<<

Pollination Biology in a Tropical High-Altitude Grassland in : Interactions at the Community Level Author(s): Leandro Freitas and Marlies Sazima Source: Annals of the Missouri Botanical Garden, Vol. 93, No. 3 (Oct., 2006), pp. 465-516 Published by: Missouri Botanical Garden Press Stable URL: http://www.jstor.org/stable/40035488 Accessed: 20-08-2014 17:23 UTC

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at http://www.jstor.org/page/info/about/policies/terms.jsp

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected].

Missouri Botanical Garden Press is collaborating with JSTOR to digitize, preserve and extend access to Annals of the Missouri Botanical Garden. http://www.jstor.org

This content downloaded from 143.106.1.143 on Wed, 20 Aug 2014 17:23:24 UTC All use subject to JSTOR Terms and Conditions POLLINATIONBIOLOGY IN A Leandro Freitas3 and Marlies Sazima4 TROPICAL HIGH-ALTITUDE GRASSLAND IN BRAZIL: INTERACTIONSAT THE COMMUNITYLEVEL1-2

Abstract

Surveysof local assemblagesof and their pollinatorsare amongthe most useful ways to evaluate specializationin pollinationand to discuss the patternsof -pollinatorinteractions among ecosystems. The high-altitudegrasslands from southeasternBrazil constitutediminutive island-like formationssurrounded by montanerainforests. We registeredthe floral traits of 124 species from the Serra da Bocaina grasslands(about 60% of the -pollinatedspecies of this flora), and determinedthe pollinatorsof 106 of them. (40 species) and Melastomataceae(10 species) were prominent,while most families were representedby few species. The predominantfloral traits were: dish or short-tubularshape; nectar as a reward;and greenishor violaceouscolors. Pollinatorswere divided into eight functionalgroups (small bees, syrphids,other dipterans, etc.) and small bees, wasps, and large bees were the most importantpollinators. Butterflies, , and hummingbirdswere poorlyrepresented, and no bats, hawkmoths,or odor-collectingbees were detected. Plants were grouped in nine pollinationsystems, amongwhich nectar-flowerspollinated by bees (28%), by wasps or wasps and flies (21%), or by several groups (19%) were the most representative.With regardto the degree of specialization,plant species were classified according to their number of pollinator groups. About 33% of the species were monophilousand 30% were oligophilous(i.e., pollinated by one or two functionalgroups, respectively).The remainingspecies were either polyphilous (17%)or holophilous(19%), a highly generalistsystem in which at least three groupsact as indistinctpollinators. The general trends of the floral traits and plant-pollinatorinteractions at the Bocaina grasslands resemble those of biogeographic- connected ecosystems, such as the Venezuelan arbustal, and the Braziliancampo rupestreand cerrado.However, in the Bocainagrasslands, the mean numberof pollinatortypes per plant was 2.09, one of the highest values obtainedfor worldwide floras.The originof the high-altitudegrasslands is linked to episodes of expansionand retractiondue to glacial events. Such a situationmay have favoredspecies able to quickly occupy new habitats,including those that do not depend on a few highly specialized pollinators. The prevalence of Asteraceae may also be linked to more generalized pollination systems. Alternatively,some floraltraits, such as spontaneousself-pollination and long-livedflowers, may be advantageousfor species with more specialized systems in these grasslandswith harsh climatic conditionsand low rates of pollinatorvisitation. Key words: Asteraceae,bees, biogeography,floral biology, generalization,pollination ecology, specialization,syndrome, wasps.

1 This paper is part of the first author'sdoctoral dissertation submitted to the Programade Pos-Graduacaoem Biologia Vegetal,Universidade Estadual de Campinas.We wouldlike to thankI. Alves-dos-Santos,A. A. A. Barbosa,P. E. Berry,D. R. Campbell,P. Feinsinger,R. A. Figueiredo,J. Semir,and S. Vogel for their commentson the early versionsof this manuscript; R. Goldenbergand two anonymousreviewers for their carefulrevision of the manuscript;A. Francoisfor the English revision; I. F. Bressan,E. Gross,R. Ramos,A. L. Ravetta,and I. San Martin-Gajardofor field and laboratoryassistance; A. AmaralJr., M. C. E. Amaral,N. M. Bacigalupo,V. Bittrich, M. Bovini, E. Cabral,L. CapellariJr., T. B. Cavalcanti,A. Chautems,N. Chukr,I. Cordeiro,I. P. Correa,J. H. A. Dutilh, G. L. Esteves, R. L. Esteves, A. D. Faria, M. A. Farinaccio,A. Fillietaz, A. Flores, M. Groppo,A. Guglieri, P. J. Guimaraes,R. M. Harley, N. Hind, S. L. Jung-Mendacoli,M. L. Kawasaki,L. S. Kinoshita,H. M. Longhi-Wagner,M. C. H. Mamede,M. C. M. Marques,A. B. Martins,V. F. Miranda,R. Monteiro,M. D. Moraes,J. N. Nakajima,R. C. Oliveira,E. R. Pansarin,J. R. Pirani,M. F. Salimena,P. T. Sano, J. Semir, G. J. Shepherd,R. Simao-Bianchini,A. 0. SimSes,R. B. Singer,J. P. Souza,V. C. Souza,J. R. Stehmann,J. Tamashiro,A. M. G. Tozzi,A. 0. S. Vieira, F. A. Vitta, M. G. L. Wanderley,K. Yamamotoand C. S. Zickel for plant identifications;I. Alves dos Santos,S. T. P. Amarante,K. S. Brown,J. M. F. de Camargo,S. A. Casari,B. W. Coelho,A. V. L. Freitas,E. Giannotti,A. X. Linhares,U. R. Martins,G. A. R. Melo, L. C. Passos, S. R. M. Pedro, I. Sazima, C. Schlindwein,0. T. Silveira, and D. Urbanfor animal identifications;J. B. Allen and R. C. Forzzafor the nomenclaturalverification of plant taxa; and IBAM A for allowingus to workin the PARNA Serrada Bocaina (throughM. A. B. Rondon).CNPq, CAPES,and FAEP-UNICAMPprovided financial support. 2 The editors of the Annals thank Sophia Balcombfor her editorialcontribution to this article. 3JardimBotanico do ,Rua Pacheco Leao 915, 22460-030, Rio de Janeiro-RJ,Brasil; [email protected]. 4 Departamentode Botanica, Caixa Postal 6109, UniversidadeEstadual de Campinas,13083-970, Campinas-SP,Brasil; [email protected]. Ann. Missouri Bot. Gard. 93: 465-516. Published on 24 October 2006.

This content downloaded from 143.106.1.143 on Wed, 20 Aug 2014 17:23:24 UTC All use subject to JSTOR Terms and Conditions 466 Annals of the Missouri Botanical Garden

Plant reproductiveprocesses are believed to help Brazilian high-altitude grasslands are interesting determinethe compositionand structureof commu- communities in which to study plant-pollinator nities (Heithaus,1974, 1979; Bawa, 1990; Oliveira& interactions. Gibbs, 2000). Among such processes, the plant- In pollinationbiology, the degree of specialization pollinatorinteractions form a dynamic,yet somewhat in plant-pollinatorrelationships has generateda de- cohesive, ecological subunit of a community,which bate in which the traditional view based on the can be studied in terms of species diversity and syndromeconcept (see Faegri & van der Pijl, 1979) distribution,resource utilization,and niche packing, has been counterbalancedby skepticism about the for instance (Moldenke & Lincoln, 1979). In the specializednature of pollinationsystems (Waser et al., Neotropics,pollination biology at the communitylevel 1996; Herrera, 1996; see also Johnson & Steiner, has been studied in forest areas (Bawa et al., 1985; 2000 and Fenster et al., 2004 for reviews of both Kress & Beach, 1994), savanna-like vegetation views). Surveysof entire local assemblagesof plants (Silberbauer-Gottsberger& Gottsberger,1988; Ra- and their pollinators provide the best evidence to mirez, 1989, 2004; Barbosa,1997; Oliveira& Gibbs, assess the degree of specialization in pollination 2000; see also Ramirez & Brito, 1992) and sub- (Waser et al., 1996), because particular plant- Andean vegetation (Arroyo et al., 1982). In the pollinatorinteractions may be examined on a broad Brazilian Atlantic Forest Domain, there are several ecological scale, reducingthe samplingbias inherent studies on pollinationbiology in forest areas on both to the comparisonof data originatingfrom small scale individualspecies and flowerassemblages (Sazima et studies, such as those concerning particular plant al., 1995, 1996, 1999; Buzato et al., 2000 and taxonomicgroups. This paper describes the commu- references therein). However, for the high-altitude nity-level interactionsbetween plants and pollinators grasslands or campos de altitude- a subtype of the of high-altitudegrasslands at the Serra da Bocaina, Atlantic Forest Domain- informationon plant-polli- southeasternBrazil. We studied the floral biology of nator interactions is limited to a few case studies plant species, recordingtheir sexual systems, (Martinelli,1997; Freitas & Sazima,2001, 2003a, b; shapes, dimensions,resources, colors, and pollinators. Freitas et al., 2006). Combiningthe data on floral features and functional The high-altitudegrasslands in southeasternBrazil groups of pollinatorsallowed us to characterizenine comprise an archipelago of mountaintopformations pollination systems in these grasslandsand to group (ca. 1500 m elevation or higher) along the main plant species in four categories that express the mountain ranges, with a North/South orientation. degree of specialization/generalizationof plant-polli- Their origin is connected to glaciationepisodes since nator interactionsin this community.In addition,we at least the Late Pleistocene, when these grassland monitoredthe flowering phenology so as to assess areas are supposed to have experienced periods of flower resource availability, throughoutthe year, for expansion and retraction (Ledru et al., 1998). The the main pollinatorgroups. We provide a discussion present high-altitude grasslands have strong bio- about general patternsof plant-pollinatorinteractions geographic connections with other high-altitude at the Serrada Bocainagrasslands in relationto other ecosystems in , such as the Andean ecosystemswith strongbiogeographic connections, as and Central-American(sub-) alpine habitats, as well well as in relation to the current debate about the as with the Central-Eastern Brazilian rocky and prevalence of generalizationversus specializationin grassy habitats (campo rupestre)(Giulietti & Pirani, the plant-pollinatorinteractions. 1988; Safford, 1999a and references therein). They are also connected to the central Braziliansavannas Area and Vegetation Studied (cerrado)(Silveira & Cure, 1993; see also Modenesi, 1988 for cerradocomponents in palinoflorasfrom the The Parque Nacional da Serrada Bocaina (PNSB) grasslands).These grasslands present- for such re- - ca. 100,000 hectares- is located between the stricted areas of habitat- extraordinarilyrich floras States of Rio de Janeiro and Sao Paulo, in with many endemics. About one third of the ca. 400 southeasternBrazil. Since it extends from sea level species of the Itatiaia plateau (< 50 km2), for to the highest peaks in the Serra da Bocaina (ca. instance, appear to be endemic to the high-altitude 2100 m), the PNSB shelters many vegetationtypes of grasslands(Martinelli, 1989). However,many species the BrazilianAtlantic Forest Domain. The lowerareas on the high-altitudegrasslands have small populations are covered mainly by lowland and submontane with few individuals. Owingto such features as high rainforest, followed by montane and high-altitude species richness, high endemism, small populations, forest. Originally, these high-altitude grasslands "island-like"occurrence, biogeographic connections, probably occurred only above 1700-1800 m, but and harshclimatic conditionsfor a tropicalplace, the nowadays,due to human disturbanceof the forested

This content downloaded from 143.106.1.143 on Wed, 20 Aug 2014 17:23:24 UTC All use subject to JSTOR Terms and Conditions Volume 93, Number 3 Freitas & Sazima 467 2006 Pollination in a High-Altitude Grassland

areas, they can be found from as low as ca. 1450 m. ently as a pre-adaptationagainst frost, that also seems These grasslandsare composedof a matrixof Poaceae to be efficient against fires. and Cyperaceaespecies mixed with shrubsand herbs This study was carried out in three grasslandsites from many families, mainly Asteraceae and Melasto- that experienceddifferent fire regimes over the years. mataceae. Scattered small , mainly species Since there is no official register of fire occurrences, belonging to the EremanthusLess. (Astera- informationfor the last 45-50 years was obtainedfrom ceae), can also be found in more protectedfields. both the PNSB staff and the inhabitantsof the region. In southeasternBrazil, the montaneareas encom- The first area (22°43'57"S, 44°37'06"W) is adjacent pass two climate regions (Kb'ppen,1948). Grasslands to a privatefarm (Fazenda Mariana) and used to burn fromthe highest summits(above ca. 2000 m) are Cwb, every two to three years. The last fire in that area had with < 12°C mean annualtemperature, cool summers, occurredca. five monthsbefore the beginningof this moderatelycold winters with frequentfrost, and rare study (Dec. 1997). The second area (22°44'50"S, snow. Montane areas at lower elevations can be 44°36'57"W) is close to Santo Izidro waterfall and classified as Cfb (Cwband Cfb as in Koppen, 1948), used to burn at intervalsof five to 10 years. The last i.e., mesothermic,with 12°C-20°C average annual fire in this areahad occurredfive years beforethe first temperature,moderate winters, and mild, wet sum- year of this study, and the area burnedagain in 1998. mers (Segadas-Vianna& Dau, 1965; Nimer, 1977; Since it presentedmany treelets and large ,the Safford,1999a, b). Most grasslandareas at Serra da third area (22°44'12"S,44°36'55"W) showed marked Bocaina are Cfb. Annual precipitation is up to structuraldifferences. There was no recordof fire here 2100 mm. Rains are concentratedin summer,mainly when the project started, but the area burned in from December to March. The dry season (Apr. to September1999. Sept.) is driestfrom June to August,with precipitation About 260 plant species belonging to 46 families < 50 mm per month.The effects of seasonal drought were recordedduring monthly collections in the three on the vegetationare partiallyoffset by thick banks of study areas and throughnon-systematic collections in orographicfog that shroud most Bocaina grasslands, other grassland areas located between 1450 m and mainly in the early morning,for much of the year. 2100 m. Althoughmost families were only represent- Annual mean temperature is ca. 15°C at 1500- ed by a few species, some families, such as Asteraceae 1600 m, and minimumtemperatures around 0°C are (ca. 70 spp.), Poaceae (ca. 35 spp.), Cyperaceae(ca. commonduring winter (L. Freitas, pers. obs.). During 15 spp.), Melastomataceae(18 spp.), Orchidaceae, the studyperiod (1998-1999), frostsoccurred on more ,, and Solanaceae,were common. than 30 days per year (L. Freitas,pers. obs.). General Most species were herbs or (sub)shrubs,with small, informationabout the vegetation,climate, and topog- often lignified and hairy . Many species had raphy of southeasternBrazil is available in Segadas- leaves with a cupressoid or rosette arrangement,and Vianna & Dau (1965), Eiten (1970, 1992), Alonso several perennial plants showed well-developed un- (1977), Nimer(1977), Moreira& Camelier(1977), and dergroundorgans (e.g., Tibouchinaminor, Microlicia Safford(1999a, b). isophylla, Esterhazyamacrodonta, Escallonia farina- Despite its long history in Bocaina highlands, cea). Such vegetative features are typical of plants humandisturbance was low until the beginningof the living in high-altitude tropics (Camerik& Werger, 20th century. The primarysource of revenue of the 1981; Smith & Young, 1987; Safford,1999a). inhabitants of this region comes from livestock. Regular fires and grazing are supposed to have Material and Methods negative impacts on the physical environment,the vegetationstructure and dynamics,and the welfareof Between December 1997 and February2000, we some of the small and endemic plant populationsof made 25 field trips to Serrada Bocaina,for a total of these grasslands (Safford, 1999a, but see Safford, 211 days of fieldwork. Pollination biology at the 2001). Natural fires are apparently rare on these community level was studied along a transect ca. grasslandsand are restrictedto the wet season, when 1 km long in each of the three study sites. We studied lightningoccurrence is frequent.Anthropogenic fires, all plant species located within 5 m of the transects, however, are frequently set during the dry season, except Cyperaceae and Poaceae (Table 1). The because such an action favorsthe sproutingof grasses following floral attributes were noted in the field: after the rains begin. Some plants may have been shape, symmetry,size, color, and rewards(Table 2). eliminated by these regular fires, in particular, FollowingFaegri and van der Pijl (1979) and Endress species restricted to mesic habitats. However,many (1994), the flowerswere classified into eight different of the typical high-altitude taxa, such as the types: inconspicuous,dish (bowl),brush, bell (funnel), geophytes, have a strong capacity to sprout, appar- flag, gullet, tube, and revolver. Tube types include

This content downloaded from 143.106.1.143 on Wed, 20 Aug 2014 17:23:24 UTC All use subject to JSTOR Terms and Conditions 468 Annals of the Missouri Botanical Garden

Table 1. Plant species, their habit, floweringtime, and floweringpeak at the Serrada Bocainagrasslands. All collections by the authors are deposited at the UEC, and the collection number (L. Freitas) follows the taxon name in parentheses. Phenologicalobservations occurred from January to December 1999. Species that were found only in the area that burntin September1999 are indicated by the symbol (f).

Family/Species Habit Floweringtime Floweringpeak Amaryllidaceae Hippeastrumglaucescens Mart. (773) herb Oct.-Nov. Nov. Apiaceae Eryngiumcanaliculatum Cham. & Schltdl. (792) herb Nov.-Dec. Dec. Eryngiumhorridum Malme (140) herb Jan.-Mar. Jan. Gonioanthelahilariana (E. Fourn.)Malme (329) vine Mar.-July(f) Mar.-May Mandevillaerecta (Veil.) Woodson(86, 509) sub- Jan.-Feb. Jan. Oxypetalumappendiculatum Mart. (123) vine Jan.-Aug. (f) Apr.-June Oxypetalumsublanatum Malme (13, 491) vine Jan., Mar.-May, Jan., Apr., Nov.-Dec. Oct.-Dec. Tassadiasubulata (Veil.) Fontella & E. A. Schwarz(686) vine May-Aug. June Aquifoliaceae Ilex amara (Veil.) Loes. (656, 797) sub-shrub Nov.-Dec. Nov.-Dec. Asteraceae Achyroclinesatureioides (Lam.) DC. (276, 330) shrub Jan.-Aug., Dec. Mar.-June Baccharisaphylla (Veil.) DC. (707) sub-shrub Sep.-Nov. Sep.-Oct. Bacchariscuritybensis Heering & Dusen (498) sub-shrub Jan.-Mar.,Dec. Jan. Baccharisdracunculifolia DC. (372) shrub Mar.-Oct. Apr.-May,Aug. Baccharisintermixta Gardner (418) shrub May-Aug. July Baccharisleptocephala DC. (392, 443, 688) herb Apr.-Sep. May-Aug. Baccharispentziifolia Sch. Bip. ex Baker (589, 691, 728) shrub Jan.-Dec. Feb., May-Aug.,Oct.- Nov. Baccharisplatypoda DC. (334, 863) shrub Mar.-June May Baccharistarchonanthoides Baker (733) sub-shrub Oct.-Nov. Oct.-Nov. BaccharisL. sp. indet. 1 (407) shrub July-Aug. July Baccharissp. indet. 2 (441, 689) shrub May-July June-July Baccharissp. indet. 3 (108) shrub Jan.-Mar. Jan. Barrosoabetoniciiformis (DC.) R. M. King & H. Rob. herb Jan.-May Feb.-Mar., May (512, 591) Campulocliniummegacephalum (Mart, ex Baker)R. M. herb Jan.-May Feb.-Apr. King & H. Rob. (101, 596) Chaptaliaintegerrima (Veil.) Burkart(744) herb Jan.-Feb., Oct.- Oct.-Nov. Dec. Chaptaliaruncinata var. graminifolia(Dusen) Burkart herb May-Nov. July-Aug. (406, 697) Chromolaenacf. decumbensGardner (100) herb Jan.-July Jan.-Mar. Chromolaenaxylorrhiza (Sch. Bip. ex Baker)R. M. herb Jan.-Feb. Jan. King & H. Rob. (835) Eremanthuserythropappus (DC.) MacLeish(403) treelet Jul.-Aug. (f) Aug. Erigeronmaximus (D. Don) Otto ex DC. (64) shrub Jan.-Mar.,Oct.- Jan.-Feb., Dec. Dec. EupatoriumL. sp. indet. 1 (60, 61) shrub Jan., Nov.-Dec. Nov. Eupatoriumsp. indet. 2 (99) herb Jan.-July, Dec. Jan.-Feb. Eupatoriumsp. indet. 3 (326) herb Mar.- June Apr.-May Eupatoriumsp. indet. 4 (374) herb Apr. Apr. Gochnatiapaniculata (Less.) Cabrera(438, 705) sub-shrub Aug.-Nov. Sep.-Oct. Grazieliagaudichaudeana (DC.) R. M. King & H. Rob. shrub Jan.-June Jan.-Feb. (226, 590) Hypochaerisgardneri Baker (481) herb Jan.-Feb., Oct.- Jan.-Feb., Nov.-Dec. Dec. Lucilia lycopodioides(Less.) S. E. Freire (398, 838) herb June-Sep. July-Aug. lundiana DC. (393) vine May-Jul. May

This content downloaded from 143.106.1.143 on Wed, 20 Aug 2014 17:23:24 UTC All use subject to JSTOR Terms and Conditions Volume 93, Number 3 Freitas & Sazima 469 2006 Pollination in a High-Altitude Grassland

Table 1. Continued.

