Selbyana 15: 1-7

HOW MUCH IS KNOWN ABOUT BROMELIACEAE IN 1994?

DAVID H. BENZING Department of Biology, Oberlin College, Oberlin, Ohio 44074 USA

If access and utility alone dictated scientific no exhaustive review of the immense literature interest, much less would be known about Bro­ is attempted, only highlights of accomplish­ meliaceae today. Fully one half of all its species ments, sprinkled here and there with exhorta­ grow in trees, many in remote rain forests. An­ tions and a bit of guarded speculation. But first other sizable group of terrestrials similarly dis­ we need some statistics, current and projected, courages study by occurring in roadless, bleak, and then some thoughts about the magnitude of upper montane habitats (e.g., Navia, Puya, many the task and the nature of the goals. Areas of Tillandsia spp.) especially in South America. inquiry follow, each updated to reflect current Commercial promise has not provided much en­ status and lastly, a concluding remark. couragement for researchers either. Beyond the pineapple, some widely marketed ornamentals, SYSTEMATICS, TAXONOMY AND and a few minor fiber products, bromeliads sim­ EVOLUTION ply offer too little potential to merit funding any­ where near that invested to study the more wide­ Bromeliaceae is the largest, almost exclusively ly useful Poaceae, Fabaceae, Rosaceae and their American family (one African Pitcairnia sp.) of kind. Nevertheless, bromeliads have attracted flowering plants. Likewise, its basic tropical char­ more than their share of scholarship beginning acter and relative youth are beyond dispute. with the attentions of an impressive array of However, many additional points, including nineteenth century, comparative and functional much of its taxonomy, are more equivocal. Ge­ morphologists (e.g., Haberlandt, Mez) and plant neric alignments continue to shift and the num­ geographers (e.g., A. F. W. Schimper). Brome­ ber of described species keeps growing, neither liaceae continues to attract an ever-widening with much likelihood of closure any time soon. group ofspecialists, just a small fraction ofwhom Family size has already expanded from the 2110 have authored the contents of this entire volume binomials recognized in Smith and Down's three of Selbyana. volume monograph (1974, 1977, 1979) to about Enough information has been compiled on 2700 (Luther, pers. comm.). More additions Bromeliaceae to warrant a comprehensive, tech­ should be expected, many contributed by intrep­ nical monograph, something not yet available, id amateurs seeking to embellish private collec­ but coming soon (Benzing in prep). While noth­ tions or horticulturists intent on augmenting ing published in 1994 could even approach a commercial stocks. Three thousand is a reason­ complete synthesis, Bromeliaceae is fast emerg­ able projection for the final tally, but the exact ing as one of the better known families of flow­ number of species will depend also on the judg­ ering plants. Specifically, perspectives on its or­ ments of numerous specialists, specifically on how igin, major radiations, and designs for life under often decisions are made to elevate the status of an exceptional variety of often stressful growing segregates of the many polymorphic, currently conditions are expanding rapidly-faster than recognized species (e.g., Tillandsia fasciculata). progress toward the same goal for most other, Genera will also continue to multiply as so often comparably large and ecologically diverse clades. occurs as complex groups become better known. However, knowledge of some aspects (e.g., the For example, Aechmea (currently 10 recognized nature of the many and often important roles subgenera) and Tillandsia (7 subgenera) will al­ these plants play in communities) seems less like­ most certainly experience this fate. Such move­ ly to develop apace due in part to greater im­ ments are afoot already as are others to combine pediments to researchers and sometimes simply established genera (e.g., Aechmea and Strepto­ because interest continues to be too low. Most calyx, Smith and Spencer 1993). Other author­ of the bromeliad literature concerns evolution