combined (Gm-forest, Gm-stream, Gm-fence, and Gmelina), of which 40 (28%) were Neotropical migrants (Appendix). Most of the 170 upland bird species that appeared in the forest-plantarion-pasture landscape were either generalists (61 species, 35.9%), or edge-tolerant forest species (59 species, 34.7%); only 11.8% were forest specialists (20 species) (Fig. 2). Generalists were also the most common group within each vegetation type, followed by edge-tolerant forest species. Forest specialists ranged from a low of 3.2% in Gm-fence to a high of 15.6% in forest. Only five forest specialist species were found in pure Gmelina. Although this number doubles in mixed plantation habitat types, sensitive species such as those from families Furnariidae and Dendrocolaptidae were not present.

Bird species richness varied among habitats within the plantation landscape even after adjusting for number of surveys as a covariate (F = 15.025, df = 5 and 128, P

The Gm-stream and Gm-forest habitats were relatively more species-rich than either Gm-fence or pure Gmelina, having 77.8% and 84.7% of the forest average, respectively. Gmelina was relatively depauperate in comparison, with only 41.7% of the number of species surveyed at forest points. Pasture points were the most species-poor, supporting only 29.2% of the species found at the average forest point.

Species-richness ratios for various land uses reported in the literature and for those observed here clearly show variation in their ability to support forest bird species. Shade coffee under a canopy of Erythrina-Inga spp. in Panama supported 100% of local forest species (Petit et al. 1999), Gm-forest most closely resembled rustic shade cardamom plantations and matorral habitats in central Guatemala (86.2%; Greenberg et al. 1997), and Gm-stream fell between matorral and Hevea spp. agroforest (70.1%; Thiollay 1995). Shade coffee under a canopy of Inga spp. (62.1%; Greenberg et al. 1997), as well as damar (Shorea javanica; 58.3%) and durian (Durio zibethinus; 57.2%) agroforests (Thiollay 1995), were lower in species richness than Gm-stream. The Gm-fence habitat was equal to fallow pastures found in Panama (55.6%; Petit et al. 1999), which supported just slightly fewer species than shade coffee grown under a canopy of Gliricidia spp. (56.3%; Greenberg et al. 1997). Pure Gmelina was closest to sun coffee plantations (44.8%; Greenberg et al. 1997) but supported 20% more species than Panamanian pine plantations (11.1%; Petit et al. 1999). Bird species richness for grazed pastures in Guatemala closely matched levels found in grazed pastures in Panama (31.1%; Petit et al. 1999).


Overall, only the mixed Gmelina habitats within the plantation supported considerably more bird species than reported for several other exotic, tropical monocultures; however, pure stands of G. arborea did not have this attractive capacity. The mixed Gmelina habitats supported a richer bird community than the other plantation habitats, largely owing to the heterogeneous landscape created by the practice of retaining pre-existing natural habitat features. Increasing bird species richness was positively associated with forest-like structural factors such as increased tree height, understory height, DBH, canopy cover, and canopy size. The amount of natural vegetation also played a role; the more speciose Gm-stream and Gm-forest habitats have greater amounts of natural vegetation (19% and 31%) than other plantation habitats. In addition, these two mixed habitats did not differ from forest in bird species richness. The Gm-fence and Gmelina habitats, on the other hand, had significantly lower BSR than these more complex habitats, but not from pasture. This result suggests that neither a single line of native trees nor G. arborea alone were sufficient to attract the variety of bird species found in more vegetatively complex habitats. Although the average point in any plantation vegetation type supported more bird species than the average pasture point, the former still fell significantly short of forest habitat.

The positive relationship between increased habitat heterogeneity and bird species richness is not unique to this plantation landscape, but is common to many systems, including comparisons of natural forests, forests and fields, and a few previously studied plantations. For example, temperate forest stands with more vertical vegetation layers tend to support richer bird communities (MacArthur and MacArthur 1961, Willson 1974, Roth 1976, James and Wamer 1982). Comparisons between tropical forest and nearby pastures in Panama found that forests supported greater numbers of both resident and migratory bird species (Saab and Petit 1992, Petit and Petit 2003), and plantations with higher-than-average understory complexity had a greater tendency to attract bird species than plantations without vegetative complexity (Cruz 1987, Curry 1991, Mitra and Sheldon 1993, Hanowski et al. 1997).


