High but not dry: diverse epiphytic bromeliad adaptations to exposure within a seasonally dry tropical forest community


  • C. Reyes-García,

    1. Centro de Investigación Científica de Yucatán S.A., Calle 43, Num. 130 Chuburná de Hidalgo 97200, Mérida, México
    2. Department of Plant Sciences, Downing Street, University of Cambridge, Cambridge CB2 3EA, UK
    Search for more papers by this author
  • M. Mejia-Chang,

    1. Department of Plant Sciences, Downing Street, University of Cambridge, Cambridge CB2 3EA, UK
    2. Centre de Recerca Ecològica i Aplicacions Forestals (CREAF-CEAB-CSIC), Campus de Bellaterra Universitat Autónoma de Barcelona, Cerdanyola del Vallès, Barcelona, 08193, Spain
    Search for more papers by this author
  • H. Griffiths

    1. Department of Plant Sciences, Downing Street, University of Cambridge, Cambridge CB2 3EA, UK
    Search for more papers by this author

Author for correspondence:
Casandra Reyes-García
Tel: +52 999 942 8330
Email: creyes@cicy.mx


  • Vascular epiphytes have developed distinct lifeforms to maximize water uptake and storage, particularly when delivered as pulses of precipitation, dewfall or fog. The seasonally dry forest of Chamela, Mexico, has a community of epiphytic bromeliads with Crassulacean acid metabolism showing diverse morphologies and stratification within the canopy. We hypothesize that niche differentiation may be related to the capacity to use fog and dew effectively to perform photosynthesis and to maintain water status.
  • Four Tillandsia species with either ‘tank’ or ‘atmospheric’ lifeforms were studied using seasonal field data and glasshouse experimentation, and compared on the basis of water use, leaf water δ18O, photosynthetic and morphological traits.
  • The atmospheric species, Tillandsia eistetteri, with narrow leaves and the lowest succulence, was restricted to the upper canopy, but displayed the widest range of physiological responses to pulses of precipitation and fog, and was a fog-catching ‘nebulophyte’. The other atmospheric species, Tillandsia intermedia, was highly succulent, restricted to the lower canopy and with a narrower range of physiological responses. Both upper canopy tank species relied on tank water and stomatal closure to avoid desiccation.
  • Niche differentiation was related to capacity for water storage, dependence on fog or dewfall and physiological plasticity.


Seasonally dry tropical forests (SDTFs) comprise 40% of tropical forest ecosystems, and are under threat from agricultural incursions, urbanization and a changing climate (Murphy & Lugo, 1986; Miles et al., 2006). Plants inhabiting semi-arid or arid ecosystems optimize the use of short, scattered (discontinuous) precipitation events (Noy-Meir, 1974; Goldberg & Novoplansky, 1997; Chesson et al., 2004; Reynolds et al., 2004; Schwinning & Sala, 2004; Reyes-García & Griffiths, 2009; Yepez & Williams, 2009). Vascular epiphytes are one key component of such systems, representing sensitive markers of climatic conditions and the extent of precipitation inputs. The Bromeliaceae family has epiphytic and terrestrial lifeforms distributed throughout the neotropics. The distribution of particular epiphytic bromeliad species can be related to narrow rainfall zones, as well as vertically within forest canopies following microclimatic constraints (Pittendrigh, 1948; Griffiths & Smith, 1983; Griffiths et al., 1986; Griffiths & Maxwell, 1999; Benzing, 2000; Graham & Andrade, 2004; Reyes-García et al., 2008a; Zotz & Schultz, 2008; Bader et al., 2009; Benavides et al., 2011).

The Bromeliaceae display a range of lifeforms and associated photosynthetic pathways, whether C3 or with Crassulacean acid metabolism (CAM: engaging in CO2 uptake and water loss primarily at night; Griffiths & Smith, 1983). Large rosette-forming bromeliad species, with overlapping leaf bases forming a distinct ‘tank’ for water storage and nutrient acquisition, are usually restricted to moister forest formations (Pittendrigh, 1948; Benzing, 2000). In the more arid, dry forest regions, where host trees (phorophytes) generally form lower canopies, there is a tendency for CAM to become the dominant photosynthetic pathway of the epiphytic community (Griffiths & Smith, 1983; Reyes-García et al., 2008a). In addition, such species show a morphological progression towards more ‘atmospheric’ lifeforms which completely lack tank water storage capacity (Benzing, 2000). Atmospheric lifeforms, exemplified by Spanish moss (Tillandsia usneoides), possess an epidermis covered by overlapping layers of leaf trichomes that absorb water and nutrients during precipitation pulses (Benzing & Ott, 1981; Benzing, 2000, 2003), mediated by aquaporins (Ohrui et al., 2007).

