Fate of Epiphytes on Phorophytes with Different Architectural Characteristics Along the Perturbation Gradient of Sabal Mexicana Forests in Veracruz, Mexico

Journal of Vegetation Science ]]: 1–10, 2009 & 2009 International Association for Vegetation Science

Fate of epiphytes on phorophytes with different architectural characteristics along the perturbation gradient of Sabal mexicana forests in Veracruz, Mexico

Aguirre, A.1,2; Guevara, R.1,3,4Ã; Garcı´a, M.5,6 &Lo´pez, J.C.7,8

1Instituto de Ecologıá, A.C. Departamento de Biologıá Evolutiva, Apartado Postal 63, Xalapa, Veracruz 91000, Me´xico; 2E-mail: [email protected]; 3E-mail: [email protected]; 4Current address: Department of Biology, Stanford University, 371 Serra Mall, Stanford, CA 94305, USA; 5Universidad Michoacana de San Nicola´sde Hidalgo, Facultad de Biologıá, Ciudad Universitaria, Morelia, Michoacań, Me´xico; 6E-mail: [email protected]; 7Centro de Investigaciones Tropicales, Universidad Veracruzana; 8E-mail: [email protected]; ÃCorresponding author; E-mail [email protected]

Abstract Keywords: Beta diversity; Epiphytic ferns; Forest fragmentation; Ophioglossum palmatum; Phylogenetic Question: Vascular epiphytes and hemiepiphytes (E/HE) diversity. in neotropical forests account for a large fraction of plant richness, but little is known of how the interplay between Nomenclature: Hietz & Hietz-Seifert (1995a). phorophyte architectural characteristics and habitat perturbation affect communities of E/HE.

