Selbyana 9: 52-60

COPING WITH THE EPIPHYTIC EXISTENCE: POLLINATION STRATEGIES

JAMES D. ACKERMAN Department of Biology, University of Puerto Rico, Rio Piedras, Puerto Rico 00931

ABSTRACT. Populations of epiphytic flowering plants are often composed of scattered individuals or small, hyperdispersed clusters. These characteristics create conditions for pollination' somewhat different from that encountered by many terrestrial plants. For example, plant distribution and size constraints imposed by the epiphytic habit may have limited floral apparency or competitiveness for adequate pollinator service. To cope with these problems, epiphytes employ one of several specialized pollinatipn strategies which involve deception, or exclusive or unique rewards. Each of the specialized strategies is probably most efficient for diffusely distributed species and may have either a terrestrial or an epiphytic origin. In one system, plants produce few flowers per day for long periods. The blossoms contain a high quality reward which is exclusively accessible to large traplining pollinators. In another, pollinator attraction is based on deceit which by-passes constraints directly imposed by pollinator foraging energetics. The decep­ tion may exploit sexual, feeding or egg-laying behaviors of their pollinators. A third strategy employed by many neotropical epiphytes is pollination by male euglossine bees. The flowers produce fragrances which serve as specific attmctants and perhaps precursors to sex pheromones of the bees. Pollination occurs when male bees arrive and collect the fragrance compounds. All three strategies are not unique to epiphytes. Both terrestrials and epiphytes with shared size and dispersion constraints seem to have more specialized pollination biologies than their more densely populated and floriferous neighbors. The dispersion-special­ ization hypothesis genemtes testable predictions, some of which are presented.

Epiphytic flowering plants possess an incred­ Obviously, the weight of a large epiphyte could ible array of morphological, anatomical and be more than a phorophyte Oive plant substrate) physiological adaptations associated with the could bear without falling or breaking. Secondly, stresses imposed by their habitat (Madison, 1977). resources might be better allocated towards sex­ Although considerable vegetative diversity exists ual reproduction rather than vegetative struc­ among epiphytes, much evolution ofspecies-rich tures for plants growing in an unstable, ephem­ groups has been attributed to adaptive radiation eral habitat (Schaffer, 1974; Benzing, 1976) in via zoophilous pollination mechanisms (Pij1 & which bark exfoliates, branches are shed, and Dodson, 1966; Ashton, 1984). Nevertheless, the trees die. Epiphytes attached to tree trunks and role of pollination biology in the evolution and major limbs may be old, but others, especially adaptation of epiphytes is rarely mentioned in twig epiphytes, grow on short-lived substrates the literature (e.g., Richards, 1952; Johansson, where life expectancy is brief. Precocious flow­ 1974; Madison, 1977; Benzing, 1981; Liittge, ering is one way epiphytes may cope with the 1985; but discussed in Benzing & Atwood, 1984). vagaries oftheir substrate, even ifit means small­ If pollination biology is of special significance to er, suboptimal floral displays (Schemske, 1980). the evolution of epiphytic groups, then there Epiphyte size may also be limited because many should exist patterns in pollination strategies that are stressed (Janzen, 1975; Benzing, 1984; Mon­ relate to shared environmental conditions. I shall talvo & Ackerman, 1986) and lack resources (nu­ attempt to identify common pollination strate­ trients, water, light) for rapid growth and main­ gies among epiphytes and those factors that may tenance of large plant bodies. Indeed, some of affect their evolution. Does the epiphytic habit the more bizarre vegetative structures of epi­ impart unusual constraints on plant-pollinator phytes are special adaptations for water and nu­ interactions? If so, then how do epiphytes cope trient procurement (e.g., Benzing, 1976; Janzen, with these conditions? 1974a). A second probable constraint on the pollina­ CONSTRAINTS tion biology of epiphytes is their <.iis1:riJ)Utipn within a community. Plant dispersion affects the '\ Th~.!t!latively small size ofepiphytes may lim­ availability ()fpollinator rewards which can have . , it flower production. Consequently, floral dis­ a profound influence on the evolution of polli­ prays rimy be insufficiently large to compete for nation strategies (Janzen, 1971; Regal, 1977; pollinator attentions on the same basis as the Stiles, 1978a; Augspurger, 1981; Burger, 1981; more floriferous trees, shrubs and lianas. There Ackerman, 1986). Chamcteristics of the phoro­ are several ecological reasons for small plant size. phyte can strongly influence vertical and hori- 52 1986] ACKERMAN: EPIPHYTE POLLINATION 53

