Introduction
The extent of mangroves in Sri Lanka is about 15,981 ha, covering only about 0.24% of the land in the coastal districts (Arulnayagam et al., Reference Arulnayagam, Khim and Park2021). Mangrove distribution in Sri Lanka is usually limited to narrow belts along the coastline because of the low (75 cm) tidal amplitude (Ranawana, Reference Ranawana2017). However, mangroves provide invaluable ecosystem services to the human community by economic and environmental means (Lee et al., Reference Lee, Primavera, Dahdouh-Guebas, McKee, Bosire, Cannicci, Diele, Fromard, Koedam, Mendelssohn, Mukherjee and Record2014; Carugati et al., Reference Carugati, Gatto, Rastelli, Martire, Coral, Greco and Danovaro2018; Getzner and Islam, Reference Getzner and Islam2020; Takahashi et al., Reference Takahashi, Hashimoto, Wanhui, Dasgupta, Hashimoto and Saito2022), especially for the people in coastal communities that heavily depend on the ecosystem services provided by mangroves directly or indirectly (Katupotha, Reference Katupotha2016). Despite their significance, mangrove ecosystems are at the brink of deterioration in Sri Lanka similar to many other locations in the world (Wickramasinghe, Reference Wickramasinghe, DasGupta and Shaw2017; Leal and Spalding, Reference Leal and Spalding2022). Furthermore, natural regeneration of mangroves is hampered when they experience unfavourable conditions for their seed dispersal, germination and seedling growth and establishment due to natural (climate change and natural disasters) and anthropogenic activities (urban development, encroachment). For instance, after the 2004 tsunami in the Indian Ocean, 1200 km of coastline in Sri Lanka was damaged, as a consequence the natural habitat of the mangroves was altered and did not recover well (Kodikara et al., Reference Kodikara, Mukherjee, Jayatissa, Dahdouh-Guebas and Koedam2017).
Thus, assisting in mangrove regeneration and rehabilitation is important to continuously receive their ecosystem services. As such, a large number of restoration attempts have been conducted in Sri Lanka to conserve mangrove ecosystems. Unfortunately, most of these attempts have failed (Kodikara et al., Reference Kodikara, Mukherjee, Jayatissa, Dahdouh-Guebas and Koedam2017). Furthermore, most of these restoration attempts have used only few mangrove species, especially targeting those with readily available propagules (Kodikara et al., Reference Kodikara, Mukherjee, Jayatissa, Dahdouh-Guebas and Koedam2017). However, restoring monoculture mangroves or mangroves with few species does not support restoration of the dynamic nature of mangroves to provide ecosystem services to the same extent as diverse mangroves.
Selection of plant species or species composition is very important in restoration when returning wetlands to their previous non-disturbed state and this can be done in two ways: native species can be selected that historically existed on the site or native species can be selected that occur in nearby ecosystems with similar environmental conditions as the targeted mangrove that needs restoration (Allen et al., Reference Allen, Ewel and Jack2001; Kettenring and Tarsa, Reference Kettenring and Tarsa2020). However, reintroduction of historical species at a site does not guarantee that they will succeed there again if the environmental degradation has exceeded the level of the species’ tolerance. In general, a sound knowledge on the major ecological components of the restoration sites, scientific understanding on mangrove ecosystem and selection of suitable plant species composition are some of the important factors that need consideration in the restoration of mangroves.
Information on seed germination behaviour (dormancy, dormancy-breaking requirements and germination requirements) for a particular mangrove species is essential for their restoration in both direct seeding and seedling transplanting methods (Baskin and Baskin, Reference Baskin and Baskin2004). Even though the direct seeding in the field is the most cost-effective method of restoration over transplanting juvenile plants to the site, seeds with dormancy might cause a negative impact on restoration. For instance, direct seeding mixtures of species might contain species with dormant seeds and others with non-dormant seeds, in which case non-dormant seeds will germinate quickly and dominate the restoration site (Baskin and Baskin, Reference Baskin and Baskin2020). This might be the same scenario in mangrove nurseries developed for restoration purposes. Therefore, information on seed dormancy and germination of species in mangrove ecosystems is crucial for restoration practitioners when conducting restoration projects (Duke et al., Reference Duke, Meynecke, Dittmann, Ellison, Anger, Berger, Cannicci, Diele, Ewel, Field, Koedam, Lee, Marchand, Nordhaus and Dahdouh-Guebas2007; Polidoro et al., Reference Polidoro, Carpenter, Collins, Duke, Ellison, Ellison, Farnsworth, Fernando, Kathiresan, Koedam and Livingstone2010; Van Lavieren et al., Reference Van Lavieren, Spalding, Alongi, Kainuma, Clüsener-Godt and Adeel2012; Baskin and Baskin, Reference Baskin and Baskin2014).
