Introduction
Subfamily Cypripedioideae (Orchidaceae; the slipper orchids) consists of five genera: Selenipedium, Phragmipedium, Mexipedium, Paphiopedilum and Cypripedium (Cox et al., Reference Cox, Pridgeon, Albert and Chase1997). China has the greatest species diversity in Paphiopedilum and Cypripedium (Cribb, Reference Cribb1997, Reference Cribb1998). The plants are adapted to a wide variety of habitats, ranging from high elevations in the Himalayas to lowland tropical regions (Guo et al., Reference Guo, Luo, Liu and Wang2012). Species and hybrids in this subfamily are very popular in floriculture because of their large, peculiar and beautiful flowers, with a pouch-like lip, two fertile stamens, a shield-like staminode and a synsepal composed of fused lateral sepals (Bream, Reference Bream1988; Cribb, Reference Cribb1997, Reference Cribb1998; Averyanov et al., Reference Averyanov, Cribb, Loc and Hiep2003; Koopowitz et al., Reference Koopowitz, Comstock and Woodin2008; Liu et al., Reference Liu, Chen, Chen and Lei2009; Lan and Albert, Reference Lan and Albert2011). The 79 species of Paphiopedilum are distributed mainly in tropical and subtropical forests from Asia to islands of the Pacific, with 27 of them occurring in south-western China, usually on karst limestone hills (Cribb, Reference Cribb1998; Chen et al., Reference Chen, Zhu, Ji, Lang, Luo and Cribb2005; Liu et al., Reference Liu, Chen, Chen and Lei2009). Plants of this genus display three contrasting life forms: terrestrial, facultative epiphytic and obligatory epiphytic (Liu et al., Reference Liu, Chen, Chen and Lei2009). By comparison, Cypripedium contains 51 species that are widely distributed in temperate and subtropical regions (Chen et al., Reference Chen, Liu and Chen2013). The 37 species in China mainly grow at elevations above 1800 m in the south-western region (Cribb, Reference Cribb1997; Chen et al., Reference Chen, Liu and Chen2013). All Cypripedium species are terrestrial (Chen et al., Reference Chen, Liu and Chen2013). Although both genera are closely related (Cox et al., Reference Cox, Pridgeon, Albert and Chase1997), the two exhibit significant differences in distribution, developmental habitat, leaf morphology, physiology and life history strategies (Liu et al., Reference Liu, Chen, Chen and Lei2009; Guo et al., Reference Guo, Luo, Liu and Wang2012; Zhang et al., Reference Zhang, Guan, Sun, Zhang, Cao and Hu2012) thereby making them an ideal system for studying possible ecological adaptations by plants. Their previous taxonomic placement was generally based on floral characters (Liu et al., Reference Liu, Chen, Chen and Lei2009). However, the seed micromorphological traits of Cypripedioideae have received little scientific attention (Arditti et al., Reference Arditti, Michaud and Healey1979) and the ecological significance of those characters remains unclear.
Orchid seeds are extremely small and numerous, and some species produce more than 1 million seeds per capsule (Arditti et al., Reference Arditti, Michaud and Healey1979). They comprise a frequently sculpted seed coat for protection as well as a simple embryo, which can vary significantly in volume but usually lacks an endosperm, the key source of nutrients to support germination (Dressler, Reference Dressler1993; Clements, Reference Clements, Pridgeon, Cribb, Chase and Rasmussen1999; Molvray and Chase, Reference Molvray, Chase, Pridgeon, Cribb, Chase and Rasmussen1999; Swamy et al., Reference Swamy, Kumar, Ramakrishna and Ramaswamy2004). Seeds are usually dispersed by wind, but patterns can vary significantly across genera and species due to morphological characters, size, shape, colour and structure (Molvray and Kores, Reference Molvray and Kores1995; Aybeke, Reference Aybeke2007).
