BIOTROPICA 38(2): 183–195 2006 10.1111/j.1744-7429.2006.00124.x Distribution and Flowering Ecology of Bromeliads along Two Climatically Contrasting Elevational Transects in the Bolivian Andes1 Thorsten Krömer2 , Michael Kessler Institute of Plant Sciences, Department of Systematic Botany, University of Göttingen, Untere Karspüle 2, 37073 Göttingen, Germany and Sebastian K. Herzog3 Institut für Vogelforschung “Vogelwarte Helgoland,” An der Vogelwarte 21, 26386 Wilhelmshaven, Germany ABSTRACT We compared the diversity, taxonomic composition, and pollination syndromes of bromeliad assemblages and the diversity and abundance of hummingbirds along two climatically contrasting elevational gradients in Bolivia. Elevational patterns of bromeliad species richness differed noticeably between transects. Along the continuously wet Carrasco transect, species richness peaked at mid-elevations, whereas at Masicurı́ most species were found in the hot, semiarid lowlands. Bromeliad assemblages were dominated by large epiphytic tank bromeliads at Carrasco and by small epiphytic, atmospheric tillandsias at Masicurı́. In contrast to the epiphytic taxa, terrestrial bromeliads showed similar distributions across both transects. At Carrasco, hummingbird-pollination was the most common pollination mode, whereas at Masicurı́ most species were entomophilous. The proportion of ornithophilous species increased with elevation on both transects, whereas entomophily showed the opposite pattern. At Carrasco, the percentage of ornithophilous bromeliad species was significantly correlated with hummingbird abundance but not with hummingbird species richness. Bat-pollination was linked to humid, tropical conditions in accordance with the high species richness of bats in tropical lowlands. At Carrasco, mixed hummingbird/bat-pollination was found especially at mid-elevations, i.e., on the transition between preferential bat-pollination in the lowlands and preferential hummingbird-pollination in the highlands. In conclusion, both richness patterns and pollination syndromes of bromeliad assemblages varied in distinct and readily interpretable ways in relation to environmental humidity and temperature, and bromeliad pollination syndromes appear to follow the elevational gradients exhibited by their pollinators. RESUMEN Comparamos la diversidad, composición taxonómica y sı́ndromes de polinización de comunidades de bromeliáceas con la diversidad y abundancia de colibrı́es a lo largo de dos gradientes altitudinales climáticamente diferentes en Bolivia. Los patrones de elevación de riqueza de especies de bromeliáceas difirieron notoriamente entre los transectos. A lo largo del transecto permanentemente húmedo de Carrasco, la riqueza de las especies alcanzó su punto máximo a elevaciones intermedias, mientras que en Masicurı́ la mayorı́a de las especies se encontraron en tierras bajas, calientes y semiáridas. Las comunidades de bromeliáceas estuvieron dominadas por grandes bromeliáceas epifiticas tanque en Carrasco y por pequeñas tillandsias atmosféricas epifiticas en Masicurı́. En contraste a los taxones epifiticos, la distribución de bromeliáceas terrestres fue bastante similar en ambos transectos. En Carrasco, la polinización por colibrı́es fue la manera de polinización más común, mientras que en Masicurı́ la mayorı́a de las especies fueron entomófilas. La proporción de especies ornitófilas aumentó con la elevación en ambos transectos, mientras que la entomófila mostró el patrón contrario. En Carrasco, el porcentaje de especies de bromeliáceas ornitófilas estuvo correlacionado significativamente con la abundancia de colibrı́es pero no con la riqueza de especies de colibrı́es. La polinización por murciélagos se relacionó con condiciones húmedas y tropicales, coincidiendo con la alta riqueza de especies de murciélagos en las tierras bajas tropicales. En Carrasco, polinización mixta de colibrı́es/murciélagos fue encontrada principalmente a elevaciones intermedias, es decir, en la transición entre polinización preferencial por murciélagos en tierras bajas y polinización preferencial por colibrı́es en las montañas. En conclusión, los patrones de riqueza y sı́ndromes de polinización de comunidades de bromeliáceas varı́an de maneras distintas y fácilmente interpretables en relación a condiciones de humedad y temperatura. Los sı́ndromes de polinización de bromeliáceas parecen seguir los patrones exhibidos por sus polinizadores preferenciales. Key words: bats; Bolivia; Bromeliaceae; diversity; hummingbirds; pollination; tropical montane forest. THE TROPICAL ANDES ARE AMONG THE BOTANICALLY MOST DIVERSE regions worldwide (Barthlott et al. 1996). Due to the difficulties of studying species-rich tropical plant communities, however, our knowledge of the magnitude and distribution of this diversity is still fragmentary (Gentry 1995). The most conspicuous changes of community composition and richness in tropical mountains perhaps are 1 Received 11 February 2005; revision accepted 26 April 2005. Corresponding author. Current address: Estación de Biologı́a Tropical “Los Tuxtlas,” Universidad Nacional Autónoma de México, Apartado Postal 94, San Andrés Tuxtla, Veracruz 95701, México; e-mail: tkroemer@ibiologia.unam.mx 3 Current address: Asociación Armonı́a, BirdLife International, Casilla 3566, Santa Cruz de la Sierra, Bolivia. 2 related to differences in elevation, but few studies have covered entire elevational gradients in enough detail to describe with precision how tropical forest characteristics change with elevation (Lieberman et al. 1996). The majority of elevational studies have focused on species richness (e.g., Gentry 1995, Lieberman et al. 1996, Vázquez & Givnish 1998, Krömer et al. 2005) and elevational zonation of vegetation types (see Frahm & Gradstein 1991). Also, most studies have been based on woody plants (e.g., Gentry 1988, 1995; Kitayama 1992; Lieberman et al. 1996), even though the majority of plant species in humid tropical forests belong to nontree life forms (Gentry & Dodson 1987, Ibisch 1996, Balslev et al. 1998). As a result, it is not yet known whether a general relationship exists
C 2005 The Author(s) C 2005 by The Association for Tropical Biology and Conservation Journal compilation
183 184 Krömer, Kessler, and Herzog between species richness and elevation, or even whether a universal explanation or model exists (Colwell & Hurtt 1994; Rahbek 1995, 1997). The family Bromeliaceae is almost exclusively restricted to the New World tropics. Nearly 50 percent of the estimated 3000 bromeliad species are epiphytic (Benzing 2000). These have leaftrichomes of varied forms which function as moisture- and nutrientabsorptive appendages (Benzing 1990, 2000). Studies on the elevational distribution and ecology of bromeliads are rare: Pittendrigh (1948) examined the vertical distribution of epiphytic bromeliads in Trinidad, Gilmartin (1973) related the distribution of 17 species growing on both slopes of the Ecuadorian Andes to meteorological data, and Garcı́a-Franco and Peters (1987) and CastañoMeneses et al. (2003) analyzed the spatial distribution patterns of selected species of Tillandsia along an elevational gradient in Mexico. Bromeliads have a wide range of pollinators, including bats, birds, and insects, and also include autogamous taxa (Gardner 1986, Till 1992, Benzing 2000, Kessler & Krömer 2000). Furthermore, bromeliads are one of the most important food sources for hummingbirds in many Neotropical forest regions (Cruden 1972; Araujo et al. 1994; Sazima et al. 1995a, 1996; Dziedzioch et al. 2003). In the present study, we compared the diversity, taxonomic composition, and pollination syndromes of bromeliad assemblages as well as the diversity and abundance of hummingbirds along two climatically contrasting elevational gradients in Bolivia. Our primary aim was to study qualitatively the relationship of bromeliad pollination syndromes and one pollinator group (hummingbirds) to temperature and rainfall. By using two transects differing strikingly in humidity, we hoped to be able to disentangle the relative effects of elevation as such from that of humidity, which often also varies with elevation (Kessler et al. 2001). We did not specifically study bats and insects as pollinators due to the difficulties of sampling them quantitatively. METHODS STUDY AREA AND THE ENVIRONMENT.