Connectance of Brazilian social bee: food plant networks is influenced by habitat, but not by latitude, altitude or network size

ARTICLES

Connectance of Brazilian social bee — food plant networks is influenced by habitat, but not by latitude, altitude or network size

Jacobus C. Biesmeijer^I,II,1; E. Judith Slaa^II, Marina Siqueira de Castro^III; Blandina Felipe Viana^IV; Astrid de M. P. Kleinert^V; Vera L. Imperatriz-Fonseca^V

^ICentre for Biodiversity and Conservation, School of Biology, University of Leeds, Leeds LS2 9JT, UK

^IINeurobiology and Behavior, Mudd Hall, Cornell University, Ithaca NY 14853, USA

^IIIDepartamento de Ciências Biológicas, Universidade Estadual de Feira de Santana, Av. Universitária, Feira de Santana, Bahia, Brazil

IVDepartamento de Zoologia, Instituto de Biologia, Universidade Federal da Bahia, Ondina, 41320-940, Salvador, Bahia, Brazil

^VDepartamentode Ecologia, Instituto de Biociências, Universidade de São Paulo, Rua do Matão, travessa 14, 321, 05508-900 São Paulo, São Paulo, Brazil

ABSTRACT

Several recent studies suggest that the level of generalization (measured as percentage connectance) of plant-pollinator networks has several ecological correlates, e.g. latitude and altitude. Here we report on levels of generalization in 27 two-mode networks of social bees and their food plants in various Brazilian habitats and urban environments. Social bees are generalist foragers and are among the most abundant flower visitors in Brazil. They probably account for 30-50% of all plant — flower visitor interactions. Connectance was significantly influenced by habitat. Cerrado forests showed lower connectance than the dry dune habitats, with Atlantic rain forest and urban sites taking intermediate position and arid Caatinga being similar to dunes. This shows that generalization in a plant — flower visitor community can be influenced by habitat even within a group of generalist flower visitors, in our case social bees. We show that the strength of the interactions is not different between Cerrado and semi-arid habitats (dunes and Caatinga) and discuss other explanations for our findings.

Key words: Pollination, stingless bees, Meliponini, Apis, Bombus, food plants.

RESUMO

Vários estudos recentes mostram que o nível de generalização (medido como porcentagem de conectância) das redes plantas-polinizadores apresentam várias correlações ecológicas, como latitude e longitude. Neste trabalho, apresentamos os níveis de generalização em 27 redes de dois-modos de abelhas sociais e das plantas que lhes servem de alimento em vários habitats e ambientes urbanos brasileiros. As abelhas sociais são forrageiras generalistas e estão entre os visitantes florais mais abundantes no Brasil. Provavelmente, são responsáveis por 30 a 50% de todas as interações entre plantas e visitantes florais. A conectância é significativamente influenciada pelo habitat. Os cerrados apresentaram uma menor conectância do que as dunas, a Mata Atlântica e os ambientes urbanos ocuparam uma posição intermediária, e a caatinga árida apresentou um grau de conectância semelhante às dunas. Isso mostra claramente que a generalização nas comunidades de visitantes florais e plantas pode ser influenciada pelo habitat, mesmo em um mesmo grupo de visitantes, como, no nosso caso, as abelhas sociais. Nós provamos que a força das interações não é diferente entre o cerrado e os habitats semi-áridos (dunas e caatinga) e discutimos outras explicações para nossas interpretações.

Palavras-chave: Polinização, abelhas sem ferrão, Meliponini, Apis, Bombus, recursos alimentares

Introduction

Plant-pollinator networks have been suggested to show trends towards higher levels of generalization with decreasing species richness, with increasing latitude and in lowlands compared to highlands (Olesen and Jordano 2002). Two earlier reviews, however, concluded that the frequency of occurrence of specialized pollination syndromes is about the same at different latitudes (Kevan and Baker 1983, Kevan and Baker 1999). In addition, generalization seems lower for some pollinator taxa (stingless bee, euglossine bees, and fig wasps) than for others (hummingbirds and bumblebees) (Jordano 1987). How these findings relate to each other is not directly clear, because some groups of pollinators (and food plants) are largely temperate whereas others are exclusively tropical. Also, data from lowland tropical pollination networks are largely lacking, which Olesen and Jordano (2002) regard a serious shortcoming for their review.

