Orchid bees (Hymenoptera: Apidae: Euglossini) are abundant and important pollinators of Neotropical forests, but orchid bee diversity is still poorly known in much of Mesoamerica, particularly in cloud forests. This paper presents results of a 2012 to 2013 survey of orchid bee diversity in Cusuco National Park, a cloud forest environment in northwest Honduras. Orchid bees were collected using insect nets at bait stations with chemical attractants. Bait stations were located at 68 sample sites throughout the park. We collected 4,293 orchid bees representing 24 species and 4 genera. One species, Euglossa imperialis Cockerell, accounted for 67.6% of the total individuals collected. A substantial easterly range extension was established for Eufriesea pallida (Kimsey). Our study provides the first intensive inventory of cloud forest orchid bee diversity in the region. Furthermore, it provides baseline data for future studies of orchid bees in a key biodiversity area that is threatened by human population growth and associated land use changes.
Orchid bees (Hymenoptera: Apidae: Euglossini) are ubiquitous and important pollinators that can comprise up to 25% of Neotropical forest bee communities (Roubik & Hanson 2004). There are over 200 known species of orchid bees, and the tribe ranges from Mexico to Argentina (Roubik & Hanson 2004; Nemésio & Silveira 2007) as well as southern North America (Minckley & Reyes 1996; Skov & Wiley 2005; Pemberton & Wheeler 2006; Eltz et al. 2011). However, knowledge of orchid bee diversity varies greatly with location. In Mesoamerica, for instance, Ramírez et al. (2002), in their catalog of Neotropical orchid bee species, list about 69 species records for Panama, but only about 10 for Honduras. A more recently published catalog of bee species and their geographic ranges published by Moure et al. (2012) lists about 66 orchid bee species records for Panama, but only about 20 for Honduras. This latter number is undoubtedly a function of the dearth of orchid bee studies in Honduras. Ascher & Pickering's (2015) Discover Life database has 29 species listed for Honduras, although this may be less accurate due to the broad and under-verified source data used by this resource.
Much of their ecological importance stems from the behavior of male orchid bees, which are among the few pollinators that visit flowers to collect aromatic compounds or associated chemicals (Roubik & Hanson 2004). These fragrances appear to be associated with species recognition, competition, and mate choice (Zimmermann et al. 2009). Orchid bees are best known for their associations with orchids, and are important pollinators of roughly 650 species in this plant family. However, they are associated with many other plant species as well, spanning about 30 families (Dressler 1982; Ackerman 1983; Roubik & Hanson 2004; Ackerman & Roubik 2012). Gilbert (1980) cited orchid bees as particularly important Neotropical “link organisms,” stating that “a single euglossine species may link plant species from all stages and strata of a forest into a system of indirect mutualism.”
Mesoamerica is one of the world's biodiversity hotspots (Mittermeier et al. 1999). Cusuco National Park, containing primarily high elevation cloud forest in northwest Honduras, is a site of critical conservation concern based on its vertebrate fauna (Le Saout et al. 2013), but little is known of the invertebrate diversity within the park. Cusuco National Park has been designated a Key Biodiversity Area, but is threatened by human population growth and associated land cover changes driven by practices such as conversion of forests to coffee plantations (Green et al. 2012). To expand our knowledge of the orchid bee fauna of Honduras, and to provide baseline data on orchid bee diversity in a location of concern that is undergoing rapidly changing land use patterns, we conducted a survey of the orchid bees of Cusuco National Park.
Materials and Methods
The study was done from mid-Jun to early Aug in 2012 and 2013 at Cusuco National Park, about 30 km west of San Pedro Sula in the State of Cortés, in northwest Honduras (Fig. 1a). Cusuco National Park is an area of about 23,400 ha and ranges in elevation from just above sea level to 2,425 m. Habitat types include disturbed and undisturbed mature broadleaf forest, secondary broadleaf forest, disturbed pine-broadleaf forest, mature pine forest, open disturbed grassland, open disturbed logged areas, and open coffee plantations. Monthly mean rainfall for all sample sites ranged from 60.8 mm in Mar to 252.4 mm in Jun, with mean total annual rainfall of 1,786.0 mm. Mean total annual rainfall by habitat type ranged from 1,743.0 mm for disturbed broadleaf forest (Mar minimum = 60.9 mm, Jun maximum = 245.8 mm) to 1,835.5 mm for secondary broadleaf forest (Mar minimum = 61.8 mm, Jun maximum = 258.3 mm). Mean precipitation for each sample site was extracted from downsampled WorldClim data ( www.worldclim.org) interpolated from weather station data in San Pedro Sula.