—Carrasco National Park covers an area of 6226 km2 on the eastern slope in the department of Cochabamba, Bolivia (Ergueta & Gómez 1997; Fig. 1). Ranging from 300 to 4500 m, the park contains a complete set of elevational vegetation belts, many of which are in almost pristine state, especially at 1000–3000 m. The topography is extremely steep, with a horizontal distance of only 35–45 km from the highest peaks to the level lowlands of the Chapare region. The present study was conducted from 450 m near the cave “Cuevas del Repechón” to 3950 m along the old, now unused gravel road from Cochabamba to Villa Tunari (17◦ 03′ –09′ S, 65◦ 27′ –37′ W). Lowland forests were studied at 300–400 m in Parque Ecoturı́stico Machı́a and on the grounds of Hotel El Puente in the vicinity of the town of Villa Tunari (16◦ 58′ S, 65◦ 26′ W). These two sites, which are among the few forest areas remaining in a region largely devoted to coca cultivation (Henkel 1995), are hereafter referred to as Villa Tunari. Mean annual precipitation at Villa Tunari, the only reliable climatic station in the region, is 5676 mm (10 yr data, Ibisch 1996). FIGURE 1. Map of Bolivia showing the location of the study sites. Most precipitation falls between November and May, but even in the “dry” season from June to October every month receives >100 mm, thus creating a perhumid climate (sensu Lauer et al. 1996). There are no climatic stations at higher elevations, but since the steep topography concentrates the convective precipitation on a narrow belt, precipitation is undoubtedly higher than in the lowlands. Ibisch (1996) estimated >3500 mm mean annual precipitation at 2200 m elevation in the somewhat sheltered valley of Sehuencas, but it is likely that slopes directly exposed to incoming clouds at mid-elevations (1000–3000 m) in the NW of the park receive annual precipitation in excess of 8000 mm (J. Böhner, pers. comm.). Additional humidity is contributed by dew and fog condensation (Ibisch 1996). Mean annual temperature is 24.6◦ C at Villa Tunari and 12–15◦ C at 2200 m at Sehuencas (Ibisch 1996), with an annual variability of about 5◦ C. Nocturnal frosts occur down to 2000 m (Ibisch 1996, T. Krömer & M. Kessler, pers. obs.), especially during periodic influxes of southern polar winds during austral winter. The natural vegetation of the entire study area consists of evergreen forest. Navarro (1997) proposed a preliminary delimitation of seven vegetation belts in Carrasco National Park defined by characteristic tree species, but in-depth studies of the forest vegetation are lacking. Tree community composition probably corresponds to that generally observed along humid east-Andean slopes (Gentry 1988, 1995; Ibisch 1996; Navarro 1997). Because of the steep topography and high rainfall, landslides are common and a considerable portion of the vegetation consists of varying stages of natural forest succession (Ibisch 1996, Kessler 1999). Above about 3500 m forests are dominated by Polylepis racemosa (Rosaceae; 3500–3800 m) and P. pepei (3800–4200 m). These Polylepis forests have largely been destroyed through centuries of human-induced burning and today are restricted to relict patches, mostly in ravines or on boulder slopes (Kessler 1995, Fjeldså & Kessler 1996). The second transect was located in the Masicurı́ valley from the confluence of the Masicurı́ and Grande rivers (19◦ 02′ S, 63◦ 42′ W; 500 m) to the highest accessible peaks at San Lorenzo (18◦ 41′ S, 63◦ 55′ W; 2500 m) on the west side of the valley along the road Elevational Patterns of Bromeliads 185 from Vallegrande to Masicurı́ in western Departamento de Santa Cruz, Bolivia (Fig. 1). Mean annual precipitation at the village of Masicurı́ (800 m) is 1792 mm. Precipitation declines toward lower elevations, with values between 1000 and 1200 mm reported from various climatic stations along the Andean foothill base to the east of the study area. At higher elevations, precipitation and fog intensity increase noticeably (see Bianchi 1981, for data from Argentina), especially at about 1000–1200 m elevation (T. Krömer & M. Kessler, pers. obs.), but no quantitative measurements are available. Mean annual temperature is about 25◦ C at 500 m and declines by about 0.6◦ C per 100 m elevational increase (Eriksen 1986, Gerold 1987). Climate is seasonal, with about 75 percent of the precipitation falling in the austral summer (November to April), and a lower frost limit in the austral winter at 800 m (Eriksen 1986). Of particular importance is the regular influx of cold polar fronts along the Andean base in austral winter, locally known as surazos, which reduce temperatures considerably for several days and represent the main source of precipitation in this season (Fjeldså et al. 1999). In accordance with climatic changes, three main elevational vegetation zones were discernible at Masicurı́ which are typical for this biogeographic region as a whole (Cabrera 1976, Ribera et al. 1992, Navarro 1997, Schulenberg & Awbrey 1997). Up to 850–1000 m, forests were deciduous to semideciduous and composed primarily of Leguminosae such as Anadenanthera macrocarpa (Benth.) Brenan, Enterolobium contortisiliquum (Well) Morong., and Parapiptadenia excelsa (Griseb.) Baker. Between 850–1000 m and about 1800 m, forests were evergreen and contained Cedrela lilloi C. DC (Meliaceae), Chrysophyllum gonocarpum (Mart. & Eichl.) Engl. (Sapotaceae), Crinodendron tucumanum Lillo (Elaeocarpaceae), Ficus spp. (Moraceae), Miconia spp. (Melastomataceae), as well as numerous naturalized orange trees (Citrus aurantium L.). Above 1800 m, the evergreen forest was dominated by Podocarpus parlatorei Pilger (Podocarpaceae) and numerous Myrtaceae such as Blepharocalyx salicifolius (H.B.K.) O. Berg, Myrcianthes callicoma McVaugh, M. pseudomato (Legrand) McVaugh, and Siphoneugenia occidentalis Legrand. Human activity varied according to climatic and topographic conditions. The alluvial plain of the Masicurı́ valley was completely deforested and used for intensive agriculture. Agricultural areas on slopes were located primarily at 900–1200 m and at 1700–2000 m, corresponding to the elevations with the most benign climate and least steeply inclined slopes. Forests on hillsides and slopes throughout the entire transect experienced selective timber extraction (especially Cedrela, Podocarpus, Tabebuia spp., depending on vegetation zone) and extensive cattle grazing. Many higher mountain areas had been burnt to establish cattle pastures. BOTANICAL SAMPLING.—Fieldwork was conducted in 1996 from 23 June to 11 September along the Carrasco transect, from 12 to 19 September at Villa Tunari, and from 9 to 15 July 1995 and 20 May to 14 June 1996 in the Masicurı́ area. Bromeliads were studied on 283 plots of 400 m2 , mostly of square shape but occasionally in other shapes to minimize habit heterogeneity. At Carrasco, we studied 183 plots (24 at 300–500 m, 24 at 500–1000 m, 24 at 1000–1500 m, 23 at 1500–2000 m, 26 at 2000–2500 m, 25 at 2500–3000, 20 at 3000–3500 m, and 17 at 3500–4000 m), and 100 plots at Masicurı́ (27 at 450–1000 m, 21 at 1000– 1500 m, 26 at 1500–2000 m, 26 at 2000–2400 m). Plot size corresponds to the minimum area required for representative surveys of bromeliads in the vegetation types sampled and is small enough to keep environmental factors more or less uniform throughout the plots (Kessler & Bach 1999). Elevation was measured with a handheld Eschenbach altimeter to the nearest 50 m, correcting for weather-related inaccuracies by repeated measures and through reference to 1:250.000 topographic maps. Presence/absence of all species was registered in each plot, treating terrestrial and epiphytic plants separately. All species encountered in the survey area (but not in every single plot) were collected in triplicate and have been deposited at the Herbario Nacional de Bolivia (LPB, including all unicates), with the specialist H. Luther at Sarasota (SAR), and at the Herbarium Göttingen (GOET). POLLINATION MODES.