Providing the missing links in fern life history: Insights from a phenological survey of the gametophyte stage.

Premise: The entire life cycle of ferns has been documented, yet their life histories are still poorly understood. In particular, the phenology of fern gametophytes remains largely unknown. To address this issue, we demonstrated a new ecological approach to explore the phenological link between spore release and gametophyte maturation within the life history of a tree fern species.
Methods: We conducted a serial survey of Alsophila podophylla gametophyte abundance in the field, and recorded the time of its spore release. Every two months for one year, all terrestrial fern gametophytes in an unsampled subplot were collected and identified using tissue-direct PCR.
Results: We found temporal differences in gametophyte abundances, with a sevenfold difference between the highest and lowest months. The number of spores released was linked to the gametophyte abundance two months later. The switch from gametophyte to juvenile sporophyte was found to be most correlated with precipitation. Discussion: The observed fluctuation in gametophyte abundance and population structure was likely associated with the phenology of spore release and environmental factors. Importantly, these findings provide the first evidence of phenological links between different developmental stages in a fern's life history.

With approximately 10,500 extant species worldwide (PPG I, 2016), ferns are a diverse group of plants with a complex life history. Ferns are unique in that they have two independently existing generations, the haploid gametophyte and the diploid sporophyte. Researchers have long been working to unravel fern life histories; however, most studies have focused on the much larger plants of the sporophyte generation, and the gametophyte generation is often overlooked. Studies of these distinct generations have exposed their different habitat requirements, the ability of the gametophyte stage to tolerate a larger range of environmental conditions, and the potential for a “separation of generations,” where gametophytes can grow independently of their conspecific sporophytes (Dassler and Farrar, 1997; Nitta et al., 2017; Pinson et al., 2017; Ebihara et al., 2019). There is still much to uncover about the gametophyte generation of ferns, and the use of a quick and effective technique for identifying small gametophytes, which often have very similar morphological features, would unlock more opportunities for studying their biology in their natural habitats.

Among the studies of fern life histories, there are very few (e.g., Sato, 1982; Peck et al., 1990; Lindsay, 1992) that have noted gametophyte phenology. This lack of awareness leaves much unknown about seasonality and the biotic and abiotic factors that influence gametophyte growth patterns. This is not surprising, however, given that knowledge of the phenology of the sporophyte generation of ferns is also sparse. In the sporophytes, shifts in phenological patterns have been shown to influence frond emergence and senescence, spore maturation and release, and sterile vs. fertile frond production (Mehltreter, 2008; Lee et al., 2018). Some of these shifts—e.g., fertile frond production, spore release—could in turn alter the timing and nature of gametophyte and juvenile sporophyte establishment. In addition, an annual rhythm of environmental changes may influence the phenological behavior of fern gametophytes, especially by affecting the timing of development and their sexual identity, and thus the subsequent generation of sporophytes following gametophyte fertilization. Spore germination and gametophyte growth, for example, could be stimulated by light, temperature, and hydration (Life, 1907; Miller, 1968; Raghavan, 1980; Ranal, 1999; Suo et al., 2015; Brock et al., 2019). In many ferns, water‐soluble pheromones known as antheridiogens promote not only spore germination, but also the production of gametophytic male organs (Näf, 1963; Chiou and Farrar, 1997; Yamane, 1998; Schneller, 2008; Romanenko et al., 2020; Hornych et al., 2021). Flagellate sperm in ferns also rely on water to complete fertilization (but see Watts et al., 2021). These well‐known biological features of fern reproduction have been studied under lab conditions, yet the phenological influences on these features have never been investigated in situ for natural populations of fern gametophytes. In particular, the correlation between gametophyte abundance and the environmental factors mentioned above should be examined through quantitative measurements.

