Ecological and evolutionary consequences of desiccation tolerance in tropical fern gametophytes


Author for correspondence: James E. Watkins Tel: +1 617 4963580 Fax: +1 617 4965854 Email:


  • • Ferns have radiated into the same diverse environments as spermatophytes, and have done so with an independent gametophyte that is not protected by the parent plant. The degree and extent of desiccation tolerance (DT) in the gametophytes of tropical fern species was assessed to understand mechanisms that have allowed ferns to radiate into a diversity of habitats.
  • • Species from several functional groups were subjected to a series of desiccation events, including varying degrees of intensity and multiple desiccation cycles. Measurements of chlorophyll fluorescence were used to assess recovery ability and compared with species ecology and gametophyte morphology.
  • • It is shown that vegetative DT (rare in vascular plants) is widely exhibited in fern gametophytes and the degree of tolerance is linked to species habitat preference. It is proposed that gametophyte morphology influences water-holding capacity, a novel mechanism that may help to explain how ferns have radiated into drought-prone habitats.
  • • Fern gametophytes have often been portrayed as extreme mesophytes with little tolerance for desiccation. The discovery of DT in gametophytes holds potential for improving our understanding of both the controls on fern species distribution and their evolution. It also advances a new system with which to study the evolution of DT in vascular plants.


There is overwhelming evidence to show that land plants evolved from simple aquatic algal ancestors (Bold, 1957; Niklas, 1997). The radiation of once aquatic plants on to dry land required the evolution of adaptive character suites that permitted life in what was and remains deadly dry air. To live in this environment, plants have evolved essentially two mechanisms of surviving desiccating conditions. One mechanism is the avoidance of desiccation. Desiccation avoidance is demonstrated in most modern terrestrial vascular plants and has been accomplished by the development of internal water conductance and characters such as highly organized cuticles and effective stomatal control. Cacti perhaps represent the embodiment of avoidance with their succulent water-storing stems sheathed in a thick and well developed cuticle. These characters and many others associated with desiccation avoidance are all thought to be derived and early land plants most certainly did not have such structures. Early plants relied on the more radical mechanism of survival from desiccation: that of desiccation tolerance (DT).

Desiccation-tolerant plants are defined as those that can lose from their vegetative structures all internal water and enter into, and recover from, anhydrobiosis, the cessation of metabolic activity as a result of low intracellular water content (Bewley, 1979). An environment that brings a plant to an air-dried anhydrobiotic state is sufficiently lethal to kill all modern agricultural crops and > 99% of all vascular land plants (Alpert, 2000; Alpert & Oliver, 2002). The rarity of DT in vascular plants is curious given that many of the characters that facilitated desiccation tolerance in ancestral land plants are still commonly found in modern algae and bryophytes (Oliver et al., 2000; Porembski & Barthlott, 2000; Proctor & Tuba, 2002; Alpert, 2005). There is no shortage of studies demonstrating the ability of bryophytes to recover from an inactive and fully air-dried state. Fantastic stories exist in the literature of some bryophytes recovering from over 20 yr of desiccation in herbaria (Alpert, 2000). Such remarkable abilities clearly have ecological consequences and many studies have shown that more desiccation-tolerant bryophytes are often associated with xeric habitats (Bowker et al., 2000; Stark et al., 2005). The ecological consequences of desiccation tolerance exist even within bryophyte species, with one example being the overrepresentation of female gametophytes in dioecious bryophytes of xeric habitats. Such sex-based disparity is often related to greater DT in females relative to males (Stark et al., 2005).

Desiccation tolerance can exist in one stage but be completely absent in another stage of an organism's life cycle. For example, in animals where DT is even rarer than in plants, DT occurs in different generations of the fly Polypedilum vanderplanki, which exhibits tolerance in the larval stage, but not in the adult stage (Watanabe et al., 2002). Such life cycle-mediated tolerance also occurs in plants. The gametophytes of some bryophytes exhibit a much greater degree of tolerance than sporophtyes. Desiccation tolerance requires the complex and organized shut-down of metabolism (Alpert & Oliver, 2002) and the occurrence of DT in many distantly related lineages and life stages indicates that there may be significant variation in the mechanisms behind this phenomenon.

Within the vascular plants, DT of spores and seeds is well known, but much less is known about vegetative DT in lineages with two separate free-living stages. The largest and most diverse lineage of vascular plants to exhibit two separate free-living generations is the ferns. The ferns alternate between independent gametophyte and sporophyte generations. Of particular interest are the morphological and physiological differences between these two stages. The gametophyte is the site of fertilization, is relatively small, lacks vascular tissue, and either completely lacks or has a poorly developed cuticle. The sporophyte, the primary stage for dispersal, has a well developed vascular system and a waxy cuticle complete with stomata. These differences alone result in unique life-cycle-mediated ecological strategies, especially as they relate to water relations and demography (Watkins et al., 2007).