Family/Species Habit Floweringtime Floweringpeak Mikania nummulariaDC. (440, 690) shrub May-Aug. June-Aug. Mikaniasessilifolia DC. (318, 417, 871) shrub Mar.-May Apr. Seneciooleosus Veil. (391, 409) shrub Jan., Mar.-Sep., Mar.,July-Aug. Nov. Stevia myriadeniaSch. Bip. ex Baker (88, 291) shrub Jan.-May Feb.-Mar. Symphyopappuscompressus (Gardner) B. L. Rob. (57) shrub Jan.-Feb. (f) Jan. Vernoniaherbacea (Veil.) Rusby (741) herb Oct.-Dec. Nov.-Dec. Vernoniamegapotamica Spreng. (36, 536) herb Jan.-Apr. Jan.-Feb. Vernoniacf. rosea Mart,ex DC. (275, 592, 844) sub-shrub Jan.-Feb. Feb. Vernoniatomentella Mart, ex DC. (67, 588) sub-shrub Jan.-May, Oct.- Jan.-Feb., Nov.-Dec. Dec. Vernoniatragiaefolia DC. (626) sub-shrub Jan.-Feb., Apr. Feb. Vernoniawestiniana Less. (92) shrub Jan.-June Feb. Bromeliaceae Dyckia tuberosa(Veil.) Beer (495) herb Sep.-Nov. Oct.-Nov. Campanulaceae Lobeliacamporum Pohl (41) herb Jan.-June, Nov.- Jan.-Feb., June, Dec. Dec. Wahlenbergiabrasiliensis Cham. (232, 280) herb Jan.-Dec. Feb., June-Sep. Clethrascabra Pers. var. scabra (332) treelet Feb.-Aug. (f) Mar.-May Convolvulaceae Convolvuluscrenatifolius Ruiz & Pav. (37, 89) vine Jan.-Feb., Dec. Jan., Dec. Ipomoeaprocumbens Mart. (76) vine Jan.-Mar. Feb. Jacquemontiagrandiflora Meisn. (118) vine Jan.-Aug., Nov.- Jan.-Feb., May-July, Dec. Dec. Cunoniaceae Weinmanniaorganensis Gardner (54, 125) treelet Jan.-Aug. (f) Feb.-Mar., May Droseraceae Droseramontana A. St.-Hil. (127) herb Jan.-Feb., Nov.- Jan., Dec. Dec. Agaristahispidula (DC.) Hook. f. ex Nied. (23, 431) shrub June-Nov. Aug.-Sep. Gaylussaciachamissonis Meisn. (469) sub-shrub Apr.-Dec. Oct.-Dec. Gaylussaciajordanensis Sleumer (370, 432) sub-shrub Jan.-May,Aug.- Sep.-Oct. Dec. Eriocaulaceae Paepalanthuspaulensis Ruhland(706) herb Sep.-Dec. Sep.-Nov. Paepalanthuspolyanthus (Bong.) Kunth (18) herb Jan.-June, Nov.- Jan.-Apr., Dec. Dec. Erythroxylaceae Erythroxylummicrophyllum A. St.-Hil. (373, 496) shrub Jan.-May, Nov.- Jan.-Feb., Dec. Dec. Esc alloniaceae Escalloniafarinacea A. St.-Hil. (507) shrub Nov.-Dec. Nov. Euphorbiaceae Crotondichrous Mull. Arg. (17, 619) shrub Jan.-Dec. Mar.-Jul.,Nov.-Dec. Gentianaceae Calolisianthuspedunculatus (Cham. & Schltdl.) Gilg herb Jan.-Apr., Dec. Jan.-Feb. (430, 520) Calolisianthuspendulus (Mart.) Gilg (80, 264) herb Jan.-May,July Feb. Deianira nervosaCham. & Schltdl. (348) herb Feb.-Apr., June Feb.-Mar. Helia oblongifoliaMart. (1, 95) herb Jan.-Mar. Jan.-Feb. Zygostigmaaustrale (Cham.& Schltdl.) Griseb. (622) herb Feb.-Apr. Feb.-Mar. Gesneriaceae Sinningia allagophylla (Mart.)Wiehler (9) herb Jan.-Mar.,Nov.- Jan., Dec. Dec.

This content downloaded from 143.106.1.143 on Wed, 20 Aug 2014 17:23:24 UTC All use subject to JSTOR Terms and Conditions 470 Annals of the Missouri Botanical Garden

Table 1. Continued.

Family/Species Habit Floweringtime Floweringpeak Hypericaceae Hypericumbrasiliense A. St.-Hil. (218, 508) shrub Jan.-May, Dec. Jan. Hypericumternum Choisy (71, 494) herb Jan.-Mar.,May- Jan.-Feb., Nov.-Dec. Aug., Oct.-Dec. Iridaceae Alophiageniculata Klatt (376) herb Mar.-Apr.,Nov.- Mar.,Nov. Dec. AlophiaHerb. sp. indet. 1 (235) herb Jan.-Feb. Jan. Calydoreacampestris (Klatt) Baker (236, 368) herb Jan.-Mar.,Oct.- Jan., Nov.-Dec. Dec. Sisyrinchiummicranthum Cav. (762) herb Jan.-Feb., Nov.- Nov.-Dec. Dec. Sisyrinchiumvaginatum Spreng. (346, 434, 515) herb Jan., Mar.-Dec. May-Oct. Lamiaceae Hyptislippioides Pohl ex Benth. (399, 423) herb July-Oct. July Hyptisplectranthoides Benth. (5, 606) herb Jan.-May, Oct.- Jan., Dec. Dec. Hyptisumbrosa Salzm. ex Benth. (310) shrub Jan.-Apr. Feb.-Mar. Peltodonradicans Pohl (313, 625) herb Mar.-May Mar.-Apr. Leguminosae ChamaecristaMoench sp. indet. 1 (85) sub-shrub Feb.-Mar., Oct.- Nov.-Dec. Dec. Crotalariabreviflora DC. (308, 818) sub-shrub Jan.-Apr., Dec. Jan.-Mar. Lupinusvelutinus Benth. (427) sub-shrub July-Oct., Dec. Aug.-Sep. Lythraceae Cupheaglutinosa Cham.& Schltdl. (20) sub-shrub Jan.-Dec. Jan., Apr.-June,Aug., Nov.-Dec. Malpighiaceae Byrsonimavariabilis A. Juss. (112, 490) sub-shrub Jan.-Feb., May, Nov.-Dec. Oct.-Dec. Malvaceae Pavonia kleinii Krapov.& Cristobal(105) vine Jan.-Feb., Dec. Feb. Sida L. sp. indet. 1 (24, 456) herb Jan.-Mar.,Oct.- Jan., Nov.-Dec. Dec. Melastomataceae Leandraerostrata Cogn. (113) herb Jan.-Feb., Oct.- Jan., Nov.-Dec. Dec. LeandraRaddi sp. indet. 1 (19) herb Jan.-Feb., Nov.- Jan., Nov.-Dec. Dec. Leandrasp. indet. 2 (401) shrub July-Oct. Aug.-Sep. Leandrasp. indet. 3 (402) shrub July-Dec. Oct.-Nov. Microliciaisophylla DC. (103) sub-shrub Jan.-Mar.,July- Jan.-Feb. Aug., Dec. Tibouchinafrigidula (DC.) Cogn. (25, 419) shrub Jan.-Dec. Jan.-Mar.,Dec. Tibouchinamartialis Cogn. (369) shrub Apr.-June June Tibouchinaminor Cogn. (32) herb Jan.-Apr., Dec. Jan.-Feb. Trembleyaparviflora (D. Don) Cogn. (341, 420) shrub Feb.-July June-July Trembleyaphlogiformis DC. (44, 74) sub-shrub Jan.-Mar. Jan.-Mar. Ochnaceae Ourateasemiserrata (Mart. & Nees) Engl. (425) treelet June-July June Orchidaceae Epidendrumsecundum Jacq. (150) herb Jan.-Aug. (f) Jan., Apr.-May Habenariaparviflora Lindl. (82, 804) herb Jan.-Mar.,Dec. Jan.-Feb. Oncidiumbarbaceniae Lindl. (282, 305) herb Jan.-Mar.,May- Jan., May,July, Sep.- Dec. Dec. Oncidiumblanchetii Rchb. f. (611) herb Jan.-Dec. July-Nov. Dec. OncidiumSw. sp. indet. 1 (501) herb Jan., Jan.

This content downloaded from 143.106.1.143 on Wed, 20 Aug 2014 17:23:24 UTC All use subject to JSTOR Terms and Conditions Volume 93, Number 3 Freitas & Sazima 471 2006 Pollination in a High-Altitude Grassland

Table 1. Continued.

Family/Species Habit Flowering time Flowering peak Orobanchaceae Esterhazya macrodonta Cham. & Schltdl. (244) shrub Jan., June-Oct. July-Aug. Polygalaceae Polygala brasiliensis L. (309, 366, 683) herb Feb.-Sep., Nov.- Mar., June-Aug., Dec. Dec. Polygala cneorum A. St.-Hil. & Moq. (390) herb Apr.-Sep., Nov. May-Aug. Rubiaceae Borreria capitata (Ruiz & Pav.) DC. (322) herb Feb.-July Mar.-Apr. Borreria tenella Cham. & Schltdl. (192, 667) herb Mar.-July May-June Declieuxia cordigera var. angustifolia Mull. Arg. (6, herb Jan.- Dec. Jan.- Feb., June, 727) Aug.-Sep. Galianthe angustifolia (Cham. & Schltdl.) E. L. Cabral sub-shrub Jan.-May, July, Jan.-Feb., Dec. (50) Sep.-Dec. Galianthe brasiliensis Spreng. (618) shrub Jan.-Aug., Oct.- Jan.-July, Nov. Dec. Galium hypocarpium (L.) Endl. ex Griseb. (371, 802) sub-shrub Jan.-July, Oct.- Jan., Apr.-May, Nov.- Dec. Dec. Solanaceae Solanum aculeatissimum Jacq. (483) herb Jan.-Feb., May, Jan.-Feb., Oct.-Dec. Oct.-Dec. Solanum americanum Mill. (488, 675) herb Jan., Nov.-Dec. Nov. Solanum pseudocapsicum L. (768) sub-shrub Jan.-Feb., Nov.- Jan.-Feb., Nov. Dec. Solanum swartzianum Roem. & Schult. (538, 659) shrub Jan., Apr.- June, Apr.- May Aug. (f) Solanum viarum Dunal (559, 742) herb Jan.-Mar., Aug.- Feb., Sep., Nov.-Dec. Dec. Verbenaceae Verbena hirta Spreng. (31, 243) sub-shrub Jan.-Dec. Jan., Sep.-Dec. Violaceae Viola cerasifolia A. St.-Hil. (124) herb Jan.-Feb., Dec. Jan., Dec. asperula Mart. (846) herb Feb.-Mar. Feb. Xyris tortulla Mart. (2, 350) herb Jan.-May, Dec. Jan.-Mar. both salverform and spurred . Asteraceae activity test (Zeisler, 1938). In some cases, flowers florets were included in the tube type, and their head were tagged and bagged at bud stage, and nectarwas as a whole was classified as either dish or brushtype. extracted on the following day with a graduated Floralmeasurements with calipers were performedon microlitersyringe (Hamilton,U.S.A.). Nectar volume at least five fresh flowers, each from differentplants. was registeredimmediately, and nectar sugarconcen- The main flower color was determinedusing a color tration was measured with a hand refractometer guidebook (Kornerup& Wanscher, 1963). For our (Atago,Japan). analysis at the community level, the flowers were Sexual systems were determinedby the presence of groupedinto six color sets: i. violet- showy combina- both functional anthers and stigmas. Flowers were tions of blue and red, such as purple and magenta;ii. categorizedas either hermaphroditic(monoclinous) or pink- light colors based on blue or red, such as rose, unisexual (diclinous). Plants with unisexual flowers lavender, and lilac; iii. red- orange and red; iv. were further classified as monoecious, dioecious, yellow- bright and deep yellow; v. greenish- colors andromonoecious,or gymnomonoecious.Other in- with sallow tonalities such as greenish, yellowish, or formationabout sexual systems- such as the presence brownish white; and vi. white- bright, apparently of dichogamyand heteromorphy- was noted for some pure, white. For many species, pollen fertility was of the species. Spontaneous self-pollination was estimatedby cytoplasmicstainability, using the acetic verified for some species by the reportof fruit set in carmine test (Radford et al., 1974), and stigmatic unmanipulatedbagged flowers (Kearns & Inouye, receptivity was verified with the H2O2 catalase 1993). Some pistils were fixed in formaldehyde-

This content downloaded from 143.106.1.143 on Wed, 20 Aug 2014 17:23:24 UTC All use subject to JSTOR Terms and Conditions 472 Annals of the Missouri Botanical Garden

8 1 *? ^ $ - § ~o 1 - I B s §< I || 6^3 I. & £§ g - 3 c «& 2-3 23 *

'^ ^ Oh ^ Dh 0^ "§ |h II "3 I 1 8 1B h jj -Ijj

cd J5 cd T3 ^ ^ ^ ^ ^.^ ^ 3 w cried ctftfccjtfcc! cd cd cd cd cd cd cd cd cd ^^fr? S ""2 cd 00 00000 o o o o o o o o o ! £ ! J i -5 a '§ 3 ^ i §

^ - C > fju cd ud cdcd^cdcd ^ tJD J; cd ^ cd cd ^ S

§^B| 1

M .5 § ^ 1^ ^ xx xxxxx x^ x^x^x^x^x^x^x^x -O^g^^ 'O "-1 rHCSJ rHlOfNLOi- I CS|X LOrHMDXloX(roX^XfoXloX|> 111! °I QU i §,».§«§Ou en (V) j3 -m cd >~» ^ g £p> j^ ^T v r s r s ^ ^ v _, ^ ^ , s ^ ^^ ^^ X S > ^J I B T ^ ^ ^ ^ S^ -. ^ ^ ^ ^ ^ ^ rHX^H g^ ^co^, ^ o~ cj d d 0s o" o" o" cj o" o o S~g>g o 3 .5.5 ^BSS^o £ 2323SS2S33332S2 g2s-5 3 E •2 ^ • S e 1 ^ § -S S

^ J 1 I. §. ^ 1 "S"i I I -a "B a" * s* "• Hi it I I 1 § ,3 11 S 8e ? E S iS« S ISqqqqqqq «E£§8 M m a KK ffiSKKffi 6 O 3 => D D D 3 P

Hill , 1 |s I I! I 1 f 1 I 3 * 1II1! 1 4 |i 1 illi ! i I i I I t I K5!h I IB S-Si-ij ii-li « J ! ! * I * I * ^.i|M £ 1 1 sill ill i|J I illllllll iiiix% il|ll|lllll|i illSiJIII

This content downloaded from 143.106.1.143 on Wed, 20 Aug 2014 17:23:24 UTC All use subject to JSTOR Terms and Conditions Volume 93, Number 3 Freitas & Sazima 473 2006 Pollination in a High-Altitude Grassland

Cfi Cfl Cfi CO £3 3333 33 3 CD O OOOO^h O O Sh O S-i •-> C sh ^h ^ in n) j_< f^n CD >h 1) &D &D bG &O StG M) M) W) ^ £ ^ j cc ° ° ° -g 'o "o "o "o "5 "o "o *§ s s .s .s .s 3 .s .s 3 .s is

QhPh PhDhDh "§|p^ ^|zzz^;^;ZZ^;!z;!z; iz; |zT|z; |zT|z; ^ iz; Z ^^ Z iz; z

1 t I S. g I I i I 3 J U i 1 « i U i { n i S ^13^13 3^^^^33.>^.2 3 S _3 "o

^ CO 00 5 ^ CO CO i- I i- I s^^x^x^-x^x^x x ^xxxx x £ XX^ x o^ rH t~- H r-H 'O XtjXnXioX^X^ Os rHCOrfCOO VO CO CO

o b ^-s ^-^ ^-s ^^o 0s ^^ oT ^-^ ^ S" ^-s

S^^ XXh VVV^VVV V ^VVVV ^ V ^VV V rn ^Vxxxxxxxxxx x ^xxxx x x xxx x x >'cn^COCS|CO^'*^CSl(NCO -^ i- I

cd ^ ^ ^ ^ ^ ^ ^ ^

fa

2 cecda3a3a3ajaja3ctic3cd T3 ^ (N(M(N(NCSI(N £££££££££££ -go ^^^^^^^^^^^ sa iiiiiiiiiil II 1 1 1 1 1 1 ><^qqqq I, a | a £ a ^* D D P P X X OjOJffi ffi ffi Ojffi X X

« rH CSl CO J -| I "I | 1 I I I I -I § glf^lJsill 1 S. J .S .S .S -| | I | I .g ^ ^ f .1 d d d $ ^S&S-fr|J|,-S5y|||3lggg '§ -g -g -g -g o o 1 -S -S | ^ ^ ^ g .2 .2 -2 ^ fa JJJJllH^'3'3|ll§g||| e2 G§c3c§c§cqOoSSot3c5t^t^i

This content downloaded from 143.106.1.143 on Wed, 20 Aug 2014 17:23:24 UTC All use subject to JSTOR Terms and Conditions 474 Annals of the Missouri Botanical Garden

en co co co co Ph Oh Oh Oh 3 3 i, 3 3 3 £ 2 2 & S3 3gO£OMkk

c 35 5^? ^r* a;a5a5^7ia5^7'i5i O CO CO J_| J_| ^COcO^CCs-iS-ii-i \3 'rtfi #FH£ cd cd o s-, 'rtg 'rtfl cd "^fl cd cti cd cd >-. "S >~> "o >-> >~> l^a <1> "o "5 "5 "5 ^ ^ f "2 § 5 S t t 3 j "3 "3 S "2 g g S

co ^ co p-Q ^ r-Q ^ > co co co p-Q co -Q -O -Q

03 'g § Ph fc Oh Ph PhPhPh Ph Ph

> CD

I * * * I "-1 1O s, £3 jjh j--j jg ^ 1 j_ -J3 epeoco cocd ^3 °3 I .S 1 .g £ 1 1 . 1 3 . .fi | 1 i i I

3

"? - ^H O ^ ^ H ^ On On On i- i i- i i- i <- i X * J- X X X | **XX* X X ^ X X ^XX- - - - ^ •J3 ON CO t> i- I i- I O l> i- I LO ^f i I C~ C i- I r-H lO r-\ r- I ON

S? os^^v5N ^-s S* to ^ 3^ ^

v v vv^^v ^ ^ ^^v ^ v ^ v v a^^s^ ^* ^o;X X X XxXxX X X XxX X X X X X r- i ^*coO CO CO N i^ O \fl 00 ^ (N r-r^iO \O LO CSI On t^h pS o o o cJt^cJts^o' o o o >^ ^ fe < < < <

cd ^^^_^^^^^ ^^^^ ^^ ^^ ^^-^^^^ ^^^ ijsijsi^ijsijsfjji-sijsfjsi-sijiijsijsijjijBijsi 3pJ3 S pQ 3-J2 g 3^Q 3 M 3^Q 3-Q 3^ 3J2 3X! 3T3 3J3 3^Q 3p£) 3.J3 3^Q

_co

gcdcdcdcdcdcdcdcdcdcdcdcdcdcdcdcd T3-S(N(NCS|(N(N(N(N(NCSlCS|CS|(NCSIOvJCSI(N

O^cpjcdajcpjcdajcpjajcdcdojcdcdcdcdcd ^222222222222222 2 2

3« C^fl^ CriS

^^ffi ffi ffi OjOjE ffi ffi Ojffi ffi ffi K ffi ffi ffi

1 ill 1 1 i 1 1 1 . i ! 1 1 1 1 1 ° esi i -e ••! .2 J -^ e e e I* £, -3 -S -3 -3 -S e2 tl^^^^SSSl^^-^^^^^

This content downloaded from 143.106.1.143 on Wed, 20 Aug 2014 17:23:24 UTC All use subject to JSTOR Terms and Conditions Volume 93, Number 3 Freitas & Sazima 475 2006 Pollination in a High-Altitude Grassland

cc en en fl . I ft . ft ft II II I I I & I 1 1 I I

°b ^!a

^3 X pQ x ?ccj3p^!c«cc ^3J3^Qt«cn ^

^ | Z ^ ^ 2f ^ ^ Z ^ ^ Oh Z Z ^ Z ^ Z ^ ^

1 ^ § I * I lit II loo j 1 „ ;§ ir ,- s .s J I ^ El 1 3 3 1 i 1 I 1 1 I'll 1 I, I •i 8 S x ^ (N CS| VO U-T rH (N

S3X XXX XX XX XXX^X^X i- i rH i- I 1^ CM f- I (N CO COI>i- i(mXcmXlO •£ LO C\| S 1

a g ,__, ^ i- i co x :§ ^xx x^^^^o ^xx ^ S 'S VO h 00^1 CM i- l CO rH T3 -O rHOOs T3'-OFT3T3 'TS o jj >; >; cT cJ d o o o o o' cJ o" o d o" o o fa < NN<1 < < < < < < <<<<<<-^<

CO ft .... ^03 ? « sq_ sp^ sQh sft Illili 2 2 3 3 3 =1 ill .s.s.sM 4

I1 1 1 11 111 I ^22o322 ^^_^a; -* cdos t> p.ft-Cftft J3 JS J3 M (N J3

i . . 1 i . I 1 Hilt i I- 1 .1*1-1 !«.! i H i tit. I ! ii 5 f-I I11JI lill III Hi illlll I it

This content downloaded from 143.106.1.143 on Wed, 20 Aug 2014 17:23:24 UTC All use subject to JSTOR Terms and Conditions 476 Annals of the Missouri Botanical Garden