1.0 A • B e

w u Z e .. • ee • I­ 0.01 Orchidaceae) are SPECIALlZATION likely traplined. Documentation, though, is scarce, yet the occurrence oftrapline pollination Because of size and dispersion characteristics, may be inferred from floral morphology and epiphytes would appear to be weak competitors flowering behavior because the syndrome is well for the same pollinator services as larger arbo­ established for terrestrial herbs, vines, lianas and reous neighbors. However, strong selection may shrubs (Janzen, 1971; Feinsinger, 1976; Acker­ exist for more specialized pollinator relation­ man, 1985). Characteristically, these plants ships to circumvent poor apparency or compe­ bloom for long periods and maintain a consistent tition for pollination to draw pollinators to widely daily production of a few, often conspicuous and dispersed flowers. Thus, pollination strategies of fragrant flowers. A high quality pollinator reward epiphytes should be somewhat more specialized is produced and protected from exploitation by than those of more floriferous trees and shrubs, all but a few pollinator and robber species (Jan­ and specialization may be related to dispersion zen, 1974b). The reward is usually nectar or pol­ patterns. len, but oils, resins and perhaps fragrances are Supporting evidence for this hypothesis is important in some families (Armbruster & Web­ meager, but most available data are consistent. ster, 1979; Simpson & Neff, 1981; Armbruster, At the community level, the diversity of polli­ 1984; Renner, 1984; Ackerman, 1985). nation mechanisms found in canopy trees of a Traplining pollinators are robust animals, lowland Costa Rican forest was relatively less commonly long-billed birds, bats, large bees and than that found in the subcanopy and also among hawkmoths (e.g., Frankie & Baker, 1974; Linhart epiphytes, although the epiphytes were not yet & Mendenhall, 1977; Feinsinger & Colwell, funy quantified (Bawa et aI., 1985). This pattern 1978). They develop foraging routes that incor­ may reflect differences in specialization of plant­ porate scattered, but predictable, high yield, re­ pollinator interactions where the canopy trees source stops. Pollinators may follow the same have more generalized floral biologies. routes daily (Janzen, 1971), but under certain Additional support comes from species-level conditions, their foraging bouts may be quite comparisons. Of two Brazilian aroids, the dif­ plastic (Ackerman et al., 1982). Nevertheless, fusely distributed species, Philodendron selloum trapliners are generally strong fliers and probably C. Koch, has a highly specialized floral biology important long-distance pollen vectors for very whereas the densely distributed species, P. bi­ scattered individuals of epiphyte populations pinnatifidum Schott, has a relatively generalized (Janzen, 1974b). pollination system (Gottsberger & Amaral, 1984). Because of highly desirable and predictable re- 1986] ACKERMAN: EPIPHYTE POLLINATION 55 wards, pollinator visits to trapline plants are to self-compatible (Grove in Bawa et ai., 1985). Se­ a certain extent, density-independent. Within lection for physiological incompatibility mech­ some critical range of densities and rewards, vis­ anisms would be low for species which are nor­ its to scattered individuals must become ener­ mally sparsely distributed and produce few getically feasible. Trapline pollination may flowers at a time (Kress, 1983b). When these therefore release competition for pollination be­ species form dense populations, inbreeding tween epiphytes and larger, more generalist flow­ depression can occur (e.g., Schemske, 1983), ering plants with large floral displays (Bawa, which may exert strong selection for developing 1983). Nevertheless, some competition for pol­ incompatibility mechanisms or against extensive lination may have influenced flowering times of clumping. Since neither of these features appear some species. In lowland central Panama, most to be prevalent among traplined plants, terres­ traplined nectar-hosts pollinated by euglossine trial or epiphytic, the incidence of high flower bees flowered during the mid-wet season even density and consequently geitonogamy is prob­ though their pollinators were more abundant ably too unusual to strongly influence selection. during the dry and early wet seasons. During Thus, the frequency of traits associated with seasons of high bee abundance, alternative nectar trapline pollination probably increased in pop­ resources with larger floral displays were in peak ulations because they offset the disadvantages of flower: the iianas, trees and shrubs (Ackerman, small, local population sizes. As a result, trapline 1985). Competition for pollination, although not pollination may lose some of its advantages in the only plausible explanation (cf. Stiles, 197 8b), dense plant populations (cf. Rabinowitz, 1981). could have produced and maintained