In addition to restoration, seed dormancy studies are important in comparing the relative importance of seed dormancy classes in species in a particular ecosystem (Baskin and Baskin, Reference Baskin and Baskin2010) and ultimately elucidating the dynamics of ecosystems (Skoglund, Reference Skoglund1992). However, many studies on dormancy and germination have been conducted at the species level and only a few studies have been done to evaluate seed dormancy and germination at the community level (Sautu et al., Reference Sautu, Baskin, Baskin, Deago and Condit2007; Athugala et al., Reference Athugala, Jayasuriya, Gunaratne and Baskin2021; Samarasinghe et al., Reference Samarasinghe, Jayasuriya, Gunaratne, Senanayaka and Dixon2022). Although several other scientists have studied seed dormancy at the plant community level (Ng, Reference Ng1975, Reference Ng1980; Ng and Asri, Reference Ng and Asri1979; Garwood, Reference Garwood1983; Murali, Reference Murali1997; Thapliyal and Phartyal, Reference Thapliyal and Phartyal2005), these studies have not been carried out in a way to allow standard comparisons using the current dormancy classification system (sensu Baskin and Baskin, Reference Baskin and Baskin2004). Furthermore, no community-level study has been focused on seed dormancy and germination in species in mangrove ecosystems.
Seed germination and dormancy studies have mainly been conducted on viviparous (i.e. germination and subsequent development of the propagule take place while the fruit is still attached to the mother plant) mangrove species, while only a few studies have been reported on non-viviparous (i.e. germination and subsequent development of the seeds/diaspore take place after seed dispersal) species (Tomlinson, Reference Tomlinson1994; Baskin and Baskin, Reference Baskin and Baskin2014; Wijayasinghe et al., Reference Wijayasinghe, Jayasuriya, Gunatilleke, Gunatilleke and Walck2019). Thus, the main objective of this study was to evaluate the seed dormancy of selected plant species in mangrove plant communities in Sri Lanka. The information gathered was used to construct a dormancy profile for the mangrove plant community and to evaluate the relative importance of the different dormancy classes found in the community in relation to other communities. Baskin and Baskin (Reference Baskin and Baskin2004, Reference Baskin and Baskin2014) developed a dichotomous key to identify the class of seed dormancy for a particular seed lot for a plant species. For that key, we need information on embryo morphology, whether or not embryo grows prior to germination, water permeability of the dispersal unit, ability of seeds to germinate within about 4 weeks and time duration of radicle and shoot emergence. Using aforesaid information, seeds can be categorized into five dormancy classes: physical dormancy (PY), physiological dormancy (PD), morphological dormancy (MD), morphophysiological dormancy (MPD) and combinational dormancy (PY + PD).
Materials and methods
Study species
Thirty true-mangroves and mangrove associates, excluding true-viviparous species, were included in this study (Table 1). Scientific names of the species were checked with ‘Plants of the world online’ website (POWO, 2023). We purposely selected a broad diversity of species to construct a representative profile for seed dormancy of plants that closely match most mangrove communities in Sri Lanka.
a Speculated from available previous literature [(Scaevola taccada – Liang et al., Reference Liang, Xiong, Guo, Yan, Jian, Ren, Zeng, Wu, da Silva, Xiong and Ma2020); (Phoenix sp – Baskin and Baskin, Reference Baskin and Baskin2014); (Lumnitzera racemosa – Perera et al., Reference Perera, Jayasuriya, Gunaratne, Karunaratne, Damunupola and Prassanna2019); (Cayratia trifolia – Baskin and Baskin, Reference Baskin and Baskin2014); (Sesuvium portulacastrum – Baskin and Baskin, Reference Baskin and Baskin2014; Martinez and Casasolai, Reference Martinez and Casasolai1992)].