MM I I I I II-I aa^a J| a a a J1 j 1 O S-i ^ ^ Q 3^ C 0 ^) Q} ^ ^"l 1 ! ^ . . 1 1 . 1 1 | i r| 1 1 s ^ I ^ I ! I ^j g § § I § f j I § js § i I i js J8

fe § Z ZZ* zS Oh Z&i Z Ph Sh' Ph"CC Ph aT OT ^ ^

r? ^&i it* K ,^3 p^ H i_h b^ Pi i i W W W i iS^i ii i H i | | |

1 -8 I S, | | | | | |

£

"2 ^^ On O LO -^ CM

1-X X"o^XoX2 c^X X XX^oXXX •^ LO CSI CO

1- I Q\ CO CO SScq ^^CM X X X ^^T X xxiox0 ; >; a >; «" >; cf « 0" J « o o >; >^

en co co co 2 !-tJ ^2 r^ p-Q co fJ3 co co ^ ^ co f^5 <-£ ^

^ « ffi i ffi ffi ffi _Hj ffi ffi ffl ffi ffi ffi K ffi ffi ffi ffi S "i 3 « I 1 ^sg 11 ^ |> I --si! -i

2 £ ll Ii ill I If 1 1 |i |Ill| 1 I 51

WHO O E £ J

This content downloaded from 143.106.1.143 on Wed, 20 Aug 2014 17:23:24 UTC All use subject to JSTOR Terms and Conditions Volume 93, Number 3 Freitas & Sazima 477 2006 Pollination in a High-Altitude Grassland

*e 43 a3 ^3

0) 13 § Ph Oh Oh ^^OOO ^|^^ Oh 2 Oh Z O" 2f £; OhOhOhOhOh 0h0h0h0h0h Oh

>H ^> > >H ^^ Dh^(^0h> >>>0h> >H >O^^^ q3|d>>

i js? 1 I I J 1 « I Illl Jh>0)0) 0) &, .2 - « r-^^S? 0) ^>^ ^ 4; >^>r_ C >F^^ ^ >^4)-«>^>s>^ pg -3 j3 -5: •> -3 §^ •> ^ S .5 3 .S .S f S ^ ^ .S S > 3 g .>.>.> S

'o

C

If _ S§ 2 jo hh2^ VVxXX 1,^ ^rH^xX

-3sg =3 mm) _g -g S3 -S-S-g-g-g -S -S "I -S -f § J ^ -| -S -g E

& M J4. 2 ^-§.22^222 ^^0)0)b s * 2 fefe aaaaa|||||-CO)O)O)0)O0 ^t)^^^ 2 O^DiCl, J3 pJZIj^ D-, J3 -J3 cd cd d ts cs j3 J -fi -fl -fl Oh

1 8 Is ^ HNW^ ^ ?l S |§l:3i 1 1,1 t li i 1 :,^111||; lilfj I till a i «|I lillitliilllillHtiJJliilli j j S S S OO

This content downloaded from 143.106.1.143 on Wed, 20 Aug 2014 17:23:24 UTC All use subject to JSTOR Terms and Conditions 478 Annals of the Missouri Botanical Garden

s I || - ^ § -I

c ^ «§ •S ii SSI S Is 8 8 8 8 S 3 8 JS te J a ^ -S S -a « « 13 cs « 13 13 «a "o^^^^ « ^ 72 £ 72 " £>*« ^i- i S cc 2 > c S2> ex cu a Oh cx c ^-C_2^o; „ ^ ^ ^ Ph ^ flj Ph pv C3 o -fi ^ Ph a 0)^ ^ .-i M S J3 ^,£3 ^o ^ ^Q £££ JQJ2^QJ3^3 X) ^Q 5? ™ En fi S ^

fc«2 ^ ^^ ^ ^ ^ ^^^ Dh Dh fc Ph fc Z Ph Ph PhSod^^ C O £ II 2

..S^ fjq ! !! ! 3-1 |__( ,__, ffj 3h Ph W W H H Pi i_i i i H fTl n 0 1 -^ f^ >> ^ > > O^O >h^^>hO > > >"• >h ^ ^2^ h ^^ S 8 H S ° w i- i -g a^a;! i^^ 8 - •Sc/3>td

0 « xxx ii ^ -g "2 Tt VO ^Ph *- (NCO ^ CO VO CSlCOrH ^ CO ^ ^ ^ ^ ^ ^S "3 -^ Z -^ XX X 1 ^ X X X XXX ^XXX^ X X XguS^ "3 CO CSICO ^f* CO CO ^"^CSI .- Ir-H^lOr-H CO V© r? ^O'^^-S I 1g 1 «1 |? e |^|| ** :s x x x xx- 2^.=!^2 x x 2 gJc5- = "S ^ CO A >^ >^ u" ^ ^ o" J" o" J o1 o o" o" o >T, o" d § I 5?

o ' - - ^ r-Q

^ -I § § I

cd >-. <>-• >~> c 5 *^3 S II ii 11 tiJU II U 1 i;3|t flT t-t Dh |V) CO "8 11 8 jjj

^^ ffi XX X X X XXX llfl> £ nm 11

This content downloaded from 143.106.1.143 on Wed, 20 Aug 2014 17:23:24 UTC All use subject to JSTOR Terms and Conditions Volume 93, Number 3 Freitas & Sazima 479 2006 Pollination in a High-Altitude Grassland

alcohol-acetic acid (FAA) at 12 to 60 hours after self respectively; others were designated as secondary or cross hand-pollination and analyzed under fluores- pollinators. The final set of species was pollinated by cence microscopy to observe pollen tube growth three or more functional groups of apparently similar (Martin, 1959). Plants along the transects were importance (holophilous) and thus did not have any checked monthly from December 1998 to February main pollinator(s). These species were grouped in the 2000, and two phenology parameters were recorded: pollination system named "several insect groups" - flowering time months in which each species was in (SIG), and their pollinators were designated as - flower; and the activity blooming peak months in indistinct to distinguish them from the secondary which more than 50% of the individuals of each and main pollinator categories. species presented flowers (Table 1). Plant species were classified into four categories The pollinators for each species were determined representing the degree of specialization/generaliza- through direct observation of flowers for at least four tion of their interaction with their pollinators: hours (Tables 3, 4). Observation periods were one to monophily, oligophily, polyphily, and holophily (Ta- two hours, but longer periods (three to eight hours) ble 4). The definition of our first three categories were necessary for some species with very low differs from that given by Faegri and van der Pijl frequency of visits; these were usually observed for (1979), because our unity of reference was functional a total of 12 or more hours (see Table 4). We groups rather than pollinator species. Thus, here, concentrated our observations on the most probable monophily refers to species exclusively pollinated by time of visitation for each Most pollinator species. one functional group. Plant species pollinated by two observation sessions occurred in the middle of the functional groups (either one or both are their main because are more active in this day, generally pollinator[s]) are considered oligophilous. Species The total observation times include the period. only pollinated by three or more functional groups are observations under climatic conditions and good classified either as polyphilous, when one or two the of each To be during flowering peak species. groups act as their main pollinators, or holophilous, considered visitors had to have pollinators, pollen when there is no preponderance of one or two groups, attached to their bodies and contact the grains stigmas i.e., when groups act as indistinct pollinators. during their visits. Hummingbirds were identified in the field or Insects the through photographs. visiting Results flowers were collected, when possible, with an net. Insect were identified entomological specimens We studied a total of 124 species from 35 families and at MZSP, ZUEC, and by specialists deposited (Table 1). Forty species belonged to Asteraceae MPEG. insect visitors were Non-captured grouped (32%), followed by Melastomataceae (10 spp., 8%). into different hierarchical at least to order categories, On the other hand, 16 (46%) and seven (20%) and, in a few cases, to level Bombus genus (e.g., Latr.). families were represented by only one and two For the broad we community study, categorized species, respectively. captured and non-captured insect visitors into seven functional groups (after Root, 1967, Cummins, 1973): Floral Biology and Pollinator Agents at the large bees, small bees, wasps, syrphids, other CommunityLevel dipterans, butterflies, and beetles (Tables 4, 5). Large bees are those larger than 12 mm (because it is not a. floral traits indigenous to Brazil, Apis mellifera was not included in any bee group). The most common flower color groups were violet Plant species were divided into nine pollination (29.8% of the 124 species), greenish (28.2%), and systems. For those species that were pollinated by one yellow (17.0%). Pink (10.5%), white (9.7%), and red or two functional groups (monophilous and oligophi- (4.8%) were the least represented groups (Table 2). lous; see below), pollination system was based on sets Most flowers were tubular (47.6%) or dish-shaped of floral traits common to plants pollinated by each (28.2%). Since we considered the shape of the head of functional group. A second set of species were Asteraceae as a whole instead of each of its florets, pollinated by three or more functional groups of dish-shaped blossoms were the most frequent which one or two were clearly identified as main (43.6%), followed by the brush (16.9%), and tube pollinators (polyphilous), and pollination system was (15.3%) types. Bell (11.3%), and gullet, inconspicu- designated according to the main pollinator. Main ous, flag, and revolver (12.8% altogether) were the pollinators were defined based on pollinator morphol- least represented flowers (Table 2). Most plants had ogy, behavior, and frequency of visits: usually over either actinomorphic (69.4%) or zygomorphic flowers 60% or 80% of visits for one or two main groups, (25.8%), though a small percentage (4.8%) bore both

This content downloaded from 143.106.1.143 on Wed, 20 Aug 2014 17:23:24 UTC All use subject to JSTOR Terms and Conditions 480 Annals of the Missouri Botanical Garden

Table 3. Pollinator species and the plant species it was collected from or observed pollinating (in the case of hummingbirds) at the Serra da Bocaina grasslands. Pollinators that were seen but not collected are included in Table 4. List of the abbreviations of the plant families: AMA, Amaryllidaceae; API, Apiaceae; APO, Apocynaceae; AQU, Aquifoliaceae; AST, Asteraceae; BRO, Bromeliaceae; CAM, Campanulaceae; CLE, Clethraceae; CON, Convolvulaceae; CUN, Cunoniaceae; DRO, Droseraceae; ERIC, Ericaceae; ERIO, Eriocaulaceae; ERY, Erythroxylaceae; ESC, Escalloniaceae; EUP, Euphorbiaceae; GEN, Gentianaceae; GES, Gesneriaceae; HYP, Hypericaceae; IRI, Iridaceae; LAM, Lamiaceae; LEG, Leguminosae; LYT, Lythraceae; MALP, Malpighiaceae; MALV, Malvaceae; MEL, Melastomataceae; OCH, Ochnaceae; ORC, Orchidaceae; ORO, Orobanchaceae; POL, Polygalaceae; RUB, Rubiaceae; SOL, Solanaceae; VER, Verbenaceae; VIO, Violaceae; XYR, Xyridaceae.

Pollinator family and species Plant family Plant species Trochiliformes Trochilidae (hummingbirds) Chlorostilbon aureoventris berlepschi ERIC: Agarista hispidula Pinto Clytolaema rubricauda (Boddaert) GES: Sinningia allagophylla Colibri serrirostris (Vieillot) BRO: Dyckia tuberosa GES: Sinningia allagophylla Leucochloris albicollis (Vieillot) AMA: Hippeastrum glaucescens ERIC: Agarista hispidula GES: Sinningia allagophylla ORO: Esterhazya macrodonta Stephanoxis lalandi lalandi (Vieillot) GES: Sinningia allagophylla Coleoptera (beetles) Buprestidae (jewel beetles) Conognatha Cobos sp. indet. 1 ERIO: Paepalanthus paulensis Cantharidae (soldier beetles) Discodon tucumanum Pic ERIO: Paepalanthus paulensis Cantharidae sp. indet. 1 API: Eryngium canaliculatum AST: Baccharis pentziifolia, Baccharis tarchonanthoides, Barrosoa betoniciiformis, Campuloclinium megacephalum, Chaptalia integerrima, Erigeron maximus, Eupatorium sp. indet. 2, Grazielia gaudichaudeana ERIO: Paepalanthus polyanthus Cantharidae sp. indet. 2 CLE: scabra CUN: Weinmannia organensis Cerambicidae (longhorn beetles) Rhinotragus festivus Perty CLE: Clethra scabra Elateridae (click beetles) Cardiorhinus Eschscholtz sp. indet. 1 API: Eryngium canaliculatum Melyridae (softwinged flower beetles) Astylus sexmaculatus Perty AST: Eupatorium sp. indet. 4 Diptera (flies) Bombyliidae (beeflies) Euprepina Hull sp. indet. 1 AST: Campuloclinium megacephalum Exoprosopa Macquart sp. indet. 1 RUB: Borreria capitata Paravilla Painter sp. indet. 1 AST: Campuloclinium megacephalum, Chaptalia runcinata, Stevia myriadenia, Vernonia megapotamica Culicidae (biting flies) Culicidae sp. indet. 1 API: Eryngium horridum Curtonotidae (curtonotid flies) Curtonotidae sp. indet. 1 API: Eryngium horridum AST: Baccharis platypoda, Mikania sessilifolia Curtonotidae sp. indet. 2 RUB: Galianthe angustifolia Sarcophagidae (flesh flies) Sarcophagidae sp. indet. 1 RUB: Galianthe angustifolia Sarcophagidae sp. indet. 2 ERIO: Paepalanthus paulensis

This content downloaded from 143.106.1.143 on Wed, 20 Aug 2014 17:23:24 UTC All use subject to JSTOR Terms and Conditions Volume 93, Number 3 Freitas & Sazima 481 2006 Pollination in a High-Altitude Grassland

Table 3. Continued.

Pollinatorfamily and species Plant family Plant species Sciaridae (sciarid flies) Sciaridaesp. indet. 1 AST: Baccharisaphylla, Baccharistarchonanthoides ERIO: Paepalanthuspaulensis Syrphidae (hover-flies) Allograptaexotica (Wiedemann) CAM: Wahlenbergiabrasiliensis GEN: Zygostigmaaustrale IRI: sp. indet. 1, Sisyrinchiummicranthum RUB: Galiantheangustifolia Palpada rufipedesThompson APO: Tassadiasubulata Pseudodorosclavatus (Fabricius) RUB: Borreriacapitata, Galianthebrasiliensis Syrphusphaeostigma Wiedemann AST: Baccharisleptocephala Toxomeruswatsoni (Curran) APO: Tassadiasubulata AST: Achyroclinesatureioides, Baccharis leptocephala, Baccharis pentziifolia,Baccharis tarchonanthoides, Chaptalia runcinata, Gochnatiapaniculata, Mikaniasessilifolia, Senecio oleosus CAM: Wahlenbergiabrasiliensis DRO: Droseramontana ERIO: Paepalanthuspaulensis, Paepalanthuspolyanthus EUP: Crotondichrous IRI: Alophiageniculata, Sisyrinchiumvaginatum LYT: Cupheaglutinosa RUB: Borreriacapitata, Galianthebrasiliensis XYR: Xyris asperula,Xyris tortulla ToxomerusMacquart sp. indet. 1 APO: Tassadiasubulata AST: Achyroclinesatureioides, Chaptalia runcinata ERIO: Paepalanthuspaulensis IRI: Sisyrinchiumvaginatum RUB: Borreriacapitata Syrphidaesp. indet. 1 DRO: Droseramontana XYR: Xyris asperula Syrphidaesp. indet. 2 IRI: Alophiageniculata Syrphidaesp. indet. 3 AST: Baccharispentziifolia Syrphidaesp. indet. 4 XYR: Xyris tortulla Syrphidaesp. indet. 5 AST: Mikaniasessilifolia, Stevia myriadenia GEN: Deianira nervosa IRI: Calydoreacampestris Syrphidaesp. indet. 6 AST: Chaptaliaruncinata (tachinidflies) Cylindromyiadorsalis (Wiedemann) AST: Grazieliagaudichaudeana HYP: Hypericumbrasiliense Jurinella cf. corpulenta(Townsend) API: Eryngiumhorridum AST: Baccharisdracunculifolia, Baccharis platypoda, Mikania sessilifolia Tachinidaesp. indet. 1 AST: Eupatoriumsp. indet. 2 RUB: Galianthebrasiliensis Tachinidaesp. indet. 2 CLE: Clethrascabra Tachinidaesp. indet. 3 ERIO: Paepalanthuspaulensis Tachinidaesp. indet. 4 AST: Baccharissp. indet. 2 RUB: Galianthebrasiliensis Tachinidaesp. indet. 5 AST: Eremanthuserythropappus Tachinidaesp. indet. 6 AST: Vernoniawestiniana Tachinidaesp. indet. 7 RUB: Galiantheangustifolia Tachinidaesp. indet. 8 AST: Achyroclinesatureioides, Baccharis dracunculifolia, Symphyopappuscompressus Tachinidaesp. indet. 9 API: Eryngiumhorridum Tephritidae (fruitflies) TrupaneaSchrank sp. indet. 1 RUB: Galiumhypocarpium

This content downloaded from 143.106.1.143 on Wed, 20 Aug 2014 17:23:24 UTC All use subject to JSTOR Terms and Conditions 482 Annals of the Missouri Botanical Garden

Table 3. Continued.

Pollinator family and species Plant family Plant species

Family indet. (flies) Diptera sp. indet. 1 API: Eryngium horridum Diptera sp. indet. 2 AST: Chaptalia runcinata Diptera sp. indet. 3 AST: Chaptalia runcinata Diptera sp. indet. 4 AST: Baccharis dracunculifolia Diptera sp. indet. 5 ERIO: Paepalanthus polyanthus Diptera sp. indet. 6 AST: Erigeron maximus Diptera sp. indet. 7 CLE: Clethra scabra Diptera sp. indet. 8 ERIO: Paepalanthus polyanthus Andrenidae, Panurginae (dagger bees) Anthrenoides Ducke sp. indet. 1 VIO: Viola cerasifolia , Apini (honey bees) Apis mellifera Linnaeus AQU: Ilex amara AST: Baccharis dracunculifolia, Baccharis pentziifolia, Baccharis platypoda, Baccharis sp. indet. 2, Barrosoa betoniciiformis, Campuloclinium megacephalum, erythropappus, Eupatorium sp. indet. 2, Grazielia gaudichaudeana, Hypochaeris gardneri, Mikania lundiana, Stevia myriadenia, Symphyopappus compressus, Vernonia westiniana CAM: Wahlenbergia brasiliensis CLE: Clethra scabra CON: Convolvulus crenatifolius ERIC: Gaylussacia chamissonis, Gaylussacia jordanensis EUP: Croton dichrous IRI: Sisyrinchium vaginatum LAM: Hyptis lippioides, Hyptis plectranthoides, Hyptis umbrosa, Peltodon radicans LYT: Cuphea glutinosa RUB: Borreria capitata, Galianthe angustifolia, Galianthe brasiliensis VER: Verbena hirta Apidae, Bombini () Bombus atratus Franklin APO: sublanatum AST: Campuloclinium megacephalum, Chromolaena cf. decumbens, Stevia myriadenia, Symphyopappus compressus, Vernonia tomentella, Vernonia westiniana CAM: Lobelia camporum CON: Ipomoea procumbens, Jacquemontia grandiflora IRI: Alophia geniculata LAM: Hyptis lippioides, Hyptis plectranthoides LEG: Crotalaria breviflora, Lupinus velutinus LYT: Cuphea glutinosa MEL: Tibouchina frigidula, Tibouchina minor, Trembleya phlogiformis RUB: Declieuxia cordigera, Galianthe angustifolia VER: Verbena hirta Bombus brasiliensis Lepeletier AST: Stevia myriadenia, Vernonia westiniana LEG: Crotalaria breviflora LYT: Cuphea glutinosa VER: Verbena hirta Bombus morio (Swederus) AST: Vernonia westiniana LYT: Cuphea glutinosa RUB: Galianthe angustifolia VER: Verbena hirta Apidae, Centridini (digger bees) Centris burgdorfi Friese CAM: Lobelia camporum RUB: Declieuxia cordigera

This content downloaded from 143.106.1.143 on Wed, 20 Aug 2014 17:23:24 UTC All use subject to JSTOR Terms and Conditions Volume 93, Number 3 Freitas & Sazima 483 2006 Pollination in a High-Altitude Grassland

Table 3. Continued.

Pollinator family and species Plant family Plant species

Centris discolor Smith MALP: Byrsonima variabilis MEL: Tibouchina frigidula Centris cf. insularis Smith MALP: Byrsonima variabilis Centris klugi Friese AST: Hypochaeris gardneri, Vernonia cf. rosea, Vernonia tomentella LEG: Chaemaecrista sp. indet. 1 LYT: Cuphea glutinosa MALP: Byrsonima variabilis MEL: Tibouchina minor Centris tarsata Smith MALP: Byrsonima variabilis VER: Verbena hirta Centris (Melacentris) Moure sp. indet. 1 MALP: Byrsonima variabilis Apidae, Ceratinini (dwarf carpenter bees) Ceratina cf. asuncionis Strand AST: Senecio oleosus, Vernonia herbacea, Vernonia tomentella CAM: Lobelia camporum MALV: Sida sp. indet. 1 VER: Verbena hirta Ceratina Latreille sp. indet. 1 LYT: Cuphea glutinosa Ceratina sp. indet. 2 CON: Convolvulus crenatifolius Apidae, Ericrocidini Mesonychium caerulescens Lepeletier AST: Vernonia tomentella & Serville LYT: Cuphea glutinosa RUB: Galianthe angustifolia VER: Verbena hirta Apidae, Eucerini (long-horned bees) Gaesischia nigra Moure AST: Vernonia tragiaefolia Melissoptila aureocincta Urban AST: Symphyopappus compressus Apidae, Meliponini (stingless bees) Melipona bicolor bicolor Lepeletier AST: Eremanthus erythropappus, Eupatorium sp. indet. 3, Grazielia gaudichaudeana, Mikania lundiana, Vernonia megapotamica, Vernonia tomentella, Vernonia westiniana CLE: Clethra scabra ERIC: Gaylussacia chamissonis IRI: Alophia geniculata LAM: Hyptis lippioides MEL: Trembleya parviflora OCH: Ouratea semiserrata Melipona quadrifasciata anthidioides AST: Campuloclinium megacephalum, Eremanthus erythropappus, (Lepeletier) Mikania lundiana, Symphyopappus compressus Paratrigona subnuda Moure AST: Chaptalia integerrima, Eremanthus erythropappus, Mikania lundiana Plebeia droryana (Friese) AST: satureioides Plebeia saiqui (Friese) API: Eryngium horridum AQU: Ilex amara AST: Vernonia westiniana ERI: Gaylussacia jordanensis IRI: Alophia geniculata, Calydorea campestris LAM: Hyptis lippioides MALP: Byrsonima variabilis Scaptotrigona bipunctata (Lepeletier) CLE: Clethra scabra Schwarziana quadripunctata AST: Baccharis sp. indet. 2, Eremanthus erythropappus, Eupatorium sp. (Lepeletier) indet. 4, Grazielia gaudichaudeana, Mikania lundiana CLE: Clethra scabra HYP: Hypericum brasiliense

This content downloaded from 143.106.1.143 on Wed, 20 Aug 2014 17:23:24 UTC All use subject to JSTOR Terms and Conditions 484 Annals of the Missouri Botanical Garden

Table 3. Continued.

