Natural persistence of the coastal plant Glehnia littoralis along temperate sandy coasts

We studied germination behaviors and persistence mechanism of wild Glehnia littoralis, a typical coastal species at temperate sandy coasts of the North Pacific Ocean, and tested the hypothesis that the coastal plants may have evolved special seeds adapting to the coasts, by which they recruit and persist easily, occupying the coasts as ideal habitats. In the Shandong Peninsula, China, we investigated temperature and moisture conditions of coast sand in relation to germination and evaluated effects of sand burial, seawater immersion and sowing time on germination. When germination began, daily dawn temperatures of sand were about 10 °C and daily noon temperatures were about 25 °C; the temperatures were not different in the sand <8 cm deep. The sand at these depths showed a significant difference in moisture contents. The seeds exhibited large germination rates if sand burial was at depths >= 3 cm and winter freezing was kept longer than 2.5 months. Seeds experiencing seawater immersion were able to germinate well. These evidences suggest that G. littoralis has evolved special seeds adapting to seawater dispersal and specific season rhythm. By the seeds, G. littoralis occupies temperate sandy coasts as ideal habitats to persist.

Effects of burial depths, seawater immersion and sowing time on germination. The cumulative germination rates responded differently to the burial depths ( Table 1). The germination rates were significantly low at the depths < = 2 cm (Fig. 1), but were fairly high at the depths > = 3 cm. These results demonstrate that the depths favorable for germination were not less than 3 cm, perhaps because moisture contents were high and steady at these depths.
Seawater immersion might have no effects on final cumulative germination rates (Table 1), which could exceed 60%, but dynamics of the germination differed obviously among these immersion levels (Fig. 2). These results indicate that seawater immersion might not promote the maximum germination rates of G. littoralis seeds but could activate the seeds for a swift germination. The pre-germination immersion for 25 days was the best in urging the germination.
Sowing time dominated the cumulative germination rates ( Table 1). The earlier the seeds were sown in winter, the higher the final cumulative germination rates (Fig. 3). Outdoor freezing for longer than 2.5 months was  necessary to arouse G. littoralis seeds for a swift and successful germination. The finding suggests that G. littoralis seeds have the property of winter-dormancy, and most of them remain dormant until winter provides them a freezing longer than 2.5 months.