In this paper we take a slightly different approach that may extend our understanding of patterns of generalization in plant-pollinator networks. We do not aim at complete plant-pollinator networks, but restrict our analysis to a specific group of pollinators, social bees, in a restricted geographic area, Brazil. In this way our analyses are largely free of taxonomic biases, however, they do not address plant-pollinator networks in general.

Social bees of the subfamily Apinae (stingless bees, bumble bees, and honey bees) are the most abundant flower visitors in many tropical habitats (Heithaus, 1979, Roubik 1989, Bawa 1990). Their colony's life span surpasses the blooming period of most plant species, which partly explains their polylectic diet, i.e. use of multiple plant species as food source. They are generalist foragers (Roubik 1989) with wide niches and some extreme cases, e.g. honey bees, have been referred to as super generalists (e.g. Olesen et al. 2002). Therefore, although we analyze only a subset of all interactions, it comprises probably 30-50% of the complete plant-pollinator network (or more correctly flower-flower visitor network). Social bee food plant relationships have been studied extensively (reviewed in Roubik, 1989, Ramalho et al. 1990), but food web theory (e.g. Jordano 1987, Dunne et al. 2002) has rarely been used to study the characteristics of the two-mode network of social bees and their food plants. We analyzed 27 complete datasets of social bee-food plant interactions and examine (1) the patterns of generalization in these networks and (2) the influence of a series of ecological factors, altitude, latitude, habitat, on generalization levels.

2. Material and Methods

2.1 Dataset selection.

The datasets were selected according to the following criteria:

1) They had to be taxonomically well resolved. The two leading Brazilian social bee experts, Profs Padre Santiago Moure and João Camargo, identified bees in all studies and plants were identified by various leading botanical taxonomists.

2) Exhaustive sampling of a restricted area. All studies collected at least once a month for a year, but most sampled more intensively. Areas ranged from 0.3 ha to more than 200 ha, and was <100 ha in all but three studies.

3) Unrestricted sampling of bees and flowering plants. Note, however, than many studies under-sampled the higher strata of the forest.

4) Comparable sampling methods. Methodology was highly uniform among the studies. All used the Sakagami et al. (1967) sampling method or a variant of that method. In this method a fixed amount of time is spent at each flowering plant (or patch) and the coverage of transects is randomized in time, order, and direction.

5) Availability of interaction tables or raw data sheets. We compiled each data matrix from the interaction lists or raw data tables provided in the source. More than a dozen additional studies had to be excluded because they lacked sufficiently detailed information of bee-plant interactions in the published material. A complete list of the main features of all studies included is given in table 1. The studies that could not be included and the reason for their exclusion are given in table 2. Many datasets have been published only in student PhD- or MSc-theses about half of which were supervised by the latter two authors, but all are publicly available from the respective universities (for full references of all studies see ^{Appendix 1}).

2.2 Network properties

We followed methods of Jordano (1987) and Olesen and Jordano (2002). A two-way social bee — food plant network is defined as a matrix describing interactions between B flower-visiting social bees and P flowering plant species. Each cell of the matrix represents a single bee-plant interaction and can have a value of 0 if not observed or 1 if observed. Many studies gave actual numbers of bees collected from plants. All values >0 were set equal to 1 for connectance calculations, but real values were used in calculating interaction strengths (see below). Network size (M) is given by M = BP, where B and P are the number of interacting bees and plants in the habitat, respectively. M indicates the maximum possible number of interactions. As measures of generalization we use the number of observed interactions (I) and the connectance (C), where C = 100 I/M. Connectance is thus the percentage of all possible interactions actually observed within a network.