Pollinator family and species Plant family Plant species

Trigona spinipes (Fabricius) AST: Baccharis sp. indet. 2 ERIO: Paepalanthus paulensis LAM: Hyptis lippioides Apidae, Tapinotaspidini (digger bees) Monoeca Lepeletier & Serville sp. MALP: Byrsonima variabilis indet. 1 Paratetrapedia (Lophopedia) cf. AST: Stevia myriadenia pygmaea (Schrottky) XYR: Xyris tortulla Paratetrapedia (Xanthopedia) MALP: Byrsonima variabilis Michener & Moure sp. indet. 1 Paratetrapedia (Trigonopedia) Moure AST: Stevia myriadenia sp. indet. 1 Apidae, Xylocopini (giant carpenter bees) Xylocopa brasilianorum (Linnaeus) AST: Symphyopappus compressus CON: Ipomoea procumbens MEL: Tibouchina frigidula, Tibouchina martialis VER: Verbena hirta Braconidae (braconid wasps) Braconidae sp. indet. 1 AST: Baccharis platypoda Chrysididae (cuckoo wasps) Chrysididae sp. indet. 1 AST: Baccharis dracunculifolia Colletidae, Colletiiii (plasterer bees) Colletes Latreille sp. indet. 1 AST: Hypochaeris gardneri ESC : Escallonia farinacea Eurytomidae (eurytomid wasps) Eurytomidae sp. indet. 1 RUB: Galium hypocarpium Gasteruptiidae (gasteruptid wasps) Gasteruptiinae sp. indet. 1 EUP: Croton dichrous Halictidae, Augochlorini (sweat bees) Augochlorodes turrifaciens Moure VIO: Viola cerasifolia Augochloropsis cf. cognata Moure AQU: Ilex amara AST: Campuloclinium megacephalum, Vernonia tomentella CLE: Clethra scabra EUP: Croton dichrous IRI: Calydorea campestris LAM: Hyptis lippioides MALP: Byrsonima variabilis MALV: Pavonia kleinii RUB: Borreria capitata, Galianthe brasiliensis Augochloropsis cyanea (Schrottky) AST: Barrosoa betoniciiformis, Campuloclinium megacephalum, Chaptalia runcinata, Eupatorium sp. indet. 4, Grazielia gaudichaudeana, Vernonia tomentella ERIC: Gaylussacia jordanensis EUP: Croton dichrous LYT: Cuphea glutinosa RUB: Galianthe angustifolia, Galianthe brasiliensis SOL: Solanum aculeatissimum, Solanum pseudocapsicum, Solanum swartzianum, Solanum viarum XYR: Xyris asperula Augochloropsis iris (Schrottky) AST: Campuloclinium megacephalum IRI: Alophia geniculata Augochloropsis Cockerell sp. indet. 1 AST: Eupatorium sp. indet. 2, Stevia myriadenia

This content downloaded from 143.106.1.143 on Wed, 20 Aug 2014 17:23:24 UTC All use subject to JSTOR Terms and Conditions Volume 93, Number 3 Freitas & Sazima 485 2006 Pollination in a High-Altitude Grassland

Table 3. Continued.

Pollinatorfamily and species Plant family Plant species ERIC: Gaylussaciachamissonis, Gaylussacia jordanensis LAM: Hyptisumbrosa LYT: Cupheaglutinosa MEL: Trembleyaphlogiformis POL: Polygala cneorum RUB: Galianthebrasiliensis SOL: Solariumpseudocapsicum CeratalictusMoure sp. indet. 1 AQU: Ilex amara AST: Vernoniawestiniana ERIC: Gaylussaciajordanensis ERY: Erythroxylummicrophyllum GEN: Zygostigmaaustrale IRI: Alophia sp. indet. 1 LAM: Hyptislippioides, Hyptis plectranthoides, Hyptis umbrosa RUB: Borreriacapitata, Borreriatenella, Galianthebrasiliensis Paroxystoglossacf. jocasta IRI: Alophiageniculata (Schrottky) ParoxystoglossaMoure sp. indet. 1 AST: Chaptaliaintegerrima Pseudaugochloracf. graminea VER: Verbenahirta (Fabricius) Augochlorinisp. indet. 1 RUB: Declieuxiacordigera Halictidae, Halictini (sweat bees) AgapostemonGuerin-Meneville sp. AST: Symphyopappuscompressus indet. 1 Dialictus Robertsonsp. indet. 1 ERIC: Gaylussaciajordanensis IRI: Alophiageniculata, Calydoreacampestris, Sisyrinchium micranthum OCH: Ourateasemiserrata VIO: Viola cerasifolia Dialictus sp. indet. 2 AST: Chaptaliaruncinata Pseudagapostemoncyaneus Moure & AST: Mikania lundiana, Seneciooleosus Sakagami Halictini sp. indet. 1 HYP: Hypericumbrasiliense Halictini sp. indet. 2 IRI: Alophiageniculata VER: Verbenahirta Leucospidae (leucospidwasps) Leucospidaesp. indet. 1 AST: Baccharisintermixta Megachilidae, Anthidiini (mason and carderbees) Anthidiumsertanicola Moure & LEG: Crotalariabreviflora Urban LYT: Cupheaglutinosa Epanthidiumautumnale (Schrottky) AST: Stevia myriadenia Megachilidae, Megachilini (- cutter, mason and cuckoo bees) CoelioxysLatreille sp. indet. 1 AST: Erigeronmaximus, Senecio oleosus Coelioxyssp. indet. 2 AST: Stevia myriadenia Megachilecf. anthidioides AST: Symphyopappuscompressus Radoskowski LYT: Cupheaglutinosa Megachileiheringi Schrottky AST: Eupatoriumsp. indet. 1, Vernoniatomentella ERIC: Gaylussaciajordanensis HYP: Hypericumbrasiliense LEG: Lupinusvelutinus Megachilelaeta Smith LEG: Lupinusvelutinus VER: Verbenahirta

This content downloaded from 143.106.1.143 on Wed, 20 Aug 2014 17:23:24 UTC All use subject to JSTOR Terms and Conditions 486 Annals of the Missouri Botanical Garden

Table 3. Continued.

Pollinator family and species Plant family Plant species

Megachile terrestris Schrottky AST: Eupatorium sp. indet. 1, Vernonia tomentella Pompilidae (spider wasps) Pompilidae sp. indet. 1 AST: Baccharis sp. indet. 2 Pompilidae sp. indet. 2 AST: Baccharis platypoda Pompilidae sp. indet. 3 AST: Baccharis sp. indet. 2 Pompilidae sp. indet. 4 AST: Baccharis sp. indet. 2 Scoliidae (digger wasps) Scoliidae sp. indet. 1 AST: Symphyopappus compressus, Vernonia westiniana Sphecidae (mud-dauber and sphecid wasps) Bicyrtes paranae Bohart RUB: Galianthe angustifolia Pryonyx thomaz (Fabricius) AST: Baccharis dracunculifolia Sphex dorsalis Lepeletier RUB: Galianthe angustifolia Sphex opacus Dahlbom AST: Eupatorium sp. indet. 4 Cercerini sp. indet. 1 CLE: Clethra scabra Cercerini sp. indet. 2 AST: Mikania nummularia CLE: Clethra scabra Larrinae sp. indet. 1 AST: Tenthredinidae (sawflies) Tenthredinidae sp. indet. 1 RUB: Galianthe brasiliensis Tiphiidae (tiphiid wasps) Tiphiidae sp. indet. 1 AST: Baccharis sp. indet. 2 Vespidae (paper wasps) Agelaia vicina (Saussure) AST: Baccharis intermixta ERIO: Paepalanthus polyanthus Brachygastra lecheguana (Latreille) AST: Baccharis sp. indet. 3 Mischocyttarus drewseni Saussure AQU: Ilex amara AST: Achyrocline satureioides, Baccharis aphylla, Baccharis dracunculifolia, Baccharis pentziifolia, Baccharis sp. indet. 2, Chaptalia integerrima, Eupatorium sp. indet. 4, Gochnatia paniculata CLE: Clethra scabra ERY: Erythroxylum microphyllum EUP: Croton dichrous RUB: Galianthe brasiliensis Polistes billardieri Fabricius AQU: Ilex amara AST: Achyrocline satureioides, Baccharis sp. indet. 2, Baccharis sp. indet. 3, Gochnatia paniculata, Mikania sessilifolia ERY: Erythroxylum microphyllum RUB: Galianthe brasiliensis Polistes cinerascens Saussure APO: Gonioanthela hilariana AST: Baccharis dracunculifolia, Baccharis platypoda Polybia fastidiosuscula Saussure AST: Baccharis platypoda, Baccharis sp. indet. 2 CLE: Clethra scabra RUB: Galianthe angustifolia Polybia minarum Ducke CLE: Clethra scabra CUN: Weinmannia organensis Polybia scutellaris (White) AST: Baccharis tarchonanthoides EUP: Croton dichrous Polybia sericea (Olivier) APO: Gonioanthela hilariana, Oxypetalum appendiculatum RUB: Borreria tenella Protonectarina sylveirae (Saussure) AST: Baccharis curitybensis Protopolybia sedula (Saussure) AST: Baccharis sp. indet. 2 Synoeca cyanea (Fabricius) CLE: Clethra scabra Family indet. (wasps) Hymenoptera sp. indet. 1 AST: Achyrocline satureioides, Mikania nummularia Hymenoptera sp. indet. 2 AST: Lucilia lycopodioides, Mikania nummularia

This content downloaded from 143.106.1.143 on Wed, 20 Aug 2014 17:23:24 UTC All use subject to JSTOR Terms and Conditions Volume 93, Number 3 Freitas & Sazima 487 2006 Pollination in a High-Altitude Grassland

Table 3. Continued.

Pollinator family and species Plant family Plant species

ERY: Erythroxylum microphyllum RUB: Borreria capitata, Borreria tenella Hymenoptera sp. indet. 3 AST: Mikania sessilifolia LYT: Cuphea glutinosa Hymenoptera sp. indet. 4 AST: Baccharis aphylla Hymenoptera sp. indet. 5 AST: Baccharis pentziifolia Hymenoptera sp. indet. 6 AST: Eupatorium sp. indet. 2 Hymenoptera sp. indet. 7 AST: Achyrocline satureioides Hymenoptera sp. indet. 8 AST: Achyrocline satureioides Hymenoptera sp. indet. 9 AST: Achyrocline satureioides Hymenoptera sp. indet. 10 AST: Baccharis pentziifolia CLE: Clethra scabra Hymenoptera sp. indet. 11 LYT: Cuphea glutinosa Hymenoptera sp. indet. 12 RUB: Galianthe brasiliensis Hymenoptera sp. indet. 13 AST: Grazielia gaudichaudeana Hymenoptera sp. indet. 14 AST: Baccharis dracunculifolia Hymenoptera sp. indet. 15 RUB: Galianthe angustifolia (butterflies) Hesperiidae (skippers and darters) Sarbia cf. damippe (Mabille & AST: Barrosoa betoniciiformis, Campuloclinium megacephalum Boullet) Sarbia cf. Xanthippe (Latreille) AST: Vernonia westiniana Thespieus Godman sp. indet. 1 AST: Campuloclinium megacephalum Urbanus Hiibner sp. indet. 1 VER: Verbena hirta Hesperiinae sp. indet. 1 AST: Grazielia gaudichaudeana Pyrrhopyginae sp. indet. 1 AST: Chaptalia integerrima, Vernonia westiniana Lycaenidae (coppers and hair streaks) Thecla Fabricius sp. indet. 1 AST: Chaptalia runcinata, Mikania nummularia, Vernonia westiniana RUB: Galianthe brasiliensis Riodininae sp. indet. 1 API: Eryngium horridum Nymphalidae (danaids and browns) Agraulis vanillae maculosa (Stichel) AST: Vernonia megapotamica Vanessa myrinna (Doubleday) AST: Campuloclinium megacephalum, Eupatorium sp. indet. 1, Eupatorium sp. indet. 4, Vernonia tomentella Vanessa Fabricius sp. indet. 1 AST: Eupatorium sp. indet. 3 Yphthimoides ochracea (Butler) AST: Symphyopappus compressus CLE: Clethra scabra Pieridae (whites and yellows) Eurema nise tenella (Boisduval) LYT: Cuphea glutinosa VER: Verbena hirta Hesperocharis erota (Lucas) AST: Vernonia westiniana

types. Small flowerswere predominant.Out of the 65 The great majorityof species presented hermaph- species with tubular or gullet flowers, for instance, roditic flowers (82.3%), and a few were either 55.4% had corolla tubes ^ 5 mm long, whereasonly monoecious (3.2%) or dioecious (8.9%) (Table 2). 13.8% had tubes >10 mm long. Nectar was the The latter group only comprised the 11 species of exclusive or main resource of 71.0% of the species. Baccharis Mill. (Asteraceae).Seven other species of Althoughpollinators fed on or collected pollen from Asteraceae(5.6%) presented both hermaphroditicand more than half of the species, it was the exclusive unisexual (ligulate) flowers (gymnomonoecism).We resourceof only 17.7% of the species. Even thoughwe detected spontaneousself-pollination in some species only measuredthe nectarproduction of a few species, including plants that were also pollinatedby low nectarproduction (i.e., less than 2 (il) prevailed. (Table 4).

This content downloaded from 143.106.1.143 on Wed, 20 Aug 2014 17:23:24 UTC All use subject to JSTOR Terms and Conditions 488 Annals of the Missouri Botanical Garden

c« oT II ^ -k . <8 o\ t>~ d-

S ^1 a .5 C c -^

-S '3 11&£ rf 8 « 1 ^S «§li1 ] "£ * d. ? "° "2 *-§ «s3 9.5'"-a S § 4 :|j-s 1 I 'a -: =1 1 s- s m 1 \ti.i 1 I »H MHi 11

m * i inn i .,s in Hie pi

•Sgl I aliju "ej-eJJl 111 2|<4lMl*

Life =3^^ -a « 11 S § 1S| Ij S"O _§ "S "S O O OOOO Q QO

^3 O S °* S 3 CO Jh ^ ^ « "a g Oh a S H 9 £ o sh o jjh *«&» & !f -b 1 li I au * 1 ii niji j 1 11 " S « .S § 1 s ^ S3 4i li O * ^_^. 1/5 Z LO ^C^coo^LO 1/^ -- ! 2 Q .S *« § "cd'*^ 10 00 i> ifl ^ vo »Tf «r!!, i 1 I *

^ i n ^ S Scfi ,-a

m fill! lillH 52 .J ! ill 'uili 111111 1? n ii

This content downloaded from 143.106.1.143 on Wed, 20 Aug 2014 17:23:24 UTC All use subject to JSTOR Terms and Conditions Volume 93, Number 3 Freitas & Sazima 489 2006 Pollination in a High-Altitude Grassland

^ s ! * ' ^ -r •» 8 -a & •IH ^ <-^ CO <3 • p^T "£ "if S CD MC^CO^rfl^r lit I Is- *ii 3!l ilil 1 I^IJ-ll III IB lu Hi IH ^ili-: |"i!J2f;:|l , ill i\ !8M Hi fffli iiiiitiJif 111 HJfi ! lij'LII UlJiil ifiilisiii J Ml JfliiJSililJ!!!!Mmtti lilfll J[J(H

I g

i ^

O4 O, Oh g

.2 g g g

a.Oco « cocnco co «S: toco 5^^co^S > &££^ > > S> co >^^ > > co co

'I ^ ^ ^ I I I !! 1 t J il ill w -i5 -S -S coco 'S ^3 a, Cl, cC»o>':3 j I -e -b -s -e -e -b -e -b -b § 3 1 I III I III I 1 §.

This content downloaded from 143.106.1.143 on Wed, 20 Aug 2014 17:23:24 UTC All use subject to JSTOR Terms and Conditions 490 Annals of the Missouri Botanical Garden

I

fJJ^! II^I ^|^*ia^l|ljllf|i|;^ i«J !i«ii * lit!! ill??? illl UIU Ifll !°i!l| liUf Ifi Ill-ill in « !iIlaH*:!:1?4fll=|il|*i1i#ljis It ^ililfiJlJ^Mllli^'ISIil^ilg15!! PQ PQ CQ C< CQPQ^f CQpq CD

1 | §

li s § i-ig g § g ^§ ^

Oh O, O, Oh Oh ,,,3 3 3 3 3 3 £O O ^hO O^hO !hO £ 6b & I 6b 6b I 6b I 6b C/3^-1 *J Q+J +JQ+J O+J C CU QJ i"^ CU CU i"^ QJ i^ CU Oco en j_i co co^hco ^co * i I ill I I I li I

-§ ? io io in

' ' cu "-

I 3 § 1 .8 I Z Z "3

coco O ^^ 5s« eS^ ^coco Q

I 1 li Jll -III II 1

This content downloaded from 143.106.1.143 on Wed, 20 Aug 2014 17:23:24 UTC All use subject to JSTOR Terms and Conditions Volume 93, Number 3 Freitas & Sazima 491 2006 Pollination in a High-Altitude Grassland

.iii in.? si « oil "lilj *$ pyiliffiiliiteltlHil^^H-s-? I tin nis-H ii i;sliiiiii!ii«?i!i

?piiri!i!lf!-||IIiiS!tl:lil?il|l^ flilttltliy- HI il!l|tlil=tt1ll!]i

II S |g3 Sa nS S 8 Zio c § ^i2 d £ £ § § d & £

aj en en C Oh , Oh & 0) O *h >-i cdo O !i I. Bb I i 1 a. - |

OcCSh^h j_)CC CO s_i I i I|I I I ij i i I ^ i iii ! I r i i j

O U LO LO lO LO LO LO cd t+H § tO LO

§ I 1 II 41 § I I -3 i 3 1 t I 111 t 3 1 i I t "I 111 1 I I ! I I

This content downloaded from 143.106.1.143 on Wed, 20 Aug 2014 17:23:24 UTC All use subject to JSTOR Terms and Conditions 492 Annals of the Missouri Botanical Garden

111 ity 1313*111 1 la I|1 II f II ill Iffliiilitli f i Stilt i! liislijiii^li^lii * |; i.t%iiHti I l-i .5115 M^ 1H 8-J-i ! i|?l S1ili1^:| flijf'illji^il"1!!- 3 Jill lilPllP s ^a s* s % s- *

O

|j

S S

0 3 (N rf CO ^CO OI>iO O -- • r-l rH o O

H

8 1 * J I « =3a « J 1 II II t:i lii . Ii *s1 si 8 -M f "S>« 1 fa ^ "S 'S "S "3 "3 £ a 1 .§ § § S I III II !^|3^!^

This content downloaded from 143.106.1.143 on Wed, 20 Aug 2014 17:23:24 UTC All use subject to JSTOR Terms and Conditions Volume 93, Number 3 Freitas & Sazima 493 2006 Pollination in a High-Altitude Grassland

I l!« I «. lillii! jalJfji |jj i . in i i miiu mam im ! * 111 s * li ti« | *! ila j-i ii;

s 1 III! I 5 Hills sjjsiils illi

IHlii li Jisilll ! -:1-:llS llll

*S J CD O >-< W '§ 'S en O cd

11 8 ga SggS^ S S 8 i I

a S. , S. S. fa i 3 fi CU 3 3 I. 1 S ? - ^ ^ S I ^ &

.S3^^o; 13 ^^-go^ 13 13 cs | -|^ s S I "!« 1 S S § g 1

H

s •s i i sf 1 1 fill 1 £ „ I 1 *> I" llll s. ^ s^ * mi 1 1 in iiii ii i if u u Q w w w

This content downloaded from 143.106.1.143 on Wed, 20 Aug 2014 17:23:24 UTC All use subject to JSTOR Terms and Conditions 494 Annals of the Missouri Botanical Garden

s ' 1 ao - cc 3 ( ^

lit i I S- -a-15 i^

11 -a 1 1 * 1 1 1 £ H§. 1 1 § i »J1 l ! ii * it tili\

i -s -g£ T - e-J 1 ^ 1 a | » - j * fe * ^ lafl I "I .§-8 tS °:-S.l|PSil* Jil li i 1 Hi! IHillS 1 tili Hi iL 11 It f!S Jiii CQ^ C fl K fl CQ ffi CQ ^CQ P3CQ ffl

o is II I 1 -i-ii is I I i i i§ §

Is gg g|ls §§ I ! * *Mi*ii | If ! I fill! l | ) ||f=|l= j Is I j ft=f=

en ^ 'o 2 to to ^ trj uo

^gvoo oncovoooo i> co r-n r-cd1^1 O I- I CO CO I- I a, ^-*J5 - (-sq

5 |

'I ^ „1 • s ^i tl I 1 I I II '£ : I til ill! li II n j.S ii I I 1UJ 1331 ii U U ill U *