Few studies have investigated the role of dew and fog in epiphytic metabolism. Although increased water content in bromeliads was described after these events decades ago (Benzing & Dahle, 1971; Virzo de Santo et al., 1976; Martin & Schmitt, 1989), only recently have the effect of these fog and dew events been related to epiphytic metabolism, describing their key role in the nutrient acquisition of the desert bromeliad Tillandsia landbeckii (González et al., 2011). Several ecological studies in dry forests and desert environments have attributed a key role to dew and/or fog in the water budget of epiphytes and other lifeforms (Rundel & Dillon, 1998; Munne-Bosch et al., 1999; Andrade, 2003; Graham & Andrade, 2004; Westbeld et al., 2009; González et al., 2011; Guevara-Escobar et al., 2011). Martorell & Ezcurra (2007) described the abundance of a specific lifeform, which enhances interception by ‘harvesting’ fog water, in sites with low rainfall and frequent fog formation. The ‘nebulophyte lifeform’ was described in plants from various families that have long and thin leaves with small boundary layers. These leaves are displayed in the upper canopy (as epiphytes or terrestrial species with long stems) and flutter with wind currents, promoting the interception of fog droplets by the leaf blade.

Stable isotopes of water (particularly 18O) constitute a useful tool to monitor exchanges of water between the leaf and the environment. During transpiration, the lighter 16O molecules tend to preferentially evaporate, leaving the remaining leaf water enriched in the heavier 18O molecules (Craig & Gordon, 1965). At a higher leaf temperature and lower humidity, higher evaporation rates will increase enrichment, whereas precipitation inputs will reset the signal back to the original (depleted) values (Barbour et al., 2000; Barbour & Farquhar, 2004). In epiphytic bromeliads showing CAM, fog, dew or saturating atmospheric humidity during night-time gas exchange will result in depleted leaf water as water vapour uptake and exchange dominate the leaf water isotopic budget (Helliker & Griffiths, 2007; Seibt et al., 2007; Cernusak et al., 2008; Reyes-García et al., 2008b; Helliker, 2011; Kim & Lee, 2011). By following leaf water 18O in CAM epiphytes, it is possible to model the relative humidity (RH) experienced by the plant during drought, or to determine a decoupling of the leaf water from the environment as stomata close preventing enrichment during drought or depletion under fog.

The present study set out to define the physiological niche space of four epiphytic Tillandsia species with contrasting lifeforms at Chamela, Mexico, an SDTF with a high frequency of dew and fog events during the prolonged dry season (Barradas & Glez-Medellín, 1999; Reyes-García & Griffiths, 2009). The epiphytic bromeliads of this community are vertically stratified, with both the deep (Tillandsia makoyana) and shallow (Tillandsia rothii) tank species found in the upper canopy. The atmospheric species are distributed either in the upper (Tillandsia eistetteri) or lower (Tillandsia intermedia) canopy (Reyes-García et al., 2008a). The atmospheric epiphytes also show contrasting morphologies: the upper canopy T. eistetteri shows reduced succulence and a nebulophytic lifeform, whereas T. intermedia shows high succulence.

In the current study, we investigate how the morphology and stratification of these four sympatric species within the SDTF can be related to physiological differences and niche differentiation through the reliance on alternative water sources (fog and dew). We set out to test the hypotheses that the two atmospheric epiphytes occupy contrasting niches, with the two species representing a trade-off between high leaf succulence and lower exposure (T. intermedia) and more frequent recharge by fog and dewfall consistent with nebulophyte traits (T. eistetteri). We also hypothesized that the water supply and physiological activity of the water-storing tank species would be more buffered against environmental extremes. Comparative field and experimental glasshouse studies were performed. In the field component, physiological performances were compared across contrasting wet and dry seasons, whereas control, drought, fogging and watering treatments were used in the glasshouse study. Relative water content (RWC), CAM activity (ΔH+), light use (photosystem II fluorescence) and leaf water isotope composition (δ18O) were followed throughout.

Materials and Methods

Study site and plant material

Field studies were conducted in an SDTF at the Biological Station of Chamela, near the Pacific Coast of Mexico, at 19°30′N, 105°03′W. The average annual rainfall is 746 mm (data from the meteorological station at the Biological Station), with 80% of these rain events occurring during the short wet season from July to October (Bullock, 1986). The mean annual minimum and maximum temperatures are 20 and 30°C, respectively. The RH at night is high throughout the year (77–92%), but drops at midday to as low as c. 30% during the dry season. The perennial bromeliads are under contrasting light environments during the wet and dry seasons because of the deciduous tree canopy (see Reyes-García et al., 2008a). During the months of January and February, there is a high frequency of dew and fog events, as well as small rain pulses, that are not recorded at the weather station, as they measure < 1 mm (Bullock, 1986; Barradas & Glez-Medellín, 1999; Reyes-García et al., 2008a).