b Imbibition test was not conducted since a high percentage of NS seeds germinated on distilled water.
c Imbibition test was not conducted because of a limited number of seeds or due to the very small size of the seed.
d Germination test was not conducted since a high percentage of NS seeds germinated on distilled water.
e Germination test was not conducted since a high percentage of NS/MS seeds germinated on distilled water.
f ND, non-dormant; MPD, morphophysiological dormancy; PD, physiological dormancy; PY, physical dormancy.
g GA3 test was not conducted due to the limited number of seeds.
h Superscript FD – Fully developed embryo; Superscript UD – Underdeveloped embryo.
Seed collection
Mature ripened dispersal-ready fruits of the study species were collected in 2012, 2013 and 2014 from at least five individuals (except for Xylocarpus rumphii) of each species at mangrove sites in Sri Lanka. Fruits were stored in labelled polythene bags in the field and brought to the Department of Botany, University of Peradeniya, Sri Lanka. Diaspores (i.e. seeds or fruits depending on the species, hereafter referred to as seeds) and seeds were extracted from fruits, and air-dried for about 2–3 h. Afterwards, they were stored in plastic bottles at ambient room temperature (28°C) until used in experiments which were initiated at least within a week after collection.
Standard germination test
A standard germination test was conducted on intact seeds to identify the presence of non-dormancy (ND) or dormancy (D). Three replicates of 15–25 intact-fresh seeds of each species were placed on tissue papers moistened with distilled water in 9-cm diameter Petri dishes and incubated in light/dark (14 h/10 h) at 25°C. Seeds were monitored for germination (radicle emergence to >1 mm) and cotyledon emergence at 3-d intervals for 30 d. Seeds that germinated to high percentages were considered as non-dormant, while those that did so to low or moderate percentages (or did not germinate at all) were considered as dormant. Non-germinated seeds were dissected to check viability of the embryo at the end of the experiment and presence of white firm embryo was considered to be potentially viable. Seeds classified as dormant were manually scarified (individually with a razor blade) and the same standard germination test was done on them, as described above.
Imbibition test
An imbibition test was conducted to identify the species with physical dormancy (PY). If seeds have PY, intact seeds will not imbibe water, while scarified seeds will imbibe water and germinate within 30 d. To test for PY in seeds, two samples of 15 non-scarified (intact-fresh) seeds and 15 manually scarified (individually with a razor blade) seeds were weighed individually with a digital analytical balance to nearest 0.0001 g, and placed individually on tissue papers moistened with distilled water in Petri dishes at 25°C. Seeds were retrieved, reweighed and returned to the Petri dishes after 2, 5, 7, and 24 h and then at 1-d intervals until all the scarified seeds were fully imbibed. Imbibition tests were not conducted for five species because of a limited number of seeds or due to the very small size of the seed (Table 1).
Gibberellic acid treatment
For seed lots that showed dormancy under the standard germination test, gibberellic acid (GA3) was applied to them to test whether GA3 could overcome dormancy. A sample of three replicates with 15 or 25 seeds each were placed on tissue papers moistened with 100 ppm GA3 solution in 9-cm diameter Petri dishes and incubated in light/dark (14 h/10 h) at 25°C. Seeds were observed for germination (radicle emergence to >1 mm) at 3-d intervals for 30 d.
Embryo type and embryo (E)/seed (S) length ratio
Fresh seeds of each species were cut longitudinally in half and the embryo was observed. The type of embryo in each species was identified according to a modified version of Martin's (Reference Martin1946) classification system (Baskin and Baskin, Reference Baskin and Baskin2007). Ten fresh seeds were selected randomly from each species and dissected. Length of the embryos and length of the seeds were measured to nearest 0.01 mm using a Vernier caliper. Embryo (E):seed (S) length ratio was determined for each species.
Since Annona glabra seeds had a low E:S ratio, embryo development was monitored during germination. Two samples of 20 non-scarified seeds each of A. glabra were placed on tissue paper moistened with distilled water or 100 ppm GA3 solution in Petri dishes separately at ambient laboratory conditions. Five seeds were retrieved after 7, 30, 52 and 60 d from each treatment (distilled water, GA3) and were dissected; E:S ratio was determined at each time.