or ities wisely council patience until changes can be systematics and this pattern seems likely to con­ better informed by more comprehensive plant tinue. collections and deeper analysis (Brown et al. What follows is a progress report-a briefap­ 1993). Today, we stand on a threshold. Molec­ praisal of where we stand along the route to even­ ular data, which can more unambiguously re­ tual synthesis. Coverage is purposely selective; solve phylogeny than most other kinds, could 2 SELBYANA [Volume 15 easily undermine hasty decisions that will further ing tillandsioids in particular match growing con­ clutter the literature and burden users with su­ ditions (humidity, light and temperature) in perfluous nomenclature. native habitats. Inquiry on systematics has not been appor­ Especially exciting and central to that even­ tioned evenly across Bromeliaceae. Most inten­ tual, evolutionary synthesis is the progress of sively studied is Tillandsioideae among the three molecular systematists, several of whom work generally recognized subfamilies, with Pitcair­ with Bromeliaceae. Ranker et al.'s (1990) anal­ nioideae second ahead of Bromelioideae, the ysis of restriction site polymorphisms (cpDNA) greatest challenge of the three. The compara­ among species selected to represent all three sub­ tively greater structural diversity and presum­ families heralded what promises to be the first ably broader genetic variety of the bromelioids of a series of penetrating reports on phylogeny. is underscored by inclusion of over half of all the They questioned both the reputed basic position bromeliad genera in this taxon. of Pitcairnioideae within the family and the le­ Recent progress toward an evolutionary tax­ gitimacy of assigning Glomeropitcairnia to Til­ onomy ofbromeliads has been assisted by three landsioideae. Coming soon are more complete developments: (I) a more thorough evaluation reconstructions of relationships based on se­ of traditional characters and employment of quence analyses of the gene coding the large sub­ newer ones from the same organs (e.g., flowers) unit (rbeL) of rubisco and a second, typically (2) the emergence of new and more powerful data more variable, and less often used plastid gene from previously inaccessible sources (e.g., chlo­ (ndhF) that should allow greater resolution per­ roplast genomes) and (3) the application of phe­ haps to genera (Clark et al. in press, Brown and netic, and particularly cladistic, analyses that, Randall pers. comm.). We can also expect prog­ compared to older methodologies, extract more ress shortly on the question of where Bromeli­ biological information from all kinds of valid aceae lies within Liliopsida, specifically its re­ data. lationships to suggested sister taxa (e.g., Traditional reliance on dried, herbarium ma­ Rapateaceae, Velloziaceae). terials has been relieved significantly by the es­ Cladograms not only display relationships tablishments .of extensive, documented living among lineages, they can help resolve the origins collections (e.g., at the Marie Selby Botanical (often multiple in Bromeliaceae) and modifica­ Gardens). Supplies of fresh and wet-preserved tion over time of ecologically important plant materials to examine delicate organs (e.g., stig­ characteristics such as CAM, epiphytism and the mas) have been further supplemented by more utilization of anta for dispersal and nutrition (e.g., comprehensive field work, aided recently by a Chase and Hills 1992). Gilmartin and Brown network of on-site collectors (e.g., Gilmartin and (1986) recognized this potential in their attempt Brown 1986). Fresh material has also provided to determine the status (apomorphic or pleiso­ a fairly complete picture of the bromeliad karyo­ morphic) of xerophytism and mesophytism in .type and its evolution (Brown and Gilmartin subgenus Phytarrhiza of Tillandsia. Both con­ 1989). Particularly revealing is Brown and Ter­ ditions occur in other parts of Tillandsia and ry's(1992) study of petal scale ontogeny that Vriesea and repeatedly elsewhere in the family. casts serious doubt on the wisdom of relying so Accurate alignments within Bromeliaceae of ad­ heavily