Introduction
The more than 3,000 species in the family Bromeliaceae can be found in diverse habitats (Benzing Reference Benzing2000), most prominently as epiphytes in many tropical and some temperate zone forests (Diaz & Alvarez Reference Diaz and Alvarez2009; Zotz & Schultz Reference Zotz and Schultz2008), but also terrestrially and rupicolously, for example, in tropical alpine vegetation up to 4000 m a.s.l. (Miller & Silander Reference Miller and Silander1991), on rocky outcrops in campo rupestre (Alves et al. Reference Alves, Kolbek and Becker2008) or in shrubby restinga vegetation (Cogliatti-Carvalho et al. Reference Cogliatti-Carvalho, Nunes-Freitas, Rocha and Van Sluys2001). Bromeliads can be also quite prominent in tropical coastal vegetation, for example, as epiphytes in mangroves (Díaz-Jiménez Reference Díaz-Jiménez2007; Zotz Reference Zotz2005), growing rupicolously at rocky shores (e.g., Pitcairnia or Dyckia,Grant Reference Grant1994; Lüttge et al. Reference Lüttge, Klauke, Griffiths, Smith and Stimmel1986; Pompelli et al. Reference Pompelli, Fernandes and Guerra2006), or in alluvial plains (e.g., Bromelia humilis, Medina et al. Reference Medina, Cram, Lee, Lüttge, Popp, Smith and Diaz1989). In the latter habitats, soil-rooted plants can get regularly into direct contact with seawater (Gómez & Winkler Reference Gómez and Winkler1991), or saline conditions may develop at least seasonally (Medina et al. Reference Medina, Cram, Lee, Lüttge, Popp, Smith and Diaz1989). Even epiphytically growing bromeliads may suffer from salt stress close to the coast when salt spray reaches these plants growing at some height in trees (Griffiths et al. Reference Griffiths, Smith, Lüttge, Popp, Cram, Diaz, Lee, Medina, Schäfer and Stimmel1989; Zotz & Reuter Reference Zotz and Reuter2009).
However, although there are quite a few bromeliad species that are potentially affected by salt stress in their natural environment, this aspect of their ecology has received little attention (Benzing Reference Benzing2000). Pitcairnia halophila L.B.Sm. is a particularly interesting member of the family in the context of salt tolerance, because the species is typically found on cliffs just above the high tide level (Fig. S1, Grant Reference Grant1994 and G. Zotz, pers. obs.). Given such a growing site, regular exposure to salt spray and even occasional inundation by salt water seem more than likely. There is some information on the physiology of congeneric Pitcairnia integrifolia, which is growing in a similar habitat in Trinidad (Lüttge et al. Reference Lüttge, Klauke, Griffiths, Smith and Stimmel1986). These researchers documented an osmotic pressure of only 0.9 MPa in leaves, which was interpreted as evidence against substantial accumulation of NaCl. In that study, it was also shown that the foliar trichomes were not functional in nutrient and water uptake, which allows the plants to avoid the foliar uptake of salt spray, although this does not explain why such uptake should not occur via roots.
The present study focuses on the effect of salt (mostly sodium chloride) on Pitcairnia halophila with a particular emphasis on the early phases of ontogeny that are known to be particularly sensitive to salt stress (Egan & Ungar Reference Egan and Ungar1999; Larcher Reference Larcher2003). Both germinating seeds and young plants of Pitcairnia halophila and three congenerics, which naturally grow in habitats without any input of NaCl, were subjected to different levels of salinity. We show that similar to most other plants found in saline environments (Ungar Reference Ungar1978), germination is quite susceptible to saline conditions via an osmotic effect, but vegetative growth of Pitcairnia halophila does not cease until salt concentrations in the soil exceed those equivalent to 50% seawater. To provide some ecological context for the results of these laboratory experiments, we also sampled mature Pitcairnia halophila individuals over an entire growing season. Remarkably, in spite of their growth in close proximity to the sea, these plants showed consistently low tissue concentrations of sodium.