the phe­ The evolution offioral traits in traplined plants nological segregation of the two plant groups. may have been a result of several different sources Competition for pollination via interspecific of selection other than competition for pollina­ pollen transfer is likely to have generated char­ tion. In some cases, resources may limit fruit and acter divergence and speciation (cf. Rathcke, seed production so that the evolution of floral 1983; Waser, 1983). It may have had a large role traits may be driven by sexual selection (e.g., in the evolution of epiphyte pollination systems. Stephenson & Bertin, 1983). In other situations, Because a trapliner can incorporate numerous mutualistic interactions could be important. For species in a given foraging bout, interspecific pol­ example, a group of plant species can maintain linations could occur and thereby limit conspe­ populations of long-lived pollinators with se­ cific pollinations and reproductive success. How­ quential flowering phenologies (Stiles, 1977). Ef­ ever, Haber (in Janzen, 1983: 633) noted that fective mutualism, though, is likely an outcome when trapiining hawkmoths visit several plant and not a cause of sequential flowering times of species, pollen from the different plants is de­ sympatric species (Waser & Real, 1979). posited on distinct regions of the pollinator's body. Though untested, the assemblage of spe­ Deception Pollination cific pollen deposition sites may be a conse­ quence of character divergence mediated by se­ The second specialized pollination strategy lection pressures caused by interspecific pollen common to epiphytes employs deception in which transfers. Clearly, we need thorough studies of the flowers offer little or no pollinator reward. pollen loads on both pollinators and stigmas, and This is particularly common among orchids, ter­ of fitness consequences of interspecific pollina­ restrial and epiphytic, but is also known in other tions (see Kress, 1983a; Kohn & Waser, 1985). families, mostly among terrestrial herbs, vines Traplined plants are often diffusely distributed and occasionally trees (Gentry, 1974; Baker, 1976; (e.g., Feinsinger, 1976; Stiles, 1978a) but they do Dafni, 1984). Deception mechanisms involve dif­ remain attractive to pollinators when growing in ferent types of mimicry and the best known are dense populations (Ackerman et aI., 1982; Bus­ those based on the reproductive or food foraging by, pers. comm.). However, pollinator foraging behaviors of their pollinators (Pijl & Dodson, behaviors, pollen flow characteristics and con­ 1966; Ackerman, 1986). sequently breeding systems may change (Handel, Deception that capitalizes on the reproductive 1985). In dense populations, pollinator move­ behaviors of pollinators sometimes involves ment between plants may, for example, be short­ mimicry of insect brood sites. Pollination occurs ened and pollen flow distances reduced. The when female insects visit a succession of flowers probability of geitonogamous pollinations could to lay their eggs on what appears to be an ap­ increase, especially ifthe high population density propriate substrate for larvae. Although the best was a result of clonal propagation, a common documented cases involve terrestrial species (e.g., phenomenon among epiphytes. Seed production Paphiopedilum: Atwood, 1985; Epipactis: Ivri & would still occur because many traplined plants Dafni, 1977), some epiphytes may be involved (tropical forest herbs and probably epiphytes) are as well. Good candidates are Dracula (Orchi- 56 SELBYANA [Volume 9

daceae) species, whose flowers reportedly mimic terns may be frequency-dependent. Reproduc­ fungi, the larval substrate offungus gnats (Sciar­ tive success should be inversely related to flower idae, Mycetophilidae; Vogel, 1978a), and some densities and mimicry should be less effective Araceae (Dafni, 1984). when the relative frequency of models is low. In Mate mimicry, better known in its extreme fact, because of these effects we would expect manifestation as pseudocopulation, is also more deception flowers, like those that are traplined, thoroughly documented for terrestrial orchids, to work best for species with diffuse populations especially the European Ophrys (Kullenberg & (e.g., epiphytes). Unfortunately, such ecological Bergstrom, 1976) and several Australian genera relationships are rarely studied (Bierzychudek, (Stoutamire, 1975, 1983). However, reliable pre­ 1981; Dafni & Ivri, 1981a, 1981b; Ackerman, liminary observations indicate that some neo­ 1986). tropical epiphytes may be "pseudocopulated" as The evolutionary loss of pollinator rewards in well, e.g., Trichoceros (Pijl & Dodson, 1966), and deceptive epiphytes and terrestrials presumably Oncidium henekenii Schomb. ex Lindl. in His­ occurred without the loss of reproductive suc­ paniola (Dod, 1976). Flowers that employ mate cess. This shift most likely occurred in hyper­ mimicry to