This content downloaded from 143.106.1.143 on Wed, 20 Aug 2014 17:23:24 UTC All use subject to JSTOR Terms and Conditions Volume 93, Number 3 Freitas & Sazima 495 2006 Pollination in a High-Altitude Grassland

I a t" 1 & 1 $ , W ^ -B I 1-8 | 3 fl -SJ-1 £'§ I

- « g "§ • § .§ •£ ~e -^ < +j 2 ^ § ^ ii * a ii irntJ i* si i

1 13 11 f If I, lUiH W ii i ?! iij is «i iifiii \\u n ?! SIH ^1 III ? 11-111 3t«Sl 2j

|| 1:1,11111 Jill ipill ^l|| la I |. I- |. I- ^ ^^^ I- a ££ 1

o 1 1 •§ ts o o oqooqo q oo

& E

S c

cfio, O OOOOOO O »h OO <^- ca .« CaT^« ^q^^q^^q^iiiiii o| qj3q^ - "5 .2^3 Q p_|Cu i CoCoCu^Cu^j CoJa p---i CoJi Co£ £-< . ph

^^OC s3C3COhS3O,C^C O C fl J & oPh^h; cu

a;

H

ill I Is ^3 I 1 I I l.iU I «"§ HI

" I t 11 I i J lit i it illii

This content downloaded from 143.106.1.143 on Wed, 20 Aug 2014 17:23:24 UTC All use subject to JSTOR Terms and Conditions 496 Annals of the Missouri Botanical Garden

i- i i I -g I <§ <§ •8 I 3 ^ § a is is 3 O w ° Si

£ |,-8 !¥- -s 1 ^ -a -s

1 J Hillil S III T3 ^C T3 'TD "S o ^ •&• O §- FT3FT3^F^rX3 8 ^ O £ £ £ £ Sf=!'-2a.^i-^^o ^ ££$,££ a « *o '53 'S 'S *cfi "S '^ "^ 'co »>3 Qq O ><| 3 ^^ -^ £S ^~> ^^ § ^ '> "> *> •> . . , , , > '> "> '> g ^ . ^ ^g •g "g "S O ^ iJ CD C/)C/) CD OOCOO £9 CD CD

1-2 •3 e3 OOOOOO OOO ll -i 1 1 1 i §g i§ § mii i i i

Q

1en o- o. o. o. o% O% O% O% O% O% __, Oi Oi c .2 -S -S .S ^r ^r ^r ^r ^v ^v ^ ^r ^ I el el el cl s s s s s s e ^ 3 a

DhC C S3 S3

«+h§\O rH \Q *$ Tf ^O OC^ C^ ^ fO O\ * ^f (N vd O O is 1- I i-HC<1 i- I f- I 1- I 1- I r- H rH t~- r- I

1 Ji i i t t ii |! ! lUliJ J 1

j 1 1 i 1 1 1 1 nil |IJ11istil I OO OP*

This content downloaded from 143.106.1.143 on Wed, 20 Aug 2014 17:23:24 UTC All use subject to JSTOR Terms and Conditions Volume 93, Number 3 Freitas & Sazima 497 2006 Pollination in a High-Altitude Grassland

Jill I ?a li^J.f J^lifl <

i si! iMMmmm i s

a"!i i| i §ii^£^|:|s-i ailiill & Hill

o

1•q^o 1 Q o o qooooo ^13o joO O QOOOOO O C

Q

fl Oh Oh Oh DO MO O ^hMMMM

2^2 b^2 23 ccccc

_ -^^10I 10 10101010 5+H^cd lOl^^ 00 rH ON 00 irf Tt 3 O (N 1- 1 NJ5 0) - '

III 1 ! ! I l illli I 1! fill 1 1 !!!!!!

This content downloaded from 143.106.1.143 on Wed, 20 Aug 2014 17:23:24 UTC All use subject to JSTOR Terms and Conditions 498 Annals of the Missouri Botanical Garden

Jill I ^ £-| ^ § g | 1 J-S^w^ -I ^.t % 1 * .a J. ?. & 8-JT s

1 Jiifll 5 i^ I! 1 Hill! 1. {?.i§^

tl^Jl II !WJ! «, 8

o I § •pi "cd O O O § S S II Ph |§ O O ^ I

I>^ I^ g^ Jo Jo S

° g O ^ vo ^ J

H

en § -5 I * ^ | s | ||| J g I s ^ J -g -g

This content downloaded from 143.106.1.143 on Wed, 20 Aug 2014 17:23:24 UTC All use subject to JSTOR Terms and Conditions Volume 93, Number 3 Freitas & Sazima 499 2006 Pollination in a High-Altitude Grassland

Table 5. Numberof plant species exclusively,mainly, and secondarilypollinated by each pollinatorgroup at the Serrada Bocaina grasslands.Percentages are given in parentheses(n = 106 plant species). Because many species are pollinatedby two to six pollinatorgroups the sum of the totals exceeds 100%.

Role in pollination

Pollinatorgroup Exclusive Main* Secondaryor indistinct Total Small bees § 12 (11.3%) 21 (19.8%) 38 (35.9%) 71 (67.0%) Wasps 7 (6.6%) 15 (14.2%) 26 (24.5%) 48 (45.3%) Largebees § 8 (7.6%) 18 (17.0%) 10 (9.4%) 36 (34.0%) Syrphids 3 (2.8%) 12 (11.3%) 18 (17.0%) 33 (31.1%) Otherflies 0 3 (2.8%) 26 (24.5%) 29 (27.3%) Butterflies 0 2 (1.9%) 23 (21.7%) 25 (23.6%) Beetles 0 0 13 (12.3%) 13 (12.3%) Hummingbirds 5 (4.7%) 0 0 5 (4.7%) * Includes plant species that have multiple pollinatorgroups but only one or two are the main pollinatorgroup(s) (see methodssection of text). § Apis melliferanot included.

B. POLLINATORAGENTS them. We recordedat least 46 species of wasps acting as pollinators,the most importantof which were the A wide of were recorded array pollinator agents social Vespidae, in particularPolistes Latr. species, followed (Tables3, 4). Hymenopterans, by dipterans, followed by Pompilidaeand Sphecidae. were the most and all insect importantpollinators, The introducedhoneybee, Apis mellifera,acted as bees were found more often as groups except large a pollen robber in Sinningia allagophylla but than as exclusive or main secondaryagents pollination pollinatedthe other 31 species they visited (Table 3). which were the agents (Table 5). Hummingbirds, only These bees were frequent visitors to some species vertebrate acted as exclusive group, pollinators classified as having bee nectar-flower pollination (Table 5). systems (i.e., pollinated by bees in search of nectar; Hymenoptera. Bees and/orwasps were amongthe see "PollinationSystems" below), such as Ilex amara, pollinator agents of 98 of the 101 plant species Gaylussaciachamissonis, G. jordanensis, and Cuphea pollinated by insects. In addition, hymenopterans glutinosa. Honeybees were the most important were the pollinators of three-fourthsof the plants pollinatorsof the three HyptisJacq. species (Lamia- pollinatedby one grouponly (Table 5). We recorded ceae) and the only flower visitor for the other at least 54 native species of bees (Table 3). The Lamiaceae,Peltodon radicans. families Apidae (including "Anthophoridae"),Halic- Diptera. Syrphidsconstituted a noticeable group tidae, and Megachilidaewere representedby 28, 16, of pollinators in the community, feeding on both and 8 species, respectively. Andrenidaeand Colleti- nectar- and pollen-flowers. Species of the genus dae were only represented by one species each. ToxomerusMacq. were particularlyimportant and, - Halictidae bees owing mainly to Augochlorini for instance, T. watsoni acted as a pollinatorof 21 - species pollinate about half of the species in this species (Table 3). Flies other than syrphids were community.Meliponini bees played a minor role in representedmainly by Tachinidae and Bombyliidae pollination, occasionally pollinating some generalist (beeflies) and, usually, were secondary or indistinct plant species, and acted more as pollen thieves in pollinators of polyphilous or holophilous species several species (e.g., Baccharisand Melastomataceae (Fig. 2B, Tables 5, 6). - species, Fig. II). However,Melipona bicolor which Lepidoptera. Pollinationby lepidopterans(Fig. 21) is markedlylarger than other Meliponinibees in that was poorly represented. Butterflies ("Rhopalocera") area- was an importantpollinator of several species and diurnal moths ("Heterocera")generally played (Fig. IB, Table 3). Small and large bees belongingto a secondaryrole for plants pollinatedmainly by bees Megachilidaepollinated 12 species fromsix families. (Tables 4, 5, 6). They were amongthe most frequent The large Apidae bees belonging to the genera pollinatorsof only a few Asteraceae species. Butter- Bombus,Xylocopa Latr., and CentrisFabr. also were flies belonging to the genus VanessaFabr. (Nympha- importantpollinators in the community.The three linae) were the most frequent flower visitors of this species- mainly Bombus atratus- polli- group.In spite of the scarcityof their visits to flowers, nated at least 29 species, and they were either the many butterflies were recorded during the summer most frequent or the exclusive pollinators of 12 of and fall, and small butterflies belonging to Hesper-

This content downloaded from 143.106.1.143 on Wed, 20 Aug 2014 17:23:24 UTC All use subject to JSTOR Terms and Conditions 500 Annals of the Missouri Botanical Garden

Figure 1. Flowersand pollinatorsat the Serrada Bocaina grasslands.A-C. Bee nectar-flowers.- A. Augochloropsisaff. cognata (Halictidae)visiting a flowerof Pavonia kleinii (Malvaceae).- B. A stingless bee Meliponabicolor (Apidae) entering a bell-shapedflower of Gaylussaciachamissonis (Ericaceae). - C. Bombusbrasiliensis (Apidae) visiting a head of Vernonia westiniana(Asteraceae). D, E. Wasp/flyflowers. - D. Polistes billardieri(Vespidae) visiting a head of Gochnatiapaniculata (Asteraceae),which is also pollinatedby syrphids.- E. Polybia sericea(Vespidae) visiting a flowerof Gonioanthelahilariana (Apocynaceae).F-J. Bee pollen-flowers.- F. A flower of Tibouchinaminor (Melastomataceae).Note the long poricidal anthers.- G. Augochloropsissp. indet. 1 collecting pollen by vibrationfrom a single anther of a Trembleyaphlogiformis (Melastomataceae)flower. This bee only occasionallytouches the stigma (arrow).- H. Bombusatratus collecting pollen by vibrationin Trembleyaphlogiformis, which is mainly pollinated by this bee. - I. Paratrigonasubnuda (Apidae) collecting pollen froma flowerof Trembleyaparviflora. This bee removespollen fromthe antherby insertingits proboscisand acts as pollen thief since it does not transferpollen to the stigma of this species, which is pollinated by another stingless bee, Meliponabicolor. - J. Megachileiheringi (Megachilidae)collecting pollen from a keel-shaped flower of Lupinusvelutinus (Leguminosae).Note the bee abdomentouching the keel petals (arrow).K, L. Syrphid/smallbee pollen-flowers.- K and L. Syrphidsspecies feeding directly on pollen fromflowers of Xyris tortulla (Xyridaceae)and Calydoreacampestris (Iridaceae), respectively.See Table 2 for flower measurements.

This content downloaded from 143.106.1.143 on Wed, 20 Aug 2014 17:23:24 UTC All use subject to JSTOR Terms and Conditions Volume 93, Number 3 Freitas & Sazima 501 2006 Pollination in a High-Altitude Grassland

Figure 2. Flowersand pollinatorsin the Serrada Bocainagrasslands. Flowers pollinated by several insect groups:- A-C. Species of three pollinatorgroups visiting the heads of Eupatoriumsp. indet. 4 (Asteraceae),respectively, a wasp (Sphex opacus, Sphecidae), a beefly (Bombyliidae), and a small bee (Augochloropsiscyanea, Halictidae). Oil-flower: - D. Paratetrapediasp. indet. 1 collecting pollen fromflowers of Byrsonimavariabilis (Malpighiaceae). Note the floweroil-glands or elaiophores(arrow). - E. Centriscf. insularis(Apidae) collecting oil fromflowers of Byrsonimavariabilis (Malpighiaceae). Syrphid/smallbee nectar-flowers:- F. Ceratina sp. indet. 2 (Apidae) visiting a flower of Convolvuluscrematifolius (Convolvulaceae).- G. Pseudagapostemoncyaneus (Halictidae) visiting a head of Senecio oleosus (Asteraceae). Minor pollinatorgroups: - H. A (Cantharidaesp. indet. 2) visiting a flower of the mainly wasp-pollinatedClethra scabra (Clethraceae).- I. A butterfly, Thecla sp. indet. 1 (Lycaenidae),visiting the generalist flowers of Galianthe brasiliensis (Rubiaceae).Doubtful pollination systems: - J. A female Monoecasp. indet. 1 (Apidae)with a pollinariumof a species of Oncidium(Orchidaceae) attached to its clypeus. This digger bee was capturedwhile collecting oil on flowers of Byrsonima variabilis,and its role in the pollinationof Oncidiumspecies is uncertain.- K. Flowerof Helia oblongifolia(Gentianaceae). Floral traits point to pollination by nocturnal moths, but no pollinators were observed. - L. Flower of Calolisianthus pedunculatus(Gentianaceae). The attractiveflowers of this species seem to be adaptedto pollinationby hummingbirds,but they are nectarless and spontaneouslyself-pollinated. See Table 2 for flower measurements.

This content downloaded from 143.106.1.143 on Wed, 20 Aug 2014 17:23:24 UTC All use subject to JSTOR Terms and Conditions 502 Annals of the Missouri Botanical Garden

Table 6. The number (and percentage) of plant species that are monophilous (most specialized), oligophilous, polyphilous, and holophilous (most generalized) within each pollination system and pollinator group at the Serra da Bocaina grasslands.

Monophily Oligophily Polyphily Holophily Total of species|

All species 36 (34.0%) 32 (30.2%) 18 (17.0%) 20 (18.8%) 106 Pollinationsystem Small and large bee 10 (33.3%) 12 (40.0%) 8 (26.7%) 0 30 nectar-flowers* Wasp and wasp/fly 7 (36.4%) 8 (31.8%) 7 (31.8%) 0 22 nectar-flowers Several insect 0 0 0 20 (100%) 20 groups Small and large bee 11 (73.3%) 3 (20.0%) 1 (6.7%) 0 15 pollen-flowers Syrphidand 2 (22.2%) 6 (66.7%) 1 (11.1%) 0 9 syrphid/small bee pollen- flowers Hummingbird 5 (100%) 0 0 0 5 flowers Syrphid/beenectar- 0 2 (66.7%) 1 (33.3%) 0 3 flowers Small and large bee 0 1 (100%) 0 0 1 . oil-flowers Syrphidnectar- 1 (100%) 0 0 0 1 flowers Undeterminedor - - - - 18 uncertain Pollinatorgroups! Small bees % 12 (16.9%) 25 (35.2%) 16 (22.5%) 18 (25.4%) 71 Wasps 7 (14.6%) 11 (22.9%) 14 (29.2%) 16 (33.3%) 48 Largebees $ 8 (22.2%) 12 (33.3%) 10 (27.8%) 6 (16.7%) 36 Syrphids 3 (9.1%) 10 (30.3%) 9 (27.3%) 11 (33.3%) 33 Otherflies 0 3 (10.3%) 9 (31.1%) 17 (58.6%) 29 Butterflies 0 3 (12.0%) 6 (24.0%) 16 (64.0%) 25 Beetles 0 0 2 (15.4%) 11 (84.6%) 13 Hummingbirds 5 (100%) 0 0 0 5 t Does not include species where the pollinationsystem was not determinedor was uncertain. * Includes Gaylussaciachamissonis and the four species of Lamiaceaepollinated exclusively or mainly by Apis mellifera. § Since many species are pollinatedby more than one pollinatorgroup, the sum total of plant species exceeds 106. %Apis mellifera not included. iinae (probablyspecies of CorticeaEvans) were very Colibri serrirostris,which seems to migrate to the commonin all seasons at the study sites. highlandsduring the wet season. Coleoptera. Beetles pollinated a few species and always acted as secondary or indistinct pollinators Pollination Systems (Tables 5, 6). Cantharidae(Fig. 2H), mostparticularly Cantharidaesp. indet. 1, which pollinated 10 species The nine pollination systems differed markedly (Table 3), was the only beetle family with a notable with regardto their numberof species. For instance, importancein pollinationat the communitylevel. the three most importantsystems, respectively, small Vertebrata. The interactions between humming- and large bee nectar-flowers, wasp and wasp/fly bird-pollinatedplants and their agents were the most nectar-flowers,and several insect groups,encompass specialized at the communitylevel. We recordedfive 72 species (68% of the 106 species total;Table 6). On Trochilinae species in the study areas (Fig. 3, the other hand, two of the systems, respectivelysmall Table 3). These species could be observedyear-round and large bee oil-flowersand syrphidnectar-flowers, in the high-altitude forest and grasslands, except only compriseone species each (Tables 4, 6).

This content downloaded from 143.106.1.143 on Wed, 20 Aug 2014 17:23:24 UTC All use subject to JSTOR Terms and Conditions Volume 93, Number 3 Freitas & Sazima 503 2006 Pollination in a High-Altitude Grassland

Figure 3. Hummingbirdsand their flowersin the Serrada Bocainagrasslands. - A. Leucochlorisalbicollis visiting a very large flower of Hippeastrumglaucescens (Amaryllidaceae). Note the hummingbird'sfoot grippingthe inferiortepal. - B. A female Colibriserrirostris visiting flowersof Dyckiatuberosa (Bromeliaceae). Note pollen on the bird'sbill and forehead.- C. Flowerof Esterhazyamacrodonta (Orobanchaceae), which is pollinated by L albicollis. - D. A female Stephanoxislalandi visitingflowers of Sinningiaallagophylla (Gesneriaceae). - E. A female Chlorostilbonaureoventris visiting flowersof Agarista hispidula (Ericaceae);originally published by Plant Systematics and Evolution, 2006, Springer-Verlag;reproduced with permission.See Table 2 for flower measurements.

A. NECTAR-FLOWERSPOLLINATED MAINLY BY SMALLAND species, Chromolaenacf. decumbensand Vernonia LARGE BEES cf. rosea. Likewise, Lobelia camporumand Hyptis lippioides were mainly pollinated by large bees, Plants of this had flowers group pollinated mainly although small bees acted as their secondary or small bees that were for nectar by large searching pollinators. Despite their typically melittophilous 1A- This shows in (Fig. C). group great variability flowers (sensu Faegri & van der Pijl, 1979), Cuphea floral and we observed associations traits, specific glutinosa and Verbenahirta were each pollinated by with its As a this pollinatorsamong species. whole, five differentagents, of which large bees were by far is the most of group representativepollination system the most frequent ones; for C. glutinosa, the pollen the community(Table 6). load deposited on the stigmas in a single large bee This plant group first includes five species visit was significantly higher than that deposited by pollinated exclusively by small bees. Species of small bees, wasps, or syrphids (L. Freitas, unpubl. Malvaceae(Fig. 1A) and Polygalaceaewere pollinat- data). Both C. glutinosa and V. hirta bloom contin- ed by Ceratininior Halictidae bees, and Escallonia uously (Table 1) and are among the most common farinacea was pollinatedonly by Colletessp. indet. 1 plants in the study area. (Colletidae)(Table 4). An importantsub-group com- Plant species pollinated by large bees produced prises species predominantlypollinated by small larger quantities of nectar than the other insect- - bees, in which wasps either were rare Mikania pollinated species, which typically secreted less than lundiana, Hyptisumbrosa, and Gaylussaciajordanen- 2 jll of nectar per flower per day. Oxypetalum - - sis or acted as secondary agents Ilex amara and sublanatum and Lobelia camporumflowers, for in- Gaylussacia chamissonis (Fig. IB). The last three stance, secreted 5.5-8 JLXl(n = 10) and 4-7 (ll (n = species have white, bell-shaped flowers. 17) of nectar per day, respectively. Large bees were the exclusive pollinators of Another sub-group in this system is composed the large flowers of Ipomoeaprocumbens and Oxype- of species in which both large and small bees talum sublanatum, and also, of two Asteraceae seemed to be the main pollinators- Hypochaeris