Materials and methods
Habitat and study species
Pitcairnia halophila L.B.Sm. was originally described by Smith (Reference Smith1964). For a long-time considered endemic to Costa Rica (Grant Reference Grant1994), it has since been found in neighbouring Panama as well (Cáceres et al. Reference Cáceres, Schulte, Schmidt and Zizka2011). Restricted to a narrow margin along the Pacific coast, typically found on cliffs just above the high tide level, some exposure to salt spray or even occasional inundation by salt water seems a matter of fact and was clearly the reason for the choice of epithet (Grant Reference Grant1994). For this study, both seed material and entire plants were collected at a locality near David, Panama. The local climate is characterised by 1360 mm of annual precipitation with a pronounced dry season in the first 4 months of the year and average temperatures of 28 °C with little seasonal variation (data are from a station located c. 61 km from our field site, Hidrometeorología ETESA, https://www.hidromet.com.pa/, data of 54 years since 1967). After collection of mature capsules, seeds were kept frozen at −20°C at the University of Oldenburg until the start of the experiments. Plant collections to assess tissue concentrations were also done at the same site. Exact procedures are described below.
Pitcairnia heydlauffii R. Vasguez & P. L. lbisch has been described relatively recently from Bolivia (Vasquez & Ibisch Reference Vasquez and Ibisch2000). It occurs naturally in very wet forest on the foothills of the lower Andes. Seeds came from the Botanical Garden Berlin-Dahlem.
Pitcairnia schultzei Harms occurs in lowland forests in Panama, Colombia and Venezuela (Cáceres et al. Reference Cáceres, Schulte, Schmidt and Zizka2011; WCVP 2020). The seed source was identical to P. heydlauffii.
Pitcairnia suaveolens Lindl. is endemic to southeast Brazil, where it mostly grows as a lithophyte along streams in Atlantic forest in the state of Rio de Janeiro (Saraiva et al. Reference Saraiva, Mantovani and Forzza2015). Seeds were provided by the Blijdorp Zoo, Rotterdam, NL.
Germination experiments
After surface sterilisation following the procedure described by Pickens et al. (Reference Pickens, Affolter, Wetzstein and Wolf2003), seeds of all four species were sown in disposable Petri dishes (100 × 15 mm) with filter paper (Machery-Nagel, 651mm, Ø 83mm). Each Petri dish was filled with 2 ml of solutions with increasing salt concentrations, from 10 to 50% full seawater concentration. We prepared the solutions from a commercially available concentrate (Tropic Marin®, Dr. Biener Aquarientechnik, Lauterbach, Germany), which is manufactured from chemically pure salts. The stock solution (equivalent to 100% seawater) was dominated by NaCl (10.7 g l−1 Na+, 19.3 g l−1 Cl−) with all other elements (Ca2+, K+ and SO42−) and 70 trace minerals also in natural concentrations. Controls received distilled water. Each Petri dish contained 20 seeds, and each treatment level and the control were replicated 5 times. The Petri dishes, which were sealed with Parafilm® (Bemis, Neenah, Wisconsin, USA), were kept in climate cabinets (Economic Delux, Snijders Scientific, Tilburg, the Netherlands) with a light–dark period of 12/12 h (photon flux density c. 60 µmol m−2 s−1) and a temperature of 25 °C for 28 days. Weighing the dishes before and after that period indicated minimal water loss, and hence we expect hardly any increase in concentration during the course of the experiment. Germination (= appearance of a whitish radicle contrasting with the brownish testa) was assessed every other day with a dissecting scope (Zeiss 57 50 57, Jena, Germany).
Seawater can impact biological processes via an osmotic and/or an ionic effect. To study the osmotic effect alone, we conducted a second germination experiment with P. halophila, using mannitol to establish well-defined water potentials. Mannitol concentrations were varied between 0.04, 0.1, 0.2, 0.4 and 0.6 mol L−1 to generate substrate potentials of 0.1, 0.25, 0.50, 1.0 and 1.5 MPa (Swagel et al. Reference Swagel, Bernhard and Ellmore1997). Since the seawater treatments can be expressed as water potentials in a similar way (e.g., 100% seawater has a water potential of about −2.5 MPa, Tomlinson Reference Tomlinson1986), it is possible to compare the effect of seawater and mannitol on germination, that is, distinguish osmotic and ionic effects.