attract pollinators are similar to pol­ dispersed populations, such as those common to linator mates in some critical characteristics that epiphytes. Reward flowers below a critical den­ elicit mating behaviors by stimulating the visual, sity would not meet the energetic needs of their tactile and olfactory senses of their pollinators. pollinators and consequently would receive no The deception may be sufficiently complete that more than a few cursory visits. Variants lacking pollinating insects may attempt to copulate with a reward still would receive a few exploratory, the flowers (Stoutamire, 1983). pollinator visits. Resources previously used for Deception also operates when flowers mimic reward production could be reallocated to pro­ pollinator food resources. Self-mimicry probably duce more flowers, mature more fruit or improve occurs in the monoecious terrestrial and epi­ longevity, if resources were limiting (Ackerman phytic begonias (Begoniaceae). Stigmas offemale & Montero Oliver, 1985; Ackerman, 1986). In flowers are similar to the stamens of male flow­ these cases, fruit and seed set may sometimes be ers. Pollinating bees probably visit the males to pollination limited, but fruit production over the collect pollen and occasionally mistake a female lifetime of the individual may be resource lim­ for a pollen flower (Vogel, 1978b). ited, a condition which may occur for the epi­ Many food-deceptive plants do not have spe­ phytic orchid, Ionopsis utricularioides (Sw.) Lindl. cific models and only appear as "likely" food (Montalvo & Ackerman, 1986). Hand pollina­ plants to naive pollinators (Gentry, 1974; Ack­ tions increased fruit set over natural levels but erman, 1981 b; Little, 1983). The epiphytic Coch­ the higher fruit set affected subsequent growth leanthes lipscombiae (Rolfe) Garay (Orchida­ and flowering. Nutritional constraints imposed ceae) had two sources of uninitiated pollinators: on epiphytes are well known (e.g., Benzing, 1981) 1) recently emerged, young bees who presumably and suggest that ultimate resource limits to sex­ had not yet established foraging routes, and 2) ual reproduction may occur. bees that were exploring additional resources during a period of rapid turnover in the flowering Male Euglossine Pollination of their nectar hosts (Ackerman, 1983a). Deception pollination systems can be compli­ Epiphytes throughout the tropics probably cated. For Cochleanthes lipscombiae, some level employ trapline and deception pollination strat­ of specific model-mimicry may also operate. The egies. However, some geographic regions have orchid shares pollinators, phenology, geography, unique pollination systems. In the neotropics, a habitat and general floral appearance with the large number of epiphytes are pollinated by male traplined liana, Clitoria javacensis HBK (Le­ euglossine bees as they collect floral fragrances. guminosae; Ackerman, 1983a). Combining the About 600 species of orchids, mostly epiphytes, bases for attraction may not be uncommon. The and some terrestrial and epiphytic members of very large orchid genus Epidendrum is primarily the Araceae, Gesneriaceae and other families a food-deceptive group, and most species are pol­ produce floral fragrances that serve as both an linated by moths and butterflies. Floral fra­ attractant and reward for the male bees (Dodson grances may be pheromone-like and attract only et al., 1969; Williams & Dressler, 1976; Arm­ one sex (Wagner, 1973; Adams & Goss, 1976), bruster & Webster, 1979; Croat, 1980; Dressler, but the pollination event actually involves feed­ 1982; Williams, 1982). Many of the orchids are ing behavior, the extension of the tongue and quite specialized and well known for their bizarre probing for non-existent nectar. flowers, exotic fragrances and complex pollina­ Visitation frequencies and reproductive suc­ tion mechanics. The bees possess specialized cess of species with deception pollination sys- structures for collection and storage of the fra- 1986] ACKERMAN: EPIPHYTE POLLINATION 57 grance compounds and perhaps convert the terrestrial plant, Spathiphyl/um friedrichsthalii chemicals to sex pheromones (Williams & Whit­ Schott (Araceae), was pollinated by numerous ten, 1983). The plants are clearly dependent on species offragrance-foraging euglossines and pol­ the bees for pollination but thus far there are no len-foraging stingless bees (Montalvo & Acker­ data indicating that the bees are dependent on man, unpubl.). The male euglossine visits were the flowers for fragrances (Ackerman, 1983b; daily but few, and the bees visited only a small Armbruster, unpubl.; Roubik & Ackerman, un­ fraction of the available fragrance producing publ.). spadices. Visitations by male euglossine bees may Both male euglossine and trapline pollination be somewhat independent of the number of are based on reward systems. They seem com­ available inflorescences in the population. Con­ mon among