—All recorded species (except for two Greigia species where no information was available) were grouped into four broad categories according to their main pollination mode as ornithophilous, chiropterophilous, entomophilous, mixed or unspecific (two or all of the first three categories). Classification was based on information from published sources (Gardner 1986; Bernardello et al. 1991; Till 1992; Stiles & Freeman 1993; Sazima et al. 1995a, 1996; Benzing 2000; Dziedzioch et al. 2003), personal field observations of flower visitors (assuming that regular flower visitors with appropriate morphology to come into contact with anthers and styles are pollinators), and on nectar contents analysis of 79 bromeliad species by High Performance Liquid Chromatography (Krömer 2005), which allows inferences to pollinators (Baker & Baker 1983, Bernardello et al. 1991, Schwerdtfeger 1996). Species for which no such information was available were classified based on deductions from morphology and flower coloration (Gardner 1986, Baker & Baker 1990). A more detailed classification, especially of the entomophilous taxa, was not possible since many bromeliad species are little known with respect to their pollinators and because even species with apparently clear-cut pollination syndromes often have more than one pollinator (Sazima et al. 1994). HUMMINGBIRD DATA.—Hummingbird communities were surveyed by S. K. Herzog in exactly the same areas as the botanical sampling with two exceptions on the Masicurı́ transect. At the lower end of that transect, surveys were restricted to the confluence of the Masicurı́ and Grande rivers (ca 500 m), with a survey gap extending up to an elevation of 1100 m. The accessible habitat in the intervening area was too degraded by human activities for a meaningful, comparable bird survey. At the upper end of the Masicurı́ transect (>2000 m), bird observations were concentrated on forest habitats, whereas bromeliads were sampled in all available habitats, including degraded shrubby areas. Fieldwork on the Carrasco transect was conducted between 22 June and 19 September 1996 (Herzog et al. 2005). The Masicurı́ transect was surveyed from 20 May to 13 June 1996, with a second visit to the confluence of the Masicurı́ 186 Krömer, Kessler, and Herzog and Grande rivers from 10 to 14 August 1999 (Herzog & Kessler 2002). Surveys followed the field procedures outlined in Herzog et al. (2002), which are only briefly summarized here. While walking slowly and quietly from dawn to midday and often again in late afternoon along roads, trails, and through the habitat where feasible, S. K. Herzog continuously recorded all visual and acoustical observations of all birds including hummingbirds (including numbers of individuals per species) within 50 m of the observer. The observer’s movement rate largely depended on the level of bird activity. When spending longer periods in one spot and during occasional re-sampling of an area (the latter occurred to approximately the same degree in all elevational belts and thus did not introduce a systematic error), repeated counts of obviously territorial individuals were avoided. Tape recordings were made extensively to supplement observations and to identify unknown voices. Observation time exceeded 22 h in every belt. DATA ANALYSIS.—Transects were divided into steps of 500 m, pooling all data within each step. Bromeliad richness was expressed as the total number of species recorded in each elevational belt, while the frequency of pollination modes was expressed as the percentage of species with a specific mode within a given belt. Hummingbird species richness was expressed as the total number of species recorded (i.e., raw species counts) in each elevational belt, and the number of species observed divided by the number of hours of observation time in that belt. Hummingbird abundance in each belt was expressed as the number of individuals (regardless of species identity) observed divided by the number of hours of observation time in that belt. Concordance of the richness and abundance patterns of hummingbird-pollinated bromeliads and hummingbirds was calculated via correlation analyses. RESULTS SPECIES RICHNESS AND DISTRIBUTION OF BROMELIADS.—On the Carrasco transect, a total of 49 bromeliad species in 10 genera was found (Appendix 1). Nearly 50 percent of these species belonged to the genera Guzmania and Tillandsia, which together included 23 species. Divided into life forms, there were 15 terrestrial species in 4 genera (Fosterella, Greigia, Pitcairnia, Puya) and 34 species of epiphytic tank bromeliads in 6 genera (Aechmea, Billbergia, Guzmania, Racinaea, Tillandsia, Werauhia). Highest species numbers were found in the elevational belts 1000–1500 m (23 species) and 1500–2500 m (19 species in each) (Fig. 2). Among the terrestrial genera, Pitcairnia had by far the widest elevational amplitude and was found in the lower and middle parts of the transect up to 2500 m. Fosterella was found only at mid-elevations (1100–1550 m), whereas Greigia and Puya were limited to the upper half of the transect. Epiphytic Guzmania, Tillandsia, and Racinaea covered a broad elevational range, extending from the lowlands to at least 3000 m. In contrast, Aechmea, Billbergia, and Werauhia were limited to the lower parts of the transect. On the Masicurı́ transect, we collected a total of 30 species in 8 genera (Appendix 1). The genus Tillandsia with 12 species FIGURE 2. Number of bromeliad species along the elevational transect in Carrasco (white columns) and Masicurı́ (black columns). was by far the most important, followed by Fosterella and Puya with 5 species each. The 16 terrestrial bromeliads belonged to 5 genera (Bromelia, Dyckia, Fosterella, Pitcairnia, Puya), whereas the 14 epiphytic species included 12 atmospheric tillandsias and 2 tank bromeliads (Aechmea, Vriesea). The highest number of species was found at 500–1000 m with 17 species (Fig. 2). The terrestrial genera Fosterella and Pitcairnia covered a wide elevational range of 500–1750 and 750–1950 m, respectively. Bromelia and Dyckia were limited to the lowlands, whereas Puya mostly occurred at high elevations but was also found down to 600 m. Among the epiphytic genera, Tillandsia covered the entire transect, whereas Aechmea and Vriesea were found only at mid-elevations. POLLINATION MODES.