Growth experiments and nutrient analysis
Seeds were allowed to germinate as described above for the controls of the germination experiment. Once reaching a size of a few mm, juvenile plants were transferred into planting trays (35 × 40 cm) filled with 50/50 (vol %) unfertilised potting soil and sand and were kept in the green house. Ambient conditions were 25–28 °C air temperature, 60–80 % relative humidity and natural light. After reaching a size of c. 2 cm maximum leaf length (LL), plants were transferred to individual pots, which were filled with the same substrate, for at least 1 week before the beginning of experiments. Many ecophysiological responses depend on plant size (Zotz et al. Reference Zotz, Hietz and Schmidt2001). Therefore, in both Pitcairnia halophila and P. suaveolens, growth experiments were conducted with plants of different size (c. 2 cm and c. 11–15 cm LL). Pot size was 20 cm3 for small and 700 cm3 for large plants, respectively.
Growth experiments also lasted for 28 days. Sample size was 20 per treatment level for small (c. 2 cm LL) plants and 10 per level for large (c. 11–15 cm LL) plants. Plants were watered three times a week with different solutions (10–50 or 60% full seawater concentration and controls (= distilled water)). A diluted standard fertiliser was added to all treatments once a week (Flory 3: 15-10-15, Euflor, Munich, Germany), equivalent to 0.05 mg l−1 NO3-N, 0.1 mg NH4-N l−1, 0.1 mg P l−1 (phosphate) and 0.15 mg K l−1. Before the experiment, we had quantified the amount of water needed to saturate the soil in the pots at each watering. Watering each pot with about 10 times this quantity made sure that salt concentrations would not increase over time, and no crusts developed even in the treatment with the highest salt concentrations.
To estimate the initial size of all experimental plants, we took digital photographs from above. Similar photographs were taken from an additional 10 plants per species and experiment, which covered the entire size range of the experimental plants. These 10 plants were then harvested, and fresh weights of shoots and roots were determined. After 48 h at 80°C, dry mass was determined. The photographs were used to quantify the projected area of each plant with Image J (Rasband 1997-2015). The correlation between projected area and plant dry mass was invariably very tight (r > 0.95) and could be used to estimate the initial dry mass of all experimental plants. At the end of the experiment, all surviving plants were harvested, and fresh and dry weights were determined as described above. This plant material was then used to determine the foliar Na+ and K+ concentrations for each species, plant size class and treatment (4–6 replica). After digestion of the samples following standard procedures (Allen Reference Allen2008), tissue Na+ and K+ concentrations were determined with an atomic absorption spectrophotometer (AAS, SpectrAA 300, Varian Inc., Darmstadt, Germany) at standard operating conditions.
In situ growth and nutrient concentrations
Over the course of one growing season from April 2019 to January 2020, we collected five times three entire plants at a locality near David, Panama. Plants were immediately transported to the laboratory in the Chiriquí University Science Faculty (Universidad Autónoma de Chiriquí, David, Panamá), where fresh weights of leaves and shoots were determined. Then subsamples were taken, weighed, and dried at 80 °C for 48 h. The ratio of dry and fresh weight of the subsamples allowed the calculation of the dry weight of the entire plants. Subsamples were then transported to Germany, where tissue Na+ and K+ concentrations were determined with an AAS as described above.
Statistical analysis
All analyses were conducted with R version 4.0.2 (R Development Core Team 2019).
Germination
Germination success after 28 days (= % of germinated seeds) was used as response variable. The effect of seawater and mannitol on germination was assessed by calculating the proportional reduction in germination success at varying salt or mannitol concentrations as compared to controls in distilled water. This reduction was expressed as percentage. Since the different salt and mannitol concentrations are equivalent to defined water potentials, it was possible to directly compare their effects on germination. Without clear expectation on the specific relationship between water potentials and response variable, we simply chose the regressions that gave the best fit for each data set, that is, linear regressions, a polynomial and an exponential regression.
Growth
For each plant, the relative growth rate (RGR) in % day−1 was estimated following Hunt (Reference Hunt1982) as:
$${\rm{RGR}} = \left( {{\rm{ln}}\,{\rm{D}}{{\rm{W}}_{\rm{e}}} - {\rm{ln}}\,{\rm{D}}{{\rm{W}}_{\rm{b}}}} \right)/\Delta t,$$
that is, the difference between plant dry weight at the end and the beginning of the experiment divided by Δt, that is, the number of days. The relationship of RGR and different salt concentrations in the irrigation solutions was assessed with linear regressions.
