Introduction

Rheophytic habitats are riparian environments between the lowest and the highest water levels along swift-running rivers or streams. They are often submerged under high water caused by intense rainfall. Because of the inability of normal terrestrial plants to withstand strong water currents, rheophytic habitats are usually unfavorable for their growth. However, unique plant communities consisting of rheophytes often occur in these habitats. Rheophytes usually exhibit some characteristics adaptive to their habitats, such as stenophyllous lanceolate leaves with a cuneate base, matted root systems, and flexible stems and petioles (van Steenis 1981, 1987; Imaichi and Kato 1997). Six hundred and fifty rheophytes are known from 68 angiosperm families (nine monocotyledonous and 59 dicotyledonous families) (van Steenis 1981). Most of them (75%) are narrow endemic species, and convergent adaptation to rheophytic habitats may have occurred in different areas of the world regardless of taxonomic group (van Steenis 1981).

Adaptation to rheophytic habitats is a common evolutionary trend in humid tropical areas (van Steenis 1981). Many rheophytes grow in the wet tropics where their habitat is well developed, while the number of species declines in dry or high-latitude regions with less annual rainfall. In subtropical rainy regions of Japan, such as southern Kyushu and the Ryukyu Islands, the rheophytic habitat is as well developed as in tropical regions (Yokota 2003). Many endemic rheophytes such as Farfugium japonicum var. luchuense (Asteraceae) and Viola stoloniflora (Violaceae), occur particularly in the Ryukyu Islands (Yokota 2003), which comprises a number of small islands lying in a chain between the Kyushu Island of Japan and Taiwan.

Scutellaria rubropunctata Hayata (Lamiaceae) also shows adaptations to rheophytic habitats in the Ryukyu Islands. The genus Scutellaria L., which is widely distributed in temperate regions of the world, comprises about 350 species (Mabberley 2017). Scutellaria rubropunctata is one of six species in the genus distributed in the Ryukyu Islands, and is endemic to the islands from the northern to the central part of the chain (Hatusima 1975, 2004; Shimabuku 1990) (Fig. 1). This species is thought to be a facultative rheophyte because it occurs in both terrestrial and rheophytic habitats within its range (Yoshimura et al. 2016). It is fairly common in terrestrial areas such as sunny mountain tops in limestone regions and disturbance-caused forest edges along roads, but it also occurs in rheophytic habitats of a number of rivers in the Ryukyu Islands. The rheophytic type of S. rubropunctata is often exposed to and submerged under fast-flowing rising water. It is a dwarf plant with small diamond-shaped leaves with cuneate bases (Yoshimura et al. 2016).

Fig. 1
figure 1

Map of the Ryukyu Islands and adjacent regions. The distribution range of Scutellaria rubropunctata is encircled by the broken line

Full size image

During preliminary investigations, we observed differences in seed germination characteristics between terrestrial and rheophytic types of S. rubropunctata. Seeds of the rheophytic type tended to germinate in a relatively short time after sowing, while those of the terrestrial type germinated somewhat gradually. Many studies on the morphological and anatomical features of rheophytes have evaluated stenophyllization, in both fern (e.g., Imaichi and Kato 1992a, b; Sharpe 1997) and angiosperm species (e.g., Ohga et al. 2012; Ueda et al. 2012; Matsui et al. 2013). Most of these rheophytes have smaller or narrower leaves than their terrestrial relatives, enabling them to decrease resistance against fast-flowing water. In Japan, similar studies have been conducted on facultative rheophytes, such as F. japonicum var. luchuense (Usukura et al. 1994; Nomura et al. 2006, 2010) and Adenophora triphylla var. japonica (Ohga et al. 2012). However, few studies have investigated the germination characteristics of rheophytes in relation to habitat adaptations.

The objective of this study was to elucidate differences in germination characteristics between the terrestrial and rheophytic types of S. rubropunctata based on germination experiments under different temperature treatments. The adaptive meaning of germination characteristics corresponding to habitat differences between the two types of S. rubropunctata is discussed. Germination patterns of the congener S. indica var. parvifolia were also investigated for comparison.