Establishing a plantation of G. arborea on a previously deforested cattle pasture provides new habitat mainly for birds that are generalists or edge-tolerant forest species, not necessarily for those most vulnerable to disturbance (see Petit and Petit 2003). For example, only five species of forest specialists were found in pure stands of G. arborea. Indeed, the main difference in bird species between G. arborea and forest fragments was the missing forest specialists, with a decrease of 50-70% in G. arborea as compared with forest. The difference between the plantation and surrounding pastures was the edge-tolerant forest species; about 2× as many edge-tolerant species are found, on average, in plantation habitats as in pasture. Generalists and edge-tolerant forest species dominated the plantation bird community because the generalists are opportunistic (by definition) and the edge-tolerant forest species are less sensitive to disturbance than forest specialists but still depend on forest-like structure. Studies in Costa Rica have found similar results; typically, forest bird species can be divided into two groups, those that rarely occur in disturbed habitats and those that occur regularly (Loiselle and Blake 1992).

With natural fragments and corridors providing increased vegetative complexity, there was a nearly threefold increase in forest-associated bird species (forest-specialist plus edge-tolerant forest species) in the mixed plantation habitats as compared with pasture. Although half the 40 migratory species found in plantation habitats were not found in pasture (including a rare sighting of an endangered Golden-cheeked Warbler [Dendroica chrysoparia] in the Gm-stream habitat), these encouraging results must be balanced by the reality that most of the plantation consisted of G. arborea (77%) and not the natural vegetation necessary for higher bird species diversity.

Evaluating the plantation for its conservation potential must be done by comparing it with other anthropogenically created land uses and their bird communities (e.g., Petit and Petit 2003). My data suggest that G. arborea grown in combination with natural vegetation to increase vegetative complexity can equal or surpass levels of bird species richness found in some shaded coffee plantations. Without this complexity, the pure G. arborea supported fewer species than sun coffee plantations, though more than pine or pasture. Clearly, the role native vegetation played in enhancing bird species richness and contributing to bird species composition in the plantation cannot be overstated.

Pure monocultures of G. arborea in this plantation are somewhat more species-rich than other tropical tree monocultures, such as Pinus spp. and Eucalyptus spp. (Smith 1974, Disney and Stokes 1976, Carlson 1986, Sawyer 1993). Those studies found the monocultures to be very poor habitat for most bird species. In addition, qualitative observations made in coconut palm (Cocos nucifera), oil palm (Elaeis guineensis; reviewed in Donald 2004), and rubber (Hevea brasiliensis) plantations found that these habitats, too, were depauperate of bird species (Thiollay 1995), with rarely more than three to four species observed, all of which were generalists. Conversely, tropical tree plantations of Albizia falcataria in Borneo and Theobroma cacao in Costa Rica were found to support much higher levels of bird species richness than the plantations just mentioned (Mitra and Sheldon 1993, Reitsma et al. 2001). Like shade coffee under Inga spp., some stands of A. falcataria supported ≤60% of the bird species found in surrounding forests, and this was attributed to increased vegetative complexity.

Management recommendations.-If the plantation is managed to maintain or increase habitat heterogeneity, G. arborea could support an even greater diversity of forest-associated resident species. For example, managers of this particular plantation found it advantageous to maintain pre-existing native vegetation to help protect watershed areas from runoff and erosion. In so doing, they created heterogeneity in what would have been a homogeneous, monoculture plantation, and thereby provided critical habitat to many forest-associated bird species. If an even greater effort were made to both enhance the size and connectivity of these fragments and to create new patches of natural growth within the monoculture matrix, an even greater number of forest bird specialists might be supported by mixed stands of G. arborea. Whether this management prescription is more or less economical than a landscape-scale monoculture remains unknown. However, at present, much of the G. arborea under cultivation is in smaller-sized stands (50-300 ha); in Asia, G. arborea is even used as an overstory shade tree for other commercially grown plant crops (Dvorak 2004). At that scale, a management plan consisting of a patchwork of mixed stands of G. arborea that each includes a relative coverage of ≥31% natural growth may promote a more bird-friendly plantation system in similar regions of Central America.