Because of the well known resurrection fern, the presence of DT in the vascular sporophytes of ferns has been reported to be more common than in other vascular plants. However, in a recent review on the subject, Proctor & Pence (2002) recorded that 64 species of ferns exhibited DT and estimated that fewer than 1% of all ferns possess such ability. Of the 64 ferns listed in this review, 40 were Cheilanthioid taxa that are commonly associated with desert-like habitats. Much less is known of species from tropical sites, but DT has been recorded in genera as phylogenetically disparate as Asplenium and Polypodium (Kappen, 1964; Gaff, 1987; Proctor & Pence, 2002). As with other lineages of vascular plants, DT in the sporophtyes of ferns remains a rare phenomenon and this is unlikely to change with increased sample size.

Studies on DT of the gametophyte generation of ferns are fewer in number. Some of the earliest comments were made by Goebel (1900) regarding the ability of the buried tubercles of Annogramme chaerophylla to resume growth following dry spells. A similar observation was made by Campbell (1904) on the surface-growing gametophytes of Gymnogramme triangularis that appeared to have survived a dry California summer. The first experimental evidence for DT was generated by Pickett (1913, 1914, 1931), who, through a series of excellently conceived desiccation experiments, was the first to clearly show that the gametophytes of Asplenium rhizophyllum and A. platyneuron could recover growth following extreme desiccation. He also found a greater degree of tolerance in A. rhizophyllum, a species of more exposed and drier habitats, than in A. platyneuron, a species often confined to more mesic sites. This was the first link of DT in the gametophyte generation with sporophyte distributions and species ecology. Unfortunately, Pickett's work was the last of its kind. There have been anecdotal published reports (Gilbert, 1970), observations on the ability of the gametophytes of Pyrossia pilosellodes to recover from drought (Ong & Ng, 1998), and unpublished field observations of the frequent occurrence of thriving gametophytic populations of Trichomaes and Vittaria in desiccated states (D. Farrar, pers. comm.) Yet little is known of the degree to which fern gametophytes can tolerate desiccation or their rates of recovery.

The goal of this paper is to survey a range of tropical fern gametophytes to determine the extent of DT in this phase of the fern life cycle. We also examine the ability of species to recover from a single desiccation event and from repeated desiccation cycles and desiccation intensities and relate this attribute to species ecological occurrences.

Materials and Methods

Spore material and growth conditions

Spore material from 12 species of differing ecology were collected at La Selva Biological Station in the Atlantic lowlands of north-eastern Costa Rica at 37–100 m asl. Fertile fronds were gathered in the field and placed into glassine envelopes with tape-sealed seams. Envelopes with plant material inside were stored in an air-conditioned laboratory and allowed to dry and release spores within these envelopes. Spores were taken back to the University of Florida where they were sown onto culture media. The growth of a broad sampling of species with different ecologies required extensive experimentation with culture techniques. All species were found to grow adequately on a combination of organic soil collected from canopy trees at La Selva mixed with a small amount of vermiculite. Spores were sown in this media into 60 mm × 15 mm Fisherbrand Petri dishes. These plates were stored in sealed clear plastic containers (Pioneer Plastics Model 395-c, Dixon, KY, USA). Cultures were exposed to 50 µmol m−2 s−1 for 10 h d−1 from GE fluorescent plant and aquarium 40 watt growbulbs and watered with deionized water every 20–30 d. When water was added to the Petri dished it was always added to the soil media and not on the surface of the gametophytes. This culture technique assured that gametophytes were maintained at 98–100% relative humidity (RH) and had little external water. In order to calculate relative water content (RWC), the saturated water content was determined by immediately weighing the thallus upon removal from the culture chamber. Initial experiments with centrifugation to remove external water and blotting with kimwipes indicated that this method was effective in limiting accumulation of external water on the thallus.