SAMPLING OF ORCHID BEES
Orchid bees were collected with insect nets at bait stations, which consisted of a cotton ball with a chemical attractant suspended about 1.7 m above the ground by a string. Five bait stations, about 5 m apart, were established during each collection period. Five chemical attractants were used (benzyl acetate, cineole, eugenol, methyl salicylate, and vanillin), with each cotton ball receiving 30 drops of one of these initially; re-application with 20 additional drops was done every 30 min. Occasionally, fewer bait stations or attractants were used for specific smaller-scale studies. The collection period was 9:00 a.m. to 11:00 a.m. Collections were done at 68 established sample sites, each within 2 km of 1 of the 7 research camps located throughout the park (Fig. 1b): Base Camp (15.4964°N, 88.2119°W), 14 sites; Buenos Aires (15.4997°N, 88.1794°W), 14 sites; Cantiles (15.5133°N, 88.2417°W), 4 sites; Cortecito (15.5233°N, 88.2886°W), 8 sites; El Danto (15.5283°N, 88.2775°W), 10 sites; Guanales (15.4886°N, 88.2342°W), 6 sites; and Santo Tomas (15.5611°N, 88.2975°W), 12 sites. Most sites were sampled twice. Identifications of specimens collected in 2012 were done by D. W. Roubik; specimens collected in 2013 were identified using reference specimens and Roubik & Hanson (2004). Voucher specimens were deposited in the Western Illinois University Entomology Collection and Oxford University Museum of Natural History (OUMNH).
The Chao excel sheet calculator (Ecological Archives E090-073-S1, Chao et al. 2009) was used to calculate the Chao1 asymptotic species richness estimate. The Morisita-Horn index of similarity, calculated using EstimateS (Colwell 2013), was used to compare species composition between years.
We collected 4,293 orchid bees representing 24 species during this study (Table 1). These results, combined with records in other lists and databases (Table 2), total 40 species of orchid bees recorded for Honduras. Euglossa imperialis Cockerell was the most abundant species collected in our study, comprising 67.6% of the total collection. The Chao1 estimate produced an estimated asymptotic richness of 24.67. The Morisita-Horn index of similarity in species composition between years was 0.987.
We collected 12 species that were not previously recorded for Honduras in Moure et al. (2012) (Table 1); one of these, Euglossa dilemma Bembé & Eltz, was a more recently recognized species. Eulaema marcii Nemésio is included by Moure et al. (2012) and Ramírez et al. (2002) in Eulaema cingulata F. However, E. cingulata is now considered South American, with E. marcii occurring in Central America and Mexico (D. W. Roubik, personal observation). Two of the species (Eulaema meriana [Olivier] and Euglossa townsendii Cockerell) not recorded as occurring in Honduras by Moure et al. (2012) are listed as occurring in Honduras by Ramírez et al. (2002). Based on distributions given in Moure et al. (2012), 3 of the species collected in our study represent range extensions. Two of these are northwestern extensions: Euglossa cybelia Moure, approximately 700 km from Puntarenas, Costa Rica, and Euglossa hansoni Moure, approximately 1,100 km from Colón, Panama. The third, Eufriesea pallida Kimsey, represents a nearly 400 km easterly range extension from Chiapas, Mexico. Of the 10 species not recorded as occurring in Honduras in Moure et al. (2012) or Ramírez et al. (2002), all but one (E. pallida) have specimens recorded from Honduras in Ascher & Pickering's (2015) Discover Life checklist, although many of the specimen records are incomplete. It should also be noted that the range of E. townsendi is uncertain and under study, but it has been verified in Panama (D. W. Roubik, personal observation).
Numbers, relative abundances, and locations of orchid bee species collected at Cusuco National Park, Honduras, during Jun to Aug 2012 and 2013.
Additional species of orchid bees recorded for Honduras that were not collected in our study, with literature sources.