on this appendage to distinguish tilland­ ditional, aberrant genera like Brocchinia will in­ siaid and probably some otherbromeliad genera. crease opportunities for similar sorts of deriva­ These structures may enhance the delivery of tive analysis, for example, whether the absorptive nectar, but their apparent, relatively recent evo­ capacity of the foliar trichome evolved to pro­ lution, hence marginal utility to circumscribe at mote mineral nutrition, water balance or both least some taxa, seems indisputable. phenomena. Pursuit of additional anatomical details, on seeds for example, continues (e. g., Gross 1992, CARBON AND WATER BALANCE 1993). Data on pollen morphology has been ac­ cumulating for several decades as well, but so far More data (e.g., H+ max> oH+, ol3C, patterns of withou.t the interpretation necessary to inform gas exchange) that indicate the mechanisms of systematists and other investigators interested in carbon and water balance or reveal their ecolog­ pollination syndromes and breeding systems. ical consequences have been collected for the Trichome structure probably also remains un­ bromeliads than for species in any other family. der-utilized for taxonomy and considering the Revelations include the discovery that related amount of information (e.g., Varadarajan and species with CAM and C3 photosynthesis some­ Gilmartin 1987, Benzing .and Seemann 1978) times co-occur without substantial differences in certainly offers much untapped potential to ex­ water economy (e,g., Griffiths et al. 1986, Grif­ plain how the foliar epidermis ofthe dryer-grow- fiths 198.8). Shade tolerance deeper than that pre- 4 SELBYANA [Volume 15 modern techniques have became available to sioideae, the best known of the three subfamilies plant physiologists. Radiotracers demonstrated by reproductive mechanisms, attracts a variety substantial uptake of diversF solutes through root of primary pollen vectors including birds, bats, systems and the foliage of phytotelm forms (e.g, and day- and night-flying insects. Seed dispersal Nadkarni and Primack 1989). Carnivory is now is more uniform (anemochory), but not consis­ well documented in Brocchinia reducta (e.g., tent if air worthiness and tendencies to adhere Givnish et al. 1984) and probably operates at a to specific tested substrates reflect performances less specialized level in Catopsis berteroniana. in nature (Bennett 1991) Substantially less in­ Owen and Thompson (1988) assessed the fine­ formation has been reported about Pitcairnioi­ structure of the foliar trichome of Brocchinia re­ deae, but seed size and form indicate consider­ ducta and demonstrated how these appendages able variation in mobility, probably more than probably support carnivory by absorbing the characterizes the tillandsioids. Least known by products of degraded prey. Tillandsioid tri­ floral and dispersal biology, except for those myr­ chomes need to be compared with those of Broc­ mecochorous ant-nest flora, is Bromelioideae. As chinia. Additional study is also required to es­ berry producers, members of this taxon engage tablish how the much more common in relationships with fauna beyond those arbo­ noncarnivorous phytotelm bromeliads process real ants and dispersal biology is accordingly litter rather than prey for nutrition. Diverse shoot complex. Fruits appear to be attractive to birds forms and varied microflora and microfauna in (e.g., odorless white, blue, many Aechmea spp.), phytotelmata indicate that different symbiotic bats (e. g., drab, smelling of rotten fruit, some biota and methods of processing mediate nutri­ Billbergia spp.), and nonvolant mammmais (e. ent flux for these bromeliads (e.g., Bermudes and g., yellow, orange hidden in foliose infloresences, Benzing 1991). many Nidularium spp.). Berries of some Ronn­ Two forms of myrmecotrophy occur in Bro­ bergia spp. rupture to eject seeds several meters meliaceae. Ant-house Tillandsia spp and prob­ indicating that mechanical enhancements some­ ably additional, similarly accommodating taxa times supplement or replace animal carriage. (e.g., Aechmea bracteata) are fed by ant colonies Information on the breeding systems of