hyperdispersed epiphytes, but pol­ sequently, a few plants may attract as many eu­ linator foraging behavior is probably distinct be­ glossine bees as a clump of numerous plants. In tween the two modes. With a few possible @y ~se, the euglossine pollination strategy seems exceptions (e.g., Dalechampia, Euphorbiaceae: well suited for neuti-aliziiig the constraints ofsmall Armbruster & Webster, 1979; Peristeria elata, plant size, uncertainty and dispersion imposed Orchidaceae: pers. obs.), most male euglossine­ . by the epiphytic habitat. pollinated species do not produce flowers daily for long periods of time as is typical of traplined CONCLUSIONS plants. Although individual flowers may last for long periods oftime, when pollinated, they cease Although only a few scattered reports on epi­ fragrance production and become unavailable for phyte pollination systems exist, epiphytes appear repeat visits. Unlike nectar, fragrances are prob­ to have employed relatively specialized polli­ ably not nutritional requirements (Ackerman & nation strategies. The epiphyte habitat imposes Montalvo, 1985), and the frequency by which two constraints that may have profoundly influ-" the bees must collect a particular compound is enced the evolution ofthese strategies: plant size . unknown. The need for collection may be irreg­ and plant dispersion. The former is usUally small ular and the foraging behavior may be oppor­ and affects floral display size and reward pro­ tunistic. duction. The latter is often hyperdispersed and Similar to traplining pollinators, male euglos­ influences pollinator foraging behavior. The three sines may carry pollen from different fragrance well-known pollination strategies ofneotropical species on different parts of their bodies (Ack­ epiphytes (traplining, deception and male eu­ erman, 1983b). Differential placement of pollen glossine pollination) are solutions to these prob­ by fragrance flowers may be a result of compe­ lems but none is unique to epiphytic plants. tition via interspecific pollen transfer. Closely The different strategies have some similar char­ related, sympatric species, though, do not share acteristics which Inay be explained by the dis­ pollinators (Dressler, 1968; Ackerman, 1983b) persion-specialization hypothesis. In essence, it because the initial stages of reproductive isola­ states that those species with small floral displays tion probably occur through evolutionary changes and populations of scattered individuals should in attraction specificity of the floral fragrance have specialized floral biologies. (Williams & Dodson, 1972). A slight change in Autecological studies of epiphyte dispersion a floral fragrance may attract a very different and pollination are clearly needed for all epi­ group of bees. However, higher taxa are often phyte groups but they should reach beyond the distinguished on the basis offloral characteristics task of describing and cataloging plant-pollina­ associated with pollination mechanics (e.g., tor interactions. Epiphyte pollination obviously Catasetinae: Dodson, 1975; Zygopetalinae: cannot be studied without epiphytes, but eco­ Dressler, 1981). These taxonomic categories do logical and evolutionary processes can be ex­ share pollinators which suggests that interspecific amined using more accessible terrestrial species pollen transfer may have served as a strong se­ with analogous floral and dispersion character­ lection force for differential pollen placement and istics. Employing a combination of terrestrial and for generating supraspecific evolution (Dressler, epiphytic systems, the rather complex disper­ 1981). sion-specialization hypothesis can be tested at Because individuals can, by means of their flo­ various levels. ral fragrance; draw pollinators from a kilometer A first approach would be to determine if spe­ or more away (Ackerman, 1981b), pollen flow cialization is related to floral display and plant distances may be great (Williams & Dodson, dispersion. If the hypothesis is correct, then the '1972). In fact, euglossine bee pollination, like relationship should hold for comparisons of re­ trapline and deception strategies, may not be a lated sympatric species with very different dis­ very efficient system for large, dense populations. plays or dispersion patterns. Life forms with sim­ In Panama, a large clonal population of a robust ilar pollination mechanisms or strategies should 58 SELBYANA [Volume 9 have similar display and dispersion character­ Euglossini (Hymenoptera: Apidae): vagabonds or istics. trapliners? Biotropica 14: 241-248. If the dispersion-specialization relationship --AND A. M. MONTALVO. 1985. The longevity of euglossine bees. Biotropica 17: 79-81. exists, then predictions concerning the mecha­ -- AND J. C. MONTERO OUVER. 