Discussion
Seeds of wild G. littoralis require appropriate temperatures and moisture for germination. After daily dawn and noon temperatures of sand rise to ca. 10 and 25 °C, G. littoralis seeds get ready for germination. We observed that if sown at the depths > = 3 cm, they germinate easily. The four sand layers showed a significant difference in moisture contents, other than in temperatures. Moisture contents were the lowest and most unstable at the 0-2 cm depth, which was adverse to germination. At the depth > = 3 cm, moisture contents became high and steady adequately, feasible for seed germination. It was reported on that seeds of G. littoralis have the property of winter-dormancy 13,34 . This study provides details to further understand this. In wilderness, seeds of wild G. littoralis can hardly germinate unless aroused by fierce freezing longer than 2.5 months. This property is a protective adaptation of this species to the long winter of the temperate zone. Adults of the species bear seeds in late summer, when temperatures are favorable. Nevertheless, the seeds refuse to germinate that time. Because of winter-dormancy, they choose to germinate after winter, thus new seedlings emerge in early spring, gaining more time to grow strong to survive the coming winter.  . Cumulative germination rates of G. littoralis seeds sown some months before spring germination. 0.5 month before: 0.5 month before the germination start; 1.5 month before: 1.5 month before the germination start; 2.5 month before: 2.5 month before the germination start; 3.5 month before: 3.5 month before the germination start; 4.5 month before: 4.5 month before the germination start.
Scientific RepoRts | 7:42784 | DOI: 10.1038/srep42784 Glehnia littoralis seeds can keep vigorous or turn more vigorous after seawater immersion, unlike most terrestrial plants [35][36][37] . This is a critical adaptation to seawater dispersal 33,38,39 . Supratidal zones are characterised by occasional seawater inundation 8,40 . Around the North Pacific Ocean, extreme storms often take place around winter to cause violent seawater surges 41,42 . As a result, seeds of G. littoralis at supratidal zones can be inundated and dispersed 17,43,44 . They keep afloat for many days, like seeds of Eryngium maritimum, Cakile maritime and Canavalia rosea 17,38,39 . They are immersed in seawater, whereas await a chance to land at supratidal zones again 17 . After they land at favorable environment, they may germinate and grow, continuing life cycles of the species. In case G. littoralis seeds were disabled during the immersion, this species would have difficulties in recruiting, and not persist for long along supratidal zones.
These germination behaviors confirm the hypothesis that the typical coastal specialist G. littoralis have evolved special seeds to adapt to temperate sandy coasts, at least including seawater dispersal, season rhythm and sandy matrix. By coupling the germination with the environment, we conceived the mechanism of wild G. littoralis for dispersal and persistence along temperate sandy coasts.
Adults of wild G. littoralis give birth to mature seeds before autumn. Although temperatures remain favorable that time, inherent dormancy sets back immediate germination of them. Storms can take place around winter in temperate seas, unlike in tropical ones 42,45 . They inevitably cause seawater surging, which may co-occur and overlap with high astronomical tides to make extreme surges. The surges can inundate supratidal zones for several hours. On this occasion, G. littoralis seeds are immersed in seawater and begin dispersing with seawater current 17,46 . Because of high buoyancy, they keep afloat during the immersion, and are able to return to supratidal zones with the same surge or another surge some days later 17,47 . With seawater, they disperse in short and long distances, and may settle at supratidal zones again. After settling at supratidal zones of temperate sandy coasts, they get the opportunity to began new lives. They are buried by moving sand with winds. During the burial, fierce winter provides them with effective freezing. After winter, they get aroused and ready for germination. In early spring, the daily dawn temperature rises to ca. 10 °C, and some seeds receive the signal to germinate 48 . Those on or in surface sand (< 3 cm deep) can hardly germinate, perhaps because of unfavorable moisture conditions. From spring to autumn, winds are relatively mild and temperatures are favorable, thus the seedlings grow as fast as possible. Several months later, taproots of them grow deep and strong enough to tolerate the first winter they will encounter. After the winter, they revive to grow. By perennial roots and winter-resistant buds, they repeat the revival and growth year after year, bearing seeds for species recruitment and persistence 49 .
The temperate sandy coasts around the North Pacific Ocean provide wild G. littoralis with ideal conditions, such as sand burial and freezing in windy winter, surging seawater for seed dispersal and activation, and calm growing season from spring to autumn. For such a provision, wild G. littoralis occupies temperate sandy coasts around the north Pacific Ocean as original habitats. After transplanted to inland, the species cannot disperse normally as along coasts, and has to lead a spoiled life, which may be adverse to species persistence 50,51 . Since G. littoralis is a typical coastal specialist depending on seawater dispersal, it ought to be conserved in temperate sandy coasts, where seawater dispersal is convenient, season rhythm is suitable and sand matrix is available. Additionally, with the warming of global climates, this species may shift its distribution area northward because long cold winter is necessary for it to germinate and regenerate, and it may be driven extinct from some warm temperate regions such as the Shandong Peninsula 52 .  Table 2). Seeds of the species germinated in early April. During the germination, we measured field temperatures of the sand with glass thermometers one time every 10 days, and the measurement lasted for a month until the germination almost ceased, i.e., four times in total. Daily dawn temperatures were measured just before sunrise, and daily noon temperatures were measured at 14:00. Each time of the measurement was done at three locations randomly chosen from an area of ca. 200 m 2 where seedlings emerged in clumps, and at each location, four sand layers were investigated, covering the depths of 0-2, 2-4, 4-6 and 6-8 cm.