—On the Carrasco transect, 51 percent of all recorded species were ornithophilous (Table 1), especially in the epiphytic genera Tillandsia (8 species) and Guzmania (5) and the terrestrial genera Pitcairnia (4) and Puya (4) (Appendix 1). Among the five chiropterophilous species, three were Guzmania, whereas all eight entomophilous species belonged to the genera Fosterella and Racinaea. Species with unspecialized pollination modes belonged mainly to the genera Aechmea and Tillandsia. The special case of species with mixed pollination by both bats and birds was dominated by Guzmania. On the Masicurı́ transect, 40 percent of all species were ornithophilous (Table 1), including all species of the terrestrial genera Dyckia (1 species), Pitcairnia (4), and Puya (5) (Appendix 1). None of the species were chiropterophilous. The entomophilous species (47%) belonged to the genera Fosterella and Tillandsia, and the unspecialized species belonged to Aechmea and Bromelia. Analyzing the frequency of pollination modes via the number of records within all plots revealed that on both transects ornithophilous species were less frequent than expected by their species numbers, whereas chiropterophilous species and species with mixed pollination by both bats and birds were more frequent in Carrasco (Table 1). Elevational Patterns of Bromeliads 187 TABLE 1. Species numbers and frequency (number of plot-records relative to total number of plot-records; % are given relative to total species numbers and numbers of records, respectively) of different pollination modes (orn = ornithophilous, chi = chiropterophilous, ent = entomophilous, mi = mixed or unspecific, mi (chi/orn) = mixed ornithophilous/chiropterophilous) among bromeliads along the elevational transects at Carrasco and Masicurı́. Carrasco Species number % Masicurı́ Records % Species number % Records % orn 24 51 216 40 12 40 63 35 chi ent mi 5 8 6 11 17 13 101 65 40 19 12 7 0 14 4 47 13 72 44 40 25 4 47 8 100 118 540 22 100 0 30 100 179 100 mi (chi/orn) Total Looking at the elevational distribution of pollination modes, on the Carrasco transect the relative contribution of ornithophilous species increased from 20 percent at 300 m to 100 percent above 3000 m, whereas the values for chiropterophilous species declined from 54 percent at 300 m to <10 percent above 2000 m (Fig. 3). Species with mixed bat/hummingbird-pollination showed a mixed pattern and reached their highest value of 47 percent at 2000–2500 m. Entomophilous species reached their highest contributions (14– 17%) at 1000–3000 m, whereas unspecialized species declined from 25 percent in the lowlands to 0 percent at high elevations. On the Masicurı́ transect, the relative contribution of ornithophilous species increased from 10 percent at 500 m to around 70 percent at 1500–2450 m (Fig. 3). Entomophilous species declined from 62 percent at 500 m to 8 percent at 1500–2000 m, and increased again to 31 percent above 2000 m. The unspecialized species showed a decline from 28 percent in the lowlands to 0 percent at high elevations. tion time, richness peaked at 3000–3500 m (0.36 species/h), with a minor peak at 1500–2000 m (0.28 species/h; Fig. 4). Abundance of hummingbirds, however, showed an overall increase with elevation and peaked above 3000 m (4.3 individuals/h). On the Masicurı́ transect, a total of only 12 hummingbird species in 10 genera were recorded, including 7 species and 4 genera not observed on the Carrasco transect (Appendix 2). Species richness (7), number of species per hour (0.23), and abundance (1.3 individuals/h) all peaked at 1500–2000 m (Fig. 4). At Carrasco, there was no significant correlation between species richness of hummingbird-pollinated bromeliads and hummingbirds, but highly significant correlations to the number of hummingbird species recorded per hour (r = 0.9, P < 0.01), and the number of hummingbird individuals recorded per hour (r = 0.98, P < 0.01) (Fig. 5). At Masicurı́, there were no significant correlations between the richness and abundance values of bromeliads and hummingbirds. HUMMINGBIRD RICHNESS AND ABUNDANCE.—We recorded a total of 28 hummingbird species in 22 genera along the Carrasco transect (Appendix 2). Hummingbird richness reached a maximum of 10 species at 1000–2000 m, but when standardizing values for observa- DISCUSSION SPECIES RICHNESS AND DISTRIBUTION.—Elevational patterns of species richness differed noticeably between the transects (Fig. 2). FIGURE 3. Relative representation of different pollination modes among bromeliads along the elevational transect in Carrasco and Masicurı́. 188 Krömer, Kessler, and Herzog FIGURE 4. Number of hummingbird species, number of hummingbird species recorded per hour, and number of hummingbird individuals recorded per hour along the elevational transect at Carrasco (white columns) and Masicurı́ (black columns). The perhumid Carrasco transect had its highest number of species at 1000–2500 m. This is the elevational range of the main cloud condensation level typically observed in the Bolivian Andes and is characterized by maximum humidity and moderate temperatures. Opposed to this, at Masicurı́ most species were found in the hot, semiarid lowlands, whereas the temperate, semihumid zone of the Tucumano-Boliviano forest at 1000–2000 m was relatively poor in species. These patterns correspond to those of previous studies both for Andean rain forests (Sugden & Robins 1979, Sugden 1981, Cleef et al. 1984) and for dry forests (Ibisch 1996). These contrasting patterns likely reflect the climatic conditions of the transects. At Carrasco, with its perhumid climate, bromeliad assemblages are dominated by large epiphytic tank bromeliads, whereas in semihumid Masicurı́ they are characterized by small, epiphytic, atmospheric (“gray”) tillandsias. Tank bromeliads capture rain and humidity with dense rosettes formed by the bases of their leaves and cannot survive long periods of heat and dryness. In contrast, atmospheric tillandsias capture water from dew condensa- tion and rain by leaves covered with specialized scales (Smith 1989; Benzing 1990, 2000). The paucity of tank bromeliads at higher elevations at Masicurı́ probably reflects the regular occurrence of frost in the area (Kessler 2002b). In contrast to the epiphytic taxa, the distribution of terrestrial bromeliads was much more similar among transects. Although only one species (Fosterella albicans) was common to both transects, the main genera Fosterella, Pitcairnia, and Puya were the same and showed comparable richness patterns. Whereas Fosterella and Pitcairnia were found only at low and mid-elevations, Puya was almost limited to the upper parts of both transects, which are characterized by extreme climatic conditions with high solar radiation and night temperatures below freezing. These three genera are by far the most numerous and important terrestrial bromeliad genera in Bolivia (Krömer et al. 1999, Krömer 2000) and can be found in all biogeographic regions of the country. The remaining terrestrial genera are less common and were found only at one of the two transects: Greigia is a typical element in the understory of montane FIGURE 5. Correlation between number of hummingbird species/number of hummingbird species recorded per hour/number of hummingbird individuals recorded per hour, and the relative contribution of ornithophilous bromeliads in Carrasco and Masicurı́. Elevational Patterns of Bromeliads 189 and cloud forests, whereas Bromelia and Dyckia are common in dry lowland to mid-elevation areas where they colonize open and disturbed areas (Krömer et al. 1999, Vásquez & Ibisch 2003, Will et al. in press). In conclusion, the uptake of water and nutrients through the root system clearly renders terrestrial bromeliads less dependent on climatic conditions, in particular on humidity, than their epiphytic relatives. POLLINATION MODES.