Materials and methods

Growing environment

The growing environment of S. rubropunctata was categorized into terrestrial and rheophytic habitats following Yoshimura et al. (2016). In the northern part of Okinawa-jima Island, which is located in the middle of the Ryukyu Islands chain, S. rubropunctata is relatively common in terrestrial habitats such as sunny mountain tops in limestone areas and anthropogenic forest edges (Fig. 2a, b). In such places, this species is not directly affected by flowing water caused by heavy rain. Scutellaria rubropunctata is relatively rare in rheophytic habitats, where it grows in cracks or depressions of large rocks along stream banks together with other rheophytes such as Osmolindsaea japonica (Lindsaeaceae) and Salvia pygmaea (Lamiaceae) (Fig. 2c, d).

Fig. 2
figure 2

Habitat and habit of Scutellaria rubropunctata.a Terrestrial habitat (Mt. Nago, Nago City); b terrestrial type; c rheophytic habitat (Fun River, Kunigami Village); d rheophytic type (Hiji River, Kunigami Village)

Full size image

Unlike S. rubropunctata, rheophytic adaptation has never been reported for its congener S. indica L. var. parvifolia Makino. This plant is distributed from Kanto to Kyushu regions in Japan, Taiwan, and continental China (Li and Hedge 1994). We collected S. indica var. parvifolia for this study from Oita Prefecture of the Kyushu region, where it is relatively common in open or half-shaded lowland habitats such as forest edges along coasts (Arakane and Tsuji 2011).

Seed sampling

Fruits of the genus Scutellaria are covered with a two-lipped calyx and comprise four nutlets. Because each nutlet has a single seed covered by a thin pericarp without a sclerenchyma region (Ryding 1995), we refer to them simply as seeds throughout this paper. It has been reported that S. rubropunctata bears two different flowers: chasmogamous flowers adapted for outcrossing and cleistogamous flowers for selfing (Denda 2002), as is the case in congeneric S. indica (Sun 1999). In this study, seeds of cleistogamous S. rubropunctata flowers were used for the experiments. However, only seeds of chasmogamous flowers could be obtained for S. indica var. parvifolia.

Collection data for seeds used in this study are shown in Table 1. All seeds of S. rubropunctata were collected from the northern mountainous part of Okinawa-jima Island. Mature seeds were collected from seven terrestrial habitats (2912 seeds from 226 individuals) and from seven rheophytic habitats (2280 seeds from 210 individuals) from July to August in 2013 and 2015. Seeds of the terrestrial and rheophytic types of S. rubropunctata showed similar morphology with tuberculate surface sculpturing, although the latter was slightly smaller than the former (H. Yoshimura, M. Yokota, T. Denda, unpublished data). For comparison, 2000 seeds of S. indica var. parvifolia were collected from one locality in Oita Prefecture in May 2017 (Table 1). In this population, S. indica var. parvifolia grew so densely that it was not possible to distinguish individual plants for seed collection.

Table 1 Sampling locality, number of individuals and seeds used in this study
Full size table

Germination experiment

Germination experiments were carried out as the seeds were collected. Before sowing, seeds were washed carefully in distilled water to remove as much surface dirt as possible. Germination experiments were then conducted under four temperature treatments of 15 °C, 20 °C, 25 °C, and 30 °C, in a temperature gradient incubator (TG-180CCFL-5LE, Nippon Medical and Chemical Instruments Co., Ltd., Osaka, Japan). Disposable plastic Petri dishes 90 mm in diameter and 15 mm in depth, in which three sheets of filter paper were laid on the bottom, were used for seed germination. Distilled water was added to the Petri dishes as necessary to prevent the filter papers from drying out.

After sowing, the number of germinated seeds was counted daily and the germinated seeds were then carefully removed from the Petri dishes. Germination experiments were continued for 100 days under 12 h light (06:00–18:00) and 12 h dark (18:00–06:00) conditions. The cumulative germination rate was calculated for each locality, then a mean value was obtained for each of the terrestrial and rheophytic types of S. rubropunctata. The number of days taken to begin germination, and the number of days when the cumulative germination rate exceeded 50% were compared between the terrestrial and rheophytic types in each of the four temperature treatments using the Mann–Whitney U test (P < 0.05). The mean final germination rates of the terrestrial and rheophytic types were compared between the four temperature treatments using the Steel–Dwass test for multiple comparisons (P < 0.05). Because S. indica var. parvifolia seeds were collected from only a single locality, mean values of cumulative germination rates were not obtained for this species.

Seed survival experiment

Even after an incubation period of 100 days, a large proportion of seeds remained ungerminated under certain temperature treatments. To verify whether these were dead or still alive, a seed survival experiment was conducted by changing the incubation temperature of all seeds to 15 °C without changing the filter papers following the end of the germination experiment. This temperature was chosen because it gave the highest germination rates for seeds of both terrestrial and rheophytic types in the germination experiments. The subsequent seed survival experiment was continued for a further 100 days. Mean final germination rates at the end of the seed survival experiment were compared between terrestrial and rheophytic types using the Steel–Dwass test for multiple comparisons (P < 0.05).