We used correlation to analyze relations among network properties (B+P, I, C) and multiple regressions to analyze the relation between C and several study site features: latitude, altitude (mean of upper and lower limit for larger areas), and M. Four natural habitat types were represented in multiple studies (Table 1): Atlantic rain forest, Cerrado forest (dense, dry forest of medium height), Caatinga (arid woodland), and Dunes. In addition, urban environments (parks and university campus) were sampled in five studies. The habitats of the remaining studies comprised fields, shrubland, orchard, a mix of habitats, and highland Araucaria forest. We analyzed the relation between the common habitats (Rain forest, Cerrado, Caatinga, Dunes, Urban) and connectance by means of a Kruskal-Wallis test with posthoc multiple comparisons tests (MCT) (Siegel and Castellan, 1988).

In our effort to find an explanation for the difference in connectance among habitats, we calculated interaction strengths for the Cerrado studies (Table 1: studies 15, 16, 17, and 19) and pooled dry habitat studies, Caatinga (Table 1: study 14) and Dunes (Table 1: studies 25 and 26), that provided numbers of bees visiting each flowering plant. We follow Jordano (1987) and use dependence of bees on food plants as a measure of interaction strength. Dependence here is the proportion of all interactions of the bee occurring between a particular plant and this bee. Note that this measure does not indicate the quality of the interaction and for the social bees we consider here high values may occasionally present a single or few successful group foraging events.

3. Results

Twenty-seven datasets were included in the analysis (Table 1), whereas many other studies could not be included (Table 2). The included studies covered 72 social bee taxa (66 stingless bees, 5 bumble bees, and the introduced African honey bee), 1725 plant species and a total of 6588 bee-plant interactions. Individual datasets contained 6 to 27 bee species (mean 13.5, median 12; table 1) and sampled 20 to 230 plant species (mean 92.9, median 93). The number of interactions ranged from 34 to 932 (mean 235.3, median 174) and increased linearly with network size (I = 0.16 M + 13.6; F = 114.5, p < 0.001, R² = 0.81). Connectance (C) ranged from 7.2% to 37.1% (mean 18.9%, median 17.6%) and was independent of the total number of species (B + P; F = 0.64, df=1,25, p = 0.43) in the dataset. A multiple regression analysis revealed that C was not significantly influenced by network size (M), latitude or altitude of the study site (F = 0.66, df=3,19, p = 0.59). However, C differed among the four natural habitats and the urban environments covered in the studies (Figure 1; KW-test: X² = 13.2, p = 0.01; 5 groups: Atlantic rain forest, Cerrado, Caatinga, Dunes, Urban). The MCT showed that Cerrado forest had significantly lower connectance than dunes, and that other combinations were not significantly different. If more Caatinga sites are sampled in the future it is possible that connectance in Caatinga will become different from Cerrado as well (Figure 1). Data were not sufficient to detect within-habitat trends between connectance and other parameters.

The distribution of dependence values of bee-plant links was not different between Cerrado forest and the semi-arid habitats, Caatinga and dunes (two-sample KS-test: D_max= 0.6, Z = 0.82, p = 0.41, N_cerrado= 745, N_{caatinga/dunes}= 236; Figure 2). This suggests (see Jordano 1987) that the lower C in Cerrado forest compared to that of dry habitats Caatinga and Dunes does not reflect differences in the strength of the interactions. In both cases weak interactions (dependence <0.1) are vastly more numerous than strong interactions (dependence>0.1).

4. Discussion

The average level of generalization (C = 19%) in the social bee food plant networks was higher than in complete plant-pollinator networks (11%; Olesen and Jordano 2002). This was expected because all social bees are generalists and they have broader food niches than do oligolectic bees and other non-social flower visitors. Generalization level was lower than in incomplete social bee - plant networks reported by Jordano (1987; average of 14 studies: C = 28.1%), but higher than in the tropical social bee studies he reviewed (Engel and Dingemans-Bakels 1980: C = 8.7 and 15.7%), probably reflecting the different methods used in that study.