Data analysis
All of the experiments were conducted in a completely randomized design. The equation to calculate the germination percentage (GP) was: GP = (number of seeds germinated/number of total viable seeds) × 100. Germination data were analysed using a binary logistic regression. Epicotyl dormancy experiments (time difference between root and shoot emergence) were analysed using non-parametric Moods’ median test. Minitab 14.1 (Minitab Inc., State College, PA, USA) statistical software was used to analyse the data. Regression lines were fitted for the data taken from the water imbibition test for each tested species.
Results
Germination of seeds on distilled water
Non-scarified seeds of 11 species germinated >80% within 30 d (Table 1) and the remaining (non-germinated) seeds of these species were viable. Non-scarified seeds of Dolichandron spathaceae, Parsonsonia alboflavescens and Pemphis acidula germinated to 55–60%. Less than 25% of non-scarified Canavalia cathertica, Heritiera littoralis, Hibiscus tiliaceus, Ipomoea violacea and Thespesia populnea seeds germinated, while the manually scarified seeds of these species germinated >65%. Moreover, germination of Dendrolobium umbellatum seeds was 91% after manual scarification (compared to 42% in non-scarified seeds) (Table 1). Non-scarified as well as manually scarified seeds of A. glabra, Cayratia trifolia, Luminitzera racemosa, Suaeda vermiculata, Scaveola taccada, Sesuvium portulacastrum and Premna serratifolia germinated to <30%. Non-scarified seeds of Phoenix pusilla and Phyla nodiflora germinated to <5%, while the germination test of manually scarified seeds of these species was not conducted due to the limited number of seeds or due to their small size.
Shoot emergence
Non-scarified seeds of Aegiceras corniculatum placed on tissue papers moistened with distilled water took 115.7 ± 17.3 d between radicle and shoot emergence, and those of Ardisia elliptica took 45.8 ± 10.0 d. Radicle-emerged seeds of all the other tested species had an emerged epicotyl within <7 d.
Imbibition test
A mass increase of manually scarified seeds of Canavalia cathertica, Hibiscus tiliaceus, Thespesia populnea, Dendrolobium umbellatum and Ipomoea violacea, was significantly higher than that of non-scarified seeds (P < 0.05) during imbibition. Furthermore, manually scarified seeds of these five species had a mass increase of >100%, while non-scarified seeds of these species showed a wide range of mass increase. For instance, a mass increase of non-scarified seeds of C. cathertica, H. tiliaceus and T. populnea was <10% (Fig. 1A), while those of D. umbellatum and I. violacea was ~50% (Fig. 1B). Eighteen species showed a similar water imbibition pattern in both manually and non-scarified seeds (Fig. 1C, D).
Gibberellic acid treatment
Non-scarified seeds of Premna serratifolia and Clerodendrum inerme germinated to >80% on tissue papers moistened with 100 ppm GA3 (Table 1). Seed germination of non-scarified Phyla nodiflora and Saueda vermiculata seeds on GA3 was >60%, whereas that of A. glabra seeds was 13%. Germination of seeds of all these species was significantly higher on GA3 than on distilled water (<0.05, data not shown).
Embryo type and embryo/seed length ratio (E/S)
Eight types of differentiated embryos were identified among the seeds of the study species (Table 1). None of the study species showed undifferentiated embryos. Seeds of A. glabra had a small underdeveloped, but differentiated embryo at the time of seed dispersal. Mean (±SE) initial E:S length ratio of non-scarified seeds was 0.22 ± 0.01. It increased to 0.70 ± 0.01 in seeds incubated for 2 months on GA3 and to 0.33 ± 0.01 on distilled water. Growth on GA3 was significantly greater than that in seeds on distilled water (Figs. 2, 3, F = 489.29, P < 0.001).