One of the biggest challenges of studying the gametophyte generation in its natural habitat is identification. Due to their cryptic morphology and small size, it is difficult to morphologically identify fern gametophytes to the species level, thereby impeding field investigations of gametophytes, along with ecological understanding. Fortunately, advanced molecular tools are available to tackle the difficulties in identifying fern gametophytes, which can be largely solved with DNA barcoding (Nitta and Chambers, 2022) and tissue‐direct PCR (TD‐PCR) (Li et al., 2010; Wu et al., 2022). With the assistance of these technical advances, we seek to discover whether the timing of spore maturation and release is correlated with gametophyte growth. To explore this dynamic, we chose to study Alsophila podophylla Hook. (Cyatheaceae), a subtropical tree fern usually <2 m tall, which produces terrestrial cordate gametophytes favoring outcrossing for reproduction (Shieh, 1994; Lee et al., 1999; Chiou et al., 2000). Alsophila podophylla was preferred for this study due to its common understory occurrence in Taiwan, and, like many other tree ferns, its high production of spores (Rose and Dassler, 2017). With these characteristics, we expect a considerable abundance of A. podophylla individuals to be present in the local gametophyte vegetation. Over one year, this study aims to (1) serially record the gametophyte abundance of A. podophylla while (2) simultaneously recording the spore release phenology of mature A. podophylla individuals. We also reference meteorological data to explore other factors that may influence the stages of gametophyte maturity and fertilization (e.g., the proportion of A. podophylla gametophytes producing sporophytes). Exploring the phenology of the gametophyte generation of ferns could ultimately elucidate a missing link in the fern's life history.

METHODS

Study site and meteorology

The study site was located in a natural forest at Fushan Botanical Garden (Ilan County, Taiwan; 24°343N, 121°343E), a subtropical environment with a year‐round rainy season and seasonal temperature changes. The annual mean temperature for the years 2012 and 2013 was 16.4°C, with a low monthly mean of 7.19°C in February 2013 and a high of 28.83°C in August 2013 (TFRI Metacat Data Catalog: https://metacat.tfri.gov.tw/tfri/ [accessed 16 March 2022]; Figure 1), as recorded from a local weather station at the meteorological station of Fushan Research Center, ca. 1 km from the plots. The forest experiences a mean annual rainfall of ca. 4120 mm and a relative humidity over 90% for most of the year. This wet, humid climate is an ideal environment for studying gametophyte establishment (Suo et al., 2015). We also explored abiotic seasonal patterns, including precipitation, temperature, and irradiance (derived from TFRI Metacat Data Catalog), to examine their correlations with the changes in gametophyte abundance and population structure.

Figure 1

Average monthly temperature (AMT), precipitation (AMP), and irradiance (AMI) at the meteorological station of Fushan Research Center, Taiwan, between July 2012 and August 2013.

Gametophyte survey and DNA‐based identification

To measure the seasonal gametophyte abundance of Alsophila podophylla, a plot was established with one fertile A. podophylla in the center. The plot was divided into seven 5 × 5 m subplots, and we sampled one of the subplots every two months from August 2012 to August 2013 in a counterclockwise series (Appendix S1). Because the A. podophylla individuals were located under a closed canopy, we assumed that spore release was less affected by wind and was thus dispersed evenly in nearby habitats. From October 2012 to August 2013, the six subplots were each further subdivided into 25 1 × 1 m squares, from which all terrestrial gametophytes (>1 mm in size recognizable by human eyes) were collected. The gametophytes were growing either on the soil or fallen tree trunks, and ranged in life stages from a developing gametophyte (stage “G”) to a gametophyte with a juvenile sporophyte attached (stages “S.0” to “S.4”). Stage S.0 represents gametophytes producing a juvenile sporophyte bud without expanded fronds. Depending on the number of expanded fronds of an attached sporophyte, the gametophytes were classified into S.1 (one expanded frond) to S.4 (four expanded fronds).

We used TD‐PCR approaches to identify all sampled gametophytes. For the pre‐treatment (also detailed by Wu et al., 2022), we first collected a small piece of gametophyte tissue (<1 mm2 in size) from each sample, and placed it in a PCR tube with 20 µL ddH2O. The tissues were fragmented through cycles of freezing with liquid nitrogen and sonication with a commercial sonication cleaner (Bransonic 52, Danbury, Connecticut, USA). These raw extractions were used as the templates for subsequent PCR experiments. For the PCR, we used 1× PCR buffer, 1 M betaine, 200 µM dNTP, 15 pmol of each primer, and 1 U polymerase (ExPrime Taq DNA Polymerase; Genet Bio, Nonsan‐si, Republic of Korea). For the PCR primer sets, we either used the trnL‐L‐F (including the intron‐contained trnL gene and the trnL‐F intergenic spacer) universal primer set (Li et al., 2010) or a newly designed taxon‐specific set for A. podophylla (see below).