bro­ residing within inflated leaf bases (Benzing 1991). meliads is scattered and the impacts on the ge­ Nutrient budgets and the importance of the ants netic structure, and isolation of populations and to the plants will require deeper inquiry to es­ speciation still poorly known. Some populations tablish. Another group ofbromelioids (e.g., sev­ routinely fruit autogamously; many more out­ eral Aechmea and Streptocalyx spp.) regularly cross through a variety of agencies ranging from root in ant cartons, sometimes to the near ex­ dioecism and self-incompatibility to various di­ clusion of other substrates (e.g., Davidson 1988). chogamous and herkogamous arrangements Chemical attractants on seeds assure successful among the self-fertile types. Populations of only dispersal by foraging ants. However, many other three species, predominantly epiphytic Tilland­ aspects of these systems, for example, whether sia ionantha and T. recurvata and terrestrial plant nutrition reflects reliance of ant-provided Aechmea magdelenae have been comprehen­ substrates, remain obscure. Atmospheric bro­ sively examined (> 20 loci) to determine genetic meliads continue to be employed to monitor air structures (Soltis et al. I 987, Murawski and quality (Schrimpff 1981, Benzing et al. 1992) be­ Hamrick 1990). The latter species relies more cause of their normal reliance on the atmosphere heavily on ramets to colonize understory habi­ and canopy washes for required ions. Extraor­ tats, possibly because soil, relative to arboreal dinary capacities to accumulate a variety of ad­ substrates, tends to be more expansive and per­ ditional anthropogenic substances underlies their manent. Bennett (1991) demonstrated propor­ utility for pollution survalence; these same qual­ tionally greater dependence on sexual than asex­ ities could prove useful on a broader scale to ual reproduction in epiphytes compared to measure important variables related to global lithophytes among a group of Tillandsia spp. change (Lugo and Scatena 1993). chosen to represent both habits. The bromeliad literature documents how ear­ lier interest (e.g., Downs in Smith and Downs REPRODUCTIVE BIOLOGY 1974) in the physiology of flowering and fruit development has diminished. Many populations Reproductive biology, like systematics and respond to photoperiod and very likely cycling evolution, has attracted more than its share of temperatures and rainfall also cue important life the scholarly attention devoted to Bromeliaceae. history events. Inquiry on seed physiology has Numerous reports document modes of pollina­ moved at about the same slow pace although tion of specific species (Sazima et al. 1989). Seed previous studies indicated differences in viability dispersal has received lower priority. Tilland- and dissimilar responses to light and tempera- 1994] BENZING: BROMELIADS 3 vious1y documented for most CAM plants pre­ (e.g., Vriesea josteriana, Benzing and Friedman vails in certain other brome1iads (e.g., Aechmea 1981). Additional inquiry on the carbon and wa­ magdalenae, Pfitsch and Smith 1988). Manyad­ ter balance mechanisms of seedlings and adults ditional CAM species grow better in partial shade of C3 tillandsioids could help test hypotheses than full sun (e.g., Bromelia humilis ) and may concerning the role of heterochrony and the evo­ be relatively poorly suited for the kinds of ter­ lutionary status of xerophytism in Tillandsioi­ restrial environments often associated with CAM. deae (Adams and Ma in 1986). Meanwhile, the On the other hand, some bromelioids are near controversy over the nature of ancestral envi­ record holders for H+ max (474 mol H+ /m-3 for ronments continues (e g., Smith 1989), a subject Aechmea nudicaulis). likely to defy resolutio until a phylogeny (below Ananas spp. are providing insights on how plant the genus) is availabl to polarize ecologically performances under different growing conditions decisive character stat s. have been altered by indigenous agriculture (Me­ Whether or not CA represents the most ef­ dina et al. 1991). Interactions among Nand ficacious means to fix O2 everywhere species so moisture supplies and PPFD (photosynthetic