Africanized honey bees pollinate and preempt the pollen of Spondias mombin (Anacardiaceae) flowers

Please download to get full document.

View again

of 13
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Document Description
Africanized honey bees pollinate and preempt the pollen of Spondias mombin (Anacardiaceae) flowers
Document Share
Document Tags
Document Transcript
  Africanized honey bees pollinate and preempt the pollenof   Spondias mombin  (Anacardiaceae) flowers Liedson Tavares C ARNEIRO 1 ,  Celso Feitosa M ARTINS 2 1 Programa de Pós-Graduação em Ciências Biológicas (Zoologia), Universidade Federal da Paraíba-UFPB,58051-900, João Pessoa, PB, Brazil 2 Departamento de Sistemática e Ecologia, Universidade Federal da Paraíba-UFPB, 58051-900, João Pessoa, PB,BrazilReceived 5 September 2011  –   Revised 3 December 2011  –   Accepted 25 December 2011 Abstract  –   The invasion of generalist Africanized honey bees may change certain plant   –   pollinator interactions. We evaluated the preemption by honey bees and the exploitative competition with native beeson a tree with nocturnally dehiscent small flowers. Our main objectives were to quantify pollen production andharvesting, to verify whether honey bees exploitatively compete with native bees and to identify the effective pollinators of   Spondias mombin . The nocturnally dehiscent flowers were pollen depleted by honey bees andattracted various nocturnal and diurnal bee species. A threefold increase in native bee abundance was produced by delaying pollen availability and by preventing the preemption of pollen by honey bees. The results suggest that honey bees reduce the foraging benefit of late-arriving native bees. Honey bees and  Scaptotrigona aff.tubiba  were regarded as the main effective pollinators of   S. mombin  due to their abundance, behavior, andability to visit a large number of flowers.  Apis mellifera  / pollen depletion / stingless bees / nocturnal bees / pollination 1. INTRODUCTION Trees with small flowers produce a largequantity of floral resources clustered in richinflorescences that are exploited by highlyeusocial bees in tropical rainforests (Ramalho2004). These bees have a tight association withthese plants that provide a large amount of thefood necessary to maintain their perennialcolonies. In the Atlantic Neotropical forest,these resources are harvested mainly by sting-less bees and only rarely by solitary bees(Ramalho 2004). However, after the Africanhoney bees,  Apis mellifera scutellata  Lepeletier,invaded Neotropical ecosystems, a high overlapin food plant use with certain stingless bees wasindicated (Sommeijer et al. 1983; Roubik  1989; Biesmeijer and Slaa  2006). It has also beenobserved that invasion of generalist Africanizedhoney bees can affect specific plant   –   pollinator interactions (Roubik  1996; Butz-Huryn 1997; Menezes et al. 2007). Most authors agree that honey bees are not particularly aggressive whileforaging and the impact occur primarily throughexploitative competition (Goulson 2003; Roubik 2009; Roubik and Villanueva-Gutiérrez 2009). The foraging of Africanized honey bees andstingless bees occur mainly on flowers with non-specialized morphological features (i.e., radialsymmetry, open corolla, and exposed fertilestructures) (Roubik  1989), which are commonin trees with small flowers.Anacardiaceae is a plant family which com- prises several taxa with the floral traits men- Corresponding author: C.F. Martins, Manuscript editor: James NiehApidologie (2012) 43:474  –  486  Original article *  INRA, DIB and Springer-Verlag, France, 2012DOI: 10.1007/s13592-011-0116-7  tioned above (Mitchell 1997). The pollenresources produced by this group of plant isamong the highly used by Africanized honey bees (Roubik  1988; Roubik and Villanueva-Gutiérrez 2009).  