Foliar nutrient concentrations
The relationship of foliar nutrient concentrations and the salt concentrations in the watering solutions was clearly linear in smaller and larger P. halophila and was therefore assessed with linear regression models. Due to a divergent, non-linear response of P. suaveolens to varying salt concentrations in the medium, the data of that species were analysed with two one-way ANOVAs in the case of Na+, while there was clearly no change in K+ concentrations with varying seawater concentrations.
Results
Germination
P. halophila germination of the controls was complete after c. 2 weeks (Figure 1). Maximum germination in the three congenerics was reached after a similar period, although in one species, P. suaveolens, only about a third of all seeds germinated. Both speed of germination and final germination success was a function of salt concentration in the medium, the negative effect varying with species (Figures 1 and 2). Most affected by salt was P. heydlauffii, which showed a substantial reduction in germination even at rather high water potentials. At 50% seawater (equivalent to −1.25 MPa), when germination completely ceased in this species, the average reduction reached c. 50% in P. suaveolens and P. schultzei and −84% in P. halophila (Figure 2). Expressed in terms of the resulting water potentials, salt or mannitol had very similar effects on germination in P. halophila.
Figure 1. Cumulative germination of seeds of Pitcairnia halophila, P. heydlauffii, P. schultzei and P. suaveolens. Data are means of 5 Petri dishes with 20 seeds each. Different symbols indicate the used salt concentrations: distilled water (Δ), 30% seawater (○) and 50 % seawater (Δ). Note that only selected treatment levels are shown for reason of clarity. The y-axes differ because absolute germination was low in Pitcairnia suaveolens even in the controls.
Figure 2. The impact of diluted seawater and mannitol solutions of comparable water potential on germination in Pitcairnia halophila, P. heydlauffii, P. schultzei and P. suaveolens expressed as the reduction of germinated seeds after 28 days compared to controls with distilled water. Data are means of 5 runs (seawater) and 3 runs (mannitol solutions) with 20 seeds each. The solid lines represent regression lines (seawater: power functions or linear regressions, p < 0.05) and the dashed line is a significant regression (mannitol). The explained variation (r2) of the relationships of seawater concentration and germination is indicated in each graph. Note that the y-axes vary. Arrows indicate the water potential leading to a 50% reduction in germination (see also Table 1). Note that seawater has a water potential of c. – 2.5 MPa.
Growth and survival
Under control conditions, small P. suaveolens seedlings grew more than twice as fast as P. halophila seedlings, while RGRs of larger individuals were comparable in the two species (Figure 3). Increasing salt concentrations in the watering solutions lead to a linear decrease in RGR, irrespective of species and plant size. In spite of this qualitative resemblance, the two species differed considerably in their overall response to increasing salt concentrations. For one, the linear reduction in RGR was steepest in small P. suaveolens plants (−0.17 [% d−1] [% seawater]−1) and smallest in larger P. halophila plants (−0.05 [% d−1] [% seawater]−1). Moreover, even the P. halophila individuals subjected to the highest concentrations looked entirely healthy after 28 days of treatment, whereas all P. suaveolens plants at concentrations ≥ 30% seawater developed brownish leaves. In the majority of cases, the number of necrotic leaves of plants receiving the highest salt concentrations (50 and 60%) exceeded that of healthy looking ones; a considerable number of individuals died, that is, the central part of the rosettes was rotten.
Figure 3. The relationship of relative growth rate and saltwater of varying concentrations in two Pitcairnia species. Data are means ± SD (n= 10 for larger and 20 for smaller individuals). Solid lines are significant linear regressions (p <0.001) with the following slopes: −0.05, −0.05, −0.17 and −0.08). Because of the high mortality at high salt concentrations, only values for ≤ 30% seawater were included in the regression analysis for larger P. suaveolens. Mortality rates are given as percentages.