equipped occur, drou t limits the productivity photon flux density) that promote different out­ of many Bromeliacea and probably more de­ comes (e.g., yield, water use efficiency, citric vs. cisively so than scarci ies of any other resource. malic acid synthesis) under specific conditions The importance of dr ught as a selective force in CAM plants in the field are probably best is apparent in those b meliads that demonstra­ known in Bromeliaceae. Widely noted, but not bly achieve some oft e most favorable transpi­ yet resolved to causes and biological conse­ ration ratios on record (Smith 1989). Some long­ quences are the mechanisms responsible for the standing questions ab ut water balance remain heavy and variable reliance on respired, com­ only partially resolve , for instance, whether or pared to atmospheric, CO2 for nocturnal acidi­ not water vapor can ignificantly relieve dehy­ fication that so dramatically characterizes many dration (e.g., Martin nd Schmitt 1989). Low bromeliads (Smith 1989). (compared to higher) ater vapor pressure def­ Longer-term monitoring in the field and fuller icits in surrounding a r allow all land plants to understanding of the subcellular aspects of CAM conserve moisture, ut claims continue that are needed to determine if and how this syn­ shoots of some of the eavily trichomed tilland­ drome benefits bromeliads under the many dif­ sioids hydrate directl from moist air without ferent environmental conditions prevailing where prior condensation 0 absorptive surfaces. The these plants grow. What regulatory processes and foliar trichome of Till ndsioideae merits closer environmental and internal cues determine the scrutiny to see how i aids water balance and sources, fluxes, and metabolic fates of carbon additionally, to what xtent the variously struc­ during CAM? In addition to its capacity to im­ tured and positioned shield matches humidity prove water economy, CAM may reduce vul­ and other conditions i native habitats. nerability to photodamage or promote N econ­ omy under appropriate circumstances. CAM may MINERAL UTRITION AND be most important to some bromeliads for ben­ efits other than water economy or it could affect survival only during uncommonly dry years. Fit­ Occasional reports ontinue to appear on the ness may not be strongly dependent on the type mineral nutrition of romeliaceae. Several or­ of photosynthetic pathway possessed by phyto­ namentals have prove amenable to hydroponic telm species native to ever-wet forests and in­ culture and success 0 conventional media in­ stead may reflect conditions in ancestral more dicated no unusual lant requirements (e.g., than contemporary habitats. Bromeliaceae do Schmitt 1982). Meth ds have been developed seem to exceed most other families in the variety for the routine ascepti micropropagation of ad­ of plant characters associated with CAM and the ditional taxa, includin some endangered species kinds of stresses its members encounter on often (e.g., Mercier and Ker auy 1992). Again, widely­ demanding substrates. Accordingly, more of the used basal media pro ed effective. Within the full array of benefits CAM imparts to land flora decade transgenic mat¢rials that combine desir­ may be expressed in this family compared to able horticultural qualities will probably reach most others. the commercial market. Superior shoot form and Contrary to the CAM types, little attention has desired leaf and infloresence color and texture been devoted to the C3 bromeliads, for example (historically the highest priorities for hybridizers) to their light relations. Particularly intriguing are will probably remain the major goals. the functional correlates of the uneven, but reg­ Bromeliaceae exceed most families in the va­ ular distributions of chlolorophyll and other pig­ riety of nutritional modes utilized, all of which ments in the foliage of certain phytotelm species have attracted at least passing attention since 6 SELBYANA [Volume 15 originated. Should those species (e.g., Tillandsia rbeL sequence comparisons. Ann. Missouri Bot. usneoides, Bromelia humilis) with proven value Gard.80. to the study of the widely-occuring aspects of DAVIDSON D.W. 1988. Ecological studies ofneotrop­ stress physiology continue in that service, in­ ical ant gardens. Ecology 69: 1138-1152. GENTRY A.H. AND C.H. DoDSON. 