Spondias mombin  L. is one of the Anacardiaceae species widespread in andnative to the lowland tropical rainforests of Mexico, Central America and South America.In Brazil,  S. mombin , one of the few forest treespecies used as a fruit crop with a potential for expansion, occurs in the north and northeast,where its fruits are known as  “ taperebá  ”  and “ cajá, ”  respectively. This species producesmany small flowers in large inflorescences,and bees in general, which forage primarily for  pollen, are regarded as their pollinators (Roubik 1995). Populations of   S. mombin  display sea-sonal and synchronous patterns in flowering phenology (Adler and Kielpinski 2000) that can be more efficiently exploited by bees. There isno information concerning the pollination biol-ogy of   S. mombin , and only superficial obser-vations on the matter have been made (Lozano1986; Stacy et al. 1996; Nason and Hamrick  1997).We investigated the pollination biology and breeding system of   S. mombin  within anexperimental area of the tropical Atlantic rain-forest in northeastern Brazil. We asked thefollowing questions: (1) How much pollen doandrogynous and staminate  S. mombin  flowers produce? (2) How much pollen is harvested byhoney bees? (3) Do Africanized honey beesexploitatively compete with native bees? and (4)What species are the flower visitors andeffective pollinators of   S. mombin ? 2. MATERIALS AND METHODS2.1. Studied species S. mombin  displays a diversified sexual expressionthroughout its occurrence in the Neotropical region.Populations of   S. mombin  are dioecious in Mexico(Pennington and Sarukhan 2005) and monoecious inCosta Rica (Bawa and Opler  1975). In Panamá, Croat (1978) reported generally androgynous flowers (withfew pistillates). However, in Colombia, Lozano(1986) observed four types of flowers: androgynous,staminate, and two variations of pistillate, one withanthers lacking pollen grains and the other withdeformed and sterile pollen grains. The many flower types and potential breeding systems in  S. mombin suggest that there are hybrids or multiple species anddeserve thorough studies. Nevertheless, we verified that our studied speci-mens were andromonoecious. The staminate flowers bore a pistillodium, and the androgynous flowersdisplayed conspicuous styles and stigmata (Figure 1).Both flower types were pentamerous and actino-morphic, with a white corolla and ten stamens bearing dorsifixed anthers with longitudinal dehis-cence. The nectarless flowers opened at night, a fewhours before dawn, and lasted 2 days. 2.2. Study area This study was carried out at the experimentalresearch station of Mangabeira (07°11 ′ 52 ″  N; 34°48 ′ 42 ″  W), a property of the Empresa Estadual dePesquisa Agropecuária da Paraíba (EMEPA-PB),João Pessoa, PB, Brazil. We conducted our observa-tions from 2007 to 2010 during the blooming season(Sep  –  Dec). The experimental station covers 260 ha close to a conserved fragment of tropical rainforest.The following crops are cultivated in the area:  Anacardium occidentale  L.,  Bixa orellana  L.,  Euge-nia uniflora  L.,  Hancornia speciosa  Gomes,  Heveabrasiliensis  (Willd. ex A. Juss.) Müll. Arg.,  Mal- pighia emarginata  DC.,  Mangifera indica  L.,  Psi-dium guajava  L.,  Spondias cytherea  Sonn., and  S.mombin  L.  S. mombin  is cultivated at three sites inthe experimental station and composed by a germo- plasm bank with seedlings and different clones propagated by grafting and cuttings, totaling 128 plants.The climate of the region is of the As ’  type (warmand wet), with a relative humidity of approximately80% and annual temperature of 22  –  26°C (Köppen).The mean annual rainfall ranges from 1,500 to1,700 mm (Feliciano and Melo 2003). 2.3. Flower and pollen production The number of flowers produced per inflorescencewas counted; open flowers, buds and branchesAfricanized honey bees and  Spondias mombin  475  constituted by androgynous and staminate flowerswere included in this count. Morphological andmorphometric descriptions of the floral types weremade using fresh and fixed materials in 70% alcoholunder a stereomicroscope (MZ12 Leica®).The number of pollen grains produced by theandrogynous and staminate flowers was determined by a direct count. Five anthers of each floral bud wereremoved and placed in a glycerin drop on a millimetric slide. The pollen grains were removedand counted under a stereomicroscope. The number of pollen grains produced per flower type wasdetermined by multiplying the mean number of  pollen grains per anther by ten. By multiplying thenumber of pollen grains per flower by the number of flowers per inflorescence, we estimated the number of pollen grains per whole inflorescence. 