I thank J. Rotenberry for his continued help and guidance with this project. I also thank L. Nunney and A. Gómez-Pompa for their support during the project, and M. Ramirez for his help in the field. Many thanks to Reforestadora Simpson for allowing me access to their property and the freedom to carry out this study. Many Reforestadora Simpson employees assisted; special thanks to M. Mussack, M. Fontes, and J. Rydelius. I thank M. Bryant and B. Kristan for statistical help, and A. Sanders and the University of California, Riverside (UCR) Herbarium for help with plant identification. Thanks to C. Elphick, M. Rubega, and the reviewers for comments and suggestions on the manuscript. I especially thank my wife, Vibeke. This project was funded by a J. William Fulbright Fellowship and a U.S. Department of Education Graduate Assistance in Areas of National Need (GAANN) Fellowship through the UCR Department of Biology.


AGUILAR-ORTIZ, F. 1982. Estudio ecológico de las aves del cafetal. Pages 103-128 in Estudios Ecológicos en el Agroecosistema Cafetalero (E. Jimenez-Avila and A. Gomez-Pompa, Eds.). Instituto Nacional de Investigaciones sobre Recursos Bióticos, Xalapa, Veracruz, México.

CANTERBURY, G. E., T. E. MARTIN, D. R. PETIT, L. J. PETIT, AND D. F. BRADFORD. 2000. Bird communities and habitat as ecological indicators of forest condition in regional monitoring. Conservation Biology 14:544-558.

CARDOSO DA SILVA, J. M., C. UHL, AND G. MURRAY. 1996. Plant succession, landscape management, and the ecology of frugivorous birds in abandoned Amazonian pastures. Conservation Biology 10:491-503.

CARLSON, A. 1986. A comparison of birds inhabiting pine plantation and indigenous forest patches in a tropical mountain area. Biological Conservation 35:195-204.

CRUZ, A. 1987. Avian community organization in a mahogany plantation on a Neotropical island. Caribbean Journal of Science 23: 286-296.

CURRY, G. N. 1991. The influence of proximity to plantation edge on diversity and abundance of bird species in an exotic pine plantation in north-eastern New South Wales. Wildlife Research 18:299-314.

DAILY, G. C., P. R. EHRLICH, AND G. A. SÁNCHEZ-AZOFEIFA. 2001. Countryside biogeography: Use of human-dominated habitats by the avifauna of southern Costa Rica. Ecological Applications 11:1-13.

DISNEY, H. J. DE S., AND A. STOKES. 1976. Birds in pine and native forests. Emu 76:133-138.

DONALD, P. F. 2004. Biodiversity impacts of some agricultural commodity production systems. Conservation Biology 18:17-38.

DVORAK, W. S. 2004. World view of Gmelina arborea: Opportunities and challenges. New Forests 28:111-126.

EVANS, J. 1992. Plantation Forestry in the Tropics, 2nd ed. Oxford University Press, Oxford, United Kingdom.

GAUCH, H. G., Jr. 1982. Multivariate Analysis in Community Ecology. Cambridge University Press, Cambridge, United Kingdom.

GREENBERG, R., P. BICHIER, A. C. ANGON, AND R. REITSMA. 1997. Bird populations in shade and sun coffee plantations in central Guatemala. Conservation Biology 11:448-459.

HANOWSKI, J. M., G. J. NIEMI, AND D. C. CHRISTIAN. 1997. Influence of within-plantation heterogeneity and surrounding landscape composition on avian communities in hybrid poplar plantations. Conservation Biology 11:936-944.