Variations in seed morphology serve as an important source of systematic characters for establishing relationships among species within a genus (Mathews and Levins, Reference Mathews and Levins1986; Ness, Reference Ness1989; Larry, Reference Larry1995). For example, Swamy et al. (Reference Swamy, Kumar, Ramakrishna and Ramaswamy2004) have used scanning electron microscopy to examine the morphologies of ten epiphytic orchid species and have found that the percentage of air space (AS) in seeds differs significantly among species. Chaudhary et al. (Reference Chaudhary, Chattopadhyay and Banerjee2014) have investigated the morphological characteristics of ten Dendrobium species using 13 quantitative trait descriptors, and have reported that, regardless of their phylogenetic associations, species from temperate regions have larger seed volumes and higher ratios of seed volume to embryo volume when compared with species from subtropical or tropical regions. Those researchers have suggested that the primary habitats for more phylogenetically related species are a function of climatic region as their prime habitat. However, the seed morphology in Epipactis is not correlated with biogeography (Arditti et al., Reference Arditti, Clements, Fast, Hadley, Nishimura, Ernst and Arditti1982) and Arditti et al. (Reference Arditti, Michaud and Healey1979) have proposed that the variations in micromorphology among several Cypripedium species could indicate taxonomic relationships. Hence, seed traits such as shape, colour, length and width are perhaps not only useful in taxonomy but might also be correlated with phylogenetic relationships (Clifford and Smith, Reference Clifford and Smith1969; Barthlott, Reference Barthlott1976; Healey et al., Reference Healey, Michaud and Arditti1980). Nevertheless, those aspects have not previously been fully investigated.
Due to their horticultural significance, members within the Orchidaceae family have been over-collected, and their habitats have been lost or destroyed. All species are now listed in the Convention on International Trade in Endangered Species. In general, their dispersal mode and seed size are important factors that govern the establishment of new populations (Tsutsumi et al., Reference Tsutsumi, Yukawa, Lee, Lee and Kato2007). However, little is known about how certain seed traits might be correlated with possible ecological adaptations. Therefore, we investigated the micromorphological characters of 18 Paphiopedilum species and six Cypripedium species. We hypothesized that seed characters in these two genera, such as internal air space and embryo volume, implicated as functional traits for the dispersal of dust seeds, differed between Paphiopedilum and Cypripedium and between life forms within these genera.
Materials and methods
Plant material
Wild plants of 18 Paphiopedilum and six Cypripedium species were collected from their natural environments (Table 1), then cultivated in a greenhouse at Kunming Institute of Botany, Chinese Academy of Sciences (102°41′E, 25°01′N) and pollinated manually during their normal period of flowering. Mature capsules were harvested from these greenhouse-grown plants for conducting our seed measurements. Because most Paphiopedilum species are endangered, it is difficult to collect seeds from their native habitats. Moreover, obtaining all seeds from a single location can reduce the effect of environmental heterogeneity on their development.
Micromorphometric data for seeds and embryo-related characters (mean ± SE) from 18 Paphiopedilum species and six Cypripedium species
Observations of seed morphology
Seeds were stained with safranin before being spread on a slide with a drop of water and protected with a cover slip. Samples were observed and photographed under a light microscope (Olympus-SZX 16, Olympus, Japan). Our morphological parameters included seed length (SL), width (SW), length/width (SL/SW) and volume (SV); embryo length (EL), width (EW), length/width (EL/EW) and volume (EV); and the percentage of air space (AS). All data were analysed with DP2-BSW software (Olympus). Values for SL, SW, EL and EW were recorded from approximately 20 seeds per species. Seed volume was calculated as SV = 2 × (half the width of the seed)2× (half the length of the seed) × π/3. Embryo volume was calculated as EV = 4/3π × (half the length of the embryo) × (half the width of the embryo)2 (Arditti and Ghani, Reference Arditti and Ghani2000). The amount of air space contained by the seed coat was calculated as AS = (seed volume–embryo volume)/(seed volume) × 100% (Arditti and Ghani, Reference Arditti and Ghani2000).
Statistical analysis
Phylogenetic relationships among species of Paphiopedilum and Cypripedium were determined according to the methods described by Cox et al. (Reference Cox, Pridgeon, Albert and Chase1997) and Li et al. (Reference Li, Liu, Salazar, Bernhardt, Perner, Tomohisa, Jin, Chung and Luo2011), respectively. The phylogenetic signal (K statistic) for each trait was calculated using ‘picante’ based on the R package. Such K statistics evaluate the strength of the phylogenetic signal while the P value represents the quantile of the observed phylogenetically independent contrast (PIC) variance versus the null distribution. Traits with P values < 0.05 have non-random phylogenetic signals. Cases where the K value is < 1 indicate convergent traits, whereas K>1 means that the characters are more conserved than would be presumed from a Brownian expectation (Blomberg et al., Reference Blomberg, Garland and Ives2003). Relationships among variables were examined using both pairwise Pearson and PIC correlations. All statistical analyses were performed with R software v. 2.15.0 (R Development Core Team, 2012). Differences in seed traits among life forms were determined by the post-hoc (LSD) test. All of those statistical analyses were performed using SPSS 16 (SPSS, Chicago, Illinois, USA), with measured variables presented as mean ± SE.