Desiccation experiments

For the initial survey experiment, five to 10 mature gametophytes (one gametophyte per Petri plate) of all 13 species were allowed to desiccate at 50% RH (Ψ = –94.57 MPa) in a humidity-controlled chamber that was constructed using one of the plastic growth boxes connected via Bevline tubing to a Licor dew point generator (Model 610, Lincoln, NE, USA) set to a flow rate of 0.5 l min−1. Samples were allowed to dry for 45 min and were removed from the chamber every 5 min and placed in a Sartiorus microbalance (Göttingen, Germany) where wet weight and a measurement of Fv/Fm was taken (discussed later). The samples were then placed back into the chamber. A Hobo Pro RH/Temp Data Logger (Bourne, MA, USA) was used to verify that the chamber typically regained the designated RH within 2 min of the top being replaced. To evaluate recovery upon completion of the desiccation treatment, samples were rehydrated with deionized water and measurements of Fv/Fm were again made at 5 min, 24 h, and 48 h post-rehydration. Samples were then dried for 72 h in a drying oven at 70°C to determine gametophyte dry weight. Relative water content was plotted against time and Fv/Fm.

A second desiccation experiment was designed to test the effect of drying intensity on recovery of Fv/Fm. The gametophytes of Diplazium striatastrum, Phlebodium pseudoaureum, and Microgramma reptans were chosen to represent the extremes of tolerance in the initial desiccation experiment and were dried at three different intensities, 20% RH (Ψ = –219.59 MPa), 50% RH (Ψ = –94.57 MPa) and 80% RH (Ψ = –30.44 MPa), following the methods described earlier. Gametophytes were kept at these intensities for 48 h, after which time they were rehydrated with deionized water and measurements of Fv/Fm were taken at 24, 48, and 72 h post-rehydration. These values were related to the dark-adapted value of Fv/Fm to determine the mean percentage recovery.

A third experiment was conducted to examine the influence of consecutive desiccation treatments on photochemical efficiency. Gametophytes of six species were chosen from the survey experiment to represent different recovery abilities. Thirty gametophytes of each species were selected and 10 each were dehydrated for one, two, or three cycles at 50% RH (Ψ = –94.57 MPa). Each cycle lasted for 48 h, after which time gametophytes were then rehydrated with deionized water and measurements of Fv/Fm were again made at 24, 48, and 72 h post-rehydration. In each case, 48 h was sufficient to allow material to come into equilibrium with the surrounding air. These values were related to the dark-adapted value of Fv/Fm to determine the mean percentage recovery.

Chlorophyll fluorescence measurements

Variation in photochemical efficiency (Fv/Fm) was measured as the desiccation-dependent change in the ratio of variable and maximal fluorescence Fv/Fm, where Fv is the difference between the maximum (Fm) and the minimum (Fo) fluorescence emissions measured with an Opti-Sciences pulse modulated fluorometer (Model OS-500, Hudson NH, USA). Minimal fluorescence was measured under a weak pulse of modulating light over 0.8 s, and maximal fluorescence was induced by a saturating pulse of light (8000 µmol m−2 s−1) applied over 0.8 s. The parameter Fv/Fm was first measured after 20 min dark adaptation, and this measurement was taken as the index of recovery. Dark-adapted Fv/Fm provides an estimate of the maximal quantum efficiency of photosystem II, which in unstressed material is generally c. 0.76–0.83.

Statistical analysis

For the initial desiccation survey, a series of regressions were run on arcsin square-root-transformed RWC data for each individual within a species to determine the rate of drying of gametophytes exposed to 50% RH (Ψ = –94.57MPa) over the 45 min time interval. The slopes of these regression lines were calculated to generate a species mean drying rate. These rates were then analyzed by a one-way anova followed by a post-hoc Tukey's test to determine differences between species. Linear regression analysis was used to assess the influence of both final RWC at 45 min and the slope of the individual drying curves on species recovery ability at 48 h post-rehydration. Percentage species recovery data were also arcsin square-root-transformed for these analyses. The mean species drying rates expressed as RWC and absolute water content (AWC) and the final water content reached at 45 min were then plotted against each species’ mean percentage recovery at 48 h. Depression in photochemical efficiency was also graphed as a function of thallus RWC (Fig. 1).

Figure 1.

Proportional recovery of the pre-treatment dark-adapted value of Fv/Fm, 24 h post-treatment and rate of thallus water loss expressed as relative water content (RWC) ((g fresh weight – g dry weight)/(g saturated weight – g dry weight)) × 100. Gametophytes were held at a relative humidity of 50% (Ø = –94.57 Mpa) for 24 h and then rehydrated and allowed to recover for 24 h before measurements. Species abbreviations: AdiLat, Adiantum latifolium Lam.; CycSem, Cyclopeltis semicordata (Sw.) J. Sm.; DenBip, Dennstaedtia bipinnata (Cav.) Maxon; DipStr, Diplazium striatastrum H. Christ; MicRep, Microgramma reptans (Cav.) A.R. Sm.; NepBis, Nephrolepis biserrata (Sw.) Schott; PhePse, Phlebodium pseudoaureum (Cav.) Lellinger; PitCal, Pityrogramma calomelanos (L) Link; PteAlt, Pteris altissima Poir.; TheBal, Thelypteris balbisii (Spreng.) Ching; TheCur, Thelypteris curta (H. Christ) C. F. Reed; TheNic, Thelypteris nicaraguensis (E. Fourn.) C.V. Morton.