Discussion
Class of seed dormancy
As revealed in the imbibition test, manually scarified seeds of five species (Canavalia cathertica, Hibiscus tiliaceus, Thespesia populnea, Dendrolobium umbellatum and Ipomoea violacea) imbibed a significantly high amount of water compared to non-scarified seeds, revealing that intact seeds of these species have impermeable seed coats, i.e. their seeds have physical dormancy (PY). This conclusion was confirmed by the germination test, where manually scarified seeds of these species germinated to a significantly higher percentage than non-scarified seeds. Furthermore, manually scarified seeds of these species germinated to a high percentage (>73) indicating that their embryos did not have a physiological component to dormancy, i.e. seeds of these species only have PY. Our observations confirmed the previous reports in the scientific literature that these species produce seeds with PY [T. populnea, Gupta et al. (Reference Gupta, Thapliyal and Singh2004) and Gagare and Mate (Reference Gagare and Mate2009); H. tiliaceus, Francis and Rodrguez (Reference Francis and Rodrguez1993); I. violacea, Jayasuriya et al. (Reference Jayasuriya, Baskin, Geneve and Baskin2009) and D. umbellatum, Jayasuriya et al. (Reference Jayasuriya, Wijetunga, Baskin and Baskin2013)].
On the other hand, only <5% of non-scarified seeds of C. cathertica, H. tiliaceus and T. populinea germinated at 25°C, while about 25 and 45% of I. violacea and D. umbellatum, respectively, did so. This finding revealed that these two groups of species have different germination strategies even though both groups have PY. Most of the seeds produced by C. cathertica, H. tiliaceus and T. populinea had PY, while a considerably large portion of seeds from I. violacea and D. umbellatum were non-dormant within the physically dormant seed lots. Producing dormant and non-dormant seeds in the same seed lot has been proposed to be an adaptation to survival in unpredictable environments (Venable, Reference Venable1985; Mandák, Reference Mandák1997). However, Paulsen et al. (Reference Paulsen, Colville, Kranner, Daws, Hogstedt, Vandvik and Thompson2013, Reference Paulsen, Hogstedt, Thompson, Vadvik and Eliassen2014) suggested that producing non-dormant seeds in a seed lot with PY is an adaptation for dispersal through rodents who depend on olfactory cues to detect seeds. The production of non-dormant seeds among physically dormant seed lots of I. violacea and D. umbellatum may be an adaptation for the species survival by both germinating a fraction of seeds soon after dispersal and depositing the remainder seeds in the soil seed bank for future germination. Although manual scarification promoted the germination of these species producing seeds with PY, natural cues for breaking dormancy are still to be identified.
According to the results of the germination tests, Acanthus ilicifolius, Aegiceras corniculatum, Allophylus cobbe, Ardisia elliptica, Cynometra iripa, Dalbergia candenatensis, Excoecaria agallocha, Pongamia pinnata, Sonneratia caseolaris, Xylocarpus granatum and Xylocarpus rumphii seeds showed non-dormancy (ND; Table 1). Although another three species (Parsonsia alboflavescens, Pemphis acidula and Dolichandrone spathacea) can be categorized as non-dormant since their germination was between 56 and 60%, the rest of the seeds in their seed lot had PD (40–44%).
Although both non-scarified and manually scarified seeds of Premna serratifolia, Clerodendrum inerme, Phyla nodiflora and Suaeda vermiculata germinated to <40% within 30 d on distilled water, seeds of these species germinated to higher percentages when incubated on GA3 (Table 1), revealing that these seeds have physiological dormancy (PD). Moreover, embryos in seeds of these species were fully developed, and thus, there is no morphological component to dormancy. Apparently, there have been less or no previous studies on germination/dormancy of P. serratifolia, P. nodiflora or S. vermiculata seeds. However, studies done on seeds of the other species in the same genera revealed that they have PD or ND (Baskin and Baskin, Reference Baskin and Baskin2014). None of the non-treated seeds of Heritiera littoralis germinated on distilled water. However, excised embryos of H. littoralis germinated to >67% under similar temperature and light conditions. Moreover, it was observed that H. littoralis seeds imbibed (i.e. they did not have PY) when kept in distilled water (data not shown). Hence, the fruit coat of the diaspore apparently acted as a barrier for radicle emergence, and thus, seed dormancy of H. littoralis can be categorized as PD. Ye et al. (Reference Ye, Lu, Wong and Tam2004) reported that >90% of the H. littoralis (from southern China) seeds germinated without any treatment within 1–2 months which contradicted our observation. This difference may be due to geographic variation in dormancy.