There appears to be little published information on orchid bee species richness at higher elevation cloud forest locations. In Mesoamerica, Roubik & Ackerman (1987) identified 50 orchid bee species at a 900 m elevation cloud forest location at Cerro Campana, Panama, over the course of a 7 yr study. Ackerman & Roubik (2012) reviewed 30 yr of euglossine pollination data at this and 2 other sites, and the aforementioned Cerro Campana site had 11 species of Eufriesea, far more than lower elevation sites. Janzen et al. (1982) collected 27, 18, and 13 species of orchid bees at 3 locations in Costa Rica in 1977 to 1979, with sample sites ranging from near sea level to about 300 m elevation. Our species richness of 24 appears to be relatively low compared with other studies, where Chao1 estimates at 6 sites in Panama and the Amazon region range from 37 to 52 total species, and 27 to 44 species were actually sampled (Roubik 2004). The current species listed for Nicaragua include 32 euglossines (Hinojosa-Diaz & Engel 2012). Comparisons among studies are difficult because of sample size differences, bait presentation differences, and study lengths, along with variation in surrounding habitat quality, from which bees arrive during longer placement in the field.
Our species richness was near the Chao1 estimate of total species richness present of 24.67. Even so, our inventory is likely incomplete. Use of different attractants could result in new species collections, because many orchid bee species are attracted to specific chemical attractants (Roubik & Hanson 2004). In addition, it is possible that we failed to collect rare highland Mesoamerican species, or species such as Euglossa cyanura Cockerell, a Panamanian and possibly Costa Rican species that does not visit any known chemical baits (1 undescribed species occurs in Costa Rica; D. W. Roubik, personal observation). Also, because our study was limited to a 2 mo period each year, it is possible that some seasonal orchid bee species were not collected, although in general orchid bee populations tend to be fairly stable over the course of a year (Roubik & Ackermann 1987). Members of the genus Eufriesea are known to be highly seasonal, and for this reason are poorly represented in many collections (Roubik & Hanson 2004).
The total Eufriesea species in our Honduras study was less than half the species recorded in the highland site of Cerro Campana, in Panama, and parallels the roughly 50% total species richness of the Honduran compared to the Panamanian site. Our sampling was done during the rainy season, which is the period when most species of Eufriesea are active (Janzen et al. 1982). In Janzen et al.'s (1982) Costa Rica study, which included mid-wet season sampling, Jun to Aug sampling yielded 70.4% (19 of 27 species), 83.3% (15 of 18 species), and 46.2% (6 of 13 species) of the total species collected for each of the 3 locations. This suggests that wet season sampling, or any one season where there are at least 2 (wet and dry) seasons per year, may miss some of the total species present. However, their mid-wet season sample sizes for the 3 locations (360, 391, and 265) were much smaller than ours, and this difference could affect comparisons because species richness is associated with sample size. The Morisita-Horn index of similarity of 0.987 in our study indicates a high degree of similarity in species composition between years, and suggests that species composition may not vary significantly from year to year, although more data are needed to verify this result.
Orchid bees are ecologically important components of Neotropical ecosystems (Roubik & Hanson 2004). At least one governmental agency, the Brazilian Institute of Environment and Renewable Natural Resources (IBAMA), recognizes orchid bees as key organisms for environmental assessments (Nemésio & Vasconcelos 2014). Our study is the first intensive inventory of orchid bee species in western Honduran cloud forest. Further studies of the diversity, ecology, and conservation status of orchid bees are needed in this understudied and threatened location. Our findings expand knowledge of orchid bee diversity in this region, and provide baseline data for future monitoring of this group in relation to potential impacts of deforestation and climate change.
We thank Brandon Taylor, Fabian Wagner, Harriet Downey, Katie Williams, and numerous Operation Wallacea volunteers, staff, and local guides for invaluable assistance with orchid bee collecting. We also thank Sarah Wilcer (Western Illinois University) for help with orchid bee pinning and labeling, and Shane Mason (Western Illinois University) for making and contributing the maps for Fig. 1. We thank Operation Wallacea for invaluable financial and logistical support, and the Western Illinois University Office of Sponsored Projects and Research Foundation for generous funding for this study.