Neither latitude (4ºN-30ºS) nor altitude (0-1050m) affected the generalization level. This corresponds to findings by Olesen and Jordano (2002) who show a general effect of both factors, but no clear relationships within the ranges included in our analysis (see their Fig. 2). The lack of a relationship between network size and C in our study is in contrast to both Olesen and Jordano (2002) and Jordano (1987). This may result from the taxonomic heterogeneity of the datasets analyzed by them compared to our datasets that are restricted to tropical social bees. More importantly, habitat significantly influenced the level of generalization in our analysis. The Cerrado forest habitat had lowest C levels at intermediate levels of network size and number of species (Fig. 1) and variance of C among habitats was high at low and intermediate network sizes. Olesen and Jordano (2002) report marginally different levels of generalizations among biogeographic regions. They report lowest connectance (residual values after controlling for species richness) for tropical and alpine networks and highest connectance for Mediterranean networks. Here we refine that result and show that among tropical habitats levels of generalization varies significantly. Interestingly, in both studies networks in the driest region, the Mediterranean (Olesen and Jordano 2002) and Dunes and Caatinga (our study), had highest levels of connectance. Cerrado forest is, however, much drier than rain forest and urban habitats but had lower connectance.

Why generalization levels in Cerrado forest are greatly lower than in the arid Dunes and Caatinga is not clear. First, whereas three of four of the arid studies have low species richness, overall connectance is not dependent on species richness. Second, the strength of the interactions of bees with their food plants, measured as dependence, is not different among these habitats (Fig. 2) suggesting that the bees' diets are not generally broader in the arid habitats. Third, the species composition is highly homogeneous among the habitats, the same genera of stingless bees occur in most habitats and several species were reported in all habitats. A more detailed analysis of the interactions and the floristic and faunistic composition may reveal why these habitats have intrinsically different levels of generalization.

In summary, our results indicate that 1) generalization level of a social bee — food plant network is affected by habitat; 2) Social bee foragers are not always extreme generalists; and 3) study of a homogeneous compartment of a network may reveal other patterns than study of a more heterogeneous 'complete' network. This analysis of a unique dataset of tropical bee-food plant networks adds to our incomplete understanding of lowland tropical plant (potential) pollinator networks.

Information on properties of social bee food plant networks is also fundamental in conservation and management of habitat for native pollinators (e.g. various contributions in Kevan and Imperatriz-Fonseca 2002, Kearns et al. 1998). The level of connectance of interactions in a food web may influence how the system changes following disturbance. In addition, complete information on the interactions between plants and pollinators may allow for more efficient management of biodiversity (Waser et al 1996, Kearns et al. 1998, Pinheiro-Machado et al. 2002), e.g. concentration on specific groups of food plants to conserve a target pollinator species or vice versa concentration on specific pollinators to improve pollination in threatened native plant species. Finally, the construction and in-depth analysis of networks of pollinators and their food plants would allow for a detailed analysis of the influence of non-native species of plants and pollinators on the relation among native species (e.g. Olesen et al. 2002, Memmott & Waser 2002). It is known that introduction of super-generalist pollinators, such as the honey bee Apis mellifera, may adversely affect some native plant-pollinator interactions (e.g. Butz-Huryn 1997, Gross & Mackay 1998), because they readily incorporate native food plants in their diet. On the other hand, invasive plants that provide ample reward to pollinators, e.g. Impatiens glandulifera, may disturb the pollination market and attract native pollinators away from the native plants they used to visit (Chittka & Schurkens 2001). In conclusion, we feel that food web theory is a valuable tool in analyzing patterns and monitoring change in ecological communities.