—On the Carrasco transect, hummingbirdpollination was the most common pollination mode. This corresponds to the overall situation in Bolivia, where 61 percent of all species are ornithophilous, compared to 24 percent entomophilous, 7 percent chiropterophilous, 4 percent autogamous, and 3 percent with mixed pollination modes (Kessler & Krömer 2000). Ornithophily was also common at Masicurı́, but even more species were entomophilous. On both transects, the percentage of ornithophilous species increased with elevation, although values were generally higher in perhumid Carrasco compared to semiarid Masicurı́. Bromeliad richness correlated very well with the abundance of hummingbirds along the Carrasco transect but not with hummingbird species richness. At Masicurı́, the low number of hummingbirds above 2000 m is striking, but may be partly due to a sampling bias. Whereas bromeliads were sampled in all available habitats, including degraded shrublands where most species were found, bird observations were concentrated on forest habitats, where hummingbirds and bird-pollinated bromeliads were rare (S. K. Herzog, pers. obs.). One caveat in this context is the seasonal elevational migratory movements of hummingbirds (Stiles 1977, 1978) that may lead to temporary shifts in their elevational distribution of richness and abundance. We do not believe that this would have severely affected our results because when we used the full known elevational range of each hummingbird species (based on Hennessey et al. 2003) to calculate the elevational pattern of species richness (data not shown), the resulting pattern closely corresponded to the richness pattern documented here. Nevertheless, it would certainly be interesting to study the seasonal variation of bromeliad flowering and hummingbird abundances and richness. Hummingbirds play a crucial role in the pollination of both epiphytic and terrestrial plants in Andean forests (Stiles 1978, Feinsinger 1983, Bawa 1990, Dziedzioch et al. 2003). Many epiphytic bird-pollinated bromeliads are adapted to their pollinators by producing few flowers over relatively long time periods (Ackerman 1986; Benzing 1990, 2000), thereby assuring a constant food supply for hummingbirds for which bromeliads are among the most important food plants (Cruden 1972; Araujo et al. 1994; Sazima et al. 1995a, 1996). In contrast, ornithophilous terrestrial bromeliads often tend to produce many flowers over short time periods, attracting an abundance of hummingbirds (Kraemer et al. 1993). We are unaware of any published explanation for this striking contrast in pollination strategies between epiphytic and terrestrial bromeliads, but we hypothesize that limitations in water availability may force epiphytic bromeliads to limit nectar production and to offset this disadvantage by providing a long-term reliable nectar source. For example, the rosette plant Lobelia rhynchopetalum (Hochst. ex A. Rich.) Hemsl. (Campanulaceae), which resembles large Puya species in habit and ecology, uses 0.3–0.6 liter of water per day vegetatively and up to 3.4 liter when flowering, mostly for the production of up to 2 liter of nectar (R. Zimmermann, pers. comm.). The only comparable study on epiphytic bromeliads (Ordano & Ornelas 2004) shows that nectar removal by pollinators induces higher nectar production through a dilution of the sugar contents, i.e., through higher water secretion, but the overall limited amount of water (<20 ml) does not allow an inference on whether the pollination strategy is influenced by water availability. Ecophysiological studies of the water budget of epiphytic and terrestrial bromeliads during flowering and nonflowering periods may be used to test the hypothesis presented here. Entomophily was of limited importance along the Carrasco transect and at middle elevations along the Masicurı́ transect, but was the most important pollination mode at low elevations at Masicurı́, mainly due to the prevalence of species of Tillandsia and Fosterella. Contrasting hummingbird and insect pollination, we found thus that hummingbird-pollination among the bromeliad assemblages increased with decreasing temperatures and with increasing humidity, while entomophily showed the opposite pattern. This general pattern was paralleled by the hummingbird assemblages, especially considering abundance. A probable reason for the predominance of hummingbirds as pollinators in cold and humid habitats is that they are the best-adapted major biotic pollinator group under these conditions. As warm-blooded organisms, hummingbirds can function even at low temperatures, and their ability to fall into torpor at night is highly energy-efficient. Opposed to this, insects depend on warm and dry conditions for optimal activity. In seasonal climates, such as along the lower part of the Masicurı́ transect, the seasonal shortage of nectar is deleterious to hummingbirds and is favorable to insects, which, in the case of some Hymenoptera, can store nectar or pollen and therefore represent the dominant pollinator group under these conditions (Proctor et al. 1996, Kessler & Krömer 2000). The importance of bat-pollination among bromeliads in tropical lowlands has been largely overlooked until very recently. Whereas many species of the genera Vriesea and Werauhia are known to be chiropteriphilous (Grant 1995, Sazima et al. 1995b), there are few additional records of bat-pollinated bromeliads (Benzing 2000). However, recent observations have shown that about 35 percent (8 of 18) of the Bolivian species of the genus Guzmania are very likely pollinated by small nectar-feeding bats (Phyllostomidae: Glossophaginae) (Krömer 2003a). Their floral syndrome includes small, night-blooming flowers with brown or green bracts, greenish to whitish petals, and mostly a specific smell such as garlic-like odor in Guzmania sphaeroidea. Further chiropterophilous bromeliads are found in the genera Billbergia (B. robert-readii), Pitcairnia (P. crassa), and Puya (P. ferruginea; Krömer 2003a). All these, in contrast to Guzmania, have large and nectar-rich flowers, which suggests that they are pollinated by fairly large bat species. Both Bolivian species of Werauhia are characterized by bell-shaped (campanulate) flowers that fit like a “head-mask” on the elongated rostrum of the nectarfeeding bats. Werauhia gladioliflora, a relatively frequent species in lower montane forest, primarily grows in the understory and the 190 Krömer, Kessler, and Herzog trunk area, where its flowers are projected into the open air (Krömer 2003b). This exposure, similar to cauliflory, provides space for wing movements of the bats during hovering. Generally speaking, batpollination is clearly linked to humid, tropical conditions (Kessler & Krömer 2000) in accordance with the high species richness of bats in tropical lowlands (Patterson et al. 1996). Interestingly, at 1500–2500 m on the Carracso transect, we found three Guzmania (G. danielii, G. killipiana, G. morreniana) and one Pitcairnia (P. cf. trianae) species that could not be readily assigned to any of the main pollination syndromes, but which contributed 23–49 percent of all bromeliad individuals. On the one hand, the inconspicuous, brownish flowering stands of the Guzmania species and their nocturnal anthesis (T. Krömer, pers. obs.) suggest bat-pollination. Further, they have hexose-rich nectar (T. Krömer, pers. comm.), corresponding to typical bat-pollinated species (Baker & Baker 1983). However, the flowers are also open during the day and are scentless (at least to human observers), which suggests that they are also pollinated by hummingbirds, as indeed observed in G. killipiana in Ecuador (Dziedzioch et al. 2003). Similar intermediate pollination syndromes have been found for three species of the genus Abutilon (Malvaceae) (Buzato et al. 1994) and for Siphocampylos sulfureus E. Wimm. (Lobeliaceae) (Sazima et al. 1994). Strikingly, on the Carrasco transect, mixed hummingbird/bat-pollination was found especially at mid-elevations, i.e., in the transition zone between preferential bat-pollination in the lowlands and preferential hummingbird-pollination in the highlands. Species with mixed pollination visited by hummingbirds and insects were found most frequently at low elevations along both transects. Such unspecific pollination has repeatedly been recorded among bromeliads and other tropical plants (Bernardello et al. 1991, Seres & Ramı́rez 1995, Benzing 2000). A possible explanation