In the first surveyed subplot from August 2012, we performed TD‐PCR and sequenced all gametophyte samples using the trnL‐L‐F universal primer set. Their species identification was based on the results of a BLASTN search against the National Center for Biotechnology Information GenBank nucleotide collection, and compared with a trnL‐L‐F sequence collection of local Fushan ferns (L.‐Y. Kuo et al., unpublished data). A sequence identity >99% was used as the criteria for our DNA‐based identification. Around half of the gametophyte samples were identified as A. podophylla, which was found to be the most dominant species among this gametophyte community. Based on this preliminary result, we further designed and verified an A. podophylla–specific primer set, “GP LF IGS” (5′‐ATGAGACAGATATCTTTATTTGATCC‐3′) and “GP trnL” (5′‐GCATAGAGTCGAAATTCGAG‐3′), which targets a partial trnL‐L‐F sequence, approximately 600 bp in size, which is specific to A. podophylla. The PCR cycles were as follows: initial denaturation at 94°C for 10 min; 35 cycles of denaturation at 94°C for 30 s, annealing at 50°C for 30 s, and extension at 72°C for 40 s; with a final extension at 72°C for 10 min. We also verified the specificity of this specific primer set using PCR tests with DNA from all sympatric tree fern species and all gametophytes sampled during August 2012. In these tests, we found no true negative or false positive results. We therefore used this specific primer set to identify the A. podophylla gametophytes for the subsequent subplots in this survey.

Measurement of spore release

We determined the number of fertile fronds per plant experiencing spore release as the estimate of the total released spore amount (i.e., number of spore‐releasing fronds per tree). Between July 2012 and August 2013, 25 individuals of A. podophylla ca. 50 m away from the gametophyte plot were selected for spore release sampling to explore temporal influence. The individuals were at least 10 m apart to avoid sampling from vegetative clones. The fronds were labeled with a plastic tag once they had emerged and expanded to a sufficient size (ca. 30 cm long). Spore release was monitored for the first time when >50% of sporangia were open, and continued to be monitored early each month (the 1st–10th day of the month). The release was observed from the middle pinna of each fertile frond, which released spores that were brown to black in color when open.

Statistical analyses

The correlation of gametophyte abundance and structural changes in the gametophyte population with the spore release and meteorological data was analyzed using a Spearman's rank correlation and a generalized linear model (GLM) with a Poisson regression. To estimate the proportions of successfully fertilized gametophytes, the abundance ratio between the initial S stages and stage G was calculated. For the former, we accounted for not only S.0 but also S.1, because the lifetime of S.0 was too short to be recorded within the survey interval of two months. Considering the microsite heterogeneity within a subplot, we also analyzed the trend of the seasonal dynamics of gametophyte abundance by dividing each of the subplots into differently sized smaller plots for further statistical analyses. All statistical analyses were performed in R (R Core Team, 2021).

RESULTS

Spore release time and gametophyte abundance

Similar to the previous findings of Lee et al. (2009a, 2009b), the highest spore release by Alsophila podophylla was observed in August (Figure 2). In terms of gametophyte abundance, we genetically identified a total of 119 A. podophylla gametophytes in the plot in August 2012, as well as 150 gametophytes from another 15 fern species: Diplazium pullingeri (Baker) J. Sm. (n = 105), Haplopteris yakushimensis C. W. Chen & Ebihara (n = 16), Plagiogyria adnata (Blume) Bedd. (n = 10), Dryopteris hasseltii (Blume) C. Chr. (n = 4), Diplazium petrii Tardieu (n = 3), Asplenium normale D. Don (n = 2), Metathelypteris uraiensis (Rosenst.) Ching (n = 2), Haplopteris flexuosa (Fée) E. H. Crane (n = 1), Alsophila spinulosa (Wall. ex Hook.) R. M. Tryon (n = 1), Angiopteris lygodiifolia Rosenst. (n = 1), Stegnogramma wilfordii (Hook.) Seriz. (n = 1), Dryopteris polita Rosenst. (n = 1), Microlepia obtusiloba Hayata (n = 1), Pteris bella Tagawa (n = 1), and Sphaeropteris lepifera (J. Sm. ex Hook.) R. M. Tryon (n = 1). In the following October survey, the highest abundance of A. podophylla gametophytes was detected, with 853 individuals (=34.12 individuals/m2) collected, a fourfold and sevenfold difference from the August surveys of 2012 (119 individuals) and 2013 (207 individuals), respectively (Figure 2). Compared with October, a comparable number of gametophytes without juvenile sporophyte growth (stage G) were detected in December (Figure 2). The abundance dynamics from different‐sized plots (i.e., 1 × 1 m to 3 × 3 m) all show a consistent trend that gametophyte abundance was highest in October 2012 (to February 2013) and lowest in August 2013 (with the decline beginning in April 2013) (Appendix S2). After December 2012, gametophytes in stage S.4 were detected, while the number in stage S.3 also increased significantly (Figure 2).