5. Acknowledgements

We thank Isabel Alves dos Santos and two anonymous referees for helpful suggestions for improvement. FAPESP provided financial support for this study for which we are very grateful. JCB and EJS like to thank the members of the Bee laboratory of USP for their hospitality during their stay.

6. References

Date Received 11/8/2004

Accepted 01/02/2005

ISSN 1676-0611

tables 1 and 2.

1. Aguiar, A.J.C. (2001) Comunidade de abelhas (Hymenoptera, Apoidea Apiformes) da vegetação de Tabuleiro da Reserva Biológica Guaribas (Mamanguape, Paraíba). Msc thesis Ciências Biológicas, Universidade Federal da Paraíba, Brazil.

2. Alves dos Santos, I. (1996) Bienen und melittophile Blütenpflanzen in der Küstenregion und im Atlantischen Regenwald von Rio Grande do Sul (Brasilien), mit einer Fallstudie zu langzungenbienen und tristylen Wasserhyazinthen. PhD thesis, Fakultät für Biologie der Eberhard-Karls-Universität Tübingen, Germany. 149p.

3. Barbola, I.F. (1993) A comunidade de Apoidea (Hymenoptera) da reserva Passa Dois (Lapa, Paraná): Diversidade, fenologia e relações tróficas. Msc thesis, Ciências Biológicas, da Universidade Federal do Paraná, Brazil. 103p.

4. Campos, M.J.O. (1989) Estudo das interações entre a comunidade de Apoidea, na procura de recursos alimentares, e a vegetação de cerrado da reserva de Corumbataí, S.P. PhD thesis. Departamento de Ciências Biológicas da Universidade Federal de São Carlos, S.P., Brazil. 114p.

5. Carvalho, A.M.C. (1990) Estudo das interações entre a apifauna e a flora apícola em vegetacão de cerrado Reserva Ecológica do Panga Uberlândia MG. Msc thesis Faculdade de Filosofia, Ciências e Letras de Ribeirão Preto - Universidade de São Paulo, S.P., Brazil.

6. Castro, M.S. de (2001) A comunidade de abelhas (Hymenoptera; Apoidae) de uma área de caatinga arbórea entre os inselbergs de Milagres (12º53'S; 39º51'W), Bahia. PhD thesis Instituto de Biociências da Universidade de São Paulo, Brazil.

7. Unpublished. Data available from MS de C

8. Unpublished. Data available from MS de C

9. Cortopassi-Laurino, M. (1982) Divisão de recursos tróficos entre abelhas cociais principalmente em Apismellifera Linné e Trigona (Trigona) spinipes Fabricius (Apinae, Hymenoptera). PhD thesis, Instituto de Biociências da Universidade de São Paulo, Brazil.

10. Faria, G.M. (1994) A flora e a fauna apícola de um ecossistema de campo rupestre, Serra do Cipó — MG, Brasil: Composição, fenologia e suas interações. PhD thesis, Instituto de Biociências da Universidade de São Paulo, Brazil. 239p.

11. Harter, B. (1999) Bienen und ihre Trachtpflanzen im Araukarien-Hochland von Rio Grande do Sul, mit Fallstudien zur bestäubung von Pionierpflanzen. PhD thesis, Fakultät für Biologie de Eberhard-Karls-Universität Tübingen.185p.

12. Hoffmann, M. (1990) Estrutura e importância de uma comunidade de abelhas (Hymenoptera: Apoidae) no Rio Grande do Sul, para a polinização de plantas cultivadas. PhD thesis Ciências Biológicas, Universidade Federal do Paraná, Brazil. 117p.

13. Knoll, F.R.N. (1990) Abundância relativa, sazonalidade e preferências florais de Apidae) em uma área urbana (23º33'D; 46º43'W). PhD thesis Instituto de Biociências da Universidade de São Paulo, Brazil. 127p.