for the prevalence of this mode in the lowlands could be the absence of harsh environmental conditions limiting the availability of specific pollinators. Under harsh conditions with few available pollinators, it may be advantageous for plants to adapt to specific pollinators in order to ensure reliable pollination (Kessler & Krömer 2000). In conclusion, we found that both the richness patterns and pollination syndromes of bromeliad assemblages varied in distinct and readily interpretable ways in relation to environmental humidity and temperature, as shown by the different patterns on the two climatically contrasting study transects. In all cases, the prevalence of certain pollination syndromes under specific climatic conditions was determined by the physiological adaptations and tolerances of the pollinator taxa, suggesting that in the studied plant–animal interactions the plants adapt mainly to the animals and not vice versa. This interpretation is further supported by the fact that bromeliad richness trailed hummingbird abundance at Carrasco, whereas hummingbird diversity was largely independent of bromeliad diversity. Naturally, the adaptive potential of bromeliads is limited by their ecophysiological growth adaptations to ecoclimatic conditions (Benzing 2000; Kessler 2002a,b). How the respective bromeliad genera realize the adaptation to a certain pollinator is apparently influenced by life form, and potentially by other life history attributes as well. Our study thus suggests that bromeliads strive toward an optimization of their pollination success by pollinator shifts and different nectar reward strategies within the limitations of their evolutionary and ecophysiological bounds. ACKNOWLEDGMENTS For help and good companionship during fieldwork, we thank A. Acebey, J. A. Balderrama, J. Gonzales, A. Green, B. Hibbits, I. Jimenez, J.-C. Ledesma, and M. Sonnentag. We further thank H. Luther and P. L. Ibisch for specimen identification and two anonymous reviewers for valuable comments. This study would have been impossible without the logistic support by the Herbario Nacional de Bolivia, La Paz, in particular by S. G. Beck, M. Cusicanqui, A. de Lima, R. de Michel, and M. Moraes. For working and collecting permits we thank the Dirección Nacional de Conservación de la Biodiversidad (DNCB), La Paz. This study was supported by the Deutsche Forschungsgemeinschaft, the A. F. W. Schimper-Stiftung, the DIVA project under the Danish Environmental Programme, the German Academic Exchange Service, Fauna and Flora International, and the Gesellschaft für Tropenornithologie. LITERATURE CITED ACKERMAN, J. D. 1986. Coping with the epiphytic existence: Pollination strategies. Selbyana 9: 52–60. ARAUJO, A. C., E. FISCHER, AND M. SAZIMA. 1994. Floração sequencial e polinização de três expécies de Vriesea (Bromeliaceae) na região de Juréia, sudests do Brasil. Rev. Bras. Bot. 17: 113–118. BAKER, H. G., AND I. BAKER. 1983. Floral nectar sugar constituents in relation to pollinator type. In C. E. Jones and R. J. Little (Eds.). Little handbook of experimental pollination biology, pp. 111–141. Van Nostrand Reinhold, New York. ———, AND ———. 1990. The predictive value of nectar chemistry to the recognition of pollinator types. Isr. J. Bot. 39: 157–166. BALSLEV, H., R. VALENCIA, G. PAZ Y MIÑO, H. CHRISTENSEN, AND I. NIELSEN. 1998. Species count of vascular plants in one hectare of humid lowland forest in Amazonian Ecuador. In F. Dallmeier and J. A. Comiskey (Eds.). Forest biodiversity in North, Central and South America, and the Caribbean, pp. 585–594. Man and The Biosphere Series, Vol. 21. UNESCO, Paris and Parthenon Publishing Group, New York. BARTHLOTT, W., W. LAUER, AND A. PLACKE. 1996. Global distribution of species diversity in vascular plants: Towards a map of phytodiversity. Erdkunde 50: 317–327. BAWA, K. S. 1990. Plant–pollinator interactions in tropical rain forests. Annu. Rev. Ecol. Syst. 21: 399–422. BENZING, D. H. 1990. Vascular epiphytes: General biology and related biota. Cambridge University Press, Cambridge, England. ———. 2000. Bromeliaceae. Profile of an adaptive radiation. Cambridge University Press, Cambridge, England. BERNARDELLO, L. M., L. GALETTO, AND H. R. JULIANI. 1991. Floral nectar, nectary structure and pollinators in some Argentinian Bromeliaceae. Ann. Bot. 67: 401–411. BIANCHI, R. 1981. Las precipitaciones del Noroeste Argentino. Instituto Nacional de Tecnologı́a Agropecuaria, Salta, Argentina. BUZATO, S., M. SAZIMA, AND I. SAZIMA. 1994. Pollination of three species of Abutilon (Malvaceae) intermediate between bat and hummingbird flower syndromes. Flora 189: 327–334. CABRERA, A. L. 1976. Regiones fitogeográficas Argentinas. Enciclopedia Argentina de Agricultura y Jardinerı́a, Vol. II. Editorial ACME, Buenos Aires, Argentina. Elevational Patterns of Bromeliads 191 CASTAÑO-MENESES, G., J. G. GARCÍA-FRANCO, AND J. G. PALACIOS-VARGAS. 2003. Spatial distribution patterns of Tillandsia violacea (Bromeliaceae) and support trees in an altitudinal gradient from a temperate forest in Central Mexico. Selbyana 24: 71–77. CLEEF, A. M., O. RANGEL, T. VAN DER HAMMEN, AND R. JARAMILLO. 1984. La Vegetación de las selvas del transecto Buritaca. In T. van der Hammen and P. M. Ruiz (Eds.). Studies on tropical Andean ecosystems 2, pp. 267–407. J. Cramer, Berlin, Germany. COLWELL, R. K., AND G. C. HURTT. 1994. Nonbiological gradients in species richness and a spurious Rapoport effect. Am. Nat. 144: 570– 595. CRUDEN, R. W. 1972. Pollinators in high-elevation ecosystems: Relative effectiveness of birds and bees. Science 176: 1439–1440. DZIEDZIOCH, C., A.-D. STEVENS, AND G. GOTTSBERGER. 2003. The hummingbird plant community of a tropical montane rain forest in Southern Ecuador. Plant Biol. 5: 331–337. ERGUETA, S. P., AND C. H. GÓMEZ. 1997. Directorio de Areas Protegidas de Bolivia. CDC-Bolivia, La Paz, Bolivia. ERIKSEN, W. 1986. Frostwechsel und hygrische Bedingungen in der Puna Boliviens. Ein Beitrag zur Ökoklimatologie der randtropischen Anden. Bolivien. In H. J. Buchholz (Ed.). Beiträge zur physischen Geographie eines Andenstaates, pp. 1–21. Jahrb. Geogr. Ges. Hannover 1985, Hannover, Germany. FEINSINGER, P. 1983. Community organization among neotropical nectarfeeding birds. Am. Zool. 18: 779–795. FJELDSÅ, J., AND M. KESSLER. 1996. Conserving the biological diversity of Polylepis woodlands of the highlands of Peru and Bolivia: A contribution to sustainable natural resource management in the Andes. NORDECO, Copenhagen, Denmark. ———, E. LAMBIN, AND B. MERTENS. 1999. Correlation between endemism and local ecoclimatic stability documented by comparing Andean bird distributions and remotely sensed land surface data. Ecography 22: 63– 78. FRAHM, J. P., AND S. R. GRADSTEIN. 1991. An altitudinal zonation of tropical rain forests using bryophytes. J. Biogeogr. 18: 669–678. GARCÍA-FRANCO, J. G., AND C. M. PETERS. 1987. Patrón especial y abundancia de Tillandsia spp. a través de un gradiente altitudinal en los altos de Chiapas, Mexico. Brenesia 27: 35–45. GARDNER, C. S. 1986. Inferences about pollination in Tillandsia (Bromeliaceae). Selbyana 9: 76–87. GENTRY, A. H. 1988. Changes in plant community diversity and floristic composition on environmental and geographical gradients. Ann. Mo. Bot. Gard. 75: 1–34. ———. 1995. Patterns of diversity and floristic composition in neotropical montane forests. In S. P. Churchill, H. Balslev, E. Forero, and J. L. Luteyn (Eds.). Biodiversity and conservation of neotropical montane forests, pp. 103–126. The New York Botanical Garden, New York. ———, AND C. H. DODSON. 