1987. Diversity sights on how this family has so successfully col­ and biogeography of neotropical vascular epi­ onized the canopy and other demanding sub­ phytes. Ann. Missouri Gard. 74:205-233. strates will become clearer. However, input on GILMARTIN A.J. AND G.K. BROWN. 1986. Bromeli­ a variety of important, but unique features of aceae: an international cooperative research pro­ these plants (e.g., the biology of their foliar tri­ ject. Taxon 35:107-109. chomes and the workings of phytotelmata) will --- AND ---. 1986. Cladistic tests ofhypoth­ require the attention of investigators specifically eses concerning evolution of xerophytes and mes­ interested in these plants. Finally, elucidation of ophytes within Tillandsia subgenus Phytarrhiza the many roles Bromeliaceae play in neotropical (Bromeliaceae). Amer. J. Bot. 73:387-397. GIVNISHT.J., E.L. BURKHARDT, R. HAPPEL AND J. WE­ ecosystems as resources for other biota and as INTRAUB. 1984. Carnivory in the bromeliad influences on basic ecosystem processes will re­ Broeehinia reducta with a cost/benefit model for quire more integrated and multi-disciplinary ap­ the general restriction of carnivorous plants to proaches than we have seen so far. sunny, moist nutrient-poor habitats. Amer. Nat­ uralist 124:479-497. GRIFFITHS H. 1988. Carbon balance during CAM: an assessment of respiratory COl recycling in the epi­ LITERATURE CITED phytic bromeliads Aeehmea nudieaulis and Aeeh­ mea fendleri. PI. Cell Environm. 11 :603-611. ADAMS W.W. AND C.E. MARTIN. 1986. Morpholog­ ---, U. LUTTGE, K.H. STIMMEL, C.E. COOK, N.M. ical changes accompanying the transition from ju­ GRIFFITHS AND J.A.C SMITH. 1986. Comparative veni1e(atmospheric) to adult (tank) forms of the ecophysiology of CAM and C3 bromeliads. III. Mexican epiphyte Tillandsia deppeana (Bromeli­ Environmental influences on CO2 assimilation and aceae). Amer. J. Bot. 73:1204-1214. transpiration. PI. Cell Environm. 9:385-393. BENNETT B.C. 1991. Comparative biology of Neo­ GROSS E. 1992. Die Samen der Bromeliaceae. Bro­ tropical epiphytic and saxicolous Tillandsia spe­ melie 1992(3):61-66. cies: population structure. J. Trop. Ecol. 17:361- ---. 1993. Die Samen der Bromeliaceae (2). Bro­ 371. melie 1993(1):13-18,25-26. BENZING D.H. 1981. Bark surfaces and the origin and ---. 1993. Die Samen der Bromeliaceae (3). Bro­ maintenance of diversity among angiosperm epi­ melie 1993(2):53-55. phytes: an hypothesis. Selbyana 5:248-255. LAESSLE A.M. 1961. A micro-limnological study of ---. 1991. Myrmecotrophy: origins, operation, and Jamaican bromeliads. Ecology 42.:499-517. importance. In Ant-plant interactions. c.R. Hux­ LuGO A.E. AND F.N. SCATENA. 1992. Epiphytes and ley and D.F. Cutler eds, pp 353-373. Oxford climate change research in the Caribbean: a pro­ --AND J. SEEMANN. 1978. The foliar epidermis posal. Selbyana 13: 123-130. in Tillandsioideae (Bromeliaceae) and its role in MARTIN C.E. AND A.K. SCHMITT. 1989. Unusual wa­ habitat selection. Amer. J. Bot. 65:359-365. ter relations in the CAM atmospheric epiphyte --AND W.E. FRIEDMAN. 1981. Patterns offoliar Tillandsia usneoides L. (Bromeliaceae) Bot. Gaz. pigmentation and their adaptive significance. Sel­ 150:1-8. byana 5:224-240. MEDINA E., U. LUTTGE, F. LEAL AND H. ZIEGLER. 1991. --, J. ARnlTIl, L.P. NYMAN, P.J. TEMPLE AND J.P. Carbon and hydrogen isotype ratios in bromeliads BENNETT. 1992. Effects of ozone and sulfur diox­ growing under different light environments in nat­ ode on four epiphytic bromeliads. Environm. Exp. ural conditions. Bot. Acta. 104:47-52. Bot. 32:25-32. MERCIER H. AND G.B. KERBAUY. 1992. In vitro mul­ BERMUDES D. AND D.H. BENZING. 1991. Nitrogen tiplication of Vriesea fosteriana. PI. Cell Tissue fixation in association with Ecuadorean bromeli­ Organ Cult. 30:247-249. ads. J. Trop. Ecol. 7:531-538. NADKARNI N.M. 1984. Epiphyte biomass and nutri­ BROWN G.K., H.E. LUTHER AND J.W. KREss. 1993. ent capital ofa neotropical elfin forest. Biotropica Comments on the responsibilities of taxonomists. 