This content downloaded from 143.106.1.143 on Wed, 20 Aug 2014 17:23:24 UTC All use subject to JSTOR Terms and Conditions 504 Annals of the Missouri Botanical Garden

gardneri, Vernoniaherbacea, V. tragiaefolia, Jacque- some Halictidaespecies. The remainingplant species montiagrandiflora, Hyptis plectranthoides, Crotalaria with poricidalanthers - Trembleyaparviflora (Fig. II) breviflora, and Declieuxia cordigera. The pollina- and the five SolanumL. species- have relativelysmall tion systems of some species of Asteraceae diverge flowers,which were only pollinatedby small bees. Four slightly fromthis last sub-group,because, in addition species of Solanumwere exclusively pollinatedby the to small or large bees as their main pollinators, same small bee- Augochloropsiscyanea (Halictidae). butterflieswere either importantagents - Eupatorium This bee (Fig. 2C) is endemic to the Braziliancoastal - sp. indet. 3 and Vernoniawestiniana (Fig. 1C) or rangeand is markedlylarger (mean body length 10.1 ± - secondary agents Eupatoriumsp. indet. 1 and V. 0.99 mm, n = 10) than otherHalictidae species in the tomentella. study area. Bees searching for pollen also pollinated some flowers without poricidal anthers. Viola cerasifolia B. NECTAR-FLOWERSPOLLINATED EITHER BY WASPS OR BY flowers have five anthers opening longitudinally,but WASPS AND DIPTERA their androeciumis arranged so as to function as a anther. This was This group includes species from several families, single poricidal species mainly small AnthrenoidesDucke bees such as Gonioanthelahilariana (Fig. IE), Oxypetalum pollinatedby (Andre- which showed a appendiculatum,Erythroxylum microphyllum, Croton nidae), singular pollen collecting behavioron these in additionto vibration dichrous, Clethra scabra (Fig. 2H), and Borreria flowers, (see Freitas & 2003a for tenella, for which wasps were either the exclusive Sazima, details). The nectarless flowers of pollinatoror the main pollinator,usually associated flag-shaped Lupinus velutinuswere with flies. Wasps were also either the main or the only pollinatedby large bees belonging the exclusive pollinators of many Asteraceae, e.g., to genera Bombusand Megachile Latr. (Fig. 1J). Achyrocline satureioides, Gochnatia paniculata During the bumblebee visits, the weight of the bee (Fig. ID), Lucilia lycopodioides,and Mikania num- flexed the wing petals downwardand pollen was mularia, and of all species of the genus Baccharis thereforepressed out in portions,at the tip of the keel except B. tarchonanthoidesand B. dracunculifolia. ("macaronipump" type of pollen presentation,sensu In the latter, as well as in Mikania sessilifolia and Endress, 1994), which contactedthe abdomenof the the spontaneously self-pollinated Galium hypocar- bee. However,pollen collection by bumblebees was pium, pollination by flies and wasps was equally usually improvedby vibration.Leafcutter bees seem important. to lack the necessary weight to flex the wings and Flowers of this group typically have easily press pollen out of the keel. During their visits to accessible nectar and are usually greenish. Pre- a flower,these bees used their head as a lever, and the dominantflower shapes are short tube (e.g., Mikania wings were thus depressed by their front and middle Willd.) and dish (e.g., Gonioanthela Malme and legs (Fig. 1J). The hind legs strokedthe sides of the ErythroxylumP. Browne),but flowers may be large keel, forcing a streamof pollen out of the keel beak, in width, allowing the insect to enter and reach the and pollen was then packed on the ventralabdomen of nectar (e.g., Oxypetalum appendiculatum).Wasps the bee ("milkingaction," sensu Wainwright,1978). carry few pollen grains on their bodies, and some Bombusatratus, collecting pollen by vibration,was species belonging to this pollination system may be the effective pollinatorof Alophiageniculata flowers, additionallypollinated by wind. although its visits were infrequent.Small bees and syrphidsfrequently fed on pollen of this species, but its flowers. A C. POLLEN-FLOWERSPOLLINATED BY SMALLAND LARGE BEES they only occasionally pollinated large similar pollinationsystem is expected for Alophiasp. Nectarless flowers with poricidal anthers are the indet. 1, althoughwe were unable to recordany large most characteristicof this system, which includes bees on its flowers (Table 4). species of Melastomataceaeand Solanaceae, among others(Fig. 1F-I). Plantsof this groupwere pollinated D. POLLEN-FLOWERSPOLLINATED EITHER BY SYRPHIDS OR BY by large and small bees, which collected pollen by SMALLBEES AND SYRPHIDS vibration.Large Apidae bees - belongingto the genera Bombus,Xylocopa, and Centris- were the exclusive Syrphidswere the exclusive pollinatorsof Drosera pollinatorsof Tibouchinafrigidula, T. martialis, and montanaand Deianira nervosaand the co-pollinators, Chamaecristasp. indet. 1. Large bees were also the in associationwith small bees, of Zygostigmaaustrale, pollinatorsof Tibouchinaminor (Fig. IF), Trembleya Calydoreacampestris, Sisyrinchium vaginatum, Xyris phlogiformis(Fig. 1G-H), and Ourateasemiserrata, in asperula, and Xyris tortulla (Fig. IK- L). Species in additionto such small bees as Meliponabicolor and this group typically bear small, dish-shaped,actino-

This content downloaded from 143.106.1.143 on Wed, 20 Aug 2014 17:23:24 UTC All use subject to JSTOR Terms and Conditions Volume 93, Number 3 Freitas & Sazima 505 2006 Pollination in a High-AltitudeGrassland

morphic, and vividly colored pollen-flowers.Pollen reach the nectar at the flower base (Fig. 3A). Pollen grains are easily accessible to pollinators of these was deposited mainly on their wings. On the other species because their anthers open longitudinally. hand, Esterhazyamacrodonta has protandrousflowers Flowers of Sisyrinchiummicranthum bear oil-secret- (Fig. 3C) that last up to six days. Nectar removalin ing trichomesat the base of the staminalcolumn (see the first days of anthesis strongly affects nectar Cocucci & Vogel, 2001, Truylio et al., 2002). production(see Freitas & Sazima,2001 for details). However,we failed to detect any oil-collecting bees Sinningia allagophylla bears a long hairy in- on these flowers, which were pollinated by syrphids florescence with more than 40 flowers,which last up and small pollen-collectingbees. to seven days and present partial herkogamyassoci- ated with protandry. This gesneriad was mainly E. NECTAR-FLOWERSPOLLINATED BY SEVERAL INSECT pollinated by Leucochlorisalbicollis and Stephanoxis lalandi but also Colibriserrirostris GROUPS (SIG) (Fig. 3D), by and Clytolaemarubricauda. The flower disposition along This group contains species pollinated by three or the inflorescenceof Dyckia tuberosaresembles that of more functional groups with similar importance S. allagophylla (Fig. 3B, D), and we only observed (holophily).The SIG system includes some species Colibriserrirostris females pollinating this terrestrial - with very small, greenish flowers Eryngium horri- bromeliad(Fig. 3B). Agarista hispidula bears dozens dum (and probablyE. canaliculatum),Paepalanthus of flowersarranged in dense inflorescences(Fig. 3E). paulensis, P. polyanthus, and Weinmanniaorganen- Flowers are urceolate with poricidal anthers, and sis- that fit the "diverse small insects" (d.s.i.) pollen is secondarilypresented on hairs at the corolla syndromeof Bawa et al. (1985). opening. It was pollinated very early in the morning, The SIG system also includes species from several mainly by the small Chlorostilbon aureoventris Asteraceaegenera (Baccharis tarchonanthoides, Barro- (Fig. 3E), but also by Leucochlorisalbicollis (see soa betoniciiformis,Campuloclinium megacephalum, Freitas et al., 2006 for details). In general, ornitho- Chaptaliaintegerrima, C. runcinata,Eremanthus ery- philous species were rare in these grasslands. We thropappus,Erigeron maximus, Eupatorium sp. indet.2, foundonly four,seven, and eight floweringindividuals Eupatoriumsp. indet. 4 [Fig. 2A-C], Grazieliagaudi- of, respectively,Agarista hispidula, Dyckia tuberosa, chaudeana, Stevia myriadenia,Symphyopappus com- and Hippeastrumglaucescens in the studyareas. It was pressus,and Vernoniamegapotamica) and three Rubia- commonfor there to be < 10 Esterhazyamacrodonta ceae species (Borreriacapitata, Galianthe angustifolia, plants along each 1 km transect;however, along one and G. brasiliensis[Fig. 21]). Large insects, such as exceptionaltransect we found 55 floweringindividu- bumblebees and Pompilidaewasps, were among the als, whose nectarfeatures were also studied (Freitas& pollination agents of some holophilous species of Sazima, 2001). In contrast, S. allagophylla was Asteraceaeand Rubiaceae, which thus differedfrom common in many places in the Serra da Bocaina. the species of Apiaceae, Eriocaulaceae,and Cunonia- Individualsof this species were usually distributedin ceae, which have very small flowers(Table 4). small clusters (three to eight plants) separatedfrom each other by many meters. However,S. allagophylla - F. HUMMINGBIRD-POLLINATEDFLOWERS plants occurredin large clusters in some areas with dozens of individuals per 100 m2- probablydue to Hummingbirdswere the exclusive pollinators of clonal reproduction.Esterhazya Mikan., Agarista D. five species from five families. These species bear Don, and Dyckia Schult. f. plants were rarelyvisited, tubular/urceolateflowers that make them easily receiving,respectively, 0.05, 0.18, and 0.20 visits per distinguishablefrom flowers of other species in this hour(see Table 4 for times of observation).Each plant community(Fig. 3). Except the nectarless, spontane- of Hippeastrumglaucescens received ca. one visit ously self-pollinatedflowers of Calolisianthuspedun- every two hours. Although plants of Sinningia Nees culatus (Fig. 2L), all species with red-orangeflowers growingin small clusters received ca. only one visit were pollinatedby hummingbirds. every three hours, when in large clusters, humming- Each of these plant species was pollinatedby one to birds visited each patch between three and six times four hummingbirdspecies (Tables 3, 4). Leucochloris per hour so that most individuals were visited at albicollis was the main visitor in the study area, intervalsof one to two hours. pollinating four species and acting as the single Because of the low flower availability, humming- pollinatorof Esterhazyamacrodonta and Hippeastrum bird species basically acted as low-rewardtrapliners glaucescens. This HippeastrumHerb, species bears in the studied grasslands (foraging strategies after very large flowers (tepals ca. 140 cm long), and the Feinsinger & Colwell, 1978). Territorial foraging birds needed to perch on one of the inferiortepals to behavior was recorded only in the large clusters of

This content downloaded from 143.106.1.143 on Wed, 20 Aug 2014 17:23:24 UTC All use subject to JSTOR Terms and Conditions 506 Annals of the Missouri Botanical Garden

- Sinningia allagophylla. Clytolaema rubricaudaand velutinus flowers, a female individual of Bombus - Stephanoxis lalandi behaved territorially in these after collecting the pollen of Lupinus L. flowers circumstances,and Leucochlorisalbicollis occasional- quickly visited a flower of Oncidium barbaceniae. ly acted as a parasitein territoriesset by Clytolaema Duringthis visit, the bee vibratedin a failed attempt rubricauda.Thus, Leucochloris albicollis may alternate to collect pollen. We did not collect this individual its foraging strategy from low-rewardtraplining to bumblebeebut observedthat the visited flowerhad no territory-parasitismin some cases. pollinia, indicatingthat this bee could have removed them. These observationsindicate that bees in search

G. MISCELLANEA of pollen could be acting as pollinators in these populations of Oncidium. Thus, in addition to the - Many oil-collecting bees mainly large bees be- expected system involvingoil-collecting bees, Oncid- longing to Centris- were recordedpollinating Byrso- ium flowers could be pollinated by deceit by pollen- nima variabilis (Fig. 2D-E). The mechanismfor oil- collecting bees. The following aspects point to collection in this species was similarto that recorded a pollination system by deceit in the studied for other Malpighiaceaespecies (see Vogel, 1990). populations:(1) bees in search of pollen visited their Byrsonimavariabilis was also pollinatedby small bees flowers; (2) bees of the "Anthophoridae"group belongingto Apidae (Meliponinae)and Halictidae in intensively collected oil on Byrsonima flowers but search of pollen, but to a minor degree. Some oil- were not recordedon flowersof the Oncidiumspecies collecting bees also collected pollen on this Byrso- growingin the same patch;(3) visits by bees fromany nima Rich, ex Juss. species. group were scarce on Oncidium flowers, and fruit The nectar-flowers of Tassadia subulata were productionwas very low in the three species. In short, exclusively pollinatedby syrphids,which carriedthe pollinationof the three species at Bocaina is unclear, pollinia on their heads. The elaborate, small, ruby and more comprehensive studies are required to flowersof this species are typicallymyophilous (sensu understandthe pollinationbiology of this genus. Faegri & van der Pijl, 1979). Small bees were also The tubularflowers of Calolisianthuspendulus do associated with syrphids on the nectar secreting not producefloral nectar,except for some rareflowers flowers of Senecio oleosus, Wahlenbergia brasiliensis, with minute nectarioles (sensu Vogel, 1998) located and Convolvulus crenatifolius (Fig. 2F- G). The last on petals. No visits to their flowerswere recorded,and two species bear bell-shaped, lilac/lavenderflowers. delayed self-pollination seems to ensure the repro- Pollinators first fed on the nectar of these three duction of this species, as well as that of the other species, although they also searched for pollen on four studied gentians of Bocaina, which include them. Syrphids and small bees seemed to be co- the ornithophilousCalolisianthus pedunculatus with pollinatorsof these species (syrphid/beenectar-flower nectarless flowers (Fig 4L) (see Freitas, 2004 for system). details). Although the floral features of three species H. SPECIES WHOSE POLLINATIONSYSTEMS ARE DOUBTFUL (Mandevillaerecta, Habenariaparviflora, and Helia oblongifolia [Fig. 2K]) point to phalaenophily and The flowers of 18 species were either not visited sphingophily(after Faegri & van der Pijl, 1979), no during the observationsessions or the role of their nocturnalmoths were recordedvisiting their flowers visitor in pollination was not clear. The results and (Table 4). These species flower in the middle of the brief discussions of these cases are presentedherein. wet season (Table 1). The flowers have narrow Althoughmany OncidiumSw. species bear flowers greenish or yellowish tubes (spurs in H. parviflora), that producenonvolatile oils with a chemical compo- presentingboth odor and nectar at night. The long- sition suitable for larval nurturingby "Anthophor- lived flowersof H. oblongifoliaare spontaneouslyself- idae" bees (Reis et al., 2000; A. D. Faria, pers. pollinated at the end of anthesis (Freitas,2004). The comm.),only one single observationof a bee actively other two species were rare in the studied areas. collecting oils on Oncidiumflowers has been reported Mandevilla erecta bears large inflorescences with so far (Singer& Cocucci, 1999). We did not observe dozens of flowers, but we only recorded one fruit oil-collecting bees visiting flowers of Oncidium producedduring three subsequent bloomingseasons species (total 47 hr. of observation, Table 4). (from 1998 to 2000). Several Habenariaparviflora However,we collected an individual of Monoecasp. plants produced fruits in the study areas. Although indet. 1 (Tapinotaspidini)(Fig. 2J) and an individual this species is self-compatible, spontaneous self- of Megachile aureiventrisSchrottky (Megachilidae) pollination is improbable (Singer, 2001). It is each carryingan Oncidiumpollinarium attached to its pollinated by crepuscular crane-flies and nocturnal head. During an observation session on Lupinus moths in an area at sea level by the Sao Paulo

This content downloaded from 143.106.1.143 on Wed, 20 Aug 2014 17:23:24 UTC All use subject to JSTOR Terms and Conditions Volume 93, Number 3 Freitas & Sazima 507 2006 Pollination in a High-Altitude Grassland

flowering in different months year-round. The poly- androus flowers of both species are similar (Table 2). The flowers of the fourteen individuals of Hypericum ternum in the study area did not produce pollen, but all flowers developed fruits with well-developed seeds. Similar results were found for the ca. 40 individuals of Hypericum brasiliense. However, two individuals of the latter were found to produce flowers with pollen grains (ca. 95% viability) in December 1998. One of these plants had been followed since January 1998, and, curiously, it produced only flowers without pollen until December 1998. This capacity for sexual and apomictic reproduction within the same individual is named facultativism and may be environmentally influenced (Asker & Jerling, 1992). In the pollen-producing flowers of Hypericum brasi- liense, self- and cross-pollen tubes grew down the style, but only crossed ones seemed to penetrate the ovule. We observed these pollen-producing individu- als in December 1998 and January 1999 and found that insects, mainly syrphids and small bees, searched for pollen and pollinated their flowers (Table 4). Empty anthers were also observed in herbarium specimens of Hypericum species from lowland forest areas (V. Bittrich, pers. comm.). It is expected that H. ternum also occasionally produces pollen. No floral visitors were registered during the 28 hr. of observation of the four Leandra Raddi species. There are records of apomixis and pollenless flowers in species of this genus (Goldenberg & Shepherd, 1998; R. Goldenberg, pers. comm.); thus the Leandra species at Bocaina may also be apomictic. Another Melastomataceae, Microlicia isophylla, was not visited either. This species is very common in grasslands and bears attractive flowers, with magenta petals and vivid yellow anthers. Their smooth pollen grains are highly viable and are produced in large quantity. Bees collecting pollen by vibration are the expected pollinators of this species, although it could also be apomictic. The small, greenish flowers of Eryngium canalicu- latum are very similar to those of E. horridum. Floral 4. Number of and in Figure species blooming (line) traits of E. canaliculatum point to a SIG pollination from Dec. 1998 to Feb. 2000, in the floweringpeak (bars) four visits two of Serra da Bocaina grasslands. - A. Distribution at the system, although only by species communitylevel (n = 124). - B-E. Distributionof plant beetles were recorded on these flowers. Further species of each pollination system. - B. Nectar-flowers observations are necessary to ascertain their pollination small and bees = - C. pollinated mainly by large (n 30). system. We did not observe pollinators visiting the Pollen-flowerspollinated by small and large bees (n = 15). showy flowers of Chromolaena xylorrhiza and Epiden- - D. Nectar-flowerspollinated either by wasps or by wasps drum secundum that fit in the and dipterans(n = 22). - E. Nectar-flowerspollinated by melittophilous syndrome. several insect groups(SIG) (n = 20). Degree of Specialization of Pollination Systems coastline (Singer, 2001). Results indicate that moths seldom use floral resources from grassland plants. Around one third of the plant species were The two studied species of Hypericum L. have pollinated by only one pollinator group (monophilous, extended flowering times, with individual plants Table 6). For 50 out the 70 species pollinated by two

This content downloaded from 143.106.1.143 on Wed, 20 Aug 2014 17:23:24 UTC All use subject to JSTOR Terms and Conditions 508 Annals of the Missouri Botanical Garden

Figure5. Floweringpatterns of the five hummingbird-pollinatedspecies from the Serrada Bocainagrasslands from December1998 to February2000. Thinand thicklines indicate,respectively, flowering time and floweringpeak. Plant speciesare abbreviated according to theirgenus (see Table1 forspecies epithet). to six pollinatorgroups, we were able to determineone observed in the community (Fig. 4A-C). However, or two insect groups acting as main pollen vectors there was a slight increase of species at peak (oligophilous and polyphilous). The remaining 20 flowering in June (Fig. 4B-C). In contrast, most species were holophilous, since they presented no species with nectar-flowerspollinated either by wasps dominant pollinator group. Monophily dominates or by wasps and dipterans reached their flowering among pollen-bee flowers, and oligophily among peak duringthe dry season (Fig. 4D). Wasps are very syrphidand bee pollen-flowers.Species belongingto importantpollinators during that season. For example, the bee nectar-flowersand wasp/fly flowers systems 11 out of 23 species at their floweringpeak in May, are more equitably distributed among monophily, 1999 were exclusively or mainly pollinatedby wasps oligophily,and polyphily(Table 6). (Fig. 4A, D). The flowering pattern of SIG species Morethan half of the species pollinatedby small and resembles the general patternof the community,but large bees are monophilousor oligophilous(Table 6). a few plants at their flowering peak were recorded In contrast,most species pollinatedby flies, butterflies, throughoutthe year (Fig. 4 A, E). The floweringof the and beetles are holophilous.The species pollinatedby five hummingbird-pollinatedspecies concentratedat wasps and syrphids are better distributed among the beginning of the wet season (Fig. 5). No plants oligophily,polyphily, and holophily. floweredin April and May,and no plantswere at their floweringpeaks duringfive monthsof the year. Flowering Phenology Discussion Floweringduration in grasslands was long (mean 5.5 ± 2.92 months,n = 124), and the mean flowering AlthoughMartinelli (1989) reporteda total of 215 peak duration was 2.6 ± 1.56 months (n = 124) flowering species in the Bocaina grasslands, we (Table 1). Floweringpattern in this communitywas collected ca. 260 species. About 20% of these 260 seasonal, with a peak during the middle of the wet species belong to the families Poaceae, Cyperaceae, season (Fig. 4A). Over60% of the species floweredin and Juncaceaeand are anemophilous.This studythus Januaryand February,and ca. one-thirdof the species included around 60% (124 species) of the ca. 210 were at their floweringpeak duringthese two months. animal-pollinatedspecies from the Bocaina grass- There was a drastic reductionin floweringduring the lands, and pollinatorswere determinedfor about half end of the dry'season/beginning of the wet season of these zoophilousspecies (106 species). In addition, (Aug.-Oct.). The fire that occurred in one area in ca. 80% of the 47 plant families found in the Bocaina September1999 may have only slightly affectedthese grasslandswere representedin this study, so that our generalpatterns, because only eight species fromthis samplewas fairly representativeof these high-altitude area did not occur in at least one of the other two grasslands. studied areas (see Table 1). We carriedout intensive pollinatorobservations for Blooming patterns of the species with nectar- or some plant species and brief ones for manyothers, as pollen-flowerspollinated by bees followedthe pattern expected in studies on plant-pollinatorinteractions at