Figure 2

Seasonal dynamics of gametophyte abundance and the number of spore‐releasing fronds of Alsophila podophylla. The asterisk indicates that gametophyte abundance in August 2012 was only divided into those in the G stage and those in all five S stages (S.0 to S.4) combined. The time periods highlighted with a bold line represent the gametophyte abundance survey period. The definition of the different gametophyte stages can be found in the main text, under “Gametophyte survey and DNA‐based identification.”

The correlation between the timing of spore releases and gametophyte abundance is statistically significant (Table 1). In addition, we found that gametophyte abundance in the smaller 1‐m2 squares within the subplots was not significantly correlated with the distance to the central fertile A. podophylla (Spearman rank order correlation = 0, P = 0.968; Appendix S3).

Table 1

Spearman correlation between gametophyte phenology and biotic/abiotic factors.

Factorsa	Total gametophyte abundance	G‐stage abundance	S‐stage abundance	S‐stage proportion	S.0&S.1 : G ratio	Spore release
RS2 (0–2 months ago)	−0.27	−0.27	−0.27	NA	NA	NA
RS2 + 2 months (2–4 months ago)	0.63	0.63	0.63	NA	NA	NA
RS2 + 4 months (4–6 months ago)	0.73	0.73	0.73	NA	NA	NA
RS2 + 6 months (6–8 months ago)	−0.070	−0.07	−0.07	NA	NA	NA
RS2 + 8 months (10–12 months ago)	−0.36	−0.36	−0.36	NA	NA	NA
RS5 (0–5 months ago)	0.50	0.50	0.50	NA	NA	NA
RS5 + 2 months (2–7 months ago)	0.82	0.82	0.82	NA	NA	NA
RS5 + 4 months (4–9 months ago)	−0.07	−0.07	−0.07	NA	NA	NA
Precipitationb	−0.25	−0.25	−0.25	0.21	0.99	0.53
Temperatureb	−0.43	−0.43	−0.43	0.36	0.83	0.39
Irradianceb	−0.57	−0.57	−0.57	0.21	0.60	0.50

Abbreviations: NA = not analyzed.

RS2 and RS5 indicate the released spore amount in the past two and five months. Significance (P < 0.05) in Spearman correlation and GLM with Poisson regression are indicated in italic and bold, respectively.

Daily average over the previous two months.

Relationship between gametophyte population structure and meteorological factors

The collection of meteorological data revealed a relationship between the gametophyte life stages and environmental factors (Table 1). Three different seasonal variables were identified (Figure 3): the ratio of the number of gametophytes showing juvenile sporophyte growth (stages S.0 and S.1; the S.0&S.1 : G ratio) to the number of gametophytes without juvenile sporophyte growth (stage G), the proportion of sporophyte‐producing gametophytes (the sum of stages S.0 to S.4), and the average daily rainfall over the past two months. With records taken every two months, these variables consistently increase and decrease together (Figure 3). Precipitation, temperature, and irradiance were negatively correlated with gametophyte abundances, but were positively correlated with the proportion of S‐stage gametophytes, the S.0&S.1 : G. ratio, and spore release (Table 1).

Figure 3

Seasonal dynamics of the community structure of Alsophila podophylla gametophytes. S = gametophyte with a juvenile sporophyte (S.0 to S.4), S.0 = gametophyte with a juvenile sporophyte producing no expanded frond, S.1 = gametophyte with juvenile sporophyte producing one expanded frond, G = gametophyte without a juvenile sporophyte.