16:249-256. J. Bromeliad Soc. 43:154-157. ---AND R.B. PRIMACK. 1989. The use of gamma -- AND A.J. GILMARTIN. 1989. Chromosome spectrometry to measure within-plant nutrient al­ numbers in Bromeliaceae. Amer. J. Bot. 76:657- locations of a tank bromeliad, Guzmania Ungu­ 665. lata. Selbyana 11:22-25. --AND R.G. TERRY. 1992. Petal appendages in OWEN T. AND W.W. THOMPSON. 1988. Sites of leu­ Bromeliaceae. Amer. J. Bot. 79:1051-1071. cine, arginine, and glycine accumulation in the CHASE M.W. AND H.G. HILLS. 1992. Orchid phylog­ absorptive trichomes of a carnivorous bromeliad. eny, sexuality, and fragrance seeking. Bioscience J. Ultrastruct. Molect. Struct. Res. 101:215-223. 42:43-49. PAOLETTI M.H., R.A.J. TAYLOR, B.R. STINNER, D.H. CLARK W.D., B.S. GAUT, M.R. DUVALL AND M.T. STINNER AND D.H. BENZING. 1991. Diversity of CLEGG. (in press). Phylogenetic relationships of soil fauna in the canopy and forest floor ofa Ven­ the Bromeliaceae and related monocots based on ezuelan cloud forest. J. Trop. Ecol. 7:373-384. 1994] BENZING: BROMELIADS 5

tures that probably reflect adaptations to specific through the nutrients they and associated debris growing conditions sequester, probably deny phorophytes and per­ haps other soil-rooted flora adequate nutrition ECOWGY AND IMPACTS IN ECOSYSTEMS under certain conditions (Benzing and Seemann 1978). At the same time, these and co-occurring More is known about the ecology of the epi­ epiphytes may increase system-wide capacity to phytic than terrestrial bromeliads. Co-ocurring intercept and store key elements for eventual slow species in dense forests at a variety of locations release (Nadkarni 1984). Additional benefits ac­ segregate within canopies according to exposure crue to forest residents from the humidity created and humidity more or less as Pittendrigh (1948) by evaporation from phytotelmata. Leafaxils may described as the pattern in the mountains of contain the only drinking water and provide the northern Trinidad. Ecophysiologists have iden­ most durable, moist refuges during long dry sea­ tified a number ofunderlying structural and func­ sons. No doubt, patterns of aggregate nutrient tional characters (as described earlier in the sec­ use and productivity change markedly as forests tion on carbon and water balance) that dictate become heavily colonized with bromeliads and plant occurrence at specific locations in the can­ the additional flora their presence encourages. opy. While host identity, the size of the sup­ Contributions to overall forest phytodiversity are porting axis, and other aspects ofthe substratum apparent in the lists ofadditional epiphytes, many (e.g., bare or moss-covered) sometimes influence supported by bromeliads, that constitute up to distributions secondarily, epiphytic bromeliads 35% of the total flora of some montane forests are generally less fastidious about anchorages than (Gentry and Dodson 1987). much other arboreal flora, especially the orchids. Diverse terrestrial, including lithic, substrates Many populations ofbromeliads with similar if support extensive, additional Bromeliaceae, not identical cultural requirements grow inter­ sometimes enough to dominate the associated spersed, apparently without clear spacing and may floras. Land-based communities with substantial achieve high local diversities in part in response and greater bromeliad presences occur on the to appropriately scheduled substrate turnover ground at high elevations (up to 4400 m in the (Benzing 1981). Phytotelm types furnish rooting Andes), occupy nearly rainless coastal deserts media for more demanding flora and create high­ (Peru), and those inland (Hechtia, Dyckia), grow quality habitat for myriad microflora and fauna, virtually unaccompanied by other vascular veg­ both vertebrates and invertebrates, with perva­ etation on cliff sides, form characteristic assem­ sive biological consequences. blages along humid Atlantic coasts (e.g., Brazil­ Bromeliads, more than members of other fam­ ian restingas), and form the understories in certain ilies, often provide the superstructure for exten­ humid forests (e.g., Aechmea magdalenae). Some sive