1985. Repro­ nisms for the evolution and maintenance of the ductive biology of Oncidium variegatum: moon system may be tested. For example, specialized phases, pollination and fruit set. Amer. Orchid pollination strategies of species that have char­ Soc. Bull. 54: 326-329. acteristically diffuse populations may be less ef­ ADAMS, R. M. AND G. J. Goss. 1976. The reproduc­ fective under crowded conditions perhaps be­ tive biology of the epiphytic orchids of Horida. cause of inbreeding depression (Schemske, 1983). III. Epidendrum anceps Jacquin. Amer. Orchid Conversely, the pollination strategies of those Soc. Bull. 45: 488-492. plants normally found in large, dense popula­ ARMBRUSTER, W. S. 1984. The role of resin in an­ giosperm pollination: ecological and chemical tions should be less effective in hyperdispersed considerations. Amer. J. Bot. 71: 1149-1160. populations because of competition for polli­ --AND G. L. WEBSTER. 1979. Pollination of two nation. species of Dalechampia (Euphorbiaceae) in Mex­ The pollination problems and solutions of epi­ ico by euglossine bees. Biotropica II: 278-283. phytes are not necessarily unique. In fact, some AsHTON, P. S. 1984. Biosystematics of tropical forest ofthe specialized pollination strategies may have plants: a problem ofrare species. pp. 497-518 in evolved among terrestrial ancestors that had W. F. GRANT, ed., Plant biosystematics. Academic similar size and dispersion constraints as epi­ Press, Orlando. phytes today (Ackerman, 1983c; Benzing & At­ ATWOOD, J. T. 1985. Pollination of Paphiopedilum wood, 1984). Nevertheless, pollination biology rothschildianum: brood site deception. National Geogr. Res. 1: 247-254. does· appear to be of special significance to the AUGSPURGER, C. K. 1981. Reproductive synchrony evolution of epiphytic groups. of a tropical shrub: experimental studies on effects on pollinators and seed predators on Hybanthus prunifo/ius (Violaceae). Ecology 62: 775-788. ACKNOWLEDGMENTS BAKER, H. G. 1976. "Mistake" pollination as a re­ productive system with special reference to the I thank M. Montalvo and two anonymous A. Caricaceae. Pp. 161-169 in J. BARLEY AND B. T. reviewers for their thorough critiques. Logistic STYLES, eds., Tropical trees: variation, breeding support for various phases was provided by W. and conservation. Academic Press, New York. J. Kress, Marie Selby Botanical Gardens, and N. BARKMAN, J. J. 1958. Phytosociology and ecology of M. Waser, University of California at Riverside. cryptogamic epiphytes. Van Goricum and Co., As­ Ms. Montalvo rendered the figure. sen, Netherlands. BAWA, K. S. 1983. Patterns of flowering in tropical plants. pp. 394-410 in C. E. JONES AND R. J. LITTLE, LITERATURE CITED eds., Handbook of experimental pollination bi­ ology. Van Nostrand Reinhold Co., New York. ACKERMAN, J. D. 1981a. Pollination biology of Ca­ --, S. H. BULLOCK, D. R. PERRY, R. E. COVILLE, lypso bulbosa var. occidentalis (Orchidaceae): a food AND M. H. GRAYUM. 1985. Reproductive biol­ deceptive system. Madroiio 28: 101-110. ogy of tropical lowland rainforest trees. II. Polli­ --. 1981 b. The phenological relationships ofmale nation systems. Amer. J. Bot. 72: 346-356. euglossine bees (Hymenoptera: Apidae) and their --, D. R. PERRY, AND J. H. BEACH. 1985. Re­ orchid fragrance hosts. Ph.D. thesis, Horida State productive biology of tropical lowland rain forest Univ., Tallahassee. trees. I. Sexual systems and incompatibility mech­ --. 1983a. Euglossine bee pollination of the or­ anisms. Amer. J. Bot. 72: 331-345. chid, Cochleanthes lipscombiae: a food source BENZING, D. H. 1976. Bromeliad trichomes: struc­ mimic. Amer. J. Bot. 70: 830-834. ture, function and ecological significance. Selby­ --. 1983b. Specificity and mutual dependency ana I: 330-348. of the orchid-euglossine bee interaction. BioI. J. --. 1981. Mineral nutrition of epiphytes: an ap­ Linn. Soc. 20: 301-314. praisal of adaptive features. Selbyana 5: 219-223. --. 1983c. On the evidence for a primitively epi­ --. 1984. Epiphytic vegetation: a profile and sug­ phytic habit in orchids. Syst. Bot .. 8: 474-477. gestions for future inquiries. pp. 155-171 in E. --. 1985. Euglossine bees and their nectar hosts. MEDINA, H. A. MOONEY, AND C. VAZQUEZ- Y ANES, Pp. 225-233 in W. G. D'ARcy AND M. D. CORREA eds., Physiological ecology of plants of the wet A., eds., The botany and natural history of Pan­ tropics. Dr. W. Junk Publishers, The Hague. ama. Missouri Botanical Garden, St. Louis. 2t---- AND J. T. ATWOOD. 1984. Orchidaceae: an- --. 1986. Mechanisms and evolution of food­ ., cestral habitats and current status in forest cano­ deceptive pollination systems in orchids. Lindley­ pies. Syst. Bot. 9: 155-165. ana 1: 108-113. BIERZYCHUDEK, P. 1981. Asclepias, Lantana, andEp­ --, M. R. MESLER, K. L. Lu, AND A. M. MON- idendrum: a floral mimicry complex? Biotropica 1iALvo. 1982. Food foraging behavior of male 13: 54-58. I 1986] ACKERMAN: EPIPHYTE POLLINAnON 59