Location: Sabal mexicana forests in a coastal area of Introduction Veracruz, Mexico. Water and nutrient uptake limits plant growth Methods: We compared communities of E/HE on phor- in most epiphytic environments (Benzing 1987, ophytes with different architectural characteristics – the 1990; Hietz & Briones 1998; Holbrook & Putz 1996). palm S. mexicana and non-palm phorophytes – in three Nonetheless, epiphytes/hemi-epiphytes (E/HE) ac- environments: conserved sites, perturbed sites and small count for ca. 10% of the vascular flora in the world regenerated forest fragments. We combined traditional (Gentry & Dodson, 1987). E/HE harbour a great (abundance, species richness, similarity and complemen- diversity of arthropods (Ellwood & Foster, 2004), tarity indices) and more recent (phylogenetic diversity) metrics to describe the communities of E/HE. provide abundant resources for frugivores and pol- linators (Cruz-Angoń & Greenberg, 2005; Gentry & Dodson, 1987), and take part in other important Results: Overall, we recorded 924 E/HE individuals (nine families, 16 genera and 21 species). The abundance and ecosystem processes, such as nitrogen fixation species richness of E/HE was higher on palms than on (Puente & Bashan, 1994). non-palm phorophytes. Abundance-based complementa- Microenvironmental conditions in epiphytic rities between phorophytes and sites were high. We rooting substrates are mediated not only by local- detected clear changes in community structure of E/HE scale environmental conditions but also by the ar- with habitat perturbation, but there were no effects on the chitectural characteristics of the phorophyte phylogenetic diversity of the E/HE community. Palm (Nadkarni 1984, 2002; Holbrook & Putz 1996). The phorophytes hosted a more phylogenetically diverse com- architectural characteristics of many species of munity of E/HE than did non-palm phorophytes. palms in the genera Caryota, Copernicia, Phoenix, Rhaphis, Sabal, Syagrus, Trachycarpus and Wa- Conclusions: Palm phorophytes are key elements support- shingtonia provide environments hospitable to ing the conservation of resilient communities of E/HE in epiphytes (Davis 1970; Putz & Holbrook 1989; S. mexicana forest. Habitat fragmentation has a strong effect on the structure of the E/HE community in S. Lawton & Williams-Linera 1996; Lo´pez & Dirzo mexicana forests. Ferns are the group of epiphytes most 2007). These palms retain their old leaf bases on the severely affected by habitat perturbation, but we detected trunk, creating a textured crisscross pattern, the in- no significant effect on the phylogenetic diversity of the ner part of which is protected from direct solar community. radiation and retains humidity. This thus facilitates 2 Aguirre, A. et al. the growth of decomposing microbes and detrivor- ing different architectural characteristics for the ous arthropods that break down and mix organic conservation of E/HE. matter with airborne dust, creating a fertile sub- Here, we investigated the effects of S. mexicana strate for epiphytes (Guevara & Lo´pez 2007). Lo´pez forest conversion on a community of vascular epi- (2007) estimated that a single palm of Sabal mex- phytes, considering two important components icana in central Veracruz contained up to 9 kg of relevant to the conservation of epiphytes. First, we fertile epiphytic soil, with 40-times more carbon, 14- considered two types of phorophyte with distinct ar- times more nitrogen and 12-times more phosphorus chitectures: S. mexicana palms and non-palm trees. than the soil supporting the growth of the palm. In Second, because the effects of forest conversion can contrast, in most non-palm trees, the epiphytic sub- vary among different groups of epiphytes, we ana- strate is often reduced to a layer of aerial roots of the lysed whether phylogenetic diversity of the epiphyte epiphytes, which trap airborne nutrients (Nadkarni community was affected by forest conversion and 1984; Nadkarni et al. 2002). type of phorophyte in this forest. We aimed to an- S. mexicana grows in lowlands on the Atlantic swer three questions. (1) Do sympatric phorophytes and Pacific coasts of Mexico, and is commonly as- with different trunk architectures host similar com- sociated with heavily transformed tropical forests munities of E/HE? (2) Is there a phylogenetic bias for (Zona 1990). However, in central Veracruz, S. mex- the composition of vascular epiphytes on different icana is a dominant species in remnants of primary types of phorophyte? (3) What are the effects of for- forests, a rare and endangered vegetation type in est conversion on the communities of vascular Mexico (Lo´pez & Dirzo 2007). Forests dominated epiphytes on the two types of phorophyte? by palms with textured crisscross trunks are scat- tered widely throughout the Americas: S. palmetto forests in southern Florida (Wade & Langdon Methods 1990), S. mexicana forests in central Veracruz, Mexico (Lo´pez & Dirzo 2007), Orbignya cohune Study site forests in Belize and other parts of Central America (Wright et al. 1959), and Copernicia tectorum in the The study was conducted in S. mexicana forests Venezuelan Llanos (Mayer 1933; Putz & Holbrook (1814804000-1815105100N, 9610502000-9610102500W), an 1989). In these forests, palms are accompanied by a ecosystem whose distribution in Mexico is restricted host of other tree species. For instance, Lo´pez & to the central coastal area of Veracruz (Pennington Dirzo (2007) listed 64 tree species in S. mexicana & Sarukhań 1998; Lo´pez & Dirzo 2007), although forests, including Brosimum alicastrum, Coccoloba similar formations occur in southern Florida (Wade barbadensis, Pithecellobium dulce, Nectandra salici- & Langdon 1990), Belize, other parts of Central folia, Ocotea cernua and Malvaviscus arboreus. America (Wright et al. 1959) and in the Venezuelan These palm-dominated forests offer a good oppor- Llanos (Mayer 1933; Putz & Holbrook 1989). tunity to investigate the fates of communities of The floristic richness of the S. mexicana forests is epiphytes on phorophytes having different archi- around 81 species, including trees and epiphytes, tectural characteristics and following human and has a clear affinity with the tropical dry and disturbance. tropical wet forests of Mexico, with a small re- S. mexicana forests are heavily impacted by hu- presentation of taxa from arid ecosystems (Lo´pez & man activities and have been predominantly Dirzo 2007). converted to pastures (Lo´pez & Dirzo 2007). Var- Most of the S. mexicana forests have been con- ious studies have documented the negative effects of verted to pastures and are systematically subjected natural forest conversion on the diversity of com- to dry season fires, preventing the establishment of munities of epiphytes (Hietz-Seifert et al. 1996; plants other than Sabal palms and C. barbadensis Padmawathe et al. 2004; Cascante-Marıń et al. 2006; trees. Some landowners stopped the practice of an- Flores-Palacios & Garcıá-Franco 2008). Further, nually burning the pastures more than 40 year ago, some studies have also shown that forest conversion and vegetation patches have re-grown, nourished by has different effects on different groups of epi- S. mexicana palms, but are still kept in check by phytes. For instance, in a moist lowland forest in cattle grazing (local peasants, personal communica- India, the abundance of vascular epiphytes, but not tion). The vegetation patches range in diameter of orchids, was negatively affected by selective log- from 2 to 30 m, and host a large fraction of the flora ging (Padmawathe et al. 2004). However, we know found in the conserved forest sites (Hernańdez-Her- very little about the relevance of phorophytes hav- nańdez 2009). - Fate of epiphytes on phorophytes with different architectural characteristics - 3