The epiphytic Bromeliaceae species found at Chamela are described in detail in Reyes-García et al. (2008a) and are shown in Fig. 1. All species found at the site showed CAM with no significant daytime carbon uptake, even under well-watered conditions (C. Reyes-García and H. Griffiths, unpublished). We used four species that represented the variety of lifeforms found at the site, with tank and atmospheric epiphytes (sensuPittendrigh, 1948), as well as upper and lower canopy (Table 1).

Figure 1.

Contrasting Bromeliaceae lifeforms from the dry deciduous tropical forest. Tank lifeforms: (a) Tillandsia makoyana; (b) T. rothii. Atmospheric lifeforms: (c) T. eistetteri (nebulophyte); (d) T. intermedia. Bars, 5 cm.

Table 1.   Lifeform, distribution in the canopy and water use efficiency (WUE) of the Bromeliaceae species sampled in the dry forest of Chamela, Mexico
SpeciesLifeformStrata1WUELeaf form indexLeaf succulenceETRmax dryETRmax wet
  1. 1Reyes-García et al. (2008a). ETRmax, maximum electron transport rate. WUE (g CO2 (kg H2O) −1) was measured under glasshouse conditions. Leaf form index is the ratio of leaf length and width. Leaf succulence is expressed as g m−2, = 3. The values with different letters within each column are significantly different: < 0.05.

Tillandsia makoyanaTankUpper strata26.4a,b10.3 ± 2.0b852 ± 70a35 ± 7a117 ± 15a
Tillandsia rothiiTankUpper strata14.2b16.9 ± 1.1b570 ± 97a,b54 ± 6a99 ± 13a
Tillandsia eistetteriAtmosphericUpper strata20.6a,b90 ± 17.5a279 ± 15b26 ± 8a80 ± 10a
Tillandsia intermediaAtmosphericLower strata32.2a7.7 ± 0.2b634 ± 114a32 ± 8a45 ± 13b

Field physiological and morphological studies

Light conditions and physiological responses were characterized throughout the wet and dry seasons (years 2002–2003) in Chamela. Measurements included RWC to assess hydration status, nocturnal acid accumulation (ΔH+), which represents the amount of carbon assimilated for these obligate CAM species, light response curves to assess maximum electron transport rate (ETRmax) values and predawn maximum quantum yield of photosystem II (Fv/Fm) to assess photoinhibition. The light environment (photon flux density, PFD) was characterized by taking digital hemispherical pictures above each plant during the wet and dry seasons using a fish eye lens (Nikon Coolpix 950 and FC-E lens), as described in Reyes-García et al. (2008a). The final numbers of replicates per season were variable because of the timing of different collection campaigns: June, = 4 per species; September and January, = 6–10 per species. The RWC and ΔH+ plot (Fig. 3) includes all field (1999–2003) measurements (= 78–88 per species). For three-dimensional plots involving Fv/Fm, RWC and ΔH+, all field and glasshouse data collected were used (= 78–122 per species).

CAM activity was measured as ΔH+ by collecting leaf samples at dusk and dawn to quantify minimum and maximum acid contents. Samples were stored in 90% ethanol, boiled and titrated to neutrality, as described by Hartsock & Nobel (1976).

The RWC was defined as (fresh weight – dry weight)/(turgid weight – dry weight) × 100, where fresh weight is the weight of the leaf disc at the time of collection, turgid weight was measured after 24 h of rehydration in distilled water at room temperature and dry weight was measured after 48 h of oven drying at 70°C.

Light response curves and predawn measurements of Fv/Fm were made seasonally on four individuals using a portable modulated fluorometer (MINI-PAM; Walz, Effeltrich, Germany). ETRmax was recorded at 1500 μmol m−2 s−1, the highest light intensity provided. All calculations were performed after Maxwell & Johnson (2000). For ETR calculations, we assumed 83% absorption of incident photosynthetic photon flux density (PPFD). All measurements were made in fully expanded leaves of adult individuals.

Succulence was calculated as (fresh weight – dry weight)/area, using samples collected during the wet season of 2002 (= 3). The leaf form index was measured for a typical fully expanded leaf of three adults of each species, and represents the ratio of leaf length and width (Martorell & Ezcurra, 2007).