14. Martins, C.F. (1990) Estrutura da comunidade de abelhas (Hymn., Apoidea) na caatinga (Casa nova, BA) e na Chapada Diamantina (Lençóis, BA). PhD thesis Instituto de Biociências da Universidade de São Paulo, Brazil. 159p.

15. Same as 14.

16. Mateus, S. (1998) Abundância relativa, fenologia e visita às flores pelos Apoidea do cerrado da Estação Ecológica de Jataí - Luiz Antônio SP. Msc thesis Faculdade de Filosofia, Ciências e Letras de Ribeirão Preto - Universidade de São Paulo, S.P., Brazil. 159p.

17. Pedro, S.R.M. de (1992) Sobre as abelhas (Hymenoptera, Apoidea) em um ecossistema de cerrado (Cujuru, NE do Estado de São Paulo): composição, fenologia e visita às flores. Msc thesis Faculdade de Filosofia, Ciências e Letras de Ribeirão Preto - Universidade de São Paulo, S.P., Brazil. 200p.

18. Ramalho, M. (1995) Diversidade de abelhas (Apoidea, Hymenoptera) em um remanescente de floresta atlântica em São Paulo. PhD thesis Instituto de Biociências da Universidade de São Paulo, Brazil. 148p.

19. Rêgo, M.M.C. (1998) Abelhas silvestres (Hymn. Apoidea) em um ecossistema de cerrado s.l. (Chapadinha MA, Brasil): uma abordagem biocenótica. PhD thesis Faculdade de Filosofia, Ciências e Letras de Ribeirão Preto - Universidade de São Paulo, S.P., Brazil. 203p.

20. Schlindwein, C. (1995) Wildbienen und ihre Trachtpflanzen in einer südbrasilianischen Buschlandschaft: Fallstudie Guaritas, Bestäubung bei Kakteen und Loasaceen. PhD thesis, Fakultät für Biologie der Eberhard-Karls-Universität Tübingen, Germany. 148p.

21. Silveira, F.A. da (1989) Abelhas silvestres (Hymenoptera: Apoidea) e suas fontes de alimento no cerrado da estação florestal de experimentação de Paraopeba — Minas Gerais. Msc thesis da Universidade Federal de Viçosa, MG, Brazil.

22. Sofia, S.H. (1996) As abelhas e suas visitas às flores em duas áreas urbanas. PhD thesis Instituto de Biociências do campus de Rio Claro, Universidade Estadual Paulista, Brazil. 236p.

23. Same as 22.

24. Taura, H.M. (1990) A comunidade de abelhas silvestres (Hymenoptera, Apoidea) do Passeio Público, Curitiba, Paraná, Sul do Brasil: Uma abordagem comparativa. Msc thesis, Ciências Biológicas, da Universidade Federal do Paraná, Brazil. 145p.

25. Viana, B.F. (1999) Biodiversidade da apifauna e flora apícola das dunas litorâneas da APA das Lagoas e Dunas de Abaeté, Salvador, Bahia — Composição, fenologia e suas interações. PhD thesis Instituto de Biociências da Universidade de São Paulo, Brazil. 171p.

26. Unpublished. Data available from BFV.

27. Wilms, W. (1995) Die Bienenfauna im Küstenregenwald Brasiliens und ihre Beziehungen zu Blütenpflanzen: Fallstudie Boracéia, São Paulo. PhD thesis, Fakultät für Biologie der Eberhard-Karls-Universität Tübingen, Germany. 219p.

28. Aguilar, J.B.V. (1998) A comunidade de abelhas (Hymenoptera: Apoidae) da reserva florestal de Morro Grande, Cotia, São Paulo. PhD thesis Instituto de Biociências da Universidade de São Paulo, Brazil.