Sampling design gram based on published works on their phylogenetic relationships (Pryer et al. 2004). Then, fol- In this study, we analysed the diversity of E/HE lowing the Missouri Botanical Garden phylogenetic communities in three different environments: con- tree service for vascular plants, we counted the served forest sites, perturbed sites (managed number of nodes to a common ancestor between pastures), and re-grown vegetation patches. To pairs of genera in the cladogram. We used the num- minimize the potential effects of the sizes of the ve- ber of nodes to a common ancestor as a measure of getation patches, we only considered those 5-10 m in phylogenetic divergence between pairs of genera diameter. This size range accounts for more than (Stevens 2008). In this way, we made a conservative 70% of the vegetation patches at the study site. estimate of phylogenetic divergence among taxa In each of the three environments referred to whose phylogenetic relationships have been re- above, we selected 30 pairs of phorophytes. Each constructed based on different markers and criteria. pair consisted of an S. mexicana palm and a non- As an estimate of the phylogenetic diversity of the palm tree. The tree phorophytes included C. barba- community, we used the ratio between the length of densisis (44%), Pithecellobium duce (10%) and the cladogram without terminal nodes (genera) and Daphnopsis americana (12%), together with a fur- the length of the cladogram with terminal nodes. In ther 10 different tree species. The three most this way, we attempted to describe the fraction of common tree phorophytes are of similar size, bark phylogenetic diversity, or the length of the clado- texture and branching pattern. Tree height ranges gram (Faith & Baker 2006) represented above the from 15 to 20 m, and diameter at breast height up to level of the terminal nodes. 60 cm. The bark is fissured and branches are ascen- dant and copious (Nevling Jr. & Barringer, 1988; Pennington & Sarukhań, 1998; C. barbadensis in Data analysis most cases). Both types of phorophyte were of a si- To compare the diversity of the E/HE commu- milar size, as estimated by their diameters at breast nities among environments and phorophytes, we height (20-25 cm). In total, we included 180 phor- calculated individual-based rarefaction curves (Gotelli ophytes. We included all individuals of E/EH & Colwell 2001). We also used a new abundance- observed by naked-eye while standing on the based index to estimate complementarity and similar- ground, and avoided those phorophytes hosting ity among the communities of E/HE in different ground-rooted strangler figs, where most of the environments and on different types of phorophyte. palm architecture is replaced by the fig architecture. This index weights an analogue of the Bray-Curtis The stem-to-stem distance between the paired phor- distance by the species richness in each community ophytes ranged from 2 to 10 m. We identified and and the number of shared species. Simulation studies counted all vascular E/HE on these 180 phor- showed that the proposed index is less influenced by ophytes. Special care was taken when counting superabundant species than other commonly used aroids, ferns, bromeliads and orchids. For instance, abundance-based indices, including the Bray-Curtis with aroids, a discrete mass of engrossed roots and Morisita-Horn indices (results will be published bearing leaves was counted as a single individual, elsewhere). Calculation of the proposed index was as and no attempt was made to investigate whether follows. such a mass resulted from the anastomosis of two or more genetically different individuals; when two or AS cS AS bS C ¼ b b and C ¼ c c more discrete units of engrossed roots were observed bc 2 cb 2 Sb þ cSb Sc þ bSc on the same phorophyte, we counted them as different individuals. Similar criteria were followed for X A A 2 ferns, bromeliads and orchids. ð i;b i;cÞ A ¼ 2; ðAi;b þ Ai;cÞ Phylogenetic diversity where A 5 Abundance-based distance, analogous to the Bray-Curtis distance; Cbc and Ccb 5 comple- To account for changes in the phylogenetic di- mentarity from site b to site c and from c to b, re- versity of E/HE, we constructed a cladogram based spectively; Aib and Aic 5 abundance of the ith on the PHYLOMATIC application for angios- species in samples b and c, respectively; Sb and Sc 5 perms, and used the most-resolved tree protocol total species richness in samples b and c, respec- (Webb & Donoghue 2005). PHYLOMATIC does tively; b and c 5 number of species exclusive to not include ferns, so we added them to the clado- samples b and c, respectively. If samples b and c are 4 Aguirre, A. et al.

identical, the complementarities, Cbc and Ccb, equal Palm phorophytes accumulated species of E/ zero. In contrast, if samples b and c are composed HE faster than non-palm phorophytes in all three entirely of different sets of species and the total environments (Fig. 1). E/HE species accumulation abundance is the same in each community, then Cbc rates on palms also differed signiﬁcantly among en- and Ccb both equal 0.5. Thus |1 (Cbc1Ccb)| is an vironments (Pbootstrapo0.05): the lowest species estimate of similarity between samples b and c. accumulation rate was observed in vegetation pat- To explore the overall pattern of similarities ches, followed by conserved sites, and perturbed between sites and phorophytes, we used the trian- sites had the highest E/HE species accumulation gular matrix of similarities between pairs of rate. In contrast, the highest E/HE species accumu- environments and phorophytes in a cluster analysis lation rate on non-palm phorophytes was in the based on Euclidian distances and Ward’s method. vegetation patches, followed by perturbed sites, To explore changes in the structure of the E/HE whereas at conserved sites only two species (Syngo- communities among environments and phorophytes nium chiapense and Anthurium schlechtendalii) were we also plotted rank-abundance curves. observed. Remarkably, the E/HE species accumula- To compare differences in phylogenetic di- tion rate on palm phorophytes at perturbed sites did versity of the E/HE communities on the two types of not differ signiﬁcantly from the overall E/HE species phorophyte, we generated 5000 pairs of random accumulation rate (Pbootstrap 5 0.92). cladograms with the function rtree (library ape)inR (Paradis et al. 2004). We used a random uniform distribution for branch lengths, and the number of Complementarity and similarity terminal nodes in the random cladograms for each Reciprocal complementarity indices among pair was equal to the number of terminal nodes in phorophytes and environments were biased (Table 1). the actual cladograms at the generic level. We then The overlap of the E/HE communities on non-palm calculated the frequency distribution of the differ- phorophytes with those on palm phorophytes was ences in phylogenetic diversity between random twofold higher than the reciprocal overlap. Overlaps pairs of cladograms and the empirical two-tailed of the E/HE communities at the conserved sites with probability for the actual differences in phylogenetic those at perturbed sites and vegetation patches were diversity between the E/HE communities on the two 2.6- and 1.4-times higher than their reciprocal over- types of phorophyte. The same approach was fol- laps, respectively. The same pattern was observed lowed to compare phylogenetic diversity among when we compared environments, considering the E/ sites. All statistical analyses, including phylogenetic HE communities on both types of phorophyte sepa- diversity analyses, were performed in the statistical rately. Cluster analysis based on the matrix of language R (http://www.R-project.org). similarities (Fig. 2) revealed two main clusters, which correspond to the two different types of phorophyte. Within each of the two main clusters, the E/HE com- Results