Results
Nine morphological characters of orchid seeds and embryos were examined across all studied species (Table 1). Values for SL, SW, SV and AS were significantly smaller for Paphiopedilum species than for Cypripedium species, while those for EL, EW and EV were significantly larger in Paphiopedilum species (Table 2). The ratio of SL to SW did not differ between the two genera (Z= –0.45, P= 0.650).
Micromorphometric data for seeds and embryo-related characters of Paphiopedilum and Cypripedium
A statistical analysis showed that EV and AS differed significantly among life forms for Paphiopedilum species (Fig. 1). Seeds from epiphytic plants had larger embryos and smaller percentages of air space than did terrestrial plants, while values for EV in facultative species were larger than in terrestrial species but smaller than in epiphytic species. The smaller embryos in terrestrial plants resulted in a larger amount of air space for Paphiopedilum seeds. Although AS was significantly smaller for terrestrial Paphiopedilum species than for terrestrial Cypripedium species, embryo volumes did not differ between the two.
Differences in embryo volume (A) and percentage of air space (B) for 18 Paphiopedilum species and six Cypripedium species as function of life form. TP, Terrestrial species of Paphiopedilum, n= 158 plants; FP, facultative species of Paphiopedilum, n= 161; EP, epiphytic species of Paphiopedilum, n= 40; TC, terrestrial species of Cypripedium, n= 123.
Across species, SL and AS showed strong phylogenetic conservatism whereas SW, SV, EL, EW and EV were significantly convergent for all examined species (Table 3). For example, SL was positively correlated with AS. When phylogeny was not considered, EV was negatively correlated with AS (Table 4).
Phylogenetic signal (K) of seed characters in 18 Paphiopedilum species and six Cypripedium species
Correlation of Pearson's regressions (upper right of the diagonal) and phylogenetic independent contrast correlations (lower left of the diagonal) among seed micromorphometric characters
* Correlation is significant at the 0.05 level (two-tailed).
** Correlation is significant at the 0.01 level (two-tailed).
Discussion
Seed micromorphological characters differed significantly among species sampled from Paphiopedilum and Cypripedium orchids. Paphiopedilum plants produce significantly smaller seeds with relatively larger embryos and smaller percentages of air space. The AS variable is an important parameter because the seeds of most orchids are wind dispersed, implying that seeds with higher AS float in the air for a longer time and, thus, can spread to more distant places (Swamy et al., Reference Swamy, Kumar, Ramakrishna and Ramaswamy2004, Reference Swamy, Kumar and Ramaswamy2007). Traits for SL and AS varied significantly across species, with both characters exhibiting clear phylogenetic signals that were evidence of a high level of conservatism and a distinct evolutionary shift among species. However, most of the other characters examined here showed weak signals and were clearly convergent, possibly because of a departure from Brownian motion that might have occurred due to adaptive evolution that would not have been correlated with phylogeny. Few studies of orchid seed traits have incorporated explicit phylogenetic methods (Chemisquy et al., Reference Chemisquy, Prevosti and Morrone2009). Therefore, further research using such comparative approaches should be performed with orchids.
The patterns of seed growth during maturation implied that elongation resulted from the lengthening of testa cells rather than an increase in overall cell numbers, thus improving buoyancy for wind dispersal (Arditti et al., Reference Arditti, Michaud and Healey1980). Whether or not phylogeny was considered, SL in Paphiopedilum and Cypripedium was positively correlated with AS, indicating that those two parameters have an evolutionary association. These findings support the belief that the development and function of SL and AS are coordinated and that both are important for optimizing dispersal, i.e. lighter weights and larger free air space increase the distance that can be travelled (Tsutsumi et al., Reference Tsutsumi, Yukawa, Lee, Lee and Kato2007; Chaudhary et al., Reference Chaudhary, Chattopadhyay and Banerjee2014). Our data showed that Cypripedium species from high elevations – at relatively low atmospheric pressure – produced longer seeds with larger AS values when compared with the Paphiopedilum species. Similar results were obtained when only the terrestrial life forms of each genus were analysed. Thus, we can conclude that the percentage of air space for seeds in subfamily Cypripedioideae (Orchidaceae) may be a function of elevation at which they are produced.