2.4. Breeding system The breeding system was determined throughcross- and open pollination treatments. The inflor-escences used in the cross-pollination treatments,including both pollen donors and receivers were bagged before anthesis to avoid pollinator contact and a loss of pollen grains. The crosses were performed between clone plants and between non-clone plants. The crosses between clones were madein 12 inflorescences and four plants (a total of 1,021hand-pollinated flowers), and the crosses betweennon-clones were made in six inflorescences and six plants (1,022 flowers). In open pollination treatments,11 marked inflorescences (4,580 flowers) distributedin four plants remained exposed to flower visitors.The numbers of initial (immature) fruits and maturefruits were counted 2 days and approximately90 days, respectively, after the end of anthesis. Thenumbers of immature and mature fruits per inflores-cence were used to calculate the fruit survival rate.To determine the requirements of pollinators, wind pollination treatments were performed. Eighteeninflorescences were bagged at pre-anthesis via tullecloth with openings of about one square millimeter insize. Flower visitors thus had no access to theflowers; however, there remained the possibility of wind action. The mature fruits were counted and thefruit set rate was estimated using the average number of androgynous flowers per inflorescence. 2.5. Flower visitors The number of flower visitors was recorded duringthe flowering seasons of 2007, 2008 and 2009. Atotal of 25 inflorescences distributed in 10 days andrandomly chosen in different plants were observedthroughout the first day of anthesis, starting before Figure 1.  The flowers of   S. mombin  from a tropical Atlantic rainforest area in João Pessoa, Paraíba, Brazil.  a Androgynous flower.  b  Staminate flower. 476  L.T. Carneiro and C.F. Martins  dawn at 0300 hours. The richness of the flower visitors and the frequency of each bee species wererecorded continuously from 0400 to 1700 hours daily.Voucher specimens were deposited in the Entomo-logical Collection of the Departamento de Sistemática e Ecologia, Universidade Federal da Paraíba. 2.6. Honey bee foraging and pollenharvesting The number of individuals immediately (~10 s)visiting a single inflorescence and the number of flowers visited per minute by honey bee workers at the peak of visitation were recorded. Honey bees withhuge pollen loads were captured and their pollenloads were removed to determine the number of  pollen grains collected from flowers at the peak of activity. In addition, fidelity to  S. mombin  flowerswas verified through pollen analysis of the honey bee pollen loads under a microscope. The pollen loadswere placed within a 0.5-mL glycerin lactic acidsolution (3:1) in a 1.5-mL Eppendorf tube. Thesolution was sonicated in a vortex stirrer for 5 min,and an aliquot was counted in a Neubauer chamber.The availability of pollen grains was determined before, during and at the end of honey bee foraging at four 30-min intervals: 0400  –  0430, 0430  –  0500, 0500  –  0530, and 0530  –  0600 hours. In each interval, the pollen grains of two flowers sampled in six different days from six plants (12 flowers per interval) weredirectly counted on millimetric slides under a stereomicroscope.To estimate how many honey bee pollen loads aninflorescence could provide, the following equationwas used:  N  PL  ¼  FI PG 0400    PG 0530 ð Þ PL  ; where FI is the mean number of flowers per inflorescence, PG 0400  is the mean number of  pollen grains that remained in flowers at 0400hours (when there had been no honey beevisits), PG 0530  is the mean number of pollengrains that remained in flowers at 0530 hours(after honey bees ’  visits), and PL is the meannumber of pollen grains gathered in the honey bee pollen loads. 2.7. Late pollen availability To verify the effect of the premature foraging patternof honey bees on native bee foraging, the abundanceand richness of flower visitors on  S. mombin  inflor-escences were analyzed under two simultaneoustemporal treatments of pollen availability in 2010:natural pollen availability (NPA; control) and late pollen availability (LPA). In the NPA treatment, theflower visitors were recorded in seven plants and5 days between 0300 and 0900 hours, on a total of 16inflorescences, sampling the natural visitation rate. InLPA, the flower visitors were recorded in six plantsand 5 days, on a total of 17 inflorescences bagged at  pre-anthesis with tulle cloth to avoid flower visitors inthe first hours of anthesis and delaying resourceavailability. The inflorescences were unbagged at 0600 hours, and the flower visitors were recordeduntil 0900 hours. Some samples of both treatmentswere prepared on the same plant. In both treatments,the flower visitors were counted per minute, and thedata were grouped in 15-min intervals. 