Abundance and species richness

In total, we recorded 924 E/HE individuals, grouped in nine families, 17 genera and 21 species. The best-represented groups were strangler ﬁgs (ﬁve species), ferns (four species), orchids (three species) and cacti (three species). The abundance and species richness of the E/HE were higher on palms (652 individuals and 17 species, respectively) than on non- palm phorophytes (272 individuals and nine species, respectively). Conserved sites and vegetation patches hosted 11 and 13 species of E/HE, respectively, whereas there were 18 species of E/HE at the per- Fig. 1. Individual-based rarefaction curves for the com- turbed sites. In contrast, the abundance of E/HE at munity of epiphytes and hemi-epiphytes on palm (Sabal the conserved sites (345) was, on average, 19% mexicana) and non-palm phorophytes at three different higher than that in the vegetation patches (288) or environments: conserved sites (CS), perturbed sites (PS) perturbed sites (291). and vegetation patches (VP). - Fate of epiphytes on phorophytes with different architectural characteristics - 5

Table 1. Complementarities among communities of epiphytes and hemiepiphytes on palm and non-palm phorophytes in three different environments in Sabal mexicana forests. Values are complementarities of col- umns on rows. Palms Non-palm trees

(a) Overall complementarity between types of phorophyte Palms 0 0.29 Non-palm trees 0.68 0

Conserved Perturbed Vegetation sites sites patches

(b) Overall complementarity among sites Conserved sites 0 0.57 0.38 Perturbed sites 0.22 0 0.10 Vegetation patches 0.26 0.37 0 (c) Complementarity among sites on palm phorophytes Conserved sites 0 0.51 0.19 Perturbed sites 0.28 0 0.07 Vegetation patches 0.36 0.47 0 (d) Complementarity among sites on non-palm phorophytes Conserved sites 0 0.72 0.72 Fig. 3. Rank-abundance curves of epiphytes and hemi- Perturbed sites 0.15 0 0.35 epiphytes in Sabal mexicana forests on palm (Sabal Vegetation patches o0.01 0.24 0 mexicana) and non-palm phorophytes at three different environments: conserved sites (CS), perturbed sites (PS) and vegetation patches (VP). (a) Syngonium chiapense,(b) Anthurium schlechtendalii,(c) Tillandsia recurvata,(d) Epiphyllum crenatum,(e) Phlebodium decumanum, (f) Cat- asetum integerrimum,(g) Vittaria sp., (h) Tillandsia concolor,(i) Ficus pertusa,(j) Ficus obtusifolia,(k) Ne- phrolepis sp., (l) Ophioglossum palmatum,(m) Brassavola sp, (n) Selenicereus coniﬂorus,(o) Myrmecophila tibicinis, (p) Bromelia pinguin,(q) Ficus aurea,(r) Ficus trigonata,(s) Lasiacis procerrima,(t) Ficus perforata,(u) Nopalea cochenillifera.