This content downloaded from 143.106.1.143 on Wed, 20 Aug 2014 17:23:24 UTC All use subject to JSTOR Terms and Conditions Volume 93, Number 3 Freitas & Sazima 509 2006 Pollination in a High-Altitude Grassland

the communitylevel which encompassmany species The prevalenceof nectaras the mainfloral resource with low populationdensity (see Momoseet al., 1998). at Bocaina was recordedin other grassy communities Boththe shortobservation times and the generallylow (Arroyo et al., 1982; Ramirez, 1989; Faria, 1994; flowervisitation rates explain why we may have failed Barbosa, 1997). In fact, nectar-flowersare the most to ascertain the typical pollination system of some importantflower class amongangiosperms in general, species in the Bocainagrasslands. However, since we because most pollinatorgroups are nectar consumers used broad categories to describe the pollination (Endress, 1994). The percentageof pollen-flowersat systems, we may outline general interactiontrends in Bocaina was similar to that registered in the campo this communityand establish general comparisons. rupestre(Faria, 1994) and the cerrado(Barbosa, 1997). There are no comparativedata on the floral and Melastomataceae species are the most important pollination biology of any other high-altitudegrass- pollen-flowersat Bocaina, as well as in the campo land areas in southeasternBrazil. However,commu- rupestre(Faria, 1994). Species of this family are also nity level pollinationdata from vegetational types with importantpollen sources in the arbustal (Ramirez, both abiotic similarities and strong biogeographic 1989). However, species of other families, such as connections are available. These include the high Leguminosae and Myrtaceae, constitute the chief temperate Chilean , especially in the sub- sources of pollen in the cerrado (Barbosa, 1997). Andean scrub zone (Arroyo et al., 1982), the Malpighiaceaeare representedby several species in VenezuelanGuiana high-altitude shrubland or arbus- the campo rupestreand the cerrado (Faria, 1994; tal (Ramirez,1989), the Brazilianopen savannas or Barbosa, 1997), which explains why oil-flowers are cerrado (Barbosa, 1997; see also Silberbauer-Gotts- more frequent in those communities than in the berger & Gottsberger,1988 and Oliveira & Gibbs, Bocaina grasslands. 2000 for other cerrado physiognomies), and, to Althoughsmall flowerspredominate in the arbustal some extent, data from the Braziliancampo rupestre and the cerrado(Ramirez, 1989; Barbosa,1997), the (Faria, 1994). We focus our discussion on compar- flowersof these two communitiesare largerthan those isons between these formations and the Bocaina at Bocaina.As for the floralcolor patterns,the relative grasslands. predominance of Asteraceae in the community strongly influences the floral size patterns. The Floral Traits prevalence of small tubular flowers secreting nectar at Bocaina is a direct reflection of the number of Flowers with light colors- including greenish, Asteraceae presenting these floral traits and appar- pink, and white color groups- are well represented ently not the result of selection by the local pollinator in the grasslands,resembling the arbustaland cerrado agents. communities,in which light-coloredflowers predom- The prevalenceof hermaphroditicflowers is typical inate (Silberbauer-Gottsberger& Gottsberger,1988; of angiosperms,and, consequently, characteristicof Ramirez,1989; Barbosa,1997). However,more than most communities (Ramirez, 1993), including the half of the species at Bocaina have showy flowers grassland flora at Bocaina. However, many of the (violet, yellow, and red color groups).This dominance hermaphroditicflowers exhibited spatial or temporal is due to the presence of Asteraceaeand Melastoma- separationof the male and female functions, through taceae, with their showy violet-colored flowers, in dichogamy,herkogamy, or heterostyly.In fact, a high these grasslands.Furthermore, greenish flowers were frequency of dichogamy (protandry) is expected more continuouslydistributed throughout the year, in in a flora dominated by Asteraceae, owing to the contrastto violet flowersthat are concentratedduring frequentpresence of secondarypollen presentationon the wet months.Using canonicaldiscriminant analysis the style among the (see Yeo, 1993). In (L. Freitas, unpubl. res.), we only detected a weak addition, the female ray-floretsof many Asteraceae relationshipbetween floral color and the frequencyof species impart a state of protogynyto the head as visits for each pollinatorgroup in the community.Red a whole. flowers were strongly associated with hummingbird All the dioecious species of the studied community visitationand greenishflowers weakly associatedwith belong to the genus Baccharis, which was primarily visitations primarilyby beetles and wasps. Although associated with wasp pollination. Dioecy has often we used the human color spectrum,such results for been associated with a generalistmode of pollination insect responses to flower color resemble the general by small opportunisticinsects (Bawa, 1994; but see trends detected by Chittkaet al. (1994) and Waser et Renner& Feil, 1993) or with wind pollination(Culley al. (1996), who found no statisticaldifferences among et al., 2002). Although this study did not focus on insect pollinatorgroups, at the orderlevel, in termsof abiotic pollination, some evidence suggests the the colors of the visited flowers. existence of ambophily among the studied taxa, in

This content downloaded from 143.106.1.143 on Wed, 20 Aug 2014 17:23:24 UTC All use subject to JSTOR Terms and Conditions 510 Annals of the Missouri Botanical Garden

which wind could be a pollen vector in addition to of the Brazilianspecies of Bombus(Alves-dos-Santos, insects (e.g., Croton dichrous). The species of 1999; Barbola et al., 2000). Bumblebees pollinated Baccharis may also fit in this category, as female most of the large bee-pollinated species at Bocaina heads have well-exposed stigmata and male flowers and predominatedin almost half of them, including produce small pollen grains. Furthermore,we found the richest ones in terms of nectar, such as pollen grains of Baccharis species adhered to Oxypetalumsublanatum, Lobelia camporum,Cuphea microscopeslides covered by glycerol, duringexplor- glutinosa, and Verbenahirta. Moreover,these bees atoryexperiments at the study sites, and most species showed a traplinebehavior, favoring pollen dispersal of this genus flower during the winter, when insect at great distances. Consideredtogether, the varietyof activity is reduced.Pollinator limitation, dry weather, visited flowers(including nectar- and pollen-flowers), and the presence of recurrent wind in the winter foragingbehavior, dominance of the richest resources, could favorthe evolutionof anemophilyand amboph- and year-roundactivity make bumblebees the main ily in the high-altitudegrasslands (see Culley et al., pollinators of the plant group related to large bee 2002). pollination. Bumblebees are known to be one of the most importantpollinators in cool temperateclimates Pollination Agents (Kevanet al., 1993; Nakano& Washitani,2003), and the harshclimatic conditions,for a tropicalplace, that Hymenopterans,markedly small bees, are the characterize the Brazilian high-altitude grasslands predominantpollinators in the Bocaina grasslands, may be related to the high importanceof bumblebee as well as in the cerrado,arbustal, and Subandean pollinationat Bocaina. zone (Arroyoet al., 1982, Ramirez, 1989; Barbosa, Pseudagapostemon cyaneus (Halictidae) visited 1997). Bees have been reported as the main only flowersof Mikania lundiana and Seneciooleosus pollinators for most communities (e.g., Moldenke, (Asteraceae)at Bocaina (Fig. 2G). Similarly,this bee 1976; Kevan & Baker, 1983; Roubik, 1989; Bawa, species was a narrrowoligolectic associated to S. 1990; Momose et al., 1998; Nakano & Washitani, oleosus flowers, in a grasslandarea at Lapa, State of 2003; Ramirez, 2004), and different patterns- for Parana, southern Brazil (Barbola et. al., 2000). example the dominance of dipterans- seem to be Pseudagapostemoncyaneus occurs from southernmost restrictedto very singularhabitats, such as Arctic and Brazilto the State of Sao Paulo. In this last region,it is alpine areas or oceanic islands (e.g., Kevan & Baker, only reported above 1000 m (Cure, 1989). The 1983; Primack & Inouye, 1993; Anderson et al., particular interaction between Senecio oleosus and 2001). However, certain communities with -similar Pseudagapostemoncyaneus at both Bocainaand Lapa percentages of bee pollination may exhibit marked is one of the most illustrative examples of bio- differences among their general patterns of pollina- geographicconnections between the mountainranges tion. Bees belonging to different groups show great of southeasternBrazil and the cool lowland areas in variationin, for instance, their preferences,abilities, southernmostregions (see Behling, 1997). and behavior on flowers (see Roubik, 1989), and Pollination by wasps has been observed in some such variation is expected to lead to different highly specialized and infrequentcases (e.g., Ophrys pollinationpatterns among communities with distinct L., Ficus L.), as well as among generalist plants bee faunas. pollinated by several insects (Faegri & van der Pijl, The two main functional groups of bees in the 1979). In most communities, wasps have been Bocaina grasslands are the large bumblebees and reportedas minor agents (e.g., Arroyoet al., 1982; small halictid bees. Agents of one or both of these Bawa et al., 1985; Silberbauer-Gottsberger& Gotts- groupswere amongthe pollinatorsof 67 out of the 106 berger, 1988; Momoseet al., 1998; Oliveira& Gibbs, species (63%). Bombus and Halictidae species 2000). In contrast, wasps pollinated 45% of the pollinated, respectively, 78% of the large bee- species in the Bocaina grasslands and acted as the pollinated species (28 out of 36), and 76% of the exclusive or main pollinatorsof morethan 20% of the small bee-pollinatedspecies (54 out of 71). Both bee animal-pollinated species in this community (Ta- groups exhibited a generalist foraging behavior, ble 5). Althoughwasps have their importancein the visiting flowers of differentshapes, sizes, and colors, pollination of cerrado and arbustal, they are the which, in many cases, were also pollinated by other exclusive pollinatorsof only a few species in these groups.An illustrativeexample is that of an individual communities(Ramirez, 1989; Barbosa,1997). In fact, of Bombus atratus observed visiting the violaceous as far as we know, such importance of wasp flowers of seven species, belonging to Asteraceae, pollination as observed in the Bocaina grasslands Lamiaceae,and Rubiaceae, in a single foragingtrip has never been reported for any other ecosystem. (forca. three min). Such generalistbehavior is typical Whetherthe wasps are minorpollinators or their role

This content downloaded from 143.106.1.143 on Wed, 20 Aug 2014 17:23:24 UTC All use subject to JSTOR Terms and Conditions Volume 93, Number 3 Freitas & Sazima 51 1 2006 Pollination in a High-Altitude Grassland

in pollination has been neglected in Neotropical (12%) (Ramirez, 1989) and campo rupestreareas communitiesis still an inadequatelyexplored issue (10%) (Faria, 1994). In the Bocaina grasslands, that deserves furtherstudy. However,at least in some ornithophilousspecies constitute secondary nectar subtropicalhabitats, the rates of floral visitation by sources for hummingbirdsthat find their main nectar wasps, as well as their probable importance as sources in the surrounding high-altitude forest pollinatorsat the communitylevel, seem to have been (Sazimaet al., 1996; Freitas & Sazima, 2001). Two underestimated(P. Feinsinger,pers. comm.). factors, at least, support this idea. First, in these At Bocaina, lepidopterans were generally co- grasslands,no ornithophilousspecies were flowering pollinators of species with a generalist pollination in April or Mayor were in their bloomingpeak during system or secondarypollinators of species pollinated Februaryto June. Therefore,this communitywould be mainlyby bees. Similarresults were found in cerrado unable to sustain hummingbirdsthroughout the year. and arbustal areas (Ramirez, 1989; Barbosa, 1997; Secondly,the five ornithophilousspecies were seldom Oliveira& Gibbs,2000). However,butterflies are very visited by hummingbirds,while hummingbirdsin- importantpollinators in the open vegetation of the tensively fed on ca. 30 species of the surrounding Chilean Andes, especially in the subnival zone forest areas (Sazimaet al., 1996; Buzatoet al., 2000; (Arroyoet al., 1982). In spite of the scarcity of their L. Freitas, pers. obs.). Hence, at Bocaina, the visits to flowers,butterflies (mainly Hesperiidae) may continued replacement of forest areas by grass- have an importantrole in the gene flow of many lands- mainly because of fires- is a risk factor for species at Bocaina, transferring pollen at long populationsof the five hummingbirdspecies observed distances due to their traplining behavior. Despite in the grasslands, in particular for StephanoxisI. some highly specializedcases, such as those relatedto lalandi, which lives exclusively in some high-altitude long-tonguedflies in South Africa (e.g., Goldblattet areas of southeasternBrazil (Grantsau,1989; Sazima al., 1995), fly pollinationhas been consideredmostly et al., 1996). unspecialized,since flies do not feed their young and We did not observeany pollinationby eitherbats or may have other food sources than flowers (Kearns, other mammalsin the Bocaina grasslands.Moreover, 1992; Proctoret. al., 1996). Most species pollinated no species were collected with floral features in by the three most importantfly groups at Bocaina- accordancewith the mammalpollination syndromes, Syrphidae,Tachinidae, and Bombyliidae- have small suggestingthe absence of such pollinationsystems in nectar-flowersthat were pollinatedby several agents. these grasslands. Reports of bat-pollinated species However,syrphids, associated with small bees, were from mesic habitats in high-altitude grasslands are the pollinators of many species, particularlythose restricted to three Bromeliaceaespecies (Martinelli, with pollen-flowers. These last species present 1994, 1997). In contrast, bat pollination is well monophilousor oligophilouspollination systems. This represented in the high-altitude forest areas that result shows that pollination by flies can be surroundthe grasslands(Sazima et al., 1999). specialized even in the absence of extremeproboscis The Bocaina grasslandsalso lack perfumeflowers, elongation. If we consider the species for which anotherhighly specialized pollinationsystem related syrphidsand otherdipterans are eitherthe mainor the to male euglossine bees (e.g., Sazima et al., 1993). exclusive pollinator, the percentage of species Euglossine bees have restricted altitudinal limits of pollinatedby flies at Bocaina is similar to that found distribution(Roubik, 1989), and aromaticcompounds in cerrado and arbustal areas (Ramirez, 1989; trapsdid not detect any such bee in the Bocainahigh- Barbosa,1997). Nevertheless, in the Chilean Andes, altitude grasslands. myophilyis more common(Arroyo et al., 1982). The In contrastto bats, nocturnalpollination by moths greatimportance of flies and butterfliesin the Chilean is expected since the Bocaina grasslandspresent at Andes, as compared to the other communities least three species adaptedto it. However,we did not considered here, probably reflects the unfavorable observethese pollinatoragents, and the low fruitset in climatic conditions for bee activity in the Andes Mandevilla erecta indicate low- perhaps unpredict- (Arroyo et al., 1982; see also Kevan, 1975 and able- rates of flower visits by nocturnal moths in Kearns, 1992). In the Bocaina grasslands, climatic these habitats. As suggested for the cerrado area conditions are not harsh enough to limit the de- (Oliveira & Gibbs, 2000), strong winds at the study velopmentof bees, which dominatethe pollinationin areas may be harmfulto the mothpollinating species. this community. In addition, the restricted flowering time of the At Bocaina, the percentage of hummingbird- species that might be moth-pollinatedin the Bocaina pollinated species is similar to that observed in grasslands may be linked to migrating pollinator cerradoareas (Barbosa,1997), but it is lowerthan that agents, which would increase the difficulty of of other high-altitudecommunities, such as arbustal observingpollinators.

This content downloaded from 143.106.1.143 on Wed, 20 Aug 2014 17:23:24 UTC All use subject to JSTOR Terms and Conditions 512 Annals of the Missouri Botanical Garden

Floweringin Relation to Pollinators cerrado species (Coutinho,1990), but, according to Safford(2001), based on his observationin the Serra At Bocaina, the bees of all families reduced their do Caparao,it is not characteristicof the Brazilian activity duringthe dry season. Such reductionin the high-altitude grasslands. A possible reason for the number of bee species and individuals in activity difference between Safford'sobservation and ours is during the winter was observed in connected that the ecological similarities between cerradoand ecosystems in southern Brazil (Alves-dos-Santos, high-altitudegrasslands may be more pronouncedat 1999; Barbolaet. al., 2000) and is related to several Bocaina, due to intense fire regimes, than at other factors, such as the reproductivephenology of bee high-altitude grasslands of southeasternBrazil (see species, climatic restriction for flying, and flower Safford, 1999a, 2001 and also Martinelli, 1989 for availability.The floweringpatterns of species polli- floristic comparison).In this sense, Bocaina grass- nated mainly by bees followed the pattern of the lands may be more pyrogenic than other grassland communityas a whole. Thus, flower resources for areas, such as those from Pico do Itatiaia, Serra dos bees, especially large bees, decrease during the dry Orgaos,and Serrado Caparao. season. In addition, such climatic conditions as low temperatures,low precipitation,and occasional frosts Specialistversus Generalist Pollination Systems are apparentlyunfavorable for flowering during the winter months. Moreover,the flowering pattern of The traditional view in studies based on the species with either nectar- or pollen-flowers over- syndromeconcept (see Faegri & van der Pijl, 1979) lapped, and this result indicates reciprocitybetween supportsthe idea that pollinationsystems mostlytend the availability of resources for adults (i.e., nectar) towardincreased specialization. Such a view has been and larvae (i.e., pollen) of bees at the community questioned in more recent times due to evidence of level, although it is difficult to ascertain if these morewidespread generalization in pollinationsystems patternsare not casual. (e.g., McDade, 1992; Waser et al., 1996; Herrera, In contrast to bees, social wasps showed a more 1996; Ollerton,1996; Olesen & Jordano,2002). These constant activity pattern throughoutthe year, and field observationsof regional or local flora contexts pollinationby this groupwas prominentduring the dry show that many plant species are visited by many season at Bocaina. Because, in this area, several animal species, and, as a result, the mean numberof species that are pollinated mainly by wasps are species of animal visitors per plant species is high. apparently also pollinated by wind, anemophily, Alternatively, we used the number of functional ratherthan pollinationby wasps, may be an ultimate groupsof pollinatorsper plant species as a parameter factor drivingflowering concentration during the dry of the degree of specialization of the pollination season in this group of species. If we consider the systemsat Bocaina.Functional groups are moresuited floweringphenology of species belongingto the most to this purpose than pollinator species for several representativepollination systems at Bocaina, only reasons(see Fensteret al., 2004). Differentspecies of the few species that flowerduring the dry season (wet the same functionalgroup, for instance, are expected season for wasp-pollinated species) seem to be to exert similar selection pressureson the floral traits fundamentalto support populations of insects. For related to pollinationmechanisms. Thus, considering example, the floral nectar of the nearly year-long- such species as distinct entities generates some floweringspecies Cupheaglutinosa, Galianthebrasi- artificiality. Another factor is that the number of liensis, and Verbenahirta, and the winter-flowering pollinatorspecies belongingto each functionalgroup Eremanthus erythropappusconstitutes a keystone varies drasticallywithin a local community,as does plant resource (sensu van Schaik et al., 1993) for the species richnessof a given functionalgroup among bee pollinatorsduring the winter. These four species distinct communitiesor regions. Thus, plant species are large shrubs or treelets (E. erythropappus)with pollinated by more diverse functional groups would many flowers and are particularly important for simply appear more generalized, despite the bi- bumblebees during this period of low flower avail- ological proprietiesof their pollinationsystem. ability. Ramirez(2002) used the functionalgroup approach Anthropogenicfires are common during the dry to compare the degree of specialization of nine season in the Bocaina grasslands.A few monthsafter tropical and temperate communities. The mean fires, we observed mass-floweringof several species numberof pollinatorgroups per plant species ranged such as Sinningia allagophylla, Microliciaisophylla, from 1.0, in a beach dune vegetation of northern Tibouchinafrigidula, T. minor, Galiantheangustifo- Brazil,to 1.85, in the forestof the VenezuelanCentral lia, and Xyris asperulla.Flowering controlled by fires Plain. Using the same functional groups as Ramirez is a well-knownand widespreadphenomenon among (2002), the Bocaina grasslands has a mean of 2.09

This content downloaded from 143.106.1.143 on Wed, 20 Aug 2014 17:23:24 UTC All use subject to JSTOR Terms and Conditions Volume 93, Number 3 Freitas & Sazima 513 2006 Pollination in a High-Altitude Grassland

pollinatortypes per plant, probablyone of the highest tural-urban mosaics or human surroundings, and values for worldwide floras. The higher degree of among short-lived plants (Vogel & Westerkamp, generalization at Bocaina is also evident if we 1991; Johnson& Steiner, 2000, 2003). In this sense, consider the distributionof species along the four frequentfires may be an additionalfactor that makes specificity categories (i.e., monophily to holophily). the environmentharsher for specialists (both plants Monophily, as used here, encompasses both the and animals)in the Bocaina grasslands. monophily and oligophily categories of Ramirez Generalist pollination systems have often been (2002). More than half of the species, at least, characterizedby small and white or greenish-colored belonged to these two categories in six tropical flowerspollinated by small insects, the diverse small communities (between 54% and 100%; Ramirez, insect (d.s.i) syndromeof Bawaet al. (1985). However, 2002), in contrastto the Bocainagrasslands where one some of the small greenishflowers at Bocainapresent third of the species are monophilous. highly specialized pollination systems (e.g., Go- However, the functional group approach to de- nioanthela hilariana), and the pollination of some termine the specialization degree of pollination holophilousspecies involvessuch largeinsects as large systems may also induce some artificiality. One bees, wasps, and butterflies.Therefore, the degree of well-demonstratedtheory is that differentpollinators generalizationshould not be inferredfrom the sizes of of a given plant species may vary in their pollinating flowers and pollinators,as previously postulated by abilities in such a way that this plant species may Ramirez(1989). Althoughrepresented mainly by the effectively specialize on its most efficient visitor families Asteraceae and Rubiaceae, the holophilous (Stebbins,1970; Schemske& Horvitz,1984; Ollerton, species at Bocaina vary in both their floral traits and 1996; Fenster et al., 2004). Anotherpoint is that the their floweringphenology. In addition, some genera definition of functional groups of pollinators (e.g., with holophilous species also have monophilousor small bees, moths, bats) is typically based on oligophilousspecies: for example,wasp pollinationin systematic boundaries. Thus, two pollinators in Baccharisand BorreriaG. Mey. and bee and butterfly distinct orders are in general placed in different pollinationin VernoniaSchreb. and in the Eupatorium functionalgroups, even if they probablyexert similar L. alliance. For such genera, the floral features that selection pressures on floral traits. For example, at determinewhether a few or several pollinatorgroups Bocaina, some of the pollen-flowersare pollinatedby are attractedto the flowersof specialistsor generalists, syrphids and small bees and some of the nectar- respectively, remain unclear. Moreover, since the flowers are pollinated by small bees and wasps; we degree of specialization/generalizationof pollination named this situation oligophily. If we combine systemsmay be influencedby factorssuch as plantlife monophilyand oligophily, we conclude that almost history, phylogenetic constraints, vegetation strata, two thirdsof the zoophilousspecies in this community successionalstatus, plant abundance, breeding system, present a specializationonto functionalgroups. Like- and local fauna(Stebbins, 1970; Vogel & Westerkamp, wise, a reanalysis of the dataset from the flora of 1991; Waser et al., 1996; Ollerton,1996; Johnson& Carlinville,Illinois, indicated that about 75% of the Steiner,2000, 2003; Ramirez,2002, 2004; Nakano& species showed specialized pollinationsystems (Fen- Washitani, 2003; Fenster et al., 2004), floral traits ster et al., 2004), contrasting with a previous alone maybe inadequateto addressthis issue properly. conclusion of generalization predominance in this community(Waser et al., 1996), which was based on ConcludingRemarks the numberof pollinatorspecies per plant species. Even using the above perspective of "wide The correlationbetween certain taxa and particular functional groups" (i.e., oligophily considered as ecological conditionsmay enhance the abundanceof a specialized system), a considerable degree of given plant groups. Thus, ecological attributes,un- generalization (more than one-third of species) is related to sexual reproduction,may promotecertain found in the Bocaina grasslands.Broadly, generaliza- taxonomicgroups with particularreproductive traits, tion is predicted as long as temporal and spatial and hence, may bias the frequency distributionof variancein pollinatorquality is appreciable;different these traits in the community(Vogel & Westerkamp, pollinatoragents do not fluctuate in unison and are 1991; Ramirez,1993). Apartfrom the wind-pollinated similarin their pollinationeffectiveness (Waser et al., species of Poaceae and Cyperaceae, the Bocaina 1996). Thus, the climatic conditions of the high- grasslands are dominatedby species of Asteraceae, altitude grasslands may favor the occurrence of which stronglyinfluence the floral features(i.e., short generalists at Bocaina. Generalizationin pollination tubular flowers that offer nectar and have exposed systems seems to be more frequent in naturally reproductiveparts) at the communitylevel. In these inclement or unstable areas, in the modern agricul- grasslands, general trends related to the types of