To assess the influence of consecutive desiccation cycles (Expt 3) and desiccation intensity (i.e. final water potential, Expt 2) on photochemical efficiency, a repeated-measures anova was performed with number of desiccation cycles or water potential and recovery time as the fixed main effects. Data were first examined for sphericity following the Mauchly criterion. Pairwise comparisons were made across recovery times with Bonferroni-adjusted multiple t-tests. Klockars & Sax (1986) recommend using the more stringent Bonferroni-adjusted multiple t-test when the number of planned comparisons is greater than the number of degrees of freedom for between-groups. In cases where data did not meet the sphericity criterion, P-values were adjusted using both Greenhouse–Geisser and Huynh–Feldt methods based on the respective epsilons.



For the initial desiccation survey, a series of regressions were run on each species to determine the change in RWC of gametophytes exposed to 50% RH (Ψ = −94.57 MPa) over the 45 min time interval. All species exhibited rapid rates of thallus water loss. In all species, regressions on the arcsin square-root-transformed data were linear and linear regressions were used to calculate the slopes of species’ drying curves to determine desiccation rates. These rates varied significantly among species, with the fastest dry down in the terrestrial Thelypteris curta and slowest rates in the terrestrial Cyclopeltis semicordata and the epiphyte Microgramma reptans (Table 1, Fig. 1).

Table 1.  Species, life form, and desiccation rates expressed as relative water content (RWC) and absolute water content (AWC)
SpeciesLife formChange in RWC min−1STEChange in AWC min−1STE
  • RWC = ((g fresh weight – g dry weight)/(g saturated weight – g dry weight)) × 100; AWC = (g wet weight/g dry weight)); STE, standard error.

  • Relative water content is useful in understanding the amount of water present at a given time relative to the saturated or initial water content of the tissue. Absolute water content expresses the total amount of water present at any given time.

  • 1

    [Correction added after online publication 26 September 2007: Terrestrial was corrected to Epiphyte]

  • 2

    [Correction added after online publication 26 September 2007: Epiphyte was corrected to Terrestrial]

Thelypteris curta (H. Christ) C. F. ReedTerrestrial–2.1990.061–0.1040.012
Thelypteris nicaraguensis (E. Fourn.) C. V. MortonTerrestrial–2.0330.059–0.1890.019
Pityrogramma calomelanos (L) LinkTerrestrial–2.0170.081–0.3650.026
Adiantum latifolium Lam.Terrestrial–1.8910.101–0.1400.030
Dennstaedtia bipinnata (Cav.) MaxonTerrestrial–1.6700.045–0.1380.010
Pteris altissima Poir.Terrestrial–1.6110.090–0.4250.050
Diplazium striatastrum LellingerTerrestrial–1.5120.084–0.1310.012
Nephrolepis biserrata (Sw.) SchottTerrestrial–1.4900.120–0.1330.022
Phlebodium pseudoaureum (Cav.) LellingerEpiphyte1–1.4630.116–0.2330.065
Thelypteris balbisii (Spreng.) ChingTerrestrial–1.4190.094–0.1080.013
Cyclopeltis semicordata (Sw.) J. Sm.Terrestrial2–1.3590.074–0.1700.022
Microgramma reptans Sw.Epiphyte–1.2790.203–0.2550.020

Depression in photochemical efficiency (Fv/Fm) as gametophytes desiccated was nonlinear and varied among species. Species also exhibited differential abilities to recover following desiccation. This recovery ability was more closely related to the rate at which gametophytes dried, expressed as RWC (r2 = 0.555, P = 0.0001; Fig. 1), when compared with desiccation rate expressed as AWC (r2 = 0.0187, P = 0.671), the final RWC reached (r2 = 0.193, P = 0.0008), and the final wet mass reached (r2 = 0.001, P = 0.932) after 45 min of drying.