Glasshouse manipulations to quantify the impact of rain and fog

To quantify the effect of contrasting water sources (rain and fog) on the photosynthetic activity of the epiphytic Bromeliaceae species from the seasonally dry forest, water sources were manipulated in a glasshouse at the Botanic Garden at the University of Cambridge. At the start of the experiment, ten healthy adult plants of similar size per species were acclimatized for 1 yr to glasshouse conditions and were sprayed three times a day with rain water; the RWC of these plants varied from 75% to 80% (control treatment). Subsequently, the plants were subjected to progressive drought until c. 40% RWC was obtained (drought treatment), similar to dry season field values. This represented a 15-d drought for T. eistetteri and a 45-d drought for the rest of the species. Plants were then randomly divided into two groups (= 5 per group, per species) and rehydrated either by watering as before (three times a day; rewatering treatment) or by placement in a chamber with a humidifier (Conair Cool Mist Face Sauna Model 3704, Rantoul, Illinois; see Reyes-García et al., 2008b) that saturated the air with water under closed chamber conditions (fog treatment) for intermittent periods of 15 min during 4 h around dawn. The experiment took place from May to July 2004 under natural light conditions (16 h of sunlight per day) with 10 h of additional light through fluorescent lamps (total mean PPFD between 100 and 500 μmol m−2 s−1 during the day); the day temperature was varied from 20 to 30°C and the night temperature between 17 and 19°C; RH was 75%. Tillandsia eistetteri was kept in the glasshouse (same temperature and light conditions), but, under the control treatment, had to be placed in containers at 100% humidity to maintain viability.

The physiological responses measured were ΔH+, RWC, predawn Fv/Fm and light response curves using the same methods as described in the field measurements (= 4 per species). Instantaneous gas exchange water use efficiency (WUE) was assessed under well-watered conditions using an infrared gas analyser (CIRAS-1 gas-exchange system; PP systems, Hitchin, Hertfordshire, UK). Atmospheric species were measured inside a custom-built whole plant chamber which fitted three individuals simultaneously and was connected to an external pump that circulated air at 700 ml min−1. Measurements were made in the glasshouse with reference air coming from outside via a buffering volume; CO2 concentrations oscillated between 350 and 470 μmol l−1. Plants were kept in the chamber for 24 h and an average of overnight WUE was calculated. The measurements were performed during two 24-h cycles for each species. During the measurements, daytime CO2 uptake was not observed in any of the species.

The isotopic ratio of 18O in leaf water was followed throughout the experiment in order to obtain an indicator of water exchange with the environment through the open stomata. As water exchanges through stomata (g represents stomatal conductance), the isotopic ratio of leaf water is modified according to the ratio of ambient to intercellular vapour pressure (wa/wi), the fractionation factors αk and α*, the amount of water present in the leaf (W) and the original isotopic signature of the leaf (RL) and of the atmospheric water vapour (Ra). αk is the ratio of molecular diffusivity of H216O to H218O through air, the boundary layer and stomata (Merlivat, 1978; Farquhar et al., 1989; Flanagan et al., 1991; Farquhar & Lloyd, 1993; Cappa et al., 2003; Farquhar & Cernusak, 2005) and α* represents the equilibrium fractionation caused by the change in phase from liquid to vapour. Thus, a change through time (dt) in the isotopic signature of leaf water can be described by the formula (Helliker & Griffiths, 2007; Helliker, 2011):

image(Eqn 1)

This formula predicts depleted values under saturating humidity and enriched values under low RH.

Leaf samples for water 18O were collected, taking subsamples from the base, middle and tip of four individuals per species at the end of each treatment. The samples were stored in exetainers; water was extracted using vacuum cryogenic distillation (as described in Reyes-García et al., 2008b) and processed at Cambridge University using a duel-inlet mass spectrometer (VG SIRA 10; modified by Pro Vac Ltd, Crewe, Cheshire, UK). Water used in the humidifier to produce fog had the same isotopic signal as the water used to irrigate the plants (− 6.43 ± 0.28(SE)‰ vs Standard Mean Ocean Water (SMOW)). The humidifier produced no significant isotopic fractionation during 24 h of operation; the source water was freshly recharged every morning.

Statistical analyses

A two-way analysis of variance (ANOVA) was used to determine the relationship between the physiological measurements (RWC, ΔH+, ETRmax, Fv/Fm or δ18O) and both species and seasons/treatments. One-way ANOVAs were used to evaluate the effect of species on WUE, succulence and leaf form index. Tukey tests were carried out to discern specific differences among the data; significance was determined when < 0.05. Analyses were performed using the program Statistica v. 7.0 (Statsoft Inc., Tulsa, OK, USA).