29. Kaminski, A.C. (2001) Abelhas visitants de três espécies de Clusia (Clusiaceae) para coleta de resinas florais, com ênfase en Ptilotrigona lurida (Smith 1854) (Hymenoptera, Apidae, Meliponinae), em uma floresta de terra firme na Amazônia Central. Msc thesis Biologia Tropical e Recursos Naturais do convênio INPA/UA

30. Laroca, S. (1974) Estudo feno-ecológico em Apoidea do litoral e primeiro planalto Paranaenses. Msc thesis, Instituto de Biologia, da Universidade Federal do Paraná, Brazil. 64p.

31. Orth, A.I. (1983) Estudo ecológico de abelhas silvestres (Hymenoptera, Apoidea) em Caçador, SC, com enfase em polinizadores potenciais da Macieira Pyrus malus (L.) (Rosaceae). Msc thesis, Ciências Biológicas, da Universidade Federal do Paraná, Brazil. 122p.

32. Machado, C.P. (2002) Diversidade e conservação de Apoidae. PhD thesis Instituto de Biociências da Universidade de Sao Pãulo, Brazil.

33. Zanella, F.C.V. (1991) Estrutura da comunidade de abelhas silvestres (Hymenoptera, Apoidea) da ilha do mel, planicie litorânea paranaense, Sul do Brasil, com notas comparativas. Msc thesis, Ciências Biológicas, da Universidade Federal do Paraná, Brazil. 88p.

34. Schwartz Filho, D. (1993) A comunidade de abelhas silvestres (Hymenoptera, Apoidea) da ilha das cobras (Parana, Brasil): aspectos ecológicos e biogeográficos. Msc thesis, Ciências Biológicas, da Universidade Federal do Paraná, Brazil. 83p.

35. Hakim, J.R.C. (1983) Estudo ecológico da comunidade de abelhas silvestres (Hymenoptera, Apoidea) do Parque da cidade, comparado ao de outras areas de curitiba, Paraná. Msc thesis, Ciências Biológicas, da Universidade Federal do Paraná, Brazil. 86p.

36. Albuquerque, P.M.C. (1998) Abelhas silvestres (Hymenoptera, Apoidea) e suas fonts de alimento em um ecossistema de dunas, na ilha do Maranhão, MA, Brasil: composição, fenelogia e interações. Phd thesis USP Riberão Preto, SP, Brazil. 208p.

37. Knoll, F.R.N. (1985) Abundância relativa das abelhas no campus da Universidade de São Paulo (23º33'D; 46º43'W),com especial referência a Tetragonisca angustula Latreille. Msc thesis Instituto de Biociências da Universidade de Sao Pãulo, Brazil. 78p.

38. Antonini, Y. (2002) Efeitos de variáveis ecológicas na ocorrência de Melipona quadrifasciata (Apidae: Meliponini) em fragmentos urbanos e rurais. Phd thesis Ecologia Conservação Manejo de Vida Silvestre, da Universidade Federal de Minas Gerais.

39. Castro, M.S. de (1994) Composição, fenelogia e visita às flores pelas espécies de Apidae em um ecossistema de caatinga (Nova Casa Nova — Bahia 9º26'S / 41º50'W). Msc thesis Instituto de Biociências da Universidade de Sao Pãulo, Brazil. 103p.

40. Viana, B.F. (1992) Estudo da composição da fauna de Apidae e da flora da Chapada Diamantina, Lençóis — Bahia (12º34'S/41º23'W). Msc thesis Instituto de Biociências da Universidade de Sao Pãulo, Brazil. 140p.

41. Lorenzon, M.C.A., Matrangolo, C.A.R. & Schroereder, J.H. (2004) Flora visitada pelas abelhas eussociais (Hymenoptera, Apidae) na Serra da Capivara, em Caatinga do Sul do Piauí. Neotropical Entomology 32: 27-35.