Desiccation intensity

The water potential of the different desiccation treatments significantly influenced the recovery abilities of both Diplazium striatastrum and Phlebodium pseudoaureum, but had little influence on Microgramma reptans (Table 2, Fig. 2). For the understory terrestrial D. striatastrum, the ability to recover following the –219.56 and –94.57 MPa treatments was essentially nonexistent. The Fv/Fm values reached at these water potentials are suggestive of significant photoinhibition and photodamage. The –30.44 MPa treatment also depressed Fv/Fm but to a lesser degree and gametophytes exposed to this treatment exhibited clear recovery following rehydration. Phlebodium pseudoarueum exhibited relatively less depression in Fv/Fm. At all three desiccation intensity treatments, gametophytes exhibited recovery albeit with lower rates from the –219.56 and –94.57 MPa treatments. Microgramma reptans exhibited remarkable Fv/Fm stability at all three intensities. No significant Fv/Fm depression occurred at any of the desiccation intensities (Table 2, Fig. 2).

Table 2. Fv/Fm recovery results from the repeated-measures anova for gametophytes exposed to three different desiccation intensities: 20% relative humidity (RH) (Ψ = –30.44), 50% RH (Ψ = –94.57) and 80% RH (Ψ = –219.56)
  1. Ψ, water potential; ID(Ψ), individual effect.

  2. The gametophytes of Diplazium striatastrum are often found in the understory, whereas those of Phlebodium pseudoaureum, and Microgramma reptans occur in the midcanopy and exposed canopy, respectively. Gametophytes were kept at these humidity levels for 48 h, after which time they were rehydrated with deionized water and measurements of Fv/Fm were taken at 24, 48, and 72 h post-rehydration. These values were related to the dark-adapted value of Fv/Fm to determine the mean percentage recovery.

Diplazium striatastrum Lellinger
Ψ 238.41< 0.00010.7393.240.662
ID(Ψ)12 1.86  0.0741   
Recovery time 2 7.32  0.0033   
Recovery time × RH 4 6.39  0.0012   
Phlebodium pseudoaureum (Cav.) Lellinger
Ψ 242.19< 0.00010.7163.590.61
ID(Ψ)12 2.2  0.0482   
Recovery time 224.31< 0.0001   
Recovery time × RH 4 7.81  0.0003   
Microgramma reptans Sw.
Ψ 2 1.46  0.27160.7792.670.75
ID(Ψ)12 4.53  0.3531   
Recovery time 2 0.6  0.6352   
Recovery time × RH 4 5.09  0.0041   
Figure 2.

Fv/Fm recovery graphs for gametophytes exposed to three different desiccation intensities: 20% relative humidity (RH) (Ψ = –30.44), 50% RH (Ψ = –94.57) and 80% RH (Ψ = –219.56). (a) Diplazium striatastrum; (b) Phlebodium pseudoaureum; (c) Microgramma reptans. Gametophytes were kept at the humidity levels for 48 h, after which time they were rehydrated with deionized water and measurements of Fv/Fm were taken at 24, 48, and 72 h post-rehydration. Pairwise comparisons were made across recovery times with Bonferroni-adjusted multiple t-tests. Different letters indicate significant differences vertically within a given recovery time. The gametophytes of Diplazium striatastrum (a) are often found in the understory, whereas those of Phlebodium pseudoaureum (b), and Microgramma reptans (c) were collected in the mid-canopy and exposed canopy, respectively.

Desiccation cycles

Six species representing different life histories and desiccation tolerance from the initial survey were chosen and exposed to multiple desiccation cycles (one, two, or three) (Table 3, Fig. 3). In all cases excluding Microgramma reptans, percentage recovery was greater following one vs two or three desiccation cycles. Recovery ability was closely linked to species ecology with slow to no recovery in the understory species Diplazium striatastrum, Adiantum latifolium, and Cyclopeltis semicordata. There was a much greater degree of recovery following two and three cycles in Pityrogramma calomelanos, and the two epiphytes. Unlike the other species, Microgramma reptans exhibited no significant difference in recovery between one and two cycles but experienced a much slower recovery following the third desiccation cycle (Table 3, Fig. 3).

Table 3. Fv/Fm recovery results from the repeated-measures anova for gametophytes exposed to one, two, or three desiccation cycles at 50% relative humidity (Ψ = –94.57)
 d.f.FPMauchlyX2Adjusted PAdjusted P
  1. ID(Trt), individual effect.

  2. Gametophytes were kept at this level for 48 h. Material was then rehydrated with deionized water and measurements of Fv/Fm were again made at 24 h, 48 h, and 72 h post-rehydration. These values were related to the dark-adapted value of Fv/Fm to determine the mean percentage recovery. Adjusted P-values are Greenhouse–Geisser (G-G) and Huynh–Feldt (H-F)-adjusted probabilities.

