Selbyana 13: 39-49 EFFECTS OF FIRE ON BROMELIADS IN SUBTROPICAL HAMMOCKS OF EVERGLADES NATIONAL PARK, FLORIDA KEVIN ROBERTSON AND WILLIAM J. PLATT Department of Botany, Louisiana State University, Baton Rouge, Louisiana U.S.A. 70803 ABSTRACT. Responses of five species of subcanopy bromeliads (Tillandsia balbisiana, T. fasciculata, T. setacea, T. utriculata, and T. valenzuelana) to large-scale disruption by naturally-occurring, low-intensity fire were inferred from comparisons of populations in burned and unburned subtropical hardwood ham­ mocks one year after fire on Long Pine Key in Everglades National Park, Florida. In burned hammocks, the soil humus was consumed. Many trees were defoliated and/or killed, which opened the canopy. Com­ parisons of epiphyte populations in burned and unburned hammocks, which were very similar in pre-fire characteristics, indicate that fire had few direct effects on these species. Nonetheless, by altering the envi­ ronment, fire changed the basic demography (increased mortality, growth, and flowering) of bromeliad popUlations. Field study of epiphyte seeds on different species of host trees also indicated that short-term adherence on trunks differed between two species (T. fasciculata and T. utriculata), and between burned and unburned hammocks. These results indicate that large-scale disturbances of the overstory can result in rapid changes in population dynamics of epiphytes. Such changes can potentially influence long-term composition and dynamics of populations of subcanopy epiphytes. Efectos de incendio sobre bromelias de los bosques subtropicales de madera dura del Parque Nacional Everglades, Florida. RESUMEN. Las respuestas de cinco especies de bromelias (Tillandsia balbisiana, T.fasciculata, T. setacea, T. utriculata, y T. venezuelana) de la canopia inferior del bosque subtropical de madera dura ("hammock") a las perturbaciones de gran escala debidas a incendios de baja intensidad, fueron inferidas a partir de comparaciones entre las poblaciones existentes en areas quemadas y no quemadas, observadas un ano despues de un incendio, en Long Pine Key en el Parque Nacional Everglades, Florida. En las areas quemadas se consumi6 el humus del suelo. Muchos arboles fueron desfoliados y/o murieron, 10 que determin6 la apertura de la canopia. Las comparaciones entre las poblaciones de epifitas de areas de "hammock" quemadas y no quemadas, muy similares antes del incendio, indican que el fuego tuvo poco efecto directo sobre las especies estudiadas. Sin embargo, a traves del cambio del ambiente, el incendio determin6 modificaciones en la demografia (incremento la mortalidad, crecimiento y floracion) de las poblaciones de bromelias. EI estudio de campo de la adherencia de semillas de epifitas sobre los troncos de arboles hue sped de diferentes especies indico que, en el corto plazo, la adherencia difiere entre dos especies (T. fasciculata y T. utriculata) y entre bosques quemados y no quemados. Estos resultados indican que los disturbios de gran escala en la canopia pueden resultar en cambios rapidos en la dinamica poblacional de las epifitas. Tales cambios pueden influir, potencialmente, en la composici6n y en las dinamica de largo plazo de las poblaciones de epifitas de la canopia inferior. INTRODUCTION of tolerance of adverse environmental condi­ tions. Studies of the ecology of vascular epiphytes Disturbances are widely recognized as pro­ have focused primarily on the harsh nature of ducing important changes in the environments the physical environment in the crowns of trees. experienced by tropical plants (Brokaw, 1985; Epiphyte morphology and physiology have been Platt & Strong, 1989). Responses of plants to related to specialization for photosynthesis, ab­ these changes may be as important as tolerance sorption of water, and uptake of nutrients in ae­ of adverse conditions (Grime, 1977). Responses rial environments that differ in light levels, mois­ of epiphytes (particularly those that inhabit ture availability, and mineral nutrients (Benzing, closed-canopy forests) to changes in environ­ 1990). Epiphytes have been categorized by the mental conditions produced by natural distur­ relative tolerance of different light levels, desic­ bances have not been studied. In this study, we cation, and by mechanisms for obtaining nutri­ explored changes in populations of subcanopy ents (e.g., Pittendrigh, 1948; Benzing & Renfrow, bromeliads in subtropical hardwood hammocks 1971; Griffiths & Smith, 1983; Gentry & Dod­ on Long Pine Key in Everglades National Park son, 1987). Species have been labeled as canopy following a naturally occurring fire. Populations or subcanopy, based in large part on the degree in burned hammocks were censused one year 39 40 SELBYANA [Volume 13 after the fire and