Foliar nutrient concentrations
Under control conditions, average tissue Na+ concentrations [Na+] ranged from 0.32 to 0.38 mM gDW−1 (Figure 4a), while K+ concentrations [K+] ranged from 0.9 to 1.2 mM gDW−1 (Figure 4b), resulting in quite similar Na/K ratios of 0.26–0.44 in both species. Irrigation with different concentrations of seawater invariably led to an increase in [Na+], although the two species showed different trends: there was a tight linear relationship between % seawater and tissue [Na+] in P. halophila, while in P. suaveolens increasing concentrations in the irrigation solution of ≤ 30% led to a relatively steep increase in tissue [Na+] without any further change at higher concentrations of seawater. The response of [K+] was inconsistent in the four experiments with a slight decrease, an increase and twice no change (Figure 4b). In relative terms, the resulting changes in the Na/K ratio were strongest in 2-cm small P. halophila seedlings (Na/K ratio at 50% seawater = 2.03; + 500% compared to the control), in absolute terms the highest Na/K ratio was observed in P. suaveolens seedlings with Na/K > 3.1.
Figure 4. The relationship of foliar Na (a) and K (b) concentrations and salt water of varying concentrations in the rooting medium of two species of Pitcairnia. Data are means ± SD (n = 10 for P. halophila or fewer (2–9) in those P. suaveolens treatments with mortality, see Figure 3). Solid lines represent significant linear regressions (p <0.001), different small letters indicate significant differences between treatments (ANOVA, post hoc Tukey’s test) and dashed horizontal lines are grand means when groups did not differ.
Seasonal changes in above-ground biomass and tissue nutrient concentrations in situ
The natural growing site in Panama is characterised by a strong seasonality in precipitation and some variation in temperature (Figure 5). Plants bolted with the beginning of the rainy season and increased in biomass until the end of that season, while at the last collecting date in the early dry season in January 2020 seeds had already dispersed, and leaves and stalks had already started to desiccate. Tissue concentrations of sodium and potassium varied very little during the 9-month study period with average tissue [Na+] of c. 0.15 mM gDW−1 in leaves and c. 0.30 mM gDW−1 in stems. The concentrations of K+ averaged 0.5 mM gDW−1 in leaves and 0.3 mM gDW−1 in stems. The Na/K ratio showed no clear seasonal variation neither in stems nor leaves but was consistently higher in the former (1.0 ± 0.6 vs. 0.4 ± 0.2, means ± SD).
Figure 5. Seasonal changes in air temperature (°C, monthly averages ± SD), in rainfall (monthly totals in mm), in leaf and stem concentrations of Na and K (means ± SD, n = 3), and in leaf, stem and total plant biomass (means ± 1SD, n = 3) at a natural growing site in Panama. The year 2019 was relatively dry with 945 mm, which amounts to c. 70% of the long-term average.
Discussion
There is no universally accepted definition of ‘halophyte’ (Flowers & Colmer Reference Flowers and Colmer2008), and any attempt to categorise plants in a dichotomous way will result in legitimate criticism. Nevertheless, it is often inevitable to conceptualise potentially continuous biological variation as categories, compare, for example, Bennett et al. (Reference Bennett, Flowers and Bromham2013) for halophytic versus glycophytic Poaceae, Smith and Winter (Reference Smith, Winter, Winter and Smith1996) for CAM versus C3 plants or Zotz et al. (Reference Zotz, Weigelt, Kessler, Kreft and Taylor2021) for epiphytes versus terrestrially rooted plants. Thus, acknowledging this caveat and embracing the definition proposed by Flowers and Colmer (Reference Flowers and Colmer2008): ‘plants with the ability to complete the life cycle in a salt concentration of at least 200 mM NaCl (= 1/3 seawater) under conditions similar to those that might be encountered in the natural environment’, the results of this study provide some arguments that P. halophila may be categorised as halophyte, although we note that we have only studied early-life stages.