This content downloaded from 143.106.1.143 on Wed, 20 Aug 2014 17:23:24 UTC All use subject to JSTOR Terms and Conditions 514 Annals of the Missouri Botanical Garden

pollinationsystems and degree of generalizationmay Barbola,I. F., S. Laroca& M. C. Almeida. 2000. Utilizacao be connected to the phylogenetic floral trends of de recursos florais por abelhas silvestres (Hymenoptera, da FlorestaEstadual Passa Dois Parana, species with certain abilities- not necessarilyrelated Apoidea) (Lapa, Brasil). Revista Brasil. Entomol.44: 9-19. to - to these such as short pollination occupy habitats, Barbosa, A. A. A. 1997. Biologia Reprodutivade uma living-cycles and adaptations to frost, low mean Comunidade de Campo Sujo, Uberlandia/MG.Ph.D. temperature,high daily temperaturevariation, acidic thesis. Univ. Estadualde Campinas,Campinas, Sao Paulo. K. S. 1990. interactionsin soils, and, more recently, some kind of adaptationto Bawa, Plant-pollinator tropical rain forests. Annual Rev. Ecol. Syst. 21: 399-422. fire Safford, (see 1999a). . 1994. Pollinatorsof tropicaldioecious angiosperms: Biogeographic and palynologic studies indicate a reassessment?No, not yet. Amer. J. Bot. 81: 456-460. that high-altitude grasslands in southern and , S. H. Bullock, D. R. Perry, R. E. Coville & M. H. southeastern Brazil have to be understood as Grayum.1985. Reproductivebiology of tropical lowland relicts of the Pleistocene cold and climates rain forest trees. II. Pollinationsystems. Amer.J. Bot. 72: dry 346-356. and of the and mid-Holocene warm and early- dry Behling, H. 1997. Late Quartenaryvegetation, climate and climates (Behling, 1997). Because of their rela- fire historyof the Araucariaforest and camposregion from tively recent origin, these grasslands are unlikely Serra Campos Gerais, Parana State (South Brazil). Rev. to have coevolved commu- Paleobot.Palynol. 97: 109-121. produced highly plant M. Sazima & I. Sazima. 2000. some relictual Buzato, S., Hummingbird- nities, although plant-pollinator pollinatedfloras at three Atlantic Forest sites. Biotropica interactions could have been retained in cases 32: 824-841. where both the plant and its pollinator colonized Camerik,A. M. & W. J. A. Werger. 1981. Leaf character- the mountain tops. These grasslands are linked to istics of the flora of the high plateau of Itatiaia, Brazil. of and retraction due to Biotropica13: 39-48. episodes expansion glacial Chittka, L., A. Shmida, M. & R. Menzel. 1994. events Vorobyev during the Quaternary (Behling, 1997; Ultravioletas a componentof flower reflections, and the Safford, 1999a). Such a situation may have favored colour perception of Hymenoptera. Vision Res. 34: species with the ability to occupy new habitats 1489-1508. A. A. & S. 2001. flowers of quickly. Dependence on a few specialized polli- Cocucci, Vogel. Oil-producing Sisyrinchiumspecies (Iridaceae)and their pollinatorsin nators may be less suitable than more generalist southernSouth America.Flora 196: 26-46. pollination systems in the occupation of a new Coutinho,L. M. 1990. Fire in the ecology of the Brazilian habitat, because previous pollinator agents may be cerrado.Pp. 82-105 in J. G. Goldhammer(editor), Fire in left behind. In fact, generalized pollination the TropicalBiota. Springer-Verlag, Berlin. T. M., S. G. Weller & A. K. Sakai. 2002. The systems have been an Culley, may ecological advantage evolution of wind pollination in angiosperms. 17: for plants colonizing post-glacial landscapes 361-369. (Johnson & Steiner, 2000, 2003). Alternatively, Cummins,K. W. 1973. Trophicrelations of aquatic insects. for those species with more specialized pollination Annual Rev. Entomol.18: 183-206. some floral traits that are at Cure,J. R. 1989. Revisao de PseudagapostemonSchrottky e systems, frequent de Ha- such as descrigao Oragapostemon,gen.n. (Hymenoptera, Bocaina, spontaneous self-pollination lictidae). Revista Brasil. Entomol.33: 229-335. mechanisms and long-lived flowers, may be Eiten, G. 1970. A vegetacao do Estado de Sao Paulo. Bol. advantageous to both plant establishment and Inst. Bot. (Sao Paulo) 7: 1-147. maintenance in these grasslands characterized by . 1992. NaturalBrazilian vegetation types and their causes. Anais. Acad. Brasil. Ci. 64: 35-65. unpredictable and scarce pollinators. Endress, P. K. 1994. Diversity and evolutionarybiology of tropicalflowers. CambridgeUniv. Press, Cambridge. LiteratureCited Faegri, K. & L. van der Pijl. 1979. The principles of pollinationecology. PergamonPress, New York. M. T. A. 1977. Alonso, Vegetacao da regiao sudeste do Faria, G. M. 1994. A Flora e a Fauna Apicola de um Brasil. Revista Brasil. Geogr.34: 91-118. Ecossistema de Campo Rupestre, Serra do Cipo - MG, Alves-dos-Santos,I. 1999. Abelhas e plantas meliferas da Brasil: Composigao, Fenologia e suas Interacoes. mata atlantica,restinga e dunas do litoral norte do estado Ph.D. thesis. Univ. Estadual Paulista., Rio Claro, Sao do Rio Grandedo Sul, Brasil. Revista Brasil. Entomol.43: Paulo. 191-223. Feinsinger,P. & R. K. Colwell. 1978. Communityorganiza- Anderson, G. J., G. Bernardello,T. F. Stuessy & D. J. tion amongNeotropical feeding birds. Amer.Zoologist 18: Crawford.2001. Breeding system and pollination of 779-795. selected plants endemic to Juan FernandezIslands. Amer. Fenster,C. B., W. S. Armbruster,P. Wilson, M. R. Dudash& J. Bot. 88: 220-233. J. D. Thomson. 2004. Pollination syndromesand floral Arroyo, M. T. K., R. Primack & J. Armesto. 1982. specialization.Annual Rev. Ecol. Evol. Syst. 35: 375-403. Communitystudies in pollination ecology in the high Freitas, L. 2004. Mecanismosde auto-polinizagaoesponta- temperateAndes of CentralChile. I. Pollinationmechan- nea em cinco especies de Gentianaceaeem uma area de isms and altitudinalvariation. Amer. J. Bot. 69: 82-97. campo de altitude. Pp. 1-8 in Proceedings of the XV Asker, S. E. & L. Jerling. 1992. Apomixis in Plants. CRC Congressoda Sociedade Botanicade Sao Paulo. Ubatuba, Press, Boca Raton. Sao Paulo.

This content downloaded from 143.106.1.143 on Wed, 20 Aug 2014 17:23:24 UTC All use subject to JSTOR Terms and Conditions Volume 93, Number 3 Freitas & Sazima 515 2006 Pollination in a High-Altitude Grassland

& M. Sazima. 2001. Nectar features in Esterhazya Ledru, M., M. L. Salgado-Laboriau& M. L. Lorscheitter. macrodonta,a hummingbird-pollinatedScrophulariaceae 1998. Vegetation dynamics in south and central Brazil in southeasternBrazil. J. PL Res. 114: 187-191. duringthe last 10,000 years BP. Rev. Palaeobot.Palynol. & . 2003a. Floral biology and pollination 99: 131-142. mechanismsin two Viola species- from nectar to pollen Martin,F. N. 1959. Staining and observingpollen tubes in flowers?Ann. Bot. (London)91: 311-317. the style by means of fluorescence. Stain Tech. 34: & . 2003b. Daily blooming pattern and 125-128. pollinationby syrphidsin Sisyrinchiumvaginatum (Irida- Martinelli,G. 1989. Camposde Altitude. EditoraIndex, Rio ceae) in southeastern Brazil. J. Torrey Bot. Soc. 130: de Janeiro. 55-61. . 1994. ReproductiveBiology of Bromeliaceaein the , L. Galetto & M. Sazima. 2006. Pollination by Atlantic Rainforestof SoutheasternBrazil. Ph.D. thesis. hummingbirdsand bees in eight syntopic species and Univ. Saint Andrews,Saint Andrews. a putativehybrid of Ericaceae in southeasternBrazil. PL . 1997. Biologia reprodutivade Bromeliaceae na Syst. Evol. 258: 49-61. ReservaEcologica de Macae de Cima.Pp. 213-250 in H. Giulietti,A. M. & J. R. Pirani. 1988. Patternsof geographic C. Lima & R. R. Guedes-Bruni(editors), Serra de Macae distributionof some plant species from the Espinhaco de Cima: DiversidadeFloristica e Conservacaoem Mata Range, and , Brazil. Pp. 39-69 in W. Atlantica. Jardim Botanico do Rio de Janeiro, Rio de R. Hever & P. F. Vanzolini (editors), Proceedings Janeiro. of a Workshop on Neotropical Distribution Patterns. McDade,L. A. 1992. Pollinatorrelationships, biogeography, AcademiaBrasileira de Ciencias, Rio de Janeiro. and phylogenetics.BioScience 42: 21-26. Goldblatt, P., J. C. Manning & P. Bernhardt. 1995. Modenesi,M. C. 1988. Significadodos depositoscorrelativos Pollination biology of Lapeirousiasubgenus Lapeirousia quaternarios em Campos do Jordao - Sao Paulo: (Iridaceae) in southern Africa; floral divergence and implicates paleoclimaticase paleoecologicas.Bol. Inst. adaptationfor long-tonguedfly pollination.Ann. Missouri Geol. 7: 1-155. Bot. Gard.82: 517-534. Moldenke, A. R. 1976. Californiapollination ecology and Goldenberg,R. & G. J. Shepherd. 1998. Studies on the vegetationtypes. Phytologia34: 305-361. reproductive biology of Melastomataceaein "cerrado" & P. G. Lincoln. 1979. Pollination ecology in vegetation.PL Syst. Evol. 211: 13-29. montaneColorado. Phytologia 42: 349-379. Grantsau, R. 1989. Os beija-flores do Brasil. Momose, K., T. Yumoto, T. Nagamitsu, M. Kato, H. Expressaoe Cultura,Rio de Janeiro. Nagamasu, S. Sakai, R. D. Harrison, T. Itioka, A. A. Heithaus, E. R. 1974. The role of plant-pollinatorinterac- Hamid& T. Inoue. 1998. Pollinationbiology in a lowland tions in determiningcommunity structure. Ann. Missouri dipterocarpforest in Sarawak,Malaysia. I. Characteristics Bot. Gard.61: 675-691. of the plant-pollinatorcommunity in a lowlanddipterocarp . 1979. Flowervisitation records and resourceoverlap forest. Amer. J. Bot. 85: 1477-1501. of bees and wasps in NorthwestCosta Rica. Brenesia 16: Moreira,A. A. N. & C. Camalier.1977. Relevo da regiao 9-52. sudeste do Brasil. Rev. Brasil. Geogr.34: 1-50. Herrera,CM. 1996. Floral traits and plant adaptationto Nakano, C. & I. Washitani.2003. Variabilityand special- insect pollinators:a devils's advocateapproach. Pp. 65-87 ization of plant-pollinatorsystems in a northernmaritime in D. G. Lloyd & S. C. H. Barret(editors), Floral Biology: grassland.Ecol. Res. 18: 221-246. Studies on Floral Evolutionin Animal-pollinatedPlants. Nimer, E. 1977. Climatologiada regiao sudeste do Brasil. Chapman& Hall, New York. Rev. Brasil. Geogr.34: 3-48. Johnson,S. D. & K. E. Steiner. 2000. Generalizationversus Olesen, J. M. & P. Jordano.2002. Geographicpatterns in specialization in plant pollinator systems. TREE 15: plant-pollinator mutualistic networks. Ecology 83: 140-143. 2416-2424. & . 2003. Specialized pollinationsystems in Oliveira,P. E. & P. E. Gibbs. 2000. Reproductivebiology of southernAfrica. S. AfricanJ. Sci. 99: 345-348. woody plants in a cerrado communityof Central Brazil. Kearns,C. A. 1992. Anthophilousfly distributionacross an Flora 195: 311-329. elevation gradient.Amer. Midi. Naturalist127: 172-182. Ollerton, J. 1996. Reconciling ecological processes with & D. W. Inouye. 1993. Techniques for Pollination phylogenetic patterns: the apparent paradox of plant- Biologists. Univ. ColoradoPress, Niwot. pollinatorsystems. J. Ecol. 84: 767-769. Kevan,P. G. 1975. Sun trackingsolar furnacesin high arctic Pnmack, R. B. & D. W. Inouye. 1993. Factors affecting flowers:significance for pollination and insects. Science pollinator visitation rates: a biogeographiccomparison. 189: 723-726. Curr.Sci. 65: 257-262. & H. G. Baker. 1983. Insects as flowervisitors and Proctor,M., P. F. Yeo & A. Lack. 1996. The NaturalHistory pollinators.Annual Rev. Entomol.28: 407-453. of Pollination.Timber Press, Portland. , E. A. Tikhmenev & M. Usui. 1993. Insects and Radford,A. E., W. C. Dickinson,J. R. Massey& C. R. Bell. plants in the pollinationecology of the boreal zone. Ecol. 1974. VascularPlant Systematics.Harper & Row Publish- Res. 8: 247-267. ers, New York. Koppen,W. 1948. Climatologia:con un estudio de los climas Ramirez, N. 1989. Biologia de pohmzacion en una de la tierra.F. C. E., Ciudadde . comunidad arbustiva tropical de la Alta Guayana Kornerup,A. & J. H. Wanscher.1963. Taschenlexikonder Venezolana.Biotropica 21: 319-330. Farben.Musterschmidt, Zurich. . 1993. Reproductivebiology in a tropicalshrubland Kress, J. W. & J. H. Beach. 1994. of VenezuelanGuayana. J. Veg. Sci. 4: 5-12. reproduction.Pp. 161-182 in L. A. McDade,K. S. Bawa, . 2002. Pollination specialization and time of H. A. Hespenheide& G. S. Hartshorn(editors), La Selva. pollination on a tropical Venezuelan plain: variationsin Univ. ChicagoPress, Chicago. time and space. Bot. J. LinneanSoc. 145: 1-16.

This content downloaded from 143.106.1.143 on Wed, 20 Aug 2014 17:23:24 UTC All use subject to JSTOR Terms and Conditions 516 Annals of the Missouri Botanical Garden

. 2004. Ecology of pollinationin a tropicalVenezue- Silveira,F. A. & J. R. Cure.1993. High-altitudebee faunaof lan savanna.PL Ecol. 173: 171-189. southeasternBrazil: implications for biogeographicpat- & Y. Brito. 1992. Pollination biology in a palm terns (Hymenoptera:Apoidea). Stud. Neotrop. Fauna swampcommunity in the VenezuelanCentral Plains. Bot. Environm.28: 47-55. J. LinneanSoc. 110: 277-302. Singer, R. B. 2001. The pollination biology of Habenaria Reis, M. G., A. D. de Faria,V. Bittrich,M. C. E. Amaral& A. J. parviflora (Orchidaceae:Habenariinae) in southeastern Marsaioli.2000. Thechemistry of flowerrewards - Oncidium Brazil. Darwiniana39: 201-207. (Orchidaceae).J. Brazil.Chem. Soc. 11: 600-608. & A. A. Cocucci. 1999. Pollinationmechanisms in Renner, S. S. & J. P. Feil. 1993. Pollinators of tropical four sympatricsouthern Brazilian Epidendroideae orchids. dioecious angiosperms.Amer. J. Bot. 80: 1100-1107. Lindleyana14: 47-56. Root, R. B. 1967. The niche exploitationpattern of the blue- Smith, A. P. & T. P. Young. 1987. Tropical alpine plant gray gnatcatcher.Ecol. Monogr.37: 317-350. ecology. Annual Rev. Ecol. Syst. 18: 137-158. D. W. 1989. and Natural of Roubik, Ecology History Tropical Stebbins, G. L. 1970. Adaptive radiation of reproductive Bees. Univ. Cambridge Press, Cambridge. characteristicsin angiosperms,I: pollinationmechanisms. H. D. 1999a. BrazilianParamos. I. An introduction Safford, Annual Rev. Ecol. Syst. 1: 307-326. to the environmentand of the physical vegetation campos Truylio,B., B. Harter-Marques& W. Engels. 2002. Biologia de altitude.J. Biogeogr.26: 693-712. floral e polinizagao de Sisyrinchiummicranthum (Irida- . 1999b. BrazilianParamos. II. Macro-and mesocli- na do Planaltodas Araucariasdo Rio Grande mate of the de altitude and affinities with ceae) Regiao campos high do Sul, Brasil. Biociencias 10: 11-24. mountainclimates of the Andes and CostaRica. tropical J. Van C. J. W. & S. J. 1993. The 26: 713-737. Schaik, P., Terborgh Wright. Biogeogr. of forests: and . 2001. BrazilianParamos. III. Patternsand rates of phenology tropical adaptive significance for consumers.Annual Rev. Ecol. in the de altitude. consequences primary postfireregeneration campos Biotropica. 24: 353-377. 33: 282-302. Syst. S. 1990. of the in the of S. Buzato & M. Sazima. 1995. The saw-billed Vogel, History Malpighiaceae light Sazima, I., Mem. New York Bot. Gard. 55: hermit naeviusand its flowersin southeastern pollination ecology. Ramphodon 130-142. Brazil.J. Ornithol.136: 195-206. . 1998. Remarkable nectaries: , & . 1996. An of structure, ecology, assemblage II. Nectarioles. Flora 193: hummingbird-pollinatedflowers in a montane forest in organophyleticperspectives. southeasternBrazil. Bot. Acta 109: 149-160. 1-29. & C. 1991. Pollination:an , & . 1999. Bat-pollinated flower Westerkamp. integrating factor of biocenoses. 159-170 in A. Seitz & V. assemblagesand bat visitors at two Atlanticforest sites in Pp. Brazil. Ann. Bot. 83: 705-712. Loeschcke (editors),Species Conservation:a Population- Birkhauser Basel. , S. Vogel, A. Cocucci & G. Hausner. 1993. The biological Approach. Verlag, and perfumeflowers of Cyphomandra(Solanaceae): pollination Wainwright,C. M. 1978. The floral biology pollination 105: by euglossine bees, bellows mechanism,osmophores, and ecology of two desert lupines. Bull. TorreyBot. Club volatiles. PL Syst. Evol. 187: 51-88. 24-38. Schemske, D. W. & C. C. Horvitz. 1984. Variationamong Waser, N. M., L. Chittka,M. V. Price, N. M. Williams & J. floral visitors in pollination ability: a precondition for Ollerton.1996. Generalizationin pollinationsystems, and mutualismspecialization. Science 225: 519-521. why it matters.Ecology 77: 1043-1060. Segadas-Vianna,F. & L. Dau. 1965. Ecology of the Itatiaia Yeo, P. F. 1993. Secondary Pollen Presentation:Form, range, southeasternBrazil. II. Climates and altitudinal Functionand Evolution.Springer (Plant Syst. Evol. Suppl. climatic zonation.Arq. Mus. Nacional 53: 31-53. 6.), Vienna. Silberbauer-Gottsberger,I. & G. Gottsberger. 1988. A Zeisler, M. 1938. Uber die Abgrenzungder eigentlichen polinizagaode plantas do cerrado. Revista Brasil. Biol. Narbenflache mit Hilfe von Reaktionen. Beih. Bot. 48: 651-663. Centralbl.58: 308-318.

This content downloaded from 143.106.1.143 on Wed, 20 Aug 2014 17:23:24 UTC All use subject to JSTOR Terms and Conditions