Leaf morphology

The epiphytic bromeliad species sampled showed an array of leaf morphologies, from highly succulent leaves in the upper canopy tank T. makoyana (852 g m−2) to intermediate values for the shallow tank T. rothii and the lower canopy atmospheric T. intermedia (570 and 634 g m−2, respectively; Table 1). The upper canopy atmospheric T. eistetteri showed characteristic nebulophyte (fog-catching) morphology, with long thin leaves (high leaf form index of 90) and low leaf succulence (279 g m−2), both traits being significantly different from the other coexisting bromeliads (< 0.05).

Field seasonal physiological responses

The range of physiological responses to seasonal changes in water and light is shown in Fig. 2 for the four bromeliad species from Chamela. There was a general trend in the tank species T. makoyana and T. rothii to show stress during the late dry season in June, when PFD was high (mean of 44 mol m−2 d−1, < 0.05; Fig. 2a) and precipitation was absent. During this month, the tank species showed the lowest mean RWC recorded (38%, < 0.05; Fig. 2c), together with the lowest ΔH+ as an indicator of CAM activity (mean of 3 μmol H+ g−1 FW, < 0.05; Fig. 2e), implying virtually no net carbon uptake or recycling of internal CO2 in these obligate CAM species. In accordance, in June, the lowest mean value of Fv/Fm was found (0.5), indicative of photoinhibition (< 0.05; Fig. 2g). All of these physiological parameters recovered during the wet season. During September, RWC recovered to an average of 70% and Fv/Fm to 0.81 in the tank species, with no statistical differences between the values shown in January (> 0.05). In these species, ΔH+ showed mean maximum values during September and January, with a mean of 82 μmol H+ g−1 FW.

Figure 2.

Seasonal performance of bromeliad species at the dry forest of Chamela, Mexico (field study). Rows show photosynthetic flux density (PFD; a, b), relative water content (RWC; c, d), nocturnal acid accumulation (ΔH+; e, f) and maximum quantum efficiency of photosystem II (Fv/Fm; g, h) for the tank epiphytes Tillandsia makoyana (black bars) and T. rothii (dotted bars), and for the atmospheric epiphytes T. eistetteri (white bars) and T. intermedia (hatched bars). The dark bar on the x axis indicates the length of the wet season. Means and standard errors are shown.

Although both outer canopy tank bromeliad species showed similar physiological responses to the changing seasons, contrasting responses to seasonal progression were observed in the two atmospheric bromeliad species. The atmospheric T. intermedia, located in the lower strata of the canopy, received significantly less PFD on an annual basis (< 0.05; Fig. 2b), with the light environment not changing significantly from September to January. This lower PFD incidence resulted in no significant changes in Fv/Fm in T. intermedia during any of the sampled months, with an annual mean of 0.76 (P > 0.05; Fig. 2h). The difference with respect to the upper canopy species was also reflected in lower overall ΔH+ values and less of a decrease in RWC during the first half of the dry period (January), in comparison with the atmospheric T. eistetteri. The low canopy species, T. intermedia, showed lower ETRmax values (45 μeq m−2 s−1) when well watered relative to the upper canopy species (80–117 μeq m−2 s−1, < 0.05; Table 1). All species showed similarly low ETRmax values when under drought (Table 1).

The atmospheric nebulophyte T. eistetteri shared a similar microclimate to that of the upper canopy tank species (similar annual PFD fluctuations), yet the physiological and morphological differences resulted in different seasonal performances. RWC in T. eistetteri showed the highest sensitivity, desiccating at the beginning of the dry season (values not significantly different between early or late in the dry period, 44% in January, > 0.05; Fig. 2d). Despite this reduced RWC, the species showed mean maximum ΔH+ values during the month of January (mean of 103 μmol H+ g−1 FW), when precipitation was rare, but fog and dew formation were frequent. Tillandsia eistetteri also showed the lowest values of Fv/Fm among the species, with a minimum mean value during June of 0.39 and a maximum mean value of 0.67 from September to January (< 0.05; Fig. 2h).

The analysis of day–night titratable acidity (ΔH+) is an effective integrator of photosynthetic performance (Reyes-García et al., 2008a), and a comparison of the different lifeforms, as a function of changing RWC, displayed a clear progression in ΔH+ between tank and atmospheric species in relation to the level of hydration (Fig. 3). Both epiphytic tank species had maximum ΔH+ values at high water content, between 60% and 80% RWC for the deep tank species T. makoyana, and between 50% and 80% RWC for the shallow tank species T. rothii (Fig. 3a). By contrast, the atmospheric species had high ΔH+ values across a larger range of RWC values (Fig. 3b). The atmospheric T. eistetteri (upper canopy nebulophyte; Table 1) had the maximum ΔH+ value of 103.4 μmol H+ g−1 FW under the low RWC of 41–50% (Fig. 3b). Below these RWC thresholds, very modest values of ΔH+ are observed, most probably indicating a switch to CAM-idling activity. As the data presented in this graph comprise samples from different seasons, the high photosynthetic activity of T. eistetteri during the dry season is evident, as supported by data from the glasshouse studies reported below (Fig. 4).