Results

Germination experiment

Cumulative germination curves of the terrestrial and rheophytic types of S. rubropunctata in the germination experiment under the four different temperature treatments are shown in Fig. 3. After 100 days of incubation, the final germination rate of the rheophytic type was 93.1 ± 2.1% (mean ± standard deviation) or more at temperatures from 15 to 25 °C (Table 2, Fig. 3a–c), but it decreased to 72.1 ± 13.6% at 30 °C (Table 2, Fig. 3d). In the case of the terrestrial type, the final germination rate was 92.6 ± 4.8% at 15 °C and 84.2 ± 8.8% at 20 °C (Table 2, Fig. 3a, b). However, it was below 50% at 25 °C (41.3 ± 19.3%) and extremely low (2.4 ± 2.4%) at 30 °C (Table 2, Fig. 3c, d). The final germination rates of the rheophytic type were always higher than those of the terrestrial type when compared under the same temperature treatments, although there was no significant difference between them at 15 °C (Fig. 4). The cumulative germination curves of S. indica var. parvifolia are shown in Fig. 5. Changes in the germination rate of S. indica var. parvifolia at 15 °C to 30 °C were similar to those of the terrestrial type of S. rubropunctata as shown in Fig. 4. S. indica var. parvifolia showed the highest germination rate (93.7%) at 15 °C, decreasing to the lowest value (1.67%) at 30 °C (Table 2, Fig. 5).

Fig. 3
figure 3

Cumulative germination curves for S. rubropunctata at 15–30 °C in the germination experiment and at 15 °C in the seed survival experiment. The two experiments were carried out in succession. Mean cumulative germination rates for the rheophytic and terrestrial types of S. rubropunctata seeds are indicated by black and open circles, respectively. Error bars indicate standard deviations

Full size image
Table 2 Final germination rates of S. rubropunctata and S. indica var. parvifolia in germination and seed survival experiments
Full size table
Fig. 4
figure 4

Averaged final germination rates for S. rubropunctata in the germination experiment. R and T indicate the rheophytic and terrestrial types of S. rubropunctata, respectively. Error bars indicate standard deviations. Different letters above the bars denote significant differences (Steel–Dwass, P < 0.05)

Full size image
Fig. 5
figure 5

Cumulative germination curves for S. indica var. parvifolia at 15–30 °C in the germination experiment and at 15 °C in the seed survival experiment. The two experiments were carried out in succession

Full size image

Seeds of both the rheophytic and terrestrial types of S. rubropunctata took fewer days to start germinating as the temperature rose from 15 to 30 °C, except for the terrestrial type at 25 °C (Table 3). Seeds of the rheophytic type started germinating earlier than those of the terrestrial type in all temperature treatments (Table 3). We next compared the number of days when the cumulative germination rate exceeded 50% in the 15 °C and 20 °C treatments where the final germination rate was above 50% between rheophytic and terrestrial types of S. rubropunctata (Table 3). At 15 °C, the rheophytic type took 9.3 ± 1.0 days to exceed 50% germination and the terrestrial type took 15.1 ± 2.4 days. At 20 °C, these times were 6.5 ± 0.6 days and 14.0 ± 3.3 days, respectively. In both cases, the rheophytic type of S. rubropunctata achieved 50% germination in fewer days than the terrestrial type. Scutellaria indica var. parvifolia took 4 days at 25 °C and 7 days at 15 °C and 30 °C to begin germination (Table 3), whereas the cumulative germination rate exceeded 50% during 11 days at 15 °C and 25°C, and nine days at 20 °C (Table 3).