BROKAW, N. V. L. 1982. Treefalls: frequency, timing, GRUBB, P. J., J. R. LLOYD, T. D. PENNINGTON, AND T. and consequences. Pp. 101-108 in E. G. LEIGH, C. WHITMORE. 1963. A comparison of montane A. S. RAND, AND D. M. WINDSOR, eds., The ecol­ and lowland rain forests in Ecuador. J. Ecol. 51: ogy of a tropical forest: seasonal rhythms and 567-601. long-term changes. Smithsonian Institution Press, HANDEL, S. N. 1985. The intrusion of clonal growth Washington, D.C. patterns on plant breeding systems. Amer. Nat. BURGER, W. C. 1981. Why are there so many kinds 125: 367-384. of flowering plants? BioScience 31: 572, 577-581. HUBBELL, S. P. 1979. Tree dispersion, abundance, CROAT, T. B. 1980. Flowering behavior of the neo­ and diversity in a tropical dry forest. Science 203: tropical genus Anthurium. Amer. J. Bot. 67: 888- 1299-1309. 904. IVRI, Y. AND A. DAFNI. 1977. The pollination ecology DAFNI, A. 1984. Mimicry and deception in pollina­ of Epipactis consimilis Don (Orchidaceae) in Is­ tion. Ann. Rev. Ecol. Syst. 15: 259-278. rael. New Phytol. 79: 173-177. --- AND Y. IVRI. 1981a. Floral mimicry between JANZEN, D. H. 1971. Euglossine bees as long-distance Orchis israelitica Bauman and Dafni (Orchida­ pollinators of tropical plants. Science 171: 203- ceae) and Bellevalia jlexuosa Boiss. (Liliaceae). 205. Oecologia 49: 229-232. ---. 1974a. Epiphytic myrmecophytes in Sara­ --- AND ---. 1981 b. The flower biology of wak: mutualism through the feeding of plants by Cephalanthera longifolia (Orchidaceae)-pollen ants. Biotropica 6: 237-259. imitation and facultative floral mimicry. PI. Syst. ---. 1974b. The deflowering of Central America. Evo!. 137: 229-240. Nat. Hist. 83: 48-53. DOD, D. 1976. Oncidium henekenii-bee orchid pol­ ---. 1975. Ecology of plants in the tropics. Ed­ linated by bee. Amer. Orchid Soc. Bull. 45: 792- ward Arnold, Limited, London. 794. ---. 1983. Insect-introduction. Pp. 619-645 in D. DODSON, C. H. 1975. Dressleria and Clowesia: a new H. JANZEN, ed., Costa Rican natural history. Univ. genus and an old one revived in the Catasetinae Chicago Press, Chicago. (Orchidaceae). Selbyana I: 130-137. JOHANSSON, D. 1974. Ecology of vascular epiphytes ---, R. L. DRESSLER, H. G. HILLS, R. M. ADAMS, in West African rain forest. Acta Phytogeogr. Suec. AND N. H. WILLiAMS. 1969. Biologically active 59: 1-136. compounds in orchid fragrances. Science 164: KOHN, J. R. AND N. M. WASER. 1985. The effect of 1243-1249. Delphinium nelsonii pollen on seed set in Jpo­ DRESSLER, R. L. 1968. Observations on orchids and mopsis aggregata, a competitor for hummingbird euglossine bees in Panama and Costa Rica. Rev. pollination. Amer. J. Bot. 72: 1144-1148. BioI. Trop. 15: 143-183. KRESS, W. J. 1983a. Self-compatibility in Central ---. 1981. The orchids: natural history and clas­ American Heliconia. Evolution 37: 735-744. sification. Harvard University Press, Cambridge. ---. 1983b. Crossability barriers in neotropical ---. 1982. Biology of the orchid bees (Eug10ssini). Heliconia. Ann. Bot. (London) 52: 131-147. Ann. Rev. Ecol. Syst. 13: 373-394. KULLENBERG, B. AND G. BERGSTROM. 1976. The pol­ FEINSINGER, P. 1976. Organization ofa tropical guild lination of Ophrys orchids. Bot. Not. 129: 11-19. of nectarivorous birds. Ecol. Monogr. 46: 257- LEVIN, D. A. AND H. W. KERSTER. 1973. Assortative 291. pollination for stature in Lythrum salicaria. Evo­ --- AND R. K. COLWELL. 1978. Communityor­ lution 27: 144-152. ganization among neotropical nectar-feeding birds. LINHART, Y. B. AND J. A. MENDENHALL. 1977. Pollen Amer. Zoo!. 18: 779-795. dispersal by hawkmoths in a Lindenia rivalis Benth. FRANKIE, G. W. AND H. G. BAKER. 1974. The im­ population in Belize. Biotropica 9: 143. portance of pollinator behavior in the reproduc­ LITTLE, R. J. 1983. A review of floral food deception tive biology of tropical trees. Anales Inst. BioI. mimicries with comments on floral mutualism. Univ. Nac. Mexico 45, Ser. Botimica 1: 1-10. Pp. 294-309 in C. E. JONES AND R. J. LITTLE, eds., ---AND R. COVILLE. 1979. An experimental study Handbook of experimental pollination biology. on the foraging behavior of selected solitary bee Van Nostrand Reinhold Co., New York. species in the Costa Rican dry forest (Hymenop­ LiiTTGE, U. 1985. Epiphyten: Evolution und Oko­ tera: Apoidea). J. Ka;qllas Ent. Soc. 52: 591-602. physiologie. Naturwissenschaften 72: 557-566. FREI, J. K. 1973. Orchid ecology in a cloud forest in MADISON, M. 1977. Vascular epiphytes: their system­ the mountains of Oaxaca, Mexico. Amer. Orchid atic occurrence and salient features. Selbyana 2: Soc. Bull. 42: 307-314. 1-13. ---AND C. H. DODSON. 1972. The chemical effect ---. 1979. Distribution of epiphytes in a rubber of certain bark substrates on the germination and plantation in Sarawak. Selbyana 5: 207-213. early growth of epiphytic orchids. Bull. Torrey Bot. MONTALVO, A. M. AND J. D. ACKERMAN. 1986. Lim­ Club 99: 301-307. itations to fruit production in /onopsis utricular­ GENTRY, A. H. 1974. Flowering phenology and di­ ioides (Orchidaceae). Biotropica (in press). versity in tropical Bignoniaceae. Biotropica 6: 64- PAULUS, H. F. AND C. GACK. 1980. Beobachtungen 68. und Untersuchungen zur Bestaiibungsbiologia GOTTSBEROER, G. AND A. AMARAL, JR. 1984. Polli­ siidspanische Ophyrs Arten. Pp. 55-68 in K. nation strategies in Brazilian Philodendron species. SENOHAS AND H. SUNDERMANN, eds., Probleme Bef. Deutsch. Bot. Ges. 97: 391-410. der Evolution bei europaische und mediterranen 60 SELBYANA [Volume 9