HOWELL, S. N. G., AND S. WEBB. 1995. A Guide to the Birds of Mexico and Northern Central America. Oxford University Press, New York.

HUTTO, R. L., S. M. PLETSCHET, AND P. HENDRICKS. 1986. A fixed-radius point count method for nonbreeding and breeding season use. Auk 103:593-602.

JAMES, F. C., AND H. H. SHUGART, JR. 1970. A quantitative method of habitat description. Audubon Field Notes 24:727-736.

JAMES, F. C., AND N. O. WAMER. 1982. Relationships between temperate forest bird communities and vegetation structure. Ecology 63:159-171.

LAMB, A. F. A. 1968. Fast Growing Timber Trees in the Lowland Tropics. 1. Gmelina arborea. Department of Forestry, University of Oxford, Oxford, United Kingdom.

LOISELLE, B. A., AND J. G. BLAKE. 1992. Population variation in a tropical bird community: Implications for conservation. BioScience 42:838-845.

MACARTHUR, R. H., AND J. W. MACARTHUR. 1961. On bird species diversity. Ecology 42: 594-598.

MITRA, S. S., AND F. H. SHELDON. 1993. Use of an exotic tree plantation by Bornean lowland forest birds. Auk 110:529-540.

PERFECTO, I., R. A. RICE, R. GREENBERG, AND M. E. VAN DER VOORT. 1996. Shade coffee: A disappearing refuge for biodiversity. BioScience 46:598-608.

PETIT, L. J., AND D. R. PETIT. 2003. Evaluating the importance of human-modified lands for Neotropical bird conservation. Conservation Biology 17:687-694.

PETIT, L. J., D. R. PETIT, D. G. CHRISTIAN, AND H. D. W. POWELL. 1999. Bird communities of natural and modified habitats in Panama. Ecography 22:292-304.

RALPH, C. J., G. R. GEUPEL, P. PYLE, T. E. MARTIN, AND D. F. DESANTE. 1993. Handbook of field methods for monitoring landbirds. U.S. Department of Agriculture, Forest Service General Technical Report PSW-GTR-144.

REITSMA, R., J. D. PARRISH, AND W. MCLARNEY. 2001. The role of cacao plantations in maintaining forest avian diversity in southeastern Costa Rica. Agroforestry Systems 53:185-193.

ROTH, R. R. 1976. Spatial heterogeneity and bird species diversity. Ecology 57:773-782.

SAAB, V. A., AND D. R. PETIT. 1992. Impact of pasture development on winter bird communities in Belize, Central America. Condor 94:66-71.

SAWYER, J. 1993. Plantations in the Tropics: Environmental Concerns. International Union for Conservation of Nature and Natural Resources, Gland, Switzerland and Cambridge, United Kingdom.

SHERRY, T. W. 2000. Shade coffee: A good brew even in small doses. Auk 117:563-568.

SMITH, K. D. 1974. The utilization of gum trees by birds in Africa. Ibis 116:155-164.

STILES, F. G. 1980. Evolutionary implications of habitat relations between permanent and winter resident landbirds in Costa Rica. Pages 421-435 in Migrant Birds in the Neotropics: Ecology, Behavior, and Distribution (A. Keast and E. S. Morton, Eds.). Smithsonian Institution Press, Washington, D.C.

THIOLLAY, J.-M. 1995. The role of traditional agroforests in the conservation of rain forest bird diversity in Sumatra. Conservation Biology 9:335-353.

WILLSON, M. F. 1974. Avian community organization and habitat structure. Ecology 55: 1017-1029.

Associate Editor: P. C. Stouffer


Department of Biology, University of California, Riverside, California 92522, USA

1 Present address: Department of Environmental Studies, University of North Carolina Wilmington, 601 South College Road, Wilmington, North Carolina 28403, USA. E-mail:

Copyright American Ornithologists’ Union Jan 2007

Provided by ProQuest Information and Learning Company. All rights Reserved