Methods
We sampled sand four times from these layers to measure moisture contents, one time every 10 days from the germination start. Every time, we randomly chose 10 points in the ca. 200 m 2 area to sample sand. At these points, we collected sand vertically with a drill and divided the sand into the four layers. Sand from the same layer from the 10 points was together lumped and sealed in a waterproof plastic bag as a composed sample. Within 3 h, we weighed the samples and dried them in an electronic oven (70 °C for 2 days). We weighed the dried samples and calculated their gravimetric moisture contents, dividing the moisture mass (mass of the wet sample minus that of the dried sample) by the mass of the dried sample.
Experiments. In August 2012, we collected mature seeds of the species (carrying seed capsule) from undisturbed sandy coasts and air-dried them for storage. In November, we began the experiment one at a coast (E 120.93°, N 36.41°) to identify sand depths favorable for germination (Table 2). We loaded 20 wood boxes measuring 40 cm (length) × 20 cm (width) × 20 cm (height) with drainage holes at the bases with sand from the coast. These boxes were divided into five equal groups, one group containing four boxes. We sowed the seeds in these boxes, and sowing depths were one of the levels: 1, 2, 3, 4 and 5 cm. One box group (4 boxes) corresponded to one level of the sowing depths. In each box, 30 seeds were evenly sown at the equal depth. After sowing, we gently watered the sand within the boxes with collected rainfall. Then, we left the boxes outdoor and spread a mesh 2 m above to prevent birds (other than snow, rain and light). Monitoring was performed everyday. Since seedlings emerged, we counted new ones every 2 days, which we timely marked with fine toothpicks. The counting and recording lasted for a month until germination almost ceased.
In August 2013, we collected seeds again and air-dried them, to prepare the experiments two and three ( Table 2). From October 17, we began to immerse seeds in seawater, one batch every 5 days. By November 16, we prepared seven batches of seeds ever immersed for 30, 25, 20, 15, 10, 5 and 0 days. During the immersing, we replaced the seawater every 5 days. On November 16, we fished out and sowed the seeds in the same boxes as the mentioned above, and placed them outdoor at the coast. Each immersion level included four boxes (replicates), where 30 seeds were evenly sown at the depth of 3 cm (we found it feasible for germination).
On November 20, 2013, the experiment three started to examine the effects of sowing time on germination ( Table 2). We sowed the newly collected seeds but not subjected to any seawater immersion within the same boxes every month until March 20, 2014. The experiment included five levels of sowing time: November 20, December 20, January 20, February 20 and March 20. Every level included four boxes (as replicates), and in each box, 30 seeds were evenly sown at the depth of 3 cm. After sowing, we watered and monitored the boxes as in the experiment one. On April 6, 2014, germination began. Then, new seedlings were counted and marked every 2 days for a month. The authors comply with Convention on the Trade in Endangered Species of Wild Fauna and Flora.
Data analyses. Using two-way ANOVA, we examined the differences in daily dawn temperatures, daily noon temperatures and moisture contents among the four sand layers and the changing time points. The sand layers were four depth levels: 0-2, 2-4, 4-6 and 6-8 cm. The changing time points included four levels in the germination process: start of the germination, 1/3 of the germination period (the 10th day), 2/3 of the germination period (the 20th day) and the end of germination. We calculated cumulative germination rates over the germination process, which were seedling number until a certain day divided by the number of initially sown seeds. Using one-way ANOVA, we checked effects of sowing depth, seawater immersion and sowing time on the final cumulative germination rates. Data analyses were completed using Origin 8.0 (OriginLab, Northampton, USA) and R 3.2 (R Core Team 2015).

Field surveys
During the germination period, we measured daily dawn and noon temperatures of coast sand. Measurements were conducted at four layers (covering the depths of 0-2, 2-4, 4-6 and 6-8 cm), one time every 10 days. At the same time, we sampled sand from these layers to measure gravimetric moisture contents.

Experiment one
This experiment was conducted to identify sand depths favorable for seed germination. We collected seeds in autumn and sowed them in winter in sand at the depths of 1, 2, 3, 4 and 5 cm.

Experiment two
This experiment was designed to assess effects of seawater immersion on seed germination. We prepared seven batches of seeds that experienced pre-germination seawater immersion for 0, 5, 10, 15, 20, 25 and 30 days. In winter, we sowed them in sand at the depth of 3 cm.

Experiment three
The experiment was used to examine effects of sowing time on seed germination. We sowed seeds in sand at the depth of 3 cm; the sowing time was, respectively, November 20, December 20, January 20, February 20 and March 20. Table 2. Brief description of the methods.