Orchideen. Jahresber. Naturwiss. Ver. Wuppertal among hummingbird foodplants of a tropical wet 33(supp!.). forest. Biotropica 10: 194-210. PICKETT,S. T.A. 1980. Non-equilibriumcoexistence STOUTAMIRE, W. P. 1975. Pseudocopulation in Aus­ of plants. Bull. Torrey Bot. Club 107: 238-248. tralian terrestrial orchids. Amer. Orchid Soc. Bull. PUL, L. VAN DER AND C. H. DODSON. 1966. Orchid 44: 226-233. flowers: their pollination and evolution. Univ. ---. 1983. Wasp-pollinated species of Caladenia Miami Press, Coral Gables. (Orchidaceae) in south-western Australia. Aust. 1. RABINOWITZ, D. 1981. Seven forms of rarity. Pp. 205- Bot. 31: 383-394. 217 in H. SYNGE, ed., The biological aspects of SUGDEN, A. M. 1979. Aspects of the ecology of vas­ rare plant conservation. John Wiley & Sons, New cular epiphytes in Columbian cloud forests. I. The York. distribution of the epiphytic flora. Biotropica 11: RATHCKE, B. 1983. Competition and facilitation 173-188. among plants for pollination. Pp. 305-329 in L. ---. 1981. Aspects of the ecology of vascular epi­ REAL, ed., Pollination biology. Academic Press, phytes in two Colombian cloud forests. Selbyana Orlando. 5: 264-273. REGAL, P. J. 1977. Ecology and evolution of flowering THORINGTON, R. W., JR., B. TANNENBAUM, A. TARAK, plant dominance. Science 196: 622-629. AND R. RUDRAN. 1982. Distribution of trees on RENNER, S. S. 1984. Pollination and breeding systems Barro Colorado Island: a five hectare sample. Pp. in some central Amazonian Melastomataceae. pp. 83-94 in E. G. LEIGH, A. S. RAND, AND D. M. 275-280 in Verne symposium international sur la WINDSOR, eds., The ecology of a tropical forest: pollinisation. INRA Pub!., Versailles. seasonal rhythms and long-term changes. Smith­ RICHARDS, P. W. 1952. The tropical rain forest: an sonian Institution Press, Washington, D.C. ecological study. Cambridge Univ. Press, Cam­ VOGEL, S. 1978a. Pilzmiickenblumen als Pilzmime­ bridge. ten. Aora 167: 329-366, 367-398. ROUBIK, D. W., J. D. ACKERMAN, C. COPENHAVER, AND ---. 1978b. Evolutionary shifts from reward to B. H. SMITH. 1982. Stratum, tree, and flower deception in pollen flowers. pp. 89-96 in A. J. selection by tropical bees: implications for the re­ RICHARDS, ed., The pollination of flowers by in­ productive biology of outcrossing Cochlospermum sects. Academic Press, London. vitifolium in Panama. Ecology 63: 712-720. WAGNER, W. H., JR. 1973. An orchid attractant for SANFORD, W. W. 1968. Distribution of epiphytic or­ monarch butterflies (Danaidae). J. Lepidopt. Soc. chids in semi-deciduous tropical forest in southern 27: 192-196. Nigeria. J. Ecol. 58: 697-705. WASER, N. M. 1983. Competition for pollination and SCHAFFER, W. M. 1974. Optimal reproductive effort floral character differences among sympatric plant in fluctuating environments. Am. Nat. 108: 783- species: a review of evidence. pp' 277-293 in C. 790. E. JONES AND R. J. LITTLE, eds., Handbook of SCHEMSKE, D. W. 1980. Evolution of floral display experimental pollination biology. Van Nostrand in the orchid Brassavola nodosa. Evolution 34: Reinhold Co., New York. 489-493. --- AND L. A. REAL. 1979. Effective mutualism ---. 1983. Breeding system and habitat effects on between sequentially flowering plant species. Na­ fitness components in three neotropical Costus ture 281: 670-672. (Zingiberaceae). Evolution 37: 523-539. WILLIAMS, N. H. 1982. The biology of orchids and SIMPSON, B. B. AND J. F. NEFF. 1981. Aoral rewards: euglossine bees. Pp. 119-171 in J. ARDITTI, ed., alternatives to pollen and nectar. Ann. Missouri Orchid biology: reviews and perspectives, II. Cor­ Bot. Gard. 68: 301-322. nell Univ. Press, Ithaca. STEPHENSON, A. G. AND R. I. BERTIN. 1983. Male ---AND C. H. DODSON. 1972. Selective attraction competition, female choice, and sexual selection of male euglossine bees to orchid floral fragrances in plants. Pp. 110-149 in L. REAL, ed., Pollination and its importance in long distance pollen flow. biology. Academic Press, Orlando. Evolution 26: 84-95. STILES, F. G. 1977. Coadapted competitors: the flow­ --- AND R. L. DRESSLER. 1976. Euglossinae pol­ ering seasons of hummingbird-pollinated plants in lination of Spathiphyllum (Araceae). Selbyana 1: a tropical forest. Science 198: 1177-1178. 350-356. ---. 1978a. Ecological and evolutionary impli­ --- AND M. WHITTEN. 1983. Orchid floral fra­ cations of bird pollination. Amer. Zool. 18: 715- grances and male euglossine bees: methods and 727. advances in the last sesquidecade. BioI. Bull. 164: ---. 1978b. Temporal organization of flowering 355-395.