Figure 3.

Relationship between relative water content (RWC) and nocturnal acid accumulation (ΔH+) in the bromeliad species at the seasonally dry forest of Chamela, Mexico (field study). The graphs are arranged by lifeform: (a) tank species: Tillandsia makoyana (black bars) and T. rothii (dotted bars); (b) atmospheric species: T. eistetteri (white bars) and T. intermedia (hatched bars). Means and standard errors are shown.

Figure 4.

Glasshouse responses to drought, fog and rewatering in the tank species Tillandsia makoyana (black bars) and T. rothii (dotted bars) and the atmospheric species T. eistetteri (white bars) and T. intermedia (hatched bars) from the seasonally dry forest of Chamela. The rows of graphs show: (a, b) relative water content (RWC; %); (c, d) nocturnal acid accumulation (ΔH+; μmol H+ g−1 FW); (e, f) Fv/Fm; (g, h) leaf water δ18O (‰). The bars show means and standard errors, = 4. Bars with different letters denote significant differences between treatments within each species (< 0.05). Plants were maintained in the UK.

Manipulations to simulate the impact of different water sources

A glasshouse experiment was performed in the UK to assess the effect of contrasting water sources (fog and liquid water) on the recovery from drought stress of the four epiphytic bromeliads (two tank and two atmospheric lifeforms). Fig. 4 shows the RWC, ΔH+, Fv/Fm and leaf water δ18O during the control, drought, fog and rewatering treatments. RWC fluctuated from means of 77% and 78% in all species for the control and rewatering treatments (similar to wet season data in the field; Fig. 2) to means of 41% and 43% under the drought treatment (Fig. 4a,b), showing no recovery in RWC when fog was applied as the only water source.

Although no recovery in RWC was registered after drought-treated plants were given 15 d of fog, positive metabolic responses were observed. There was a tendency to an increase in carbon assimilation (ΔH+; Fig. 4d) observed in the atmospheric nebulophyte T. eistetteri, showing mean values of 71 μmol H+ g−1 FW at the end of the drought treatment and 145 μmol H+ g−1 FW after the fog treatment (= 0.08); this ‘fogged’ value was similar to the control value of 174 μmol H+ g−1 FW. There was also a recovery in ETR and capacity for nonphotochemical quenching (NPQ) recorded in light response curves for T. eistetteri (data not shown, < 0.05). The other species showed no enhanced carbon assimilation (as ΔH+) in response to fogging. Tillandsia makoyana showed low values of Fv/Fm (0.54) at the end of the drought treatment, indicating photoinhibition, and there was a nonsignificant tendency to recover to a mean of 0.67 by the end of the fogging treatment (< 0.05; Fig. 4e). The recovery to ‘nonstressed’ values of Fv/Fm when fog was provided after drought was significant in T. eistetteri (Fig. 4f), whereas T. intermedia and T. rothii showed no photoinhibition throughout the experiment (Fig. 4e,f).

Leaf water δ18O was measured throughout the experiment in order to obtain a proxy for evaporative demand, and coupling to atmospheric conditions through open stomata. All tank species had similar values throughout the experiment (mean of 7.1 ± 0.6‰), irrespective of the treatment, consistent with the low CAM activities seen above (Fig. 4g). By contrast, atmospheric species showed significant shifts in δ18O throughout the experiment, the average values for both species being 0.1, 3.7 and 2‰ during the control, fog and rewatering treatments, respectively, consistent with isotopically depleted water vapour uptake and exchange at night (Fig. 4h). By contrast, highly enriched values were observed during drought (9.9‰), indicative of progressive enrichment as RWC declined (Fig. 4h).

Differences in instantaneous WUE were not observed between tank and atmospheric species, when measured on well-watered (control) plants under glasshouse conditions (Table 1). WUE showed high typical CAM values for all species, particularly for the atmospheric T. intermedia and the tank T. makoyana, with values of 32.2 and 26 g CO2 (kg H2O)−1, respectively (Table 1).


In accordance with our initial hypotheses, we found contrasting physiological responses between representative epiphytic species in the dry seasonal forest at Chamela, Mexico: of the two atmospheric species, T. eistetteri was more exposed to environmental fluctuations; both tank species were highly buffered against these fluctuations. To characterize the ‘physiological niche space’, it is important to understand the response of individual species to resource availability and natural stressors (Anthony & Connolly, 2004). Different physiological niches will be defined by boundaries along environmental gradients, and by responsiveness to changing environmental conditions. Having measured several morphological traits and physiological variables, both seasonally in the field and under different water treatments in the glasshouse, we can now relate the physiological sensitivity of individual species to their observed distribution and associated environmental gradients within the forest canopy.