Diplazium striatastrum Lellinger
Rate 2  3.07  0.06490.63015.0810.0788  0.00430.0016
ID(Trt)12  3.58  0.3931     
Recovery time 2 98.7< 0.0001  G-G ɛ=  0.73 
Recovery time × Trt 4  4.06  0.0118  H-F ɛ=  0.944 
Adiantum latifolium Lam.
Rate 2  1.53  0.23780.6914.060.1312  0.00730.0029
ID(Trt)12  2.09  0.4982     
Recovery time 2324.53< 0.0001  G-G ɛ=  0.764 
Recovery time × Trt 4  5.44  0.0029  H-F ɛ=  0.999 
Cyclopeltis semicordata (Sw.) J. Sm.
Rate 2600.01< 0.00010.7063.8270.1475  0.0180.0094
ID(Trt)12  1.98  0.5498     
Recovery time 2  7.33  0.0083  G-G ɛ=  0.773 
Recovery time × Trt 4  2.43  0.753  H-F ɛ=  1 
Pityrogramma calomelanos (L) Link
Rate 2 57.65< 0.00010.8571.6880.4298< 0.0001< 0.0001
ID(Trt)12  0.56  0.792     
Recovery time 2 80.09< 0.0001  G-G ɛ=  0.875 
Recovery time × Trt 4 11.67< 0.0001  H-F ɛ=  1 
Phlebodium pseudoaureum (Cav.) Lellinger
Rate 2 98.81< 0.00010.4628.480.0143< 0.0001< 0.0001
ID(Trt)12  0.38  0.8699     
Recovery time 2169.52< 0.0001  G-G ɛ=  0.65 
Recovery time × Trt 4 20.49< 0.0001  H-F ɛ=  0.818 
Microgramma reptans Sw.
Rate 2 24.82< 0.00010.4927.780.02  0.00460.0019
ID(Trt)12  1.65  0.548     
Recovery time 2 87.4< 0.0001  G-G ɛ=  0.664 
Recovery time × Trt 4  6.78  0.0008  H-F ɛ=  0.839 
Figure 3.

Proportional Fv/Fm recovery results for gametophytes exposed to one, two, or three desiccation cycles at relative humidity (RH) = 50% (Ψ = –94.57 MPa). Gametophytes were kept at this humidity for 48 h. Material was then rehydrated with deionized water and measurements of Fv/Fm were again made at 24, 48, and 72 h post-rehydration.


Desiccation tolerance has been widely studied and has been shown to be rare in the vegetative tissues of vascular plants: fewer than 1% exhibit this ability (Proctor & Pence, 2002). The ability to recover from desiccation is, however, common in the gametophytes of nonvascular plants like mosses, liverworts, and algae. The ferns pose an enigmatic intermediate system that combines an independent gametophyte that outwardly resembles some bryophytes, with a complex vascular leaf that is similar to leaves in other more derived vascular plants. We have a relatively clear understanding of DT in fern sporophytes. In general, DT in sporophtyes is rare and only slightly more common than in the sporophytes of seed plants. Yet the degree of DT in fern gametophytes remains poorly studied and thus our understating of DT in this phase of the life cycle is especially limited. A major goal of this paper is to examine how DT is expressed in a system such as the ferns and to evaluate life cycle-mediated differences in the fern response to desiccation.

Apart from work completed early in the 20th century (Pickett, 1913, 1914, 1931) we know little regarding this aspect of gametophyte biology. Our initial results, that fern gametophytes exhibit rapid rates of water loss, are perhaps not surprising given the gametophyte's lack of cuticular and stomatal systems that are normally associated with the control of water loss. With such poor control of transpiration, gametophytes face considerable variations in water content throughout the day and over the growing season. Survival would seem to depend on the ability of gametophytes to survive such conditions, that is, to rely on DT.

After exposure to a relative humidity of 50% (Ψ = –94.57 MPa) for 24 h, all species exhibited > 50% recovery of the pretreatment Fv/Fm values by 24 h after rehydration and the majority had recovered > 70% of this value (Fig. 1). The study species are all tropical in origin and while they experience various degrees of humidity in nature, some of the more exposed canopy trees for this same forest average 86% humidity (VPD = 3.0 Kpa) during the dry season (Cardelus & Chazdon, 2005). In nature, species in this study may see humidities of 50%, but the value remains on the extreme end of what they typically experience.