of the fern Vittaria graminifolia was reduced six- fold at perturbed sites compared with its abun- Fig. 2. Hierarchical grouping of the community of epiphytes and hemiepiphytes on palm (Sabal mexicana) and dance at conserved sites, and this species was not non-palm phorophytes, at three different environments: recorded in vegetation patches. Two other fern conserved sites (CS), perturbed sites (PS) and vegetation species, Nephrolepis sp. and Ophioglossum palma- patches (VP). tum, were also exclusively observed at conserved sites. In contrast, abundances of the orchid Cata- munity at the conserved sites was separate from those setum integerrimum and the cactus Epiphyllum in the other two environments. crenatum increased twofold and threefold, respectively, at perturbed sites compared with their Community structure abundances at conserved sites. It is noteworthy that a number of species were The rank-abundance curves for the E/HE restricted to perturbed sites or vegetation patches, communities on palm and non-palm phorophytes and that these species were differently accom- in all three environments were steep, indicating that modated in the E/HE communities on palm and the diversity of these communities was low (Fig. 3). non-palm phorophyte. At perturbed sites and in ve- Nonetheless, the curves showed important changes getation patches, palm phorophytes accumulated in the structures and compositions of E/HE com- rare species of E/HE (low abundance and low fre- munities on palm and non-palm phorophytes. On quency), including hemiepiphytic species of Ficus (F. palms, the abundance of S. chiapense was reduced aurea, F. trigonata and F. perforata) and accidental by 50% at perturbed sites compared with its abun- epiphytes, i.e. taxa that normally grow as terrestrial dance at conserved sites. Similarly, the abundance plants, such as the bromeliad Bromelia pinguin, the 6 Aguirre, A. et al. bamboo Lasiacis procerrima and the cactus Nopalea Lo´pez & Dirzo (2007) found that forest conver- cochenillifera. In contrast, in non-palm phor- sion had little impact on E/HE communities in S. ophytes, those species of E/HE restricted to the mexicana forests. In contrast, we found that forest perturbed sites or vegetation patches were domi- conversion had strong effects on E/HE communities nant, including the bromeliads Tillandsia recurvata in this S. mexicana forest, but that these effects were and T. concolor, and to a lesser extent the orchids phorophyte type-dependent. Species richness in- Brassavola sp. and Myrmecophila tibicinis. creased at perturbed sites compared with species richness at conserved sites. This pattern held when we looked separately at E/HE communities on palm Phylogenetic diversity and non-palm phorophytes. Our findings are con- trary to those reported for other tropical forests, in The cladogram shows that the E/HE commu- which E/HE species richness was consistently im- nity in S. mexicana forests is composed of four main poverished by forest conversion (Barthlott et al. clades (Fig. 4): ferns, monocots, cacti and strangler 2001; Cascante-Marıń et al. 2008; Flores-Palacios & figs. The clade of ferns is restricted to palm phor- Garcıá-Franco 2008). The species richness of E/HE ophytes, and the clades of cacti and Ficus species are on palm phorophytes at perturbed sites was high strongly biased toward palm phorophytes. In con- because of the accumulation of low-abundance spe- trast, bromeliads and orchids are predominantly cies (F. aurea, F. trigonata and F. perforata) and restricted to non-palm phorophytes, with the excep- accidental epiphytes (N. cochenillifera, L. procerri- tion of the orchid C. integerrimum, which was found ma and B. pinguin). Conversely, the species richness exclusively on palm phorophytes. The clade of ferns of E/HE on non-palm phorophytes at perturbed was also biased toward conserved sites, whereas the sites was high because of the accumulation of highly clades of Ficus and cacti were biased toward per- abundant species (T. recurvata, T. concolor and turbed sites. Brassavola sp.). These different patterns observed Finally, we found that the differences in phylo- on palm and non-palm phorophytes can be partly genetic diversity at the generic level between palm explained by the rich and abundant aerial substrate (0.82) and non-palm (0.52) phorophytes were un- offered by palms, which facilitates the growth of ac- likely to be attributable to chance (P 5 0.014), cidental epiphytes in open areas. In contrast, non- whereas differences in phylogenetic diversity among palm phorophytes accumulate species that are cap- sites were not statistically significant (P40.30). able of deriving nutrients from airborne dust and atmospheric nitrogen through symbiotic mutualists in their roots (Puente & Bashan 1994; Gebauer & Discussion Meyer 2003; Bayman & Otero 2006). Nonetheless, it is puzzling that no Tillandsia or orchid species, other Nine of the 21 species of E/HE associated with than C. integerrimum, were observed growing on two types of phorophyte are newly recorded species palm phorophytes. Although availability of light at the study site (cf. Lo´pez & Dirzo 2007), and the could limit germination in the inner part of the tex- observation of O. palmatum is new for the state of tured crisscross pattern formed by the old leaf Veracruz (cf. Mickel & Smith 2004). In agreement petioles of S. mexicana, the external surface is fully with Lo´pez & Dirzo (2007), we found that aroids exposed, and there seems to be no limitation to were the most abundant E/HE in S. mexicana for- the establishment of species such as T. recurvata, ests. This finding is consistent with another study of which is commonly observed growing on electrical neotropical forests (Benavides et al. 2006) but con- transmission cables (Puente & Bashan 1994). In tradicts a study in southern India, where aroids are agreement with our observations, other studies have poorly represented in a community of E/HE similar demonstrated experimentally that phorophyte ar- in total richness to the E/HE communities in S. chitectural and bark characteristics can restrict mexicana forests. Ferns were the second-most growth of some epiphyte species (Callaway et al. abundant group of E/HE in our study, which is in 2002). agreement with most studies that have investigated A salient conservation factor detected in this the diversity of E/HE communities in the neotropics study is the threat that forest conversion represents (Kreft et al. 2004; Ku¨per et al. 2004; Cardelu´s et al. for epiphytic ferns in S. mexicana forests. Whereas 2006; Zotz & Schultz 2008), Asia (Annaselvam & the abundance of Phlebodium decumanum was un- Phartasarathy 2007) and Africa (Zapfack & En- affected by habitat perturbation, only O. palmatum gwald 2008). and Nephrolepis sp. were observed at conserved - Fate of epiphytes on phorophytes with different architectural characteristics - 7