For these bromeliad epiphytes, one major axis of niche differentiation is the difference between tank and atmospheric lifeforms. Tank species have the potential to prolong the access to water for hours or days after rain events, whereas atmospheric species only have immediate access to such a precipitation event (Rundel & Dillon, 1998; Zotz & Vera, 1999; Benzing, 2000), leading to more specialized use of precipitation pulses and fog inputs in the atmospheric bromeliads. In this study, tank species only showed nocturnal acid accumulation, indicative of carbon assimilation, when under high RWC conditions (Fig. 3a), consistent with stomatal closure to limit water loss (Schwinning & Ehleringer, 2001), as has been observed previously in tank bromeliads (Adams & Martin, 1986; Zotz & Andrade, 1998; Graham & Andrade, 2004). Leaf water δ18O compositions of the tank species were relatively constant, as water sources were manipulated, consistent with stomatal closure and tank water storage (Fig. 4g).

Under field conditions, the tank species concentrate photosynthetic activity during the wet season, when a high canopy position promotes high interception of precipitation (Rundel & Dillon, 1998; Reyes-García et al., 2008a). Tillandsia makoyana had the highest tank capacity and lowest range of RWC over which ΔH+ was maintained (Fig. 3a). Succulent tissues may be recharged through dew interception during the dry season, which is promoted by lower leaf temperatures relative to air and sympatric atmospherics (Andrade, 2003; Reyes-García et al., 2008a) and by large leaf blades with horizontal angles (Rundel & Dillon, 1998; Kidron, 2005; Reyes-García et al., 2008a). Changes in Fv/Fm values in T. makoyana under fogging in controlled conditions (Fig. 4e) were consistent with such reactivation in response to fog.

Without the water reservoir provided by a tank, atmospheric species benefit from the rapid absorption of water through the abundant trichomes, as well as from physiological traits to effectively use pulsed water sources. Both atmospheric species, T. intermedia and T. eistetteri, showed the ability to maintain nocturnal acid accumulation at low RWC values under field conditions (Fig. 3b). These species also showed a tight coupling with atmospheric conditions, water supply and leaf water δ18O under experimental and field conditions. Evaporative enrichment occurred under drought stress, and lower δ18O when exchange with isotopically depleted vapour took place (Fig. 4h; Helliker & Griffiths, 2007; Helliker, 2011). Contrasting strategies were observed among the atmospherics. The low canopy T. intermedia had very succulent leaves (Table 1) that buffered water fluctuations in between rain events, but maximum nocturnal acid accumulation occurred across a narrower range of RWC than found for T. eistetteri (Fig. 3). The species benefited from the low canopy position to avoid excess light (Fig. 2b) and photoinhibition (Fig. 2h).

The morphology and physiological responses of the atmospheric T. eistetteri were the most divergent of the four species. In contrast with expectations, this species, located in the upper, most exposed part of the canopy, had nonsucculent leaves, no tank and limited stomatal control. This species also showed rapid and fatal desiccation under nonsaturating night-time RH (75%) under glasshouse conditions in the UK (data not shown). Martorell & Ezcurra (2007) introduced the term nebulophyte to describe plants from various families that have long and thin leaves with small boundary layers specialized in intercepting fog. Nebulophytes tend to display leaves in the upper canopy (epiphytes or terrestrials with long stems) closer to wind currents, and leaf fluttering creates air turbulence, promoting the interception of fog droplets by the leaf blade. This morphology corresponds to that of T. eistetteri, with the high leaf index and exposed canopy position (Table 1). In this respect, succulence would need to be reduced to increase leaf movement under wind currents. A high metabolic response to small pulse precipitation and fogging would reinforce the nebulophyte strategy.

In Reyes-García & Griffiths (2009), responses to naturally occurring precipitation events of < 1 mm each, on three consecutive days, were evaluated for epiphytic bromeliads at Chamela. The nebulophyte T. eistetteri was the most responsive species, activating carbon assimilation after trace rain events, even when the RWC was low and not significantly affected by the pulses. This physiological response was corroborated under controlled glasshouse conditions when, after the drought treatment, only T. eistetteri showed an increase in nocturnal acid accumulation in response to fog, recovering almost 100% of maximum ΔH+ (Fig. 4d). This recovery was similar to that observed for all species when rehydrated with liquid water. Changes in RWC in T. eistetteri in response to fog were not found, supporting field observations that significant rehydration is not necessary to resume/maintain photosynthesis in this species. Thus, the annual carbon budget in T. eistetteri could be highly influenced by fog and dew incidence (Fig. 2d). The limited distribution of the species to the upper canopy, where night-time RH is higher relative to the lower strata (Barradas & Glez-Medellín, 1999), also helps to avoid desiccation. This is consistent with the model calculations of sustained exposure to a mean 90% RH environment, derived from field leaf water δ18O samples collected for this species (Reyes-García et al., 2008b). This high canopy position exposes the species to higher photoinhibition during drought (low Fv/Fm; Fig. 2h), yet we observed a rapid recovery of Fv/Fm values during the early wet season (data not shown) and under experimental conditions in both the fog and rewatering treatments (Fig. 4f).