1987. Diversity and biogeography of neotropical vascular epiphytes. Ann. Mo. Bot. Gard. 74: 205–233. GEROLD, G. 1987. Untersuchungen zur Klima-, Vegetations-, Höhenstufung und Bodensequenz in SE-Bolivien. Ein randtropisches Profil vom Chaco bis zur Puna. In Beiträge zur Landeskunde Boliviens, pp. 1–70. Geographisches Institut der RWTH Aachen, Germany. GILMARTIN, A. J. 1973. Transandean distributions of Bromeliaceae in Ecuador. Ecology 54: 1389–1393. GRANT, J. R. 1995. The resurrection of Alcantarea and Werauhia, a new genus. Trop. Subtrop. Pflanzenwelt 91: 1–57. HENKEL, R. 1995. Coca (Erythroxylum coca) cultivation, cocaine production, and biodiversity in the Chapare region of Bolivia. In S. P. Churchill, H. Balslev, E. Forero, and J. L. Luteyn (Eds.). Biodiversity and conservation of Neotropical montane forests, pp. 551–560. The New York Botanical Garden, New York. HENNESSEY, A. B., S. K. HERZOG, AND F. SAGOT. 2003. Lista anotada de las Aves de Bolivia. Asociación Armonı́a/Birdlife International, Santa Cruz, Bolivia. HERZOG, S. K., AND M. KESSLER. 2002. Biogeography and composition of dry forest bird communities in Bolivia. J. Ornithol. 143: 171–204. ———, ———, AND K. BACH. 2005. The elevational gradient in Andean bird species richness at the local scale: A foothill peak and a high-elevation plateau. Ecography 28: 209–222. ———, AND T. M. CAHILL. 2002. Estimating species richness of tropical bird communities from rapid assessment data. Auk 119: 749–769. IBISCH, P. L. 1996. Neotropische Epiphytendiversität—das Beispiel Bolivien. Martina Galunder-Verlag, Wiehl, Germany. KITAYAMA, K. 1992. An altitudinal transect study of the vegetation of Mount Kinabalu, Borneo. Vegetatio 102: 149–171. KESSLER, M. 1995. Polylepis-Wälder Boliviens: Taxa, Ökologie, Verbreitung und Geschichte. Diss. Bot. 246. J. Cramer, Berlin, Germany. ———. 1999. Plant species richness and endemism during natural landslide succession in a perhumid montane forest in the Bolivian Andes. Ecotropica 5: 123–136. ———. 2002a. Environmental patterns and ecological correlates of range-size among bromeliad communities of Andean forests in Bolivia. Bot. Rev. 68: 100–127. ———. 2002b. Species richness and ecophysiological type among Bolivian bromeliad communities. Biodiv. Conserv. 11: 987–1010. ———, AND K. BACH. 1999. Using indicator families for vegetation classification in species-rich neotropical forests. Phytocoenologia 29: 485–502. ———, S. K. HERZOG, J. FJELDSA, AND K. BACH. 2001. Species richness and endemism of plant and bird communities along two gradients of elevation, humidity, and land use in the Bolivian Andes. Divers. Distrib. 7: 61– 77. ———, AND T. KRÖMER. 2000. Patterns and ecological correlates of pollination modes among Bromeliad communities of Andean forests in Bolivia. Plant Biol. 2: 659–669. KRAEMER, M., U. SCHMIDT, AND K. L. SCHUCHMANN. 1993. Notes on the organization of a neotropical high-altitude hummingbird-flower community. In W. Barthlott, C. M. Naumann, K. Schmidt-Loske, and K.-L. Schuchmann (Eds.). Animal–plant interactions in tropical environments, pp. 61–65. Zoologisches Forschungsinstitut und Museum Alexander König, Bonn, Germany. KRÖMER, T. 2000. Distribution of terrestrial bromeliads along the La Paz to Caranavi road in Bolivia. J. Brom. Soc. 50: 158–164. ———. 2003a. Fledermausbestäubung (Chiropterophilie) bei Bromelien der Gattung Guzmania in Bolivien. Die Bromelie 3/2003: 60–63. ———. 2003b. Diversität und Ökologie der vaskulären Epiphyten in primären und sekundären Bergwäldern Boliviens. Cuvillier, Göttingen, Germany. ———, M. KESSLER, S. R. GRADSTEIN, AND A. ACEBEY. 2005. Local-scale diversity patterns of vascular epiphytes along an elevational gradient in the Andes. J. Biogeogr. 32: 1799–1809. ———, ———, B. K. HOLST, H. E. LUTHER, E. J. GOUDA, P. L. IBISCH, W. TILL, AND R. VÁSQUEZ. 1999. Checklist of Bolivian Bromeliaceae with notes on species distribution and levels of endemism. Selbyana 20: 201–223. LAUER, W., M. D. RAFIQPOOR, AND P. FRANKENBERG. 1996. Die Klimate der Erde. Eine Klassifikation auf ökophysiologischer Grundlage der realen Vegetation. Erdkunde 50: 275–300. LIEBERMAN, D., M. LIEBERMAN, R. PERALTA, AND G. S. HARTSHORN. 1996. Tropical forest structure and composition on a large-scale altitudinal gradient in Costa Rica. J. Ecol. 84: 137–152. NAVARRO, G. 1997. Contribución a la clasificación ecológica y florı́stica de los bosques de Bolivia. Rev. Boliviana Ecol. Conserv. 2: 3–37. ORDANO, M., AND J. F. ORNELAS. 2004. Generous-like flowers: Nectar production in two epiphytic bromeliads and a meta-analysis of removal effects. Oecologia 140: 495–505. PATTERSON, B. D., V. PACHECO, AND S. SOLARI. 1996. Distribution of bats along an elevational gradient in the Andes of South-Eastern Peru. J. Zool. Lond. 240: 637–658. PITTENDRIGH, C. S. 1948. The bromeliad-anopheles-malaria complex in Trinidad. I. The bromeliad flora. Evolution 2: 58–89. 192 Krömer, Kessler, and Herzog PROCTOR, M., P. YEO, AND A. LACK. 1996. The natural history of pollination. Timber Press, Portland, Oregon. RAHBEK, C. 1995. The elevational gradient of species richness: A uniform pattern? Ecography 18: 200–205. ———. 1997. The relationship among area, elevation, and regional species richness in neotropical birds. Am. Nat. 149: 875–902. REMSEN, J. V., JR., A. JARAMILLO, M. NORES, J. F. PACHECO, M. B. ROBBINS, T. S. SCHULENBERG, F. G. STILES, J. M. C. DA SILVA, D. F. STOTZ, AND K. J. ZIMMER. 2005. A classification of the bird species of South America. American Ornithologists’ Union. Version 14 March 2005. http://www.museum.lsu.edu/∼Remsen/SACCBaseline.html RIBERA, M. O., LIBERMANN, M., BECK, S., AND M. MORAES. 1992. Vegetación de Bolivia. In B. K. Mihotek (Ed.). Comunidades, Teritorios Indı́genas y Biodiversidad en Bolivia, pp. 169–222. U.A.G.R.M.—CIMAR, Santa Cruz, Bolivia. SAZIMA, I., S. BUZATO, AND M. SAZIMA. 1995a. The saw-billed hermit Ramphodon naevius and its flowers in a montane forest in southeastern Brazil. J. Ornithol. 136: 195–206. ———, ———, AND ———. 1996. An assemblage of hummingbirdpollinated flowers in a montane forest in southeastern Brazil. Bot. Acta 109: 149–160. SAZIMA, M., I. SAZIMA, AND S. BUZATO. 1994. Nectar by day and night: Siphocampylus sulfureus (Lobeliaceae) pollinated by hummingbirds and bats. Plant Syst. Evol. 191: 237–246. ———, ———, AND ———. 1995b. Bat pollination of Vriesea in Southeastern Brazil. Bromélia 2: 29–37. SCHULENBERG, T. S., AND K. AWBREY. 1997. A rapid assessment of the humid forests of South Central Chuquisaca. Bolivia. RAP Working Papers 8. Conservation International, Washington, D.C. SCHWERDTFEGER, M. 1996. Die Nektarzusammensetzung der Asteridae und ihre Beziehung zu Blütenökologie und Systematik. Diss. Bot. 264. J. Cramer, Berlin, Germany. SERES, A., AND N. RAMÍREZ. 1995. Biologı́a floral y polinización de algunas monocotiledoneas de un bosque nublado venezulano. Ann. Mo. Bot. Gard. 82: 61–81. SMITH, J. A. C. 1989. Epiphytic bromeliads. In U. Lüttge (Ed.). Vascular plants as epiphytes, pp. 109–138. Springer-Verlag, Berlin, Germany. STILES, F. G. 1977. Coadapted competitors: Flowering seasons of hummingbird food plants in a tropical forest. Science 198: 1177–1178. ———. 1978. Temporal organization of flowering among the hummingbird foodplants of a tropical wet forest. Biotropica 10: 194–210. ———, AND C. E. FREEMAN. 1993. Patterns in floral nectar characteristics of some bird-visited plant species from Costa Rica. Biotropica 25: 191–205. SUGDEN, A. M. 1981. Aspects of the ecology of vascular epiphytes in two Colombian cloud forests. II. Habitat preferences of Bromeliaceae in the Serrania de Macuria. Selbyana 5: 264–273. ———, AND R. J. ROBINS. 1979. Aspects of the ecology of vascular epiphytes in Colombian cloud forests, I. The distribution of the epiphytic flora. Biotropica 11: 173–188. TILL, W. 1992. Systematics and evolution of the tropical–subtropical Tillandsia subgenus Diaphoranthema (Bromeliaceae). Selbyana 13: 88–94. VÁSQUEZ, R., AND P. L. IBISCH. 