Fig. 4. Phylogenetic relationship of epiphytes and hemiepiphytes in Sabal mexicana forests and palm and non-palm phorophytes. On the right is abundance of each species on the two types of phorophyte, S. mexicana palms (black bars) and non- palms phorophytes (grey bars) at three different environments: conserved sites (CS), perturbed sites (PS) and vegetation patches (VP). Asterisks indicate singletons. sites, and the abundance of V. graminifolia was re- perturbation on the phylogenetic diversity of E/HE, duced at perturbed sites compared with its and this was independent of the occurrence of acci- abundance at conserved sites. The case of O. palma- dental epiphytes. In contrast, palm phorophytes tum is highly relevant because this species was hosted a more phylogenetically diverse community previously known in Mexico at only a handful of of E/HE than non-palm phorophytes. Whereas non- localities in Chiapas and Oaxaca States, so our palm phorophytes mainly hosted monocots, palm finding extends the distribution range of this species. phorophytes included all the major clades of E/HE Another important issue for conservation is that found in the S. mexicana forest: strangler figs, cacti, vegetation patches resembled conserved sites in the monocots and ferns. To the best of our knowledge, composition of their E/HE communities, although no other study has addressed the phylogenetic di- the overall abundance was low, similar to that ob- versity of communities of E/HE, and we suggest that served at perturbed sites. This is an important this level of analysis has great potential to identify finding, suggesting that the E/HE community is further threats to biodiversity posed by forest con- highly resilient to perturbation. version. Despite the negative effects of forest conversion Consistent with the differences in phylogenetic on ferns, we found no significant effect of habitat diversity of the E/HE communities on both types of 8 Aguirre, A. et al. phorophyte, the similarity indices also distinguished diameter at breast height between 20 and 25 cm for the communities: first by type of phorophyte and both types of phorophyte. Lo´pez (2007) showed that second by habitat perturbation. All the evidence a single S. mexicana palm could accumulate up to gathered in this study suggests that palm phor- 9 kg of highly fertile aerial substrate. Furthermore, ophytes are essential to maintenance of E/HE the epiphytic substrate on S. mexicana harbours a community diversity in S. mexicana forests. We healthy microbial decomposer community, actively propose that the textured crisscross pattern formed releasing nutrients locked in the organic matter and by old leaf petioles, and the organic substrate that rendering them available for plant growth (Guevara accumulates in it, provides a hospitable environ- &Lo´pez 2007). Conversely, non-palm phorophytes ment for epiphytic plant growth. Remarkably, there accumulate very little or no substrate (Nadkarni were a number of species (all of the ferns, the orchid 1984; Nadkarni et al. 2002). Therefore, differences C integerrimum, the cactus Selenicereus sp. and most in architecture and especially differences in the species of Ficus) restricted to palm phorophytes. It is amount of aerial substrate (soil) accumulated on the likely that stems of S. mexicana retain humidity fa- two types of phorophyte seem to be the most likely vouring the growth of specialized epiphytes in the explanations for the observed differences in metrics same way as the stem base of trees in cloud forests of the communities. favour establishment of a specialized group of gen- In conclusion, our study shows that forest con- uine epiphytic ferns that require high atmospheric version has a strong effect on the structure of the E/ humidity (Hietz & Hietz-Seifert 1995b). The en- HE community in S. mexicana forests. Ferns are the vironment provided by the palm stem is so benign group of epiphytes most severely affected by forest that it even allows growth of accidental epiphytes, conversion, although we detected no significant ef- such as L. procerrima, B. pinguin and N. cochenilli- fect on the phylogenetic diversity of the community. fera. However, as noted above, Tillandsia and However, our analysis was conservative, insofar as orchids other than C. integerrimum were not ob- we used the number of nodes to a common ancestor served on palm phorophytes. Overall, our finding as the estimate of phylogenetic diversity, and also that the architectural characteristics of Sabal palms because we estimated phylogenetic diversity at the facilitate the occurrence of a diverse community of generic level. Our study also shows that palm E/HE is consistent with the study of Mehltreter et al. phorophytes are key components in the conserva- (2005), who found that species richness and abun- tion of a subset of the community of E/HE in S. dance of epiphytes was higher on tree-fern mexicana forests that have been heavily trans- phorophytes with a complex stem architecture than formed by human activities, where light irradiation on angiosperm phorophytes in a tropical cloud for- is high and may prevent establishment of species est. Similarly, Laube & Zotz (2006) and Zotz & with high nutrient and humidity requirements. Schultz (2008) demonstrated that palms were im- Further investigation of other tropical forests con- portant phorophytes for epiphytes in the tropical taining palm phorophytes with similar trunk rain forests of Panama. architecture to that of S. mexicana,e.g.S. palmetto The species richness, abundance and phyloge- forests in southern Florida (Wade & Langdon netic diversity of E/HE were higher on palm 1990), O. cohune forests in Belize and other parts of phorophytes than on non-palm phorophytes. S. Central America (Wright et al. 1959) and C. tector- mexicana is the most important woody species at the um in the Venezuelan Llanos (Mayer 1933; Putz & study site, in both conserved and perturbed sites Holbrook 1989), is required to fully understand (Lo´pez & Dirzo 2007). Nevertheless, if we consider the role of palms as phorophytes for vascular epi- the pooled importance of all potential non-palm phytes. phorophytes with diameters at breast height 410 cm, the relative importance of S. mexicana in the forest community is considerably diluted, especially at conserved sites and in vegetation patches. Acknowledgements. INECOL funded this project and Thus, the density- or frequency-dependent effects of supported A. Aguirre with a postdoctoral fellowship. M. the two types of phorophyte are unlikely explana- Garcıá was supported by the summer research programs tions for the observed differences in the E/HE DELFIN-2006 and AMC-2007, and R. Guevara was community metrics on the two types of phorophyte. supported by CONACYT (79846 sabbatical leave fellow- A size-related bias is equally unlikely (e.g. Hietz & ship). We thank Claudia Moreno and Jose´G. Garcıá Hietz-Seifert 1995a; Zotz & Vollrath 2003) because Franco for their constructive criticism of a previous ver- we restricted our analysis to phorophytes with a sion of this work. - Fate of epiphytes on phorophytes with different architectural characteristics - 9