The range of physiological responses under field and glasshouse manipulations is plotted in Fig. 5 to reveal the comparative physiological niche space for tank and atmospheric species. In this figure, we combine measured ΔH+, Fv/Fm and RWC (Fig. 5a–c), as well as leaf water δ18O, Fv/Fm and RWC (Fig. 5d–f), for the species. The figure allows a graphical representation of the variation of the physiological traits of each species. The low canopy atmospheric T. intermedia, growing in a more sheltered microenvironment when compared with the exposed upper canopy bromeliads, shows the smallest range of possible physiological responses (Fig. 5c,f). By contrast, the maintenance of photosynthetic activity across a wider range of conditions is revealed by the larger physiological niche space of the upper canopy nebulophyte T. eistetteri. This nebulophyte shows higher flexibility in its nocturnal acid accumulation, as it prolongs photosynthesis through most of the year: Fv/Fm values are highly variable as a result of the high PFD exposure and low water storage capacity (Fig. 5b); RWC and leaf water δ18O values are also variable, reflecting the low degree of stomatal control in T. eistetteri (Fig. 5e). Yet, all of these metabolic changes constitute the ‘normal’ physiological niche plasticity of the species. Tank species (Fig. 5a,d) show an intermediate range of physiological responses compared with the two atmospherics, their upper canopy position constituting a highly seasonal environment; yet, adaptations, such as the presence of the tank, tight stomatal control and high succulence, help to isolate the species from the changing conditions.

Figure 5.

Three-way physiological trait relationships in four epiphytic bromeliads from a seasonally dry forest. The panels show the relationships between relative water content (RWC, %), maximum quantum yield of photosystem II (Fv/Fm) and noctural acid accumulation (ΔH+, μmol H+ g−1 FW) for the tank species Tillandsia makoyana (closed circles) and T. rothii (open circles) (a) and for the atmospheric species T. eistetteri (closed squares) (b) and T. intermedia (open squares) (c). (d–f) The relationships between RWC, Fv/Fm and leaf water δ18O (‰) for the tank species (a) and the atmospherics T. eistetteri (e) and T. intermedia (f).

Having a nebulophytic lifeform is only one of the possible strategies to survive in an SDTF. In low-altitude SDTFs, where fog and dew events are infrequent, bromeliads can be present but do not show the nebulophyte lifeform (Orellana et al., 1999; Andrade, 2003; Ramírez-Morillo et al., 2004). Alternatively, nebulophytes of different families can be found at xeric sites where fog and dew can represent most of the annual water budget (Martorell & Ezcurra, 2002, 2007).

In conclusion, the epiphytic bromeliad community of Chamela exhibits an array of physiological strategies related to the morphology and microenvironment inhabited by the species. The ‘physiological niche space’ was broadest for the nebulophyte T. eistetteri, exposed in the upper canopy. The low degree of succulence in this species demanded regular recharge from fog and dew, but supported a much wider range of physiological responses throughout wet and dry seasons. By contrast, T. intermedia, a more succulent atmospheric restricted to the lower canopy, demonstrated a much narrower range of physiological tolerance. The other key species, T. makoyana and T. rothii, exhibited mechanisms that isolated them from the environment, such as water storage in tanks and succulent tissue, thick cuticles and high stomatal control. Furthermore, it would be expected that the nebulophytic species, being highly coupled to the atmospheric conditions, would be the most vulnerable under the current changing climate and habitat degradation.


This research was supported by the Consejo Nacional de Ciencia y Tecnología of Mexico (PhD scholarship 169 748), the Cambridge Overseas Trust and Cambridge Churchill College. We thank the staff of the Biological Station of Chamela, Mexico, for assistance and climate data, as well as Alex Goodall and Peter Michna, from the Botanical Gardens at Cambridge, who cared for the bromeliads. Thanks are also due to Marcia Carmona, Miguel Mejía, Zazil Reyes and Jaime Flores for field assistance, and to three anonymous referees who provided extremely helpful comments to improve the manuscript.