Table 3 The number of days for seeds to begin germination after sowing and for the cumulative germination rate to exceed 50% in S. rubropunctata and S. indica var. parvifolia at different temperatures
Full size table

Seed survival experiment

The cumulative germination curves obtained for S. rubropunctata in the seed survival experiment are shown in Fig. 3 sequentially with the results for the germination experiment. Germination rates of the rheophytic type exceeded 90% at 15 °C and 20 °C and those of the terrestrial type exceeded 90% at 15°C. In the subsequent seed survival experiment, there was thus no change in the final germination rate at 15 °C and 20 °C for the rheophytic type and at 15 °C for the terrestrial type (Table 2, Fig. 3). Seeds of the rheophytic type also showed a high germination rate (93.1 ± 2.1%) at 25 °C in the germination experiment, and only one seed newly germinated in the seed survival experiment, resulting in a slight increase in the final germination rate to 94.2 ± 2.2% (Table 2, Fig. 3). The seeds of the rheophytic and terrestrial types of S. rubropunctata which showed lower germination rates at 30 °C and 20–30 °C, respectively, in the germination experiment increased their rates of germination in the seed survival experiment. The germination rate of the rheophytic type at 30 °C increased to 89.2 ± 3.1% after the incubation temperature was reduced to 15 °C (Table 2, Fig. 3). The seeds of the terrestrial type which had the lowest germination rate at 30 °C in the germination experiment showed a remarkable increase in their germination rate (92.0 ± 6.0%) in the seed survival experiment (Table 2, Fig. 3). Similarly, the seeds of the terrestrial type incubated at 20 °C and 25 °C increased their final germination rate to more than 89% at 15 °C in the seed survival experiment (Table 2, Fig. 3). As a result, there were no significant differences in the final germination rate between all treatments of S. rubropunctata in the seed survival experiment (Fig. 6). In S. indica var. parvifolia, the final germination rates of seeds that showed high germination rates at 15 °C and 20 °C in the germination experiment increased slightly to 94.3% or kept the same value, respectively (Table 2, Fig. 5). Seeds of S. indica var. parvifolia incubated at 25 °C and 30 °C in the germination experiment had a final germination rate of more than 80% at 15 °C in the seed survival experiment (Table 2, Fig. 6).

Fig. 6
figure 6

Averaged final germination rates for S. rubropunctata in the seed survival experiment. R and T indicate the rheophytic and terrestrial types of S. rubropunctata, respectively. Error bars indicate standard deviations. No significant differences were found between all treatment combinations (Steel–Dwass, P > 0.05)

Full size image

Discussion

Seeds of the rheophytic and terrestrial types of S. rubropunctata clearly showed different responses to temperature in the present study. Seeds of the rheophytic type germinated almost simultaneously regardless of the temperature, and reached a high germination rate in a relatively short period of time. In contrast, the terrestrial type showed the highest germination rate at 15 °C, with the rate decreasing as the temperature increased. Interestingly, most seeds of the terrestrial type did not germinate and the final germination rate was extremely low at 30 °C. The germination rate of the rheophytic type also showed a slight decrease at 30 °C, but the suppression of germination because of a temperature increase was more pronounced in the terrestrial type. To our knowledge, this is the first report on intraspecific germination dimorphism in the genus Scutellaria.

In the Ryukyu Islands, including Okinawa-jima Island, rivers often rise because of intense rainfall during the rainy and typhoon seasons. Each time this occurs, rheophytic habitats are exposed to or submerged under swift water currents. Indeed, in the Hiji River where seeds of the rheophytic type of S. rubropunctata were collected, plants growing on rocks along the stream were submerged under high water at least 19 times in 2014 (Yoshimura et al. 2016). In particular, when Typhoon No. 8 (Neoguri) hit Okinawa-jima Island in July 2014, the habitat of the rheophytic type was completely submerged under intense water flow for more than 24 h (Yoshimura et al. 2016). The seeds of the rheophytic type used in this study were collected in July and August, when increased water flow caused by typhoons is most likely.

Germination characteristics of the rheophytic type of S. rubropunctata are considered adaptive in such a severe environment. Scutellaria seeds are generally dispersed by raindrops (ombrohydrochory) (e.g., Nelson and Goetze 2010), which results in a short dispersal distance and seeds tend to be scattered only around parental plants (Nakanishi 2002; Parolin 2006). If seed germination of the rheophytic type of S. rubropunctata is suppressed under high summer temperatures, as is seen in the terrestrial type, most seeds stay close to parental plants without germinating and could be swept away by the rapid water flow caused by strong rainfall. Rivers on Okinawa-jima Island are only 5–10 km in length, with steep gradients and small catchment basin areas of 20–30 km2; thus, most of the rainfall quickly flows out to the sea (Tachihara 2003). It is likely that seeds of S. rubropunctata that are carried by the river reach the sea quickly, consequently wasting the parental investment in reproduction. However, our results show that seeds of the rheophytic type of S. rubropunctata can germinate simultaneously in a short period of time regardless of temperature, suggesting that they may increase their ability to settle quickly, and potentially reduce the risk of being swept away by rising river water. In germination experiments using spores of fern species, the growth of gametophytes and reproductive maturation of rheophytic Osmunda lancea (Osmundaceae) were faster than those of its putative ancestor O. japonica (Hiyama et al. 1992). The shortening of the gametophyte generation in O. lancea may prevent the gametophytes from being swept away by the water flow. The germination characteristics found in the rheophytic type of S. rubropunctata appear to utilize essentially the same strategy.