References epiphyte community in lower montane forest, Veracruz, Mexico. Biodiversity and Conservation 17: Annaselvam, J. & Phartasarathy, N. 2007. Diversity and 191–207. distribution of herbaceous vascular epiphytes in a Gebauer, G. & Meyer, M. 2003. 15N and 13C natural tropical evergreen forest at Varagalaiar, Western abundance of autotrophic and myco-heterotrophic Ghats, India. Biodiversity and Conservation 10: 317–329. orchids provides insight into nitrogen and carbon Barthlott, W., Schmit-Neuerburg, V., Nieder, J. & gain from fungal association. New Phytologist 160: Engwald, S. 2001. Diversity and abundance of 209–223. vascular epiphytes: a comparison of secondary Gentry, A.H. & Dodson, C.H. 1987. Diversity and vegetation and primary montane rain forest in the biogeography of neotropical vascular epiphytes. Venezuelan Andes. Plant Ecology 152: 145–156. Annals of the Missouri Botanical Garden 74: 205–233. Bayman, P. & Otero, J.T. 2006. Microbial endophytes of Gotelli, N. & Colwell, R.K. 2001. Quantifying orchid roots. In: Schulz, B., Boyle, C. & Sieber, T.N. biodiversity: procedures and pitfalls in the (eds.) Microbial Root Endophytes. pp. 153–177. measurement and comparison of species richness. Springer-Verlag, Berlin and Heidelberg, DE. Ecology Letters 4: 379–391. Benavides, M.A., Wolf, J.H.D. & Duivenvoorden, J.F. Guevara, R & Lo´pez, J.C. 2007. Quality of rooting 2006. Recovery and succession of epiphytes in upper environments and patterns of root colonization by Amazonian follows. Journal of Tropical Ecology 22: arbuscular mycorrhizal fungi in strangler figs in a 705–717. Mexican palmetto woodland. Mycorrhiza 17: 589– Benzing, D.H. 1987. Vascular epiphytism: taxonomic 596. participation and adaptive diversity. Missouri Hernańdez-Hernańdez, D.E. 2009. Plant diversity co- Botanical Garden Annual Report 74: 183–204. varies with soil fertility and fragment size in Sabal Benzing, D.H. 1990. Vascular epiphytes. General biology mexicana forests in Veracruz. MSc Thesis. INECOL. and related biota. Cambridge University Press, Xalapa, Ver. MX. Cambridge, UK. Hietz, P. & Briones, O. 1998. Correlation between water Callaway, M.R., Reinhart, K.O., Moore, G.W., Moore, relations and within-canopy distribution of epiphytic D.J. & Pennings, S.C. 2002. Epiphyte host preferences ferns in a Mexican cloud forest. Oecologia 114: 305– and host traits: mechanisms for species-specific 316. interactions. Oecologia 132: 221–230. Hietz, P. & Hietz-Seifert, U. 1995a. Composition and Cardelu´s, C.L., Colwell, R.K. & Watkins, J.E. 2006. ecology of vascular epiphyte communities along an Vascular epiphyte distribution patterns: explaining altitudinal gradient in central Veracruz, Mexico. the mid-elevation richness peak. Journal of Ecology Journal of Vegetation Science 6: 487–498. 94: 144–156. Hietz, P. & Hietz-Seifert, U. 1995b. Structure and ecology Cascante-Marıń, A., Wolf, J.H.D., Gerard, J., of epiphyte communities of a cloud forest in central Oostermeijer, B. & den Nijs, J.C.M. 2008. Veracruz, Mexico. Journal of Vegetation Science 6: Establishment of epiphytic bromeliads in successional 719–728. tropical premontane forests in Costa Rica. Biotropica Hietz-Seifert, U., Hietz, P. & Guevara, S. 1996. Epiphyte 40: 441–448. vegetation and diversity on remnant trees after forest Cascante-Marıń, A., Wolf, J.H.D., Oostermeijera, clearance in southern Veracruz, Mexico. Biological J.G.B., den Nijsa, J.C.M., Sanahujab, O. & Durań- Conservation 75: 103–111. Apuy, A. 2006. Epiphytic bromeliad communities in Holbrook, N.M. & Putz, F.E. 1996. Physiology of tropical secondary and mature forest in a tropical premontane vines and hemiepiphytes: plants that climb up and area. Basic and Applied Ecology 7: 520–532. plants that climb down. In: Smith, A.P., Mulkey, S.S. Cruz-Angoń, A. & Greenberg, R. 2005. Are epiphytes & Chazdon, R.L. (eds.) Tropical Forest Plant important for birds in coffee plantations? An Ecophysiology. pp. 363–394. Chapman & Hall, experimental assessment. Journal of Applied Ecology London, UK. 42: 150–159. Kreft, H., Ko¨ster, N., Ku¨per, W., Nieder, J. & Barthlott, Davis, T.A. 1970. Epiphytes that strangulate palms. W. 2004. Diversity and biogeography of vascular Principes 14: 10–25. epiphytes in Western Amazonia, Yasunı´, Ecuador. Ellwood, M.D.F. & Foster, W.A. 2004. Doubling the Journal of Biogeography 31: 1463–1476. estimate of invertebrate biomass in a rainforest Ku¨per, W., Kreft, H., Nieder, J., Ko¨ster, N. & Barthlott, canopy. Nature 429: 549–551. W. 2004. Large-scale diversity patterns of vascular Faith, D.P. & Baker, A.M. 2006. Phylogenetic diversity (PD) epiphytes in Neotropical montane rain forests. and biodiversity conservation: some bioinformatics Journal of Biogeography 31: 1477–1487. challenges. Evolutionary Bioinformatics (online) 2: Laube, S. & Zotz, G. 2006. Neither host-specific nor 70–77. random: vascular epiphytes on three tree species in a Flores-Palacios, A. & Garcıá-Franco, J.G. 2008. Habitat Panamanian lowland forest. Annals of Botany 97: isolation changes the beta diversity of the vascular 1103–1114. 10 Aguirre, A. et al.