Desiccation comes in many forms and over the lifetime of a gametophyte, desiccation events can be of varying intensities. Desiccation intensities have been shown to have considerable influence on recovery in several bryophyte species (Proctor, 2003). To examine the influence of desiccation intensity on photochemical efficiency, three species were exposed to three different desiccation intensities chosen to reflect daily water potentials experienced by most species (80% RH, Ψ = –30.44 MPa), a typical drought event (50% RH, Ψ = –94.57 MPa) and an extreme value that species in this site rarely, if ever, experience (20% RH, Ψ = –219.59 MPa). The results from this experiment demonstrated a remarkable tolerance to desiccation intensity that is tightly linked to species ecology (Fig. 2). Diplazium striatastrum is a creek-side species and had little tolerance of desiccation intensities below 50% RH. Phebodium pseudoaureum is an epiphyte of open and exposed habitats, often found growing along roadsides and open clearings. The recovery of this species was slow, but reached 60% by 72 h post-rehydration. Microgramma reptans, which is often found growing in more highly exposed areas such as fence posts, tree trunks, and exposed canopy twigs, exhibited essentially no sensitivity to increasing desiccation intensity. These patterns correspond closely with those reported for bryophytes, with species from xeric habitats exhibiting greater DT than those from mesic habitats (Oliver et al., 1993; Deltoro et al., 1998; Proctor, 2001; Cleavitt, 2002; Alpert, 2005; Proctor et al., 2007).

Not only do species experience different intensities of desiccation in nature, they also experience multiple desiccation cycles throughout the day and/or growing season. The species in this study clearly exhibited different abilities to cope with consecutive desiccation cycles, with species of more mesic habitat exhibiting little ability to cope with more than one cycle of desiccation (Fig. 3). The more mesic creek-side Diplazium striatastrum was the most desiccation-sensitive after once cycle and had the worst recovery, whereas, the more xeric terrestrial species Adiantum latifolium, Cyclopeltis semicordata, and Pityrogramma calomelanos had higher recoveries following one cycle. Phlebodium pseudoaureum and Microgramma reptans, the species of more xeric epiphytic habitats, also exhibited depression in Fv/Fm, but showed the highest recovery from all desiccation cycles. Common for all terrestrial species is the relative tolerance to a single desiccation cycle and the great depression (< 40% recovery) caused by repeated cycles. In the short term, simple recovery of PSII function is related to the release of excess excitation energy. However, long-term and repeated desiccation cycles may induce more biological damage, and recovery may depend more on protein synthesis which is required to repair true photo-damage (Proctor & Smirnoff, 2000; Proctor et al., 2007).

While all species exhibited recovery following an extreme desiccation event, the extent of recovery differed among species and was closely linked to species’ natural ecology. The full role of the fern gametophyte in controlling recruitment remains unclear, but the data presented in this study suggest that they are relatively robust in dealing with desiccation, especially in limited cycles. While desiccation intensity clearly influenced recovery, repeated cycles of desiccation were more likely to limit recovery and, in the case of the most mesic species, likely resulted in significant photo-damage. The degree of recovery following desiccation and its relation to species ecology suggests that fern gametophytes exhibit adaptively meaningful variation in this character. The difference in DT between epiphytic and terrestrial taxa, combined with the revelation that gametophytes of tropical epiphytes may live for years (Watkins et al., 2007), suggests that gametophyte ecology plays a critical role not only in structuring populations but also in influencing species evolution.

Role of gametophyte morphology in water relations

Gametophytes clearly exhibit variation in DT and such abilities were linked to species’ habitat preference. When the ability to recover from desiccation was examined, it was found that recovery was more closely related to the rate of drying than to the final water content reached (Fig. 1) or gametophyte size (expressed as dry mass, data not shown). There was considerable variation in drying rates among species exposed to identical drying conditions in spite of the relative simplicity of the gametophyte. Mechanisms ranging from complex biochemistry to aspects of cell wall strength (Oliver et al., 2005; Moore et al., 2006) have been reported to play a role in a plant's ability to recover from desiccation. Whereas these factors may also influence drying rate, species with more complex three-dimensional morphologies exhibited significantly slower dry-down rates than gametophytes that were more planar.