DISCUSSION

Phenological links between fern life history stages

Plants undergo changes in their phenological patterns, such as seasonal growth and reproduction, due to a wide range of intrinsic and environmental factors (Mehltreter, 2008; Lee et al., 2018). As noted, the phenological differences during the mature sporophyte stage (e.g., fertile frond production, spore release) can in turn influence the growth patterns of the gametophyte stage. Among all the factors surveyed here, the amount of spores released seems to be most significant for explaining the changes in gametophyte abundance over time. We showed that the gametophyte abundance was significantly and positively correlated with the released spore amount 2–4 months earlier, a range reflecting the highs and lows of gametophyte abundance (see “Spore release time and gametophyte abundance”) (Figure 2, Table 1). From this, we can deduce that the abundance of terrestrial cordiform A. podophylla gametophytes is temporally influenced. The observed time lag of 2–4 months from spore germination to gametophyte growth is consistent with observations from lab cultures in the same species (Lee et al., 1999; Chiou et al., 2000). The gametophyte abundances themselves seem indirectly linked to the abiotic factors that we expected would enhance gametophyte growth and sporophyte development, and thus yield positive correlations with gametophyte abundance; however, a few factors (e.g., precipitation, temperature, and irradiance; Table 1) were revealed to have a negative influence. The phenological patterns of the prior life stages, i.e., spore maturation and release, were previously shown to be positively linked with temperature for A. podophylla (Lee et al., 2009a, 2009b; Table 1), suggesting that a shift in temperature could affect gametophyte abundances. Likewise, such autocorrelations can be expected for other environmental factors (e.g., irradiance and rainfall; Table 1). Another possibility could be that increases in precipitation, temperature, and irradiance lead to a higher mortality rate in gametophytes. Regarding the ranges of these factors during our survey period (Figure 1), this possibility is less likely; instead, an opposite trend is generally expected based on our understanding of fern gametophyte physiology. Unfortunately, we were unable to generate real mortality data from the same gametophyte populations to confirm this theory due to the destructive sampling approach required. Previous lab cultures have shown that A. podophylla gametophytes can generate sporophyte descendants within five months of sowing the spores, and be sustained for up to one year without the development of a sporophyte juvenile (Chiou et al., 2000). The cordate gametophytes of terrestrial ferns may have an average lifespan of four months in the field (Watkins et al., 2007), but for tree ferns it seems to be longer (Lee et al., 1999). Collectively, if the A. podophylla gametophytes in the field had a general lifespan of at least five months and then began to progressively wither, and their mortality was less influenced by the environmental changes across seasons, this could better explain why gametophyte abundance began to significantly decrease from April 2013 (Figure 2, Appendix S2).

The most notable environmental factor correlated with gametophyte population structure is precipitation, with average daily rainfall following the trends of juvenile sporophyte growth (Table 1, Figure 3). Ferns produce motile sperm that swim through water; therefore, ferns require a damp environment for successful fertilization (Greer, 1993), a characteristic that is especially important for A. podophylla due to its preference to outcross to produce sporophytes (Chiou et al., 2000). Furthermore, the production of male organs could be promoted by water‐soluble pheromones (i.e., antheridiogens), which also highlights the importance of precipitation for fertilization and the regeneration of sporophytes in many sexual fern species. However, A. podophylla is unlikely to be equipped with such an antheridiogen system, and its antheridia are neither induced by pheromones of conspecific mature gametophytes nor those from other leptosporangiate ferns (e.g., Lygodium Sw., Pteridium Gled. ex Scop., and Ceratopteris Brongn.) (Chiou et al., 2000). In addition to precipitation, temperature and irradiance were found to be positively correlated with the productivity of juvenile sporophytes (i.e., S‐stage proportions and the S.0&S.1 : G ratio values in Table 1). We can infer that these extrinsic physical factors may have promoted the development of gametangia and/or early sporophyte development following fertilization. Under lab cultures, the differentiation of sex organs and their abundances have been found to be affected by light intensity and quality, and orientation of light with respect to the gametophyte thallus (e.g., Life, 1907; Chang et al., 2007; Farrar et al., 2008); however, to our knowledge, the effect of different temperatures has yet to be tested. All these physical relationships are worthy of further exploration in future experiments.

Source of gametophyte establishment

After their release, spores can be incorporated into the soil bank while waiting for the appropriate environmental conditions (e.g., hydration and light levels) for germination into a gametophyte. In the field, the germination rates of released spores are presumably lower than in the lab because they might not fall on a suitable site, thus failing to germinate. In some cases, released spores may have been buried in the soil, although under proper conditions, the spores can remain viable in the soil bank. This is especially true for moist soil environments, where Pedrero‐López et al. (2021) found that spores in the cloud forest remained viable for more than five months, concluding that substrate moisture is a key factor in maintenance of spore viability. Fern spores can be incorporated up to 125 cm deep in the soil, although those deposited within the top 10 cm retain a higher viability for germination to gametophytes (Dyer and Lindsay, 1992; Lindsay, 1992; Ranal, 2003; Hore et al., 2016). Importantly, a safe and suitable microsite is crucial for the completion of a fern's life cycle, from spore germination to subsequent gametophyte development, to fertilization and development of the sporophyte generation (Cousens et al., 1985). We also revealed a significant positive relationship between the released spore amount over the past five months (RS5; particularly RS5 + 2 months, i.e., released spore amount over the past two to seven months) and gametophyte abundance (Table 1). Although it is possible that some of the gametophytes in this study germinated from preserved spores in the soil bank as opposed to fresh spores, we did not observe any environmental influence supporting the increase of gametophyte abundance. In other words, it is unlikely that the majority of gametophytes for this subtropical tree fern came from dormant spores stimulated by an extrinsic environmental cue. Instead, like many (sub)tropical regions, the microhabitats in the Fushan plots seem suitable for gametophyte growth during most times of the year, and spore cohorts from the current season can germinate immediately after landing on safe microsites.