None of the non-scarified or manually scarified A. glabra seeds germinated within 30 d on distilled water or on GA3 and it took more than 5 months to germinate on distilled water. However, non-scarified seeds of A. glabra germinated >85% on GA3 within 2 months, indicating that there is a physiological component to dormancy. Moreover, fresh A. glabra seeds have a linear, differentiated underdeveloped embryo with E:S ratio of 0.22. The E:S ratio increased up by 0.71 and 0.33 when seeds were placed on GA3 and on distilled water, respectively, for 60 d. The E:S ratio of A. glabra seeds just after the seed coat ruptured was ~0.71. Thus, the embryo in A. glabra seeds elongated within the seed prior to radicle emergence showing that there was a morphological component to seed dormancy. Moreover, the rate of embryo growth and germination increased with GA3 confirming a physiological component to dormancy as well. Thus, seeds of A. glabra have morphophysiological dormancy (MPD).
Although all Aegiceras corniculatum and Ardisia elliptica seeds germinated within 30 d, a substantial time delay was observed between radicle emergence and shoot emergence indicating epicotyl dormancy. Interestingly, our study is the first report of epicotyl dormancy in seeds of mangrove species.
Less than 5% of the non-scarified seeds of Phoenix pusilla, Cayratia trifolia, Lumnitzera racemosa, Scaevola taccada and Sesuvium portulacastrum germinated on distilled water revealing that they are dormant. Imbibition test was not conducted on seeds of S. portulacastrum as seeds of this species are very small. Both non-scarified and manually scarified seeds of the other four species imbibed at a similar rate, revealing that they did not have PY. Moreover, none of the tested treatments improved germination of their seeds except manual scarification which increased germination of S. taccada from 0 to 25%. Furthermore, embryos of L. racemosa, S. taccada and S. portulacastrum were fully developed and thus, they do not have a morphological component to their dormancy. Thus, seeds of L. racemosa, S. taccada and S. portulacastrum may have PD. This conclusion was supported by similar observations reported by Ye et al. (Reference Ye, Lu, Wong and Tam2004) and Martinez and Casasolai (Reference Martinez and Casasolai1992); Ye et al. (Reference Ye, Lu, Wong and Tam2004) reported that no fresh L. racemosa seeds germinated, while ~30% of 2-month dry-stored (after-ripened) seeds did so. Martinez and Casasolai (Reference Martinez and Casasolai1992) found that S. portulacastrum seeds required after-ripening or incubation at fluctuating temperatures for germination. E:S ratio in fresh seeds of P. pusilla and C. trifolia was low, and thus, it could be speculated that seeds of these species have MPD. Furthermore, P. pusilla (Arecaceae) and C. trifolia (Vitaceae) belong to families known to contain species with MPD (Baskin and Baskin, Reference Baskin and Baskin2014). However, other Phoenix species have been reported to produce seeds with PD only (Baskin and Baskin, Reference Baskin and Baskin2014). Further studies are needed to clearly identify the dormancy classes of these five species.
Ecological implications
Heretofore, seed dormancy of mangroves has not been studied at a community level, and this study is the first attempt to categorize dormancy (sensu Baskin and Baskin, Reference Baskin and Baskin2004) of mangrove plants at a community level. Most of the investigated species (18 species – 60%) had dormant seeds, while the rest had non-dormant seeds (12 species – 40%) (Table 1). Among species producing dormant seeds, most of them showed PD (44%). PY, MPD and presence of epicotyl dormancy are represented by 28, 17 and 11% of the species, respectively. For the tropics, Samarasinghe et al. (Reference Samarasinghe, Jayasuriya, Gunaratne, Senanayaka and Dixon2022) conducted a community-level study on seed dormancy of lowland rainforest tree species in Sri Lanka. They found that the identified MPD, PD, MD and PY dormancy classes among the species were related to the forest strata and dispersal time. In contrast to our mangrove ecosystem, the majority of lowland rainforest trees had non-dormant seeds (62%) and 14.3, 14.3, 7.0 and 2.3% had MPD, PD, MD and PY, respectively (Samarasinghe et al., Reference Samarasinghe, Jayasuriya, Gunaratne, Senanayaka and Dixon2022). Thus, stronger selection pressures must be present for dormancy in mangrove ecosystems than in a lowland rainforest. All of the different dormancy classes represented in our mangrove community show that various germination strategies are present enabling species to survive in this harsh environment. The different kinds of dormancy classes shown by species in the mangrove plant community are crucial for their survival as a community because they provide mechanisms for coping with the environmental challenges they face, including saltwater inundation, drought, low temperatures and low-oxygen soils. Dormancy allows mangroves to avoid or tolerate adverse conditions, conserve energy, and ensure their long-term survival, thus enabling the persistence and resilience of the entire mangrove plant community.