The larger embryos and smaller AS values for our epiphytic Paphiopedilum species versus terrestrial species are in agreement with a report by Tsutsumi et al. (Reference Tsutsumi, Yukawa, Lee, Lee and Kato2007), who used phylogeny and comparative methods to investigate the genus Liparis in Japan. That research group showed that epiphyte seeds with smaller air spaces had lower capacity for dispersal than seeds of terrestrial plants, where the percentage of air space was greater. Other data suggested that life form is also correlated with AS in Paphiopedilum. However, our study indicated that the terrestrial species P. areeanum and P. insigne did not fit closely within this overall habitat-connected classification; both had approximately 32% AS compared with 60–90% for other terrestrial species in that genus. The life form for these two species was quoted from previous taxonomic references, in which P. areeanum (P. rhizomatosum) was described as a new species based on a cultivated plant sampled from northern Myanmar (Chen and Liu, Reference Chen and Liu2002) while P. insigne was recorded in Flora of China although it still lacks a collected specimen (Lang et al., Reference Lang, Chen and Ji1999). No observations were described for the habitats of those two species. Therefore, in the absence of such details, it is possible that they are not obligate terrestrial plants as stated earlier, but instead are facultative epiphytic or epiphytic species.
The large percentage of air space in Orchidaceae is thought to be beneficial to long-distance dispersal because of enhanced seed buoyancy (Arditti and Ghani, Reference Arditti and Ghani2000; Augustine et al., Reference Augustine, Yogendrakumar and Sharma2001; Tsutsumi et al., Reference Tsutsumi, Yukawa, Lee, Lee and Kato2007, Chaudhary et al., Reference Chaudhary, Chattopadhyay and Banerjee2014). The height at which seeds are released from a plant may also be a critical factor that results in greater dispersal distances despite those seeds having smaller/fewer air spaces (Murren and Ellison, Reference Murren and Ellison1998). Compared with epiphytic species, higher AS values in seeds of terrestrial species can help spread them further along the forest floor where wind speeds are lower.
Our morphometric comparisons in Paphiopedilum showed that embryo volumes differed significantly among life forms. That is, seeds of epiphytic species had larger embryos than those of terrestrial species, allowing the former type to germinate earlier (Chaudhary et al., Reference Chaudhary, Chattopadhyay and Banerjee2014). Chase and Pippen (Reference Chase and Pippen1988) have found hooked testa extensions in some epiphytic orchids. However, we did not observe any special extensions on the seed coats of Paphiopedilum epiphytes. At the seedling development stage, adhesive organs, such as rhizoids or emerging roots, facilitate the growth of those seedlings on trees or stones. Consequently, larger embryos may play an important role in seedling establishment by epiphytic species (Chaudhary et al., Reference Chaudhary, Chattopadhyay and Banerjee2014) because they can store more nutrients that provide an advantage in poor environments. Moreover, those larger embryos mean that epiphytic Paphiopedilum species have higher establishment success even though greater seed dispersibility is sacrificed. Furthermore, having smaller AS values may not present a challenge for these epiphytes, because seed dispersal is better from greater heights. All of these characteristics probably reflect a trade-off between seedling establishment and dispersibility that comes from this divergence in embryo size and percentage of air space between epiphytic and terrestrial species.
In summary, both life form and phylogeny have significant effects on embryo volumes and the percentage of air space in Paphiopedilum and Cypripedium. To a certain extent, the importance of seed morphology in Cypripedioideae is associated with the possible ecological adaptations required for seed dispersal or establishment of new populations. More air space means that seeds of terrestrial species can travel further and land where habitats are most suitable for germination. By contrast, larger embryos in epiphytic species can benefit seedling recruitment. We noted a correlated evolution between seed length and percentage of air space in the two genera that implies an evolutionary trade-off between seedling establishment and dispersal within the Cypripedioideae subfamily. Our findings provide new insights into the development of seed micromorphology characteristics related to those two factors for orchids growing under different ecological conditions, and are useful to the understanding of possible ecological significances for seed micromorphology in Cypripedioideae.
Acknowledgements
We are grateful to M.W. Chase and M.A. Chemisquy for their valuable advice and comments on earlier versions of this manuscript and for improving the English. We thank Dr Lee Yungi for the orchid seed exchange.
Financial support
This work was supported by the National Natural Science Foundation of China (31300200, 31170315, 31370362), ‘Light of West China’ Program of the Chinese Academy of Sciences (CAS), the Natural Science Foundation of Yunnan Province (2013FA044) and the Yunnan Social Development Science Program (2007C0001Z).
Conflicts of interest
None.