Germination characteristics of the terrestrial type of S. rubropunctata differ from those of the rheophytic type. Seed germination in the terrestrial type and congeneric S. indica var. parvifolia is strongly suppressed under high temperatures. This appears to be a form of seed dormancy to avoid germination under high temperatures, which may provide an opportunity to delay germination until the temperature reaches optimal conditions for seedling survival. However, the germination suppression seen in terrestrial S. rubropunctata and S. indica var. parvifolia is actually a state of enforced dormancy as defined by Harper (1977). Bewley (1997) defined seed dormancy as the incapacity of intact viable seeds to complete germination even under favorable conditions. On the other hand, in enforced dormancy, limitations in physical environmental factors such as temperature, light, or moisture prevent non-dormant viable seeds from germinating; germination then occurs when the limitations are removed (Baskin and Baskin 2004). In the case of the terrestrial type of S. rubropunctata and S. indica var. parvifolia, the limiting factor appears to be temperature; in our study, those seeds in a state of quiescence under high temperatures were released from germination suppression soon after being transferred to optimal low temperature conditions.

Unlike the germination experiments, the natural environment shows temperature fluctuations. In July and August when seeds of S. rubropunctata were collected for this study, the daily maximum/minimum temperature over the past 3 decades (1981–2010) has ranged from 26.2 to 31.8 °C in the northern part of Okinawa-jima Island (Nago City) (Home page of Japan Meteorological Agency, http://www.jma.go.jp/jma/index.html). In such an environment, it is unlikely that the enforced dormant seeds of terrestrial S. rubropunctata would remain ungerminated throughout the summer season. Instead, the seed response to temperature fluctuations is likely to be more complex and germination may be somewhat asynchronous. This would reduce the risk of disturbance, and has often been seen in plants living in habitats with fluctuating environments (e.g., Harper 1977; Silvertown 1984; Venable 1985). Temperature and humidity at the forest edge or sunny limestone mountain tops and bedrocks, where the terrestrial type of S. rubropunctata often grows, are thought to be more variable than in the forest interior, although studies on this are limited (Washitani and Kabaya 1988; Chen et al. 1993). By varying the timing of germination, terrestrial S. rubropunctata may avoid the risk of seedling destruction if the growing environment suddenly deteriorates. Because S. indica var. parvifolia has similar habitat preference to terrestrial S. rubropunctata (Murata 1957; Arakane and Tsuji 2011), they may share the same germination strategy. However, this should be verified because there is currently insufficient information about seed germinate in the original environment.

Rheophytic species are generally thought to derive from ancestors that grew in terrestrial habitats (Okada and Hotta 1987; Imaichi and Kato 1997). According to this hypothesis, the germination characteristics of rheophytic S. rubropunctata derived from the ancestral terrestrial type, which is supported by the fact that congeneric S. indica var. parvifolia has the same germination characteristics as terrestrial S. rubropunctata. The germination characteristics of rheophytic S. rubropunctata may have differentiated as the plant adapted to its rheophytic habitat, similar to the differentiation seen in halophytes growing in deserts (e.g., Philipupillai and Ungar 1984; Mandák and Pyšek 2001; Wang et al. 2012). However, to the best of our knowledge, the differentiation of germination characteristics accompanying rheophytic adaptation has not been reported in angiosperms. Although studies on adaptation to rheophytic habitats have been carried out in many plant species, most have focused on problems associated with stenophyllization. Research into this has evolved from description of morphological variation to analysis at histological and embryological levels, incorporating genetic findings from model plants (Imaichi and Kato 1997). Little attention has been paid to characteristics other than stenophyllization, and our current findings suggest the need to approach rheophytic adaptation from a wider perspective than before. Traits related to germination and colonization undergo strong selective pressure as they differentiate within a short timescale of just a few hundred years (Angevine and Chabot 1979). Moreover, the adaptive differentiation of germination characteristics may occur in many other rheophytes that occupy niche environments under strong selective pressure.