Two mangrove species produced seeds with epicotyl dormancy, which represented the first report of this kind of dormancy in a mangrove ecosystem and more investigations should be done to identify the type of epicotyl dormancy. Athugala et al. (Reference Athugala, Jayasuriya, Gunaratne and Baskin2018) suggested that after initiating and establishing a root system, the timing of shoot emergence can be flexible for a seed with epicotyl dormancy. Although Athugala et al.'s (Reference Athugala, Jayasuriya, Gunaratne and Baskin2018) study focused on tropical montane forest species, the same reasoning might be true for mangrove ecosystems. In mangroves, establishing a root system and becoming stationary in the muddy substrate with fluctuating tides, is difficult and the timing of such would be very important for the seedling to start obtaining water and nutrients for growth. To this end, shoot emergence could be delayed until the plant is firmly established.
A relatively high SE (standard error) in germination percentages was observed among our study species. This variation suggests that germination would be spread-out temporally. Since seeds of mangrove species experience highly fluctuating environmental conditions such as water, salinity and temperature, after seed dispersal, the probability of seeds germination and seedlings becoming established would be increased by having this large variation instead of a narrow window for germination and establishment. This high-temporal variation may be critical for the continuation and survival of species in mangrove environments.
Implications for restoration
Mangrove rehabilitation and restoration is considered as one of the most effective management options globally for dealing with degradation of mangrove forests. In Sri Lanka, 80% of the restoration sites of mangroves were unsuccessful according to the study done by Kodikara et al. (Reference Kodikara, Mukherjee, Jayatissa, Dahdouh-Guebas and Koedam2017). In most of the restoration sites in Sri Lanka, a single species (mainly Rhizophoraceae species) is planted (direct seedling) without considering the biotic and abiotic factors in the site, because of ease of access to plant material as well as handling both during the nursery and planting stages. Kodikara et al. (Reference Kodikara, Mukherjee, Jayatissa, Dahdouh-Guebas and Koedam2017) recommended that direct planting practice (mainly for monospecific cultivation) should be avoided, and mixed species planting should be promoted to enhance the effectiveness of the restoration. To this end, our dormancy profile gives necessary information on seed germination and dormancy of plant species in a mangrove community so that restoration practitioners can identify the kind of dormancy (or non-dormancy) and the level of dormancy to manage nursery stock to be out-planted for restoration and conserve genetic diversity of species having deep dormancies (Table 2).
Conclusion
Three dormancy classes (PY, PD and MPD) and presence of epicotyl dormancy were identified among mangrove species showing that mangrove plants have diverse germination strategies. Furthermore, our study is the first to report the presence of epicotyl dormancy in seeds of mangroves. These findings can assist practitioners in crafting strategies to effectively break dormancy and germinate seeds in conservation and restoration activities of mangroves because failure of germination due to seed dormancy can be a major impediment in producing these plants.
Acknowledgements
The authors would like to thank Yasoja S. Athugala who helped in collecting seeds.
Author contributions
M.M.W. and K.M.G.G.J. designed the study with the help of C.V.S.G., I.A.U.N.G. and J.L.W. M.M.W. and K.M.G.G.J. collected seeds from the field. M.M.W. conducted the experiments. M.M.W. and K.M.G.G.J. analysed the data. M.M.W. wrote the manuscript with K.M.G.G.J., C.V.S.G., I.A.U.N.G. and J.L.W. All authors approved the final article.
Funding
This work was supported by the National Science Foundation in Sri Lanka (grant no. RG/2011/NRB/08).
Competing interest
The authors declare no conflict of interest.