compared to populations in minimize pre-fire differences: 1) all hammocks adjacent unburned hammocks. were in close proximity (about 1 km apart at the west end of Long Pine Key); 2) all hammocks METHODS had a similar fire history; prior to 1989, all six hammocks were last recorded as burned in 1940 Study Area (Olmsted et aI., 1983); 3) all hammocks were of similar size, from 80-100 m diameter, and were This study was conducted on Long Pine Key, more or less circular in shape. Burned hammocks a habitat island of about 8,000 ha of oolitic lime­ 9, 10, and 119 were approximately 0.8 ha, 1.1 stone outcroppings in Everglades National Park, ha, and 0.5 ha, respectively, and unburned ham­ Dade County, Florida. Long Pine Key, which has mocks 90, 101, and 113 were approximately 0.8 a maximum elevation of 5 m msl, contains ex­ ha, 0.5 ha, and 0.6 ha, respectively. tensive pine savannas dominated by Pinus el­ A transect was established through the center liotti var. densa that interdigitate with and are of each hammock and extended to the surround­ surrounded by short hydroperiod prairies and ing pinelands. Quadrats of 5 x 5 m were estab­ sawgrass savannas dominated by Cladium ja­ lished at 20 m intervals along the transects. Two maicense and Muhlenbergia jilipes (Craighead, out of three each of burned and unburned ham­ 1971; Olmsted et aI., 1983; Snyder et al., 1990). mocks contained 5 quadrats; the third hammock Hammocks imbedded within the pine savannas of both types contained four quadrats. In each contain an arborescent flora (more than 100 spe­ hammock, quadrats 1 and 5 (4) were close to the cies) that is predominantly tropical and West In­ edge (but still inside the hammock) and 2-4 (3) dian in origin (Robertson, 1953; Craighead, 1971; were in the interior. A total of 700 m 2 was sam­ Alexander & Crook, 1973; Olmsted et aI., 1983). pled for this study. Half ofthis area was in burned On Long Pine Key, hammocks typically have a hammocks, and half in unburned hammocks. closed canopy around 5 m, with scattered trees In each quadrat, the species and the sizes of emergent above the canopy (Olmsted et aI., 1980). all trees ~2 cm dbh (diameter at breast height) They are located in areas containing solution holes were measured and recorded as alive or dead filled with water. This produces a humid micro­ above ground level. In each quadrat, each bro­ environment, especially during the wet season meliad located below 5 m (canopy height) was from June-November (Phillips, 1940). Ham­ surveyed and the following data recorded: spe­ mock soils consist of peak deposits of 10-50 cm, cies, size (measured as length oflargest leaf), lo­ depending on elevation and recent fire history cation (on ground, tree trunk, treefall, branchfall, (Simpson, 1920; Alexander, 1967; Craighead, or vine), and height above ground. If present on 1974; Olmsted et aI., 1983). Hammocks on Long a tree, the size (dbh) and species of the host t~ee Pine Key have been numbered and mapped was recorded. Each plant was recorded as ahve (Olmsted et al., 1983); that numbering system or dead, and notes were taken on bolting, flow­ was used in this study. ering, and ovary/fruit development. The impact of the Ingraham Fire on survival, Survival, Growth and Flowering Following Fire growth, and flowering ofbromeliads was inferred from differences between plants in burned and In May 1989, a 50,000 ha lighting-initiated unburned hammocks. Five of the ten species of fire (the Ingraham Fire) burned pan of Long Pine Tillandsia recorded for Everglades National Park Key. Two hammocks, #8 and #119, were ob­ (Avery & Loope, 1983) were included in this served in detail during the fire. The fire was con­ study: (T. balbisiana Schult., T. fasciculata Sw., fined to the peat layer; flame lengths rarely reached T. setacea Sw., T. utriculata L., and T. valen­ 0.2 m above the substrate. Trees were top-killed zuelana A. Rich.). Of these five species, T. se­ and/or completely killed by girdling of trunks or tacea and T. valenzuelana liave been designated roots (but not by crown fires). Observations dur­ subcanopy species with atmospheric and tank ing and shortly after the fire indicated that epi­ trichomes, respectively (Snyder et al., 1990). Til- phytes were not directly harmed unless they were located on the ground or less than about 1 m landsia fasciculata and T. utriculata have been above ground. described as canopy species with tank- and at­ In the summer of 1990, approximately one mospheric-absorbing trichomes (Pittendrigh, year after the fire, six small hammocks were se­ 1948; Griffiths & Smith, 1983). Tillandsia bal­ lected for study. Three (#9, 10, and 119) had bisiana (which resembles T. jlexuosa and T. cir­ burned and three (#90, 100, and 113) had not cinnata in its morphology) has