communities, particularly in the canopies of ofthese species also occur as epiphytes and many humid montane forests occurring between about more possess similar form and function, in effect, 300-3000 m. What so far have been preliminary features that blur the distinction between what studies only, provide a glimpse of the impact of their obligate relatives suggest are more distinct this compartment on important processes in sup­ habits. One of the most interesting of the little porting ecosystems and early indications suggest researched aspects of bromeliad biology con­ substantial significance. Some bromeliads may cerns the determinants of habitat choice, which achieve keystone status as essential resources in this ecologically varied family must be ex­ (mainly housing) for some of the most aggressive traordinarily diverse and sometimes exception­ and abundant ants in Neotropical forests (Wilson ally precise. 1987). On a broader scale, the densities and di­ versities of invertebrates associated with the hu­ CONCLUDING REMARKS mus contained in leafaxils may exceed those in adjacent suspended soils and the ground below Research on the evolution, systematics, nat­ (e.g., Paoletti et al. 1991). Physical conditions in ural history, structure, and ecophysiology ofBro­ bromeliad phytotelmata have not yet received meliaceae is moving forward at a quickening pace. enough attention (e.g., Laessale 1967) to fully Soon, probably within the next two decades, bro­ explain why they support so many and such var­ meliad phylogeny will be sufficiently understood ied kinds of residents. Insights on the events to associate populations into natural genera and within phytotelmata that make them high-qual­ these taxa in turn into more natural tribes and ity substitutes for nutritive soil will require the subfamlies. This framework in turn will provide co-ordinated efforts of microbiologists and spe­ opportunity to establish where (in which lin­ cialists knowledgeable about the workings ofvar­ eages), how often, and possibly when the char­ ious groups of invertebrates. acteristics that distinguish the family and ac­ Viewed from another perspective bromeliads, count for its considerable ecological novelty 1994] BENZING: BROMELIADS 7

PFITSCH W.A. AND A.P. SMITH. 1988. Growth and SMITH L.B. AND M.A. SPENCER. 1992. Reduction of photosynthesis of Aechmea magdalenae, a terres­ Streptocalyx(Bromeliaceae: Bromelioideae). Phy­ trial CAM plant in a tropical moist forest in Pan­ tologia 72:96-98. ama. J. Trop. BioI. 4:199-207. --AND RJ. DoWNS. 1974. A. Neotrop. Monogr. PtTTENDRIGH C.S. 1948. The bromeliad-Anopheles­ No. 14, Part 1 Pitcairnioideae(Bromeliaceae) Haf­ malaria complex in Trinidad. I. The bromeliad ner Press, New York. flora. Evolution 2:58-89. --AND --. 1977. A. Neotrop. Monogr. No. RANKER T.A., D.E. SOLTIS, P.S. SOLTIS AND A.J. GIL­ 14, Part 2 Tillandsioideae (Bromeliaceae) Hafner MARTIN. 1990. Subfamilial phylogenetic rela­ Press, New York tionships of the Bromeliaceae: Evidence from --AND--. 1979. A. Neotrop. Monogr. No. chloroplast DNA restriction site variation. Syst. 14, Part 3 Bromelioideae (Bromeliaceae) New York Bot. 15:425-434. Botanical Garden, New York. SAZIMA I., S. VOGEL AND M. SAZIMA. 1989. Batpol­ SOLTIS D.E., A.J. GILMARTIN, L. RIESBERG AND S. GA­ lination of Encholirium glaziovii, a terrestrial bro­ RDNER. 1987. Genetic variation in the epiphytes meliad. PI. Syst. Evol. 168:167-179. Tillandsia ionantha and T. recurvata (Bromeli­ SCHMITT W. 1982. Aus Saat von Bromelien aufHy­ aceae) Amer. J. Bot. 74:531-539. drokultur. Bromelie 3:173-174. VARADARAJAN G.S. AND A.J. GILMARTIN. 1987. Fo­ SCHRlMPFF E. 1984. Air pollution patterns in two liar scales of the subfamily Pitcairnioideae (Bro­ cities of Colombia, S.A. according to trace sub­ meliaceae). Syst. Bot. 12:562-571. stances content of an epiphyte (Tillandsia recur­ WILSON E.O. 1987. The arboreal ant fauna of Peru­ vata). Water, Air, and Soil Pollution 21:279-315. vian Amazonian forests: a first assessment ..Bio­ SMITH J.A.C. 1989. Epiphytic bromeliads. In Vas­ tropica 19:245-282. cular plants as epiphytes. U. Liittge ed, pp 109- 134. Springer-Verlag. Berlin.