Given that drying rate was related to recovery and that gametophytes lack sophisticated water conservation tools, we propose that increasing degrees of gametophyte morphological complexity may play an important role in species’ radiation and survival in dry habitats. Gametophytes with more complex morphologies can hold external water in the folds and proliferations of the thallus and/or create small pockets of highly humidified air, thus extending the amount of time that the thallus remains wet and photosynthesizing. A similar phenomenon has been reported for the moss Grimmia pulvinata (Zotz et al., 2000). If such morphology poses some advantage to survival or fitness in drought-prone habitats, then an increase in proliferating morphologies would be expected to occur in drought-prone habitats. This hypothesis is partially substantiated by the general observation that gametophytes of species from drought-prone environments, such as those in epiphytic habitats, tend to produce thalli that often exhibit complex branching, overlapping wings, and proliferations, whereas species from more buffered terrestrial habitats have less ornamented and simple planar morphologies (Fig. 4) (Atkinson & Stokey, 1964; Nayar & Kaur, 1971; Dassler & Farrar, 2001). These patterns were clearly observed in our study when we examined morphology and recovery abilities between understory terrestrial (simple morphology) and epiphytic (complex morphology) taxa (Fig. 4).

Figure 4.

Morphology in fern gametophytes is diverse and is closely related to species ecology and phylogeny. Gametophytes of species from drought-prone epiphytic habitats tend to produce thalli that often exhibit complex branching, overlapping wings, proliferation and hairs, whereas species from more buffered terrestrial habitats have simple, less ornamented morphologies. The athyrioids, thelypteroids, onocleoids, woodsioids and blechnoids are almost entirely terrestrial and tend to have simple morphologies; perhaps < 1% of the known species are true epiphytes. On the other hand, the lomariopsoids, elaphoglossoids, oleandroid, davallioid and polypodioids have many epiphytic species; perhaps as many as 60–70% of the species in this group as a whole are epiphytic, most of which exhibit complex morphologies. Such differences in morphology are significant and may have been critical in the radiation from terrestrial species into canopy habitats. Phylogeny redrawn from Smith et al., 2006).

While morphology could act to enhance water-holding capacity, there may also be anatomical specializations that can both slow and protect desiccating cells. It has been hypothesized that cells with thicker walls will lose water more slowly and can more effectively resist collapse compared with thinner cell walls. Indeed, some xerophytic species tend to have thicker walls when compared with mesophytic species (Oertli et al., 1990; Oliver, 1996; Moore et al., 2006). Direct measurements of cell wall thickness of fern gametophyte have not been made and future work will need to examine cell wall thickness along with thallus morphology.

Dassler & Farrar (2001) suggested that gametophyte branching and perennial growth have evolved in epiphytic taxa in response to the need to grow among bryophytes in the epiphytic habitats. They also argue that such morphology enhances the probability for outcrossing provided by gametophytic perenniality in the epiphytic habitat where the opportunity for gametophyte establishment via spores is limited. Morphology likely serves a dual function and the mechanism leading to increased longevity required for effective outcrossing may be increased water-holding capacity that has evolved with changes in morphology. If differences in morphology confer increased DT and/or a greater ability to control desiccation, then such changes may have been critical in the radiation from protected terrestrial habitats into canopy and more drought-prone habitats.


In spite of the critical role of the gametophyte in the ecology of ferns, this stage of the life cycle has been frustratingly unappreciated. Modern, and indeed ancient, thinking would suggest that fern gametophytes are ephemeral, small, delicate plants that are relegated to permanently wet habitats. Such a nonsensical view has placed a considerable limitation on our understanding of the organismal biology of the ferns. Ferns have been remarkably successful in colonizing the very same habitats as their sister spermatophytes and every sporophyte population owes its origin to the one or more supposedly delicate gametophytes that preceded it in that habitat.

Given similar physical and morphological constraints, fern gametophytes are, for all intents and purposes, bryophytes. It is thus not surprising that fern gametophytes have followed a decidedly bryological trajectory to form, structure, and physiology. This is a natural enough consequence of similar limitations on this growth form. If one looks back to the bryophyte literature of the last half century (and perhaps longer), one will see that the morphology and physiology of nonvascular plants have evolved in much the same way as seed plants. The notion that these ‘lower plants’ should in some way be less fit or less able to evolve needs to be corrected in our modern thinking. Ecological and evolutionary success requires sophisticated adaptation regardless of whether the plant is a bryophyte, fern gametophyte, gymnosperm, or orchid. A broader understanding of fern gametophyte ecology and how this is linked to the sporophyte can ultimately provide a living window into many of the critical innovations that vascular plants required before they were able to dominate our modern flora.


Funding for this work was provided by an NSF DDIG to JEW and MCM, an OTS-Mellon Foundation Pilot Grant to JEW, and a Mellon Foundation Young Investigator grant to MCM. We very much appreciate comments provided by Pamela Soltis, Shimon Rahmilevich, Donald Farrar, and Michael Proctor on an earlier version of this manuscript.