Recommendations for further studies

Studying gametophyte phenology is essential to further understanding the complete life history of a fern species and uncovering population dynamics and succession. Compared with previous ex situ attempts to understand fern gametophyte phenology (Sato, 1982; Lindsay, 1992), we overcame the difficulty of fern gametophyte identification using TD‐PCR (see more applications of TD‐PCR in overcoming fern ecology issues in Wu et al., 2022) and provided the first in situ results with quantitative measurements. Although several uncertainties remain (e.g., gametophyte mortality and subplot heterogeneity), our findings have revealed that, for terrestrial ferns producing cordate gametophytes, the establishment and development of a gametophyte population seems to be temporally influenced by the timing of spore release, and is likely motivated by other environmental factors. While the influence of environmental factors on spore phenology has been previously recorded, there are few studies exploring the link between abiotic factors and gametophyte phenology. We were able to reference general climatic data from the meteorological station nearby, but it would be beneficial to install environmental loggers in field sites to more accurately measure the abiotic factors of the gametophyte microhabitat in future studies.

Additionally, this study could be applied on a larger scale to measure the diversity of fern gametophyte communities and compare the phenology of different species. Although ferns are known to have high dispersibility, each species varies in its fecundity, habitat requirements, and gametophyte ecology (Peck et al., 1990; Farrar et al., 2008). Consequently, one study of fern phenology in a subtropical species cannot be extrapolated to all other species, especially temperate ones (Sato, 1982; Peck et al., 1990; Lindsay, 1992). Furthermore, conclusions about fern phenological patterns should not be made without sufficient long‐term studies. Studies done over a longer period of time leave space for variation in the climatic conditions and fern life cycles (Mehltreter, 2008; Sharpe and Mehltreter, 2010; Lee et al., 2018). As studies of fern phenology progress, the focus should be on long‐term studies conducted across different geographical areas for both the gametophyte and sporophyte generations.

There are several factors of this current approach that should be further developed to survey the demographics of fern gametophytes in situ. First, the genetic identification of these gametophytes relies on destructive sampling; therefore, in order to design plot comparisons (e.g., revealing temporal influences in this study) in the field, the homogeneity of the microhabitats and other physical conditions should be considered and recorded prior to sampling. For instance, fern gametophytes tend to follow a patch distribution at a microscale, where species distribution may depend on microclimate or substrate preference. Second, whether and how spore dispersal is driven by wind direction is an important consideration. In particular, seasonal visits of typhoons and monsoons in some areas might drive the dispersal phenology of ferns there. Furthermore, due to the destructive sampling used, we were unable to obtain the mortality rate of the gametophyte community that was sampled. Ultimately, we are looking forward to exploring any non‐destructive approach that could be further employed to assist the in situ investigations of fern gametophyte communities (e.g., Watkins et al., 2007, and the photographic monitoring by Schneller and Farrar, 2022).

AUTHOR CONTRIBUTIONS

L.‐Y.K., Y.‐M.H., and W.‐L.C. designed the experiments and carried out the gametophyte surveys in the field. L.‐Y.K. and T.‐Y.T. completed the lab experiments for the genetic identification of the gametophytes. P.‐S.L., Y.‐M.H., and W.‐L.C carried out the spore phenology survey. P.‐S.L., L.‐Y.K., and A.Q. analyzed the data. A.Q. and L.‐Y.K. prepared the draft of the manuscript. All authors approved the final version of the manuscript.

Appendix S1. The design of the seasonal gametophyte plots in this study. The gray numbers represent the smaller plot no. of each seasonal subplot. The details about gametophyte abundance of these smaller plots are in Appendix S3.

Click here for additional data file.

Appendix S2. Statistical trends of gametophyte abundance across the different seasons.

Click here for additional data file.

Appendix S3. Gametophyte abundance in each smaller 1‐m2 plot.

Click here for additional data file.