2003. The genus Bromelia (Bromeliaceae) in Bolivia (with the description of two new species from the Santa Cruz department). Rev. Soc. Bol. Bot. 4: 51–65. VÁZQUEZ, J. A., AND T. J. GIVNISH. 1998. Altitudinal gradients in tropical forest composition, structure, and diversity in the Sierra de Manantlán. J. Ecol. 86: 999–1020. WILL, B., H. E. LUTHER, T. KRÖMER, AND M. KESSLER. Three new species and two new records of Greigia Regel (Bromeliaceae) from Bolivia. Selbyana. In press. Elevational Patterns of Bromeliads 193 APPENDIX 1. Bromeliad species observed on the Carrasco and Masicurı́ elevational transects in the Bolivian Andes. Pollination mode based on 1 published sources: a Sazima et al. (1995a), b Benzing (2000), c Dziedzioch et al. (2003); 2 personal observation; 3 nectar content: a Bernardello et al. (1991), b Stiles and Freeman (1993), c T. Krömer, pers. comm.; 4 deduction based on flower morphology. Species Corolla length (mm) Bract color Petal color Pollination mode Elevational range Carrasco Aechmea angustifolia Poepp. & Endl. 12–16 Red Yellow mi4 300–1050 Aechmea distichantha Lem. Aechmea longifolia (Rudge) L. B. Sm. & M. A. Spencer 18 25–30 Rose Pink or rose White, purple or blue White mi1a,2,3a mi3c 400–450 Billbergia jandebrabanderi R. Vásquez & P. L. Ibisch Bromelia serra Griseb. 85–100 Pink Green orn3c 850–1600 15 Red Blue-purple mi3a 5.5 4–5 Green Green Red-orange White White orn4 ent1b ent1b 5 Green Green Green White White White ent1b ent1b ent1b 32 Green Castaneous White Pale purple ent1b no data 1700–2200 Greigia cf. kessleri H. Luther Guzmania besseae H. Luther Guzmania calothyrsus Mez 20 ca 25 45 Brown Castaneous Green to brown Rosey purple White White no data orn1c chi2 2600–3500 2000–2200 300–900 Guzmania danielii L. B. Sm. Guzmania gloriosa (André) André ex Mez Guzmania killipiana L. B. Sm. ca 25 50 38 Castaneous Red Castaneous White Yellow White mi (chi/orn)3c orn1c mi (chi/orn) 1c,3c 1600–2700 1700–3000 650–2600 Guzmania marantoidea (Rusby) H. Luther Guzmania melinonis Regel Guzmania morreniana (Linden Hortus) Mez ca 25 28 22 Red Red Castaneous Violet Yellow White orn4 orn3c mi (chi/orn)3c 1000–1300 300–1050 1300–2300 Guzmania retusa L. B. Sm. Guzmania roetzlii (E. Morren) Mez Guzmania sphaeroidea (André) André ex Mez 23 23 ca 25 Green Pale green Green or brown White White Cream or green chi2 mi2 chi3c 1300–1600 500–1050 450–1550 Guzmania squarrosa (Mez & Sodiro) L. B. Sm. & Pittendrigh Mezobromelia capituligera (Grisebach) J. R. ca 50 Red to orange Yellow orn3c Dyckia cf. leptostachya Baker Fosterella albicans (Griseb.) L. B. Sm. Fosterella chaparensis Ibisch, Vásquez & Gross Fosterella cf. schidosperma (Baker) L. B. Sm. Fosterella spec. 1 MK 5284 Fosterella spec. 2 MK 5373 Fosterella spec. 3 MK 5973 Greigia cochabambae H. Luther 1500–1950 500–800 1300 1550 500 1000 1250 1600 28 Red or orange White or yellow orn4 500–2200 53 Red Red orn4 650–1300 Pitcairnia brittoniana Mez Pitcairnia divaricata Wittmack 40 50 Red Yellow or orange Red orn3b orn4 650–1750 Pitcairnia lanuginosa Ruiz & Pavón Pitcairnia paniculata (Ruiz & Pavón) Ruiz & Pavón 90 45 Green with purple lines Red orn4 orn4 300–900 300–1300 Greenish yellow mi (chi/orn)3c orn?4 orn?4 1900–2500 Castaneous White Greenish white Yellow orn4 chi1b,2 orn3c 2800–2950 2200–2550 3450–3800 Brown Blue Blue-green Dark violet orn4 orn3c orn4 ca 45 Puya atra L. B. Sm. Puya ferruginea (Ruiz & Pavón) L. B. Sm. Puya herzogii Wittmack 60 140 50 Puya leptostachya L. B. Sm. Puya nana Wittm. Puya sanctae-crucis (Baker) L. B. Sm. 30 40 40 Green 500–550 1750 1100–1300 Grant Pitcairnia amboroensis Ibisch, Vásquez, Gross & Kessler Pitcairnia cf. trianae André Pitcairnia spec. 1 MK 5956 Pitcairnia spec. 2 MK 6010 Masicurı́ 1750–1900 1900–1950 750–1200 1150 3650 2100–2400 1950–2150 194 Krömer, Kessler, and Herzog APPENDIX 1. Continued. Species Puya cf. secunda L. B. Sm. Puya serranoensis Rauh Corolla length (mm) 40 60–70 Bract color Red Rose Petal color Dark violet Blue-green Racinaea seemannii (Baker) M. A. Spencer & L. B. Sm. Racinaea spiculosa (Grisebach) M. A. Spencer & L. B. Sm. Racinaea tetrantha (Ruiz & Pavón) M. A. Spencer & L. B. Sm. orn3c orn3c Elevational range Carrasco Masicurı́ 2200 2100–2400 orn?4 orn?4 Puya spec.1 MK 5332 Puya spec.2 MK 6516 Racinaea kessleri H. Luther Racinaea schumanniana (Wittmack) J. R. Grant Pollination mode 750 2400 5–7 5 Brown Green Yellow White to yellowish ent4 ent4 2950 650–2550 12–15 Red White orn1c,3c 2000–3150 5–10 Greenish white ent4 650–1400 14 Yellow ent1c 2500–2950 Yellow Violet ent?4 orn4 orn4 1600–2550 1300–2500 2200 mi4 mi4 , auto1 orn1c,3a 1800–2550 Racinaea spec. MK 6976 Tillandsia asplundii L. B. Sm. Tillandsia australis Mez ? 20 20–40 Tillandsia bryoides Griseb. Tillandsia capillaris Ruiz & Pavón Tillandsia complanata Bentham 5–9 5 20–25 Red or purple Sulphur-yellow White, yellow or brown Rose, purple or blue Tillandsia didisticha (E. Morren) Baker Tillandsia fendleri Grisebach 15–20 25–45 Red Yellow or red White Blue-violet mi2,3a orn4 650–1400 Tillandsia ionochroma André ex Mez Tillandsia kessleri H. Luther Tillandsia cf. kuntzeana Mez 20 35 20–25 Red Yellow Blue or violet Violet Violet orn4 orn4 orn3c 2700–3400 2100–3000 1700–3050 Tillandsia krukoffiana L. B. Sm. Tillandsia loliacea Martius ex Schultes f. Tillandsia pohliana Mez 10 6–10 18–25 Blue Pale violet to yellow White orn4 mi4 mi4 Tillandsia recurvata (L.) L. Tillandsia cf. reichenbachii Baker Tillandsia cf. rusbyi Baker 10–13 7 30 Green Pale violet or white Blue-violet White mi4 , auto1 mi4 mi4 Tillandsia spiralipetala Gouda Tillandsia streptocarpa Baker Tillandsia tenuifolia L. 12 18–25 20 Green Yellow-brown Blue or purple White, rose or blue mi4 mi4 orn3a 500–1250 500–1550 800–2300 Tillandsia tricholepis Baker Tillandsia usneoides (L.) L. Tillandsia violascens Mez 7 9–11 7 Green Yellow greenish Pale green Violet mi4 mi4 mi4 500–1000 1200–2450 1300–1550 Vriesea maxoniana (L.B. Sm.) L. B. Sm. Werauhia gladioliflora (Wendland) J. R. Grant 45 40–70 Greenish yellow Green Yellow Greenish white orn4 chi1b 300–1500 Red Green to pale rose Red Pink to red 500 1250–2450 500–1800 2200 1100 500–1000 750–800 500–1000 500–600 500–1050 1000–1900 Elevational Patterns of Bromeliads 195 APPENDIX 2. Elevational distribution of hummingbird species observed along the study transects. Taxonomy and species sequence follow Remsen et al. (2005). Species Carrasco Phaethornis ruber (Linnaeus) P. stuarti Hartert 300–500 500–1000 P. pretrei (Lesson & DelLattre) P. malaris (Nordmann) Campylopterus largipennis (Boddaert) 300–2000 300–1500 Colibri delphinae (Lesson) C. thalassinus (Swainson) Anthracothorax nigricollis (Vieillot) 1500–2000 1000–3000 300–500 Klais guimeti (Bourcier) Lophornis delattrei (Lesson) 500–1500 1500–2000 Chlorostilbon mellisugus (Linnaeus) C. aureoventris (d’Orbigny & Lafresnaye) Thalurania furcata (J. F. Gmelin) 2000–2500 300–1000 Hylocharis chrysura (Shaw) Chrysuronia oenone (Lesson) Amazilia chionogaster (Tschudi) 300–1000 500–1950 A. fimbriata (J. F. Gmelin) Adelomyia melanogenys (Fraser) Heliodoxa leadbeateri (Bourcier) 300–500 1100–3000 1500–2000 Aglaeactis pamela (d’Orbigny) Coeligena coeligena (Lesson) C. torquata (Boissonneau) 3000–3500 1100–2000 2000–3500 C. violifer (Gould) Pterophanes cyanopterus (Fraser) Patagona gigas (Vieillot) 2500–3500 2500–3500 3000–3500 Heliangelus amethysticollis (d’Orbigny & Lafresnaye) Eriocnemis glaucopoides (d’Orbigny & Lafresnaye) 2000–3500 Ocreatus underwoodii (Lesson) Sappho sparganura (Shaw) Metallura tyrianthina (Loddiges) 1100–2000 2000–3500 M. aeneocauda (Gould) Algaiocercus kingi (Lesson) Schistes geoffroyi (Bourcier) 2500–3500 1100–2500 300–2000 Heliomaster longirostris (Audebert & Vieillot) Microstilbon burmeisteri (P. L. Sclater) Masicurı́ 1100–1950 1100–1500 1500–1950 1100–1950 500 500–2500 500 1100–1950 1500–1950 2000–2500 300–500 500