Lawton, R.O. & Williams-Linera, G. 1996. early leptosporangiate divergences. American Journal Hemiepiphyte-host relationships: research problems of Botany 91: 1582–1598. and prospects. Selbyana 17: 71–74. Puente, M.E. & Bashan, E. 1994. The desert epiphyte Lo´pez, J.C. 2007. Variacioń ontogeńica en la estrategia de Tillandsia recurvata harbours the nitrogen-fixing defensas antı´-herbı´voro en plantas hemiepı´fitas: un bacterium Pseudomonas stutzeri. Canadian Journal of estudio con Ficus obtusifolia en un palmar biodiverso Botany 72: 406–408. y amenazado. Tesis de Doctorado. Instituto de Putz, F.E. & Holbrook, N.M. 1989. Strangler Fig Rooting Ecologıá, A.C, MX. Habits and Nutrient Relations in the Llanos of Lo´pez, J.C. & Dirzo, R. 2007. Floristic diversity of Sabal Venezuela. American Journal of Botany 76: 781–788. palmetto woodland: an endemic and endangered Stevens, P.F. 2008. Angiosperm Phylogeny Website. vegetation type from Mexico. Biodiversity and Version 9. Available at: http://www.mobot.org/ Conservation 16: 807–825. MOBOT/research/APweb/ Accessed 16 April 2009. Mayer, J.G. 1933. Notes on the vegetation of Venezuelan Wade, D.D. & Langdon, G. 1990. Sabal palmetto (Walt.) llanos. Journal of Ecology 21: 335–349. Lodd. ex J. A. & J. H. Schult. In Burns, R.M. & Mehltreter, K., Flores-Palacios, A. & Garcıá-Franco, J.G. Honkala, B.H. (eds.) Silvics of North America: 2005. Host preferences of low-trunk vascular Hardwoods. pp. 1466–1474. United States epiphytes in a cloud forest of Veracruz, Mexico. Department of Agriculture (USDA), Forest Service, Journal of Tropical Ecology 21: 651–660. Agriculture Handbook, Washington, DC, US. Mickel, J.T. & Smith, A.R. 2004. The Pteridophytes of Webb, C.O. & Donoghue, M.J. 2005. Phylomatic: tree Mexico. Memoirs of the New York Botanical Garden assembly for applied phylogenetics. Molecular 88: 1–1054. Ecology Notes 5: 181–183. Nadkarni, N.M. 1984. The biomass and nutrient capital Wright, A.C.S., Romney, D.H., Arbuckle, R.H. & Vial, of epiphytes in a neotropical cloud forest, V.E. 1959. Land in British Honduras. Colonial Monteverde. Biotropica 15: 1–9. Research Publications #24, London, UK. Nadkarni, N.M., Schaefer, D., Matelson, T.J. & Solano, Zapfack, L. & Engwald, S. 2008. Biodiversity and spatial R. 2002. Comparison of arboreal and terrestrial soil distribution of vascular epiphytes in two biotopes of characteristics in a lower montane forest in the Cameroonian semi-deciduous rain forest. Plant Monteverde, Costa Rica. Pedobiologia 46: 24–33. Ecology 195: 117–130. Nevling Jr, L.I. & Barringer, K. 1988. Flora de Veracruz: Zona, S. 1990. A monograph of Sabal (Arecaceae: Thymelaeaceae. Fascıćulo 59. Instituto Nacional de Coryphoideae). Aliso 12: 583–666. Investigaciones Sobre Recursos Bio´ticos, MX. Zotz, G. & Schultz, S. 2008. The vascular epiphytes of a Padmawathe, R., Qureshi, Q. & Rawat, G.S. 2004. Effects lowland forest in Panama: species composition and of selective logging on vascular epiphyte diversity in a spatial structure. Plant Ecology 195: 131–141. moist lowland forest of Eastern Himalaya, India. Zotz, G. & Vollrath, B. 2003. The epiphyte vegetation of Biological Conservation 119: 81–92. the palm Socratea exorrhiza correlations with tree size, Paradis, E., Claude, J. & Strimmer, K. 2004. APE: tree age and bryophyte cover. Journal of Tropical analyses of phylogenetics and evolution in R Ecology 19: 81–90. language. Bioinformatics 20: 289–290. Pennington, T.D. & Sarukhań, J. 1998. A´rboles tropicales de Me´xico. Universidad Nacional Autońoma de Me´xico y Fondo de Cultura Econo´mica, 2nd edicioń. MX. Pryer, K.M., Schuettpelz, E., Wolf, P.G., Schneider, H., Received 26 May 2009; Smith, A.R. & Cranfill, R. 2004. Phylogeny and Accepted 18 September 2009. evolution of ferns (Monilophytes), with a focus on the Co-ordinating Editor: O. Wildi.