The last statement is seemingly at odds with the results of the germination experiment because germination of P. halophila was strongly affected by relatively low levels of salinity, and half-strength seawater almost completely inhibited germination (Figures 1 and 2). However, a majority of halophytes germinate better in the absence of salt, and a compilation of NaCl concentrations inhibiting germination in a range of beach, cliff and foredune species by Baskin and Baskin (Reference Baskin and Baskin2001) lists quantitatively comparable inhibitory effects to those observed in P. halophila. Our comparison of germination success in salt water versus mannitol solutions (Figure 2) indicates that this inhibition in the tested range of concentrations is not related to ionic toxicity but simply reflects an osmotic effect (Lambers et al. Reference Lambers, Chapin and Pons2008). The severity of the observed inhibition can also be compared with the available data for 14 species of bromeliads (Table 1). The water potential leading to 50% reduction in germination success (Ψ−50%) in P. halophila is comparable to Ψ−50% of three congenerics (P. schultzei and P. suaveolens, this study, and P. flammea, Pereira et al. Reference Pereira, Andrade, Pereira, Forzza and Rodrigues2009) and Ψ−50% of some species in other genera (Vriesea and Alcantarea). In most tested bromeliads, germination is more severely affected by low water potentials.
Table 1. Effect of the water potential (Ψ) of test solutions on the germination success of bromeliad seeds. Given is the Ψ that led to a 50% reduction in germination compared to controls (distilled water). Values were obtained from the figures or tables in the cited publications.
1 Range of four germination experiments at different temperatures.
2 Range of three different populations.
While P. halophila was not particularly tolerant to lower water potentials during germination, plant growth was much less affected by low Ψ in this species than in the studied congeneric (Figure 3). At the cut-off defined by Flowers and Colmer (Reference Flowers and Colmer2008), the RGRs still reached about 30% of the maximum value. Moreover and in contrast to P. suaveolens plantlets, which showed visual signs of damage and quite high foliar Na+ concentrations (Figure 4), P. halophila looked invariably healthy during the entire course of the experiment and was capable of maintaining relatively low Na+ concentrations in their leaves, particular in larger plants. Notably, growth was low in P. halophila even under control conditions (Figure 3). Such a low growth potential could make this species a poor competitor, which is possibly one reason why it is restricted to rock outcrops and is not found in adjacent vegetation.
The seeds of P. halophila are released with the beginning of the dry season (pers. obs.). Under natural conditions, germination will probably not occur until the onset of the rainy season, which should assure relatively low salinity in soil during germination and initial establishment. Thus, the fact that germination of this species was rather susceptible to lower water potentials may not be very relevant ecologically.
The field observations (Figure 5) also suggest that the demonstrated ability of P. halophila to grow under somewhat saline conditions may actually not be tested in situ even at a growing site in immediate vicinity to the sea. The tissue concentrations of sodium in both leaves and stems were consistently low and comparable to those of the controls in the greenhouse experiment. Given that plants rest during the dry season, the frequent and substantial precipitation during the rainy season when plants actively grow seems to dilute any input from salt spray to a degree that the plants do apparently not have to deal with elevated soil salinity in situ.
In conclusion, we present the results of an experimental/observational study that explored the physiological and ecological basis of the occurrence of P. halophila in the spray zone of cliffs in Central America. Our study gives a mixed picture: while the species is quite susceptible to moderate salinity during germination, growth of plantlets was still possible when watering solutions reached half-strength seawater, which is typically taken as a characteristic of halophytes. However, in situ observations suggest that the conditions at the natural growing sites with abundant precipitation during the growing season are not saline, and thus, ecologically, P. halophila does not face the challenges of a halophyte.
Supplementary material
To view supplementary material for this article, please visit https://doi.org/10.1017/S0266467422000116
Acknowledgements
Assistance with the measurements by Lars Vormstein, Ingeborg Eden, Silvia Kempen, Norbert Wagner and Brigitte Rieger is gratefully acknowledged. Seed material was obtained from natural populations in Panama (P. halophila – SEX/P-42-11) and from cultivated plants: P. suaveolens from Blijdorp Zoo, Rotterdam, NL and P. schultzei and P. heydlauffii from the Botanical Garden Berlin-Dahlem. We also acknowledge the permission to collect and export plant material of mature plants of P. halophila – (PA-01-ARB-008-2021), and field assistance by Diana Gómez González, Eyvar Rodríguez-Quiel and Zabdy Samudio (Universidad Autónoma de Chiriquí, Panama).
Financial support
This research received no specific grant from any funding agency, commercial or not-for-profit sectors.
Competing interests
The author(s) declare none.
Open access