• Bawa, K.S. 1990. Plant-pollinator interactions in tropical rain forests. Ann. Rev. Ecol. System. 21:399-422.
• Butz-Huryn, V.N. 1997. Ecological impacts of introduced honeybees. Q. Rev. Biol. 72:275-297.
• Chittka, L. & Schurkens, S. 2001. Successful invasion of a floral market. Nature 411:653.
• Dunne, J.A., Williams, R.J. & Martinez, N.D. 2002. Network structure and biodiversity loss in food webs: robustness increases with connectance. Ecol. Lett. 5:558-567.
• Engel, M.S. & Dingemans-Bakels, F. 1980. Nectar and pollen sources for stingless bees (Meliponinae, Hymenoptera) in Surinam. Apidologie 11:341-350.
• Gross, C.L. & Mackay, D. 1998. Honeybees reduce fitness in the pioneer shrub Melastoma affine (Melastomataceae). Biol. Cons. 86:169-178.
• Heithaus, E.R. 1979. Community structure of neotropical flower visiting bess and wasps: diversity and phenology. Ecology 60:190-202.
• Jordano, P. 1987. Patterns of mutualistic interactions in pollination and seed dispersal: connectance, dependence, and coevolution. Amer. Nat. 129:657-677.
• Kearns, C.A., Inouye, D.W. & Waser, N.M. 1998. Endangered mutualisms: the conservation of plant-pollinator interactions. Ann. Rev. Ecol. Syst. 29:83-112.
• Kevan, P.G. & Baker, H.G. 1983. Insects as flower visitors and pollinators. Ann. Rev. Entomology 28:407-453.
• Kevan, P.G. & Baker, H.G. 1999. Insects and flowers. In C.B. Huffaker & A.P. Guttierrez (eds.) Ecological Entomology. 2^nd edition. New York, Wiley and sons. pp. 553-584.
• Kevan, P.G. & Imperatriz-Fonseca, V.L. (eds.). 2002. Pollinating bees The conservation link between agriculture and nature. Brasilia: Ministry of Environment.
• Memmott, J. & Waser.M. 2002. Integration of alien plants into a native flower-pollinator visitation web. Proc. Roy. Soc. Lond. B. 269:2395-2399.
• Olesen, J.M. & Jordano, P. 2002. Geographic patterns in plant-pollinator mutualistic networks. Ecology 83:2416-2424.
• Olesen, J.M., Eskildsen, L.I. & Venkatasamy, S. 2002. Invasion of pollination networks on oceanic islands: importance of invader complexes and endemic super generalists. Diversity and Distributions 8:181-192.
• Pinheiro-Machado, C., Alves dos Santos, I., Imperatriz-Fonseca, V.L., Kleinert, A.M.P. & Silveira, F.A. 2002. In P.G. Kevan & V.L. Imperatriz-Fonseca (eds.) Pollinating bees The conservation link between agriculture and nature. Brasilia: Ministry of Environment. pp115-130.
• Ramalho, M.A., Kleinert-Giovannini, A. & Imperatriz-Fonseca, V.L. 1990. Important bee plants for stingless bees (Melipona and Trigonini) and Africanized honeybees (Apis mellifera) in neotropical habitats: a review. Apidologie 21:469-488.
• Roubik, D.W. 1989. The ecology and natural history of tropical bees. Cambridge University Press, Cambridge Massachusetts.
• Sakagami, S.F., Laroca, S. & Moure, J.S. 1967. Wild bee biocenotics in São José dos Pinhais (PR), south Brasil. Preliminary report. J. Fac. Sci. Hokkaido Univ., Ser. VI, Zool., 16:253-291.
• Siegel, S. & Castellan, N.J. 1988. Nonparametric statistics for the behavioral sciences. MacGraw Hill Singapore.
• Waser, N.M., Chittka, L., Price, M.V., Williams, N.M. & Ollerton, J. 1996. Generalization in pollination systems, and why it matters. Ecology 77:1043-1060.

Appendix 1:  References of all studies included in and excluded from our analysis. Numbers correspond to those given in

Publication Dates

• Publication in this collection
  23 Oct 2006
• Date of issue
  2005

History

• Received
  11 Aug 2004
• Accepted
  01 Feb 2005