2.8. Statistical analysis To compare the number of androgynous andstaminate flowers produced per inflorescence, thenumber of pollen grains produced per androgynousand staminate flower and the size of their floral parts,the  t   test for independent samples was used. AShapiro  –  Wilk test was used to check data normality.A non-parametric Mann  –  Whitney test was used tocompare the non-normal data for pistil and pistillo-dium length. The mean number of pollen grains inthe different anthesis phases (bud, 0400, 0430, 0500,and 0530 hours) was compared using a one-wayanalysis of variance (ANOVA). The outliers wereremoved from the sampling for this analysis. Tocompare the abundance of honey bees, native beesand Diptera in relation to the temporal treatments of  pollen availability, the two-way ANOVA followed byTukey ’ s post hoc test was applied. Homogeneity of variance and residual normality was reached bytransforming the data into natural logarithms. Chi-square test ( χ ²) with Yates correction was used tocompare the initial fruit set, mature fruit set and fruit survival rate between all the three treatments (open pollination, non-clone crosses, and clone crosses).Africanized honey bees and  Spondias mombin  477  The number of fruits was used in the calculations.Statistical analyses were performed with STATIS-TICA 6.0 (Statsoft Inc.). 3. RESULTS3.1. Flower and pollen production S. mombin  produced androgynous and stami-nate flowers on the same inflorescence. Stami-nate flowers were generally distributed in the basal branches of the inflorescence; the androg-ynous flowers were distributed from the middleto the apex. Inflorescences presented a mean of 1,708 flowers (SD±587,  n =13), with 935±275(55%) androgynous and 773±426 (45%) stami-nate flowers. There was no significant differ-ence between the number of androgynous andstaminate flowers ( t  =1.154;  P  =0.262). In addi-tion, the size of the floral parts of the androg-ynous and staminate flowers did not differ significantly, but the pistils were significantlylonger than the pistillodia (Table I).Androgynous flowers produced an averageof 12,357 (±2016;  n =18) pollen grains, andstaminate flowers produced 12,234 (±1523;  n =18) (Figure 2). No significant difference wasfound ( t  =0.205; g.l.=34;  P  =0.838). Thus, a   S.mombin  inflorescence produced 21×10 6  pollengrains on average. 3.2. Breeding system S. mombin  is self-incompatible and depen-dent on pollinators to set its fruit. The initialand mature fruit set were significantly higher in the non-clone cross treatment (n) than inthe open pollination treatment (o) and in thesewere higher than in the clone cross-treatment (c) (Table II). However, the fruit survival ratewas significantly higher in o followed by n and by c.Only 0.02% of the mature fruit set wasobtained from wind pollination treatments. 3.3. Flower visitors During 3 years of observation (2007  –  2009), S. mombin  flowers attracted mainly bees, whichforaged for pollen as a floral resource. Amongthe flower visitors, 92.7% were Apoidea (bees),4.5% Diptera, 1.2% Formicidae, 0.9% Vespi-dae, and 0.8% Curculionidae. Seventeen beespecies were recorded during the entire study period, and the Africanized honey bee  A.mellifera  showed an extraordinarily high rela-tive frequency (71%) (Table III). Table I.  Number and dimension of the floral structures of androgynous and staminate flowers of   Spondiasmombin  from a tropical Atlantic rainforest area in João Pessoa, Paraíba, Brazil.Floral structures Androgynous flowers Staminate flowers Statistics Number Dimension(mm;  n =10) Number Dimension(mm;  n =10) t P  Calyx (sepals) 5 5Corolla (petals) a  5 4.1±0.24 5 4.0±0.15 1.326 0.205Stamens  b 10 3.3±0.14 10 3.2±0.11 1.054 0.306Anthers  b 1.3±0 1.3±0Pistil/  pistillodium  b 1.9±0.08 1.2±0.13 100.00 0.0001Carpels/locules 3  –  5Ovules 3  –  5 a  Width  b Length 478  L.T. Carneiro and C.F. Martins
Similar documents
View more...
Search Related
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks