INTRODUCTION
Australian species of Euphorbia are represented by seven endemic species in three subgenera, Eremophyton (5 species), Esula (1 species) and Euphorbia (1 species) (Hassall, 1976). Euphorbia boöphthona C. A. Gardner is a perennial species of Euphorbia sect. Eremophyton, and is widely distributed in the Northern Territory and Western Australia (Hassall, 1976). Euphorbia boöphthona, which inhabits a variety of soils, is morphologically distinct from the rest of the Australian sect. Eremophyton species by its thickened tap-root, large, tuberculate seeds, and downward-curving fruits.
The section consists of five Australian endemic species, and has been treated as a natural group (Hassall, 1976). Of the five Australian Euphorbias, E. tannensis Spreng., E. parvicaruncula D. C. Hassall and E. planiticola D. C. Hassall are diploid species with chromosome number 2n = 14 while the remaining species were polyploidy reaching the octoploid (E. boöphthona, 2n = 56) and hexa-/dodecaploid (E. stevenii F. M. Bailey, 2n = 42, 84) (Hassall, 1976). Euphorbia stevenii is distinct from E. boöphthona by its ovoid and ecarunculate seeds, and is found mainly along flooded roadsides and in ponded areas. Polyploidization in sect. Eremophyton species is thought to be an important evolutionary mechanism for adaptation to diverse habitats in inland Australia (Hassall, 1977).
Recent phylogenetic analyses using molecular data have failed to provide consistent results regarding the relationships and placement of E. boöphthona. Zimmermann et al. (2010) hypothesized the monophyly of three diploid species, and proposed the sister-group relationship between E. tannensis and E. boöphthona based on the phylogenetic analyses using nuclear ribosomal internal transcribed spacer and trnL-trnF IGS data. However, phylogenetic trees using internal transcribed spacer (ITS), ndhF, and matK sequences did not support the monophyly of five Australian species, and placed them two independent subgenera, Euphorbia and Chamaesyce (Yang et al., 2012; Dorsey et al., 2013). In particular, E. boöphthona was included in subg. Euphorbia sect. Pacificae together with E. stevenii, suggesting a close relationship between the two species (Dorsey et al., 2013).
Recently, a genetic study was conducted to identify the ancestor of E. boöphthona using isozyme data. The results of this study suggested that E. stevenii is the ancestor of E. boöphthona, and the remaining ancestors may be unknown or extinct (Park, 2022).
The approximately 2,000 species belonging to the genus Euphorbia have a similar cyathium structure, so they have recently tended to be grouped together as a single genus Euphorbia. However, the seed characters of Euphorbia are very diverse, providing useful information for classifying the genus as several sections and elucidating the relationships among species (Park, 2000; Jin and Park, 2008; Na and Park, 2010; Salmaki et al., 2011; Pahlevani and Akhani, 2011; Pahlevani et al., 2015). This study aimed to test previously proposed conflicting hypotheses about the origin and relationship of E. boöphthona using seed morphological data of five Australian Euphorbia species.
MATERIALS AND METHODS
This study is based on the plant materials collected from the natural habitats of five Australian Euphorbia species. To investigate seed coat and sculpturing and caruncle morphology using SEM, mature seeds were washed with acetone, dried in the air, sputter-coated with Au-Pd for 60 seconds under 0.1 Torr, and photographed with SEM (field emission scanning electron microscope: S-4200). The length of the seeds was measured under a dissecting microscope (Olympus SZ-LGR, Tokyo, Japan). More than 10 seeds per species were selected and measured, and the average was calculated, and the maximum and minimum values were also recorded. The shape of the caruncle was also observed under a dissecting microscope, and compared with photographs taken with a SEM. Collection data and voucher deposition for Euphorbia seed sources were listed in Table 1.
RESULTS
The seed morphological characters of five Euphorbia species were measured and presented in Table 2, and the seed images taken by SEM are presented in Figs. 1 and 2. An identification key based on the seed characters is provided.
According to the measurements, the largest seed is those of E. stevenii (2.60–3.20 mm long, 1.60–3.20 mm wide), and the smallest seeds are those of E. planiticola (2.00–2.30 mm long, 1.40–1.60 mm wide) (Table 2). The seeds of E. planiticola and E. stevenii are ovate in shape, while the other species are quadrangular. In most species, the apex of the seed is pointed and the base is flat. Except for E. stevenii, the dorsal view of the seeds of the remaining species is vertically grooved, and the ventral view is distinctly raised (Fig. 1).
Except for E. stevenii (Fig. 1E), the remaining four species have a caruncle at the tip of the seed. Euphorbia stevenii has no caruncle, and only traces of the stalk of the caruncle remain under SEM. Comparing the shape of the caruncle, E. boöphthona and E. tannensis have a deeply folded hat shape (Fig. 1C, F), but E. planiticola and E. parvicaruncula have a crescent-shaped caruncle (Fig. 1B, D). In particular, the stalk of the caruncle appears very short in E. planiticola, but is distinct in E. tannensis and E. parvicaruncula (Fig. 1B, D, F).
The surface ornamentation of seeds observed by SEM showed that E. parvicaruncula and E. tannensis were similar in seed surface pattern that they were tuberculate (Fig. 1D, F), but the other species showed very diverse forms. Testa cells of E. tannensis and E. parvicanuncula are circular in shape and have triangular intercellular spaces filled with particles (Fig. 2E, F).
Euphorbia. boöphthona had a mamillate form with papillae developed along the slits (Fig. 2B, C). In particular, one papilla develops for each testa cell and is densely distributed between the protruding round projections (mamilla) on the surface (Fig. 2C).
The seed surface of E. stevenii shows papillate characteristics, with each protruding projection being formed by the aggregation of several testa cells (Figs. 1E, 2A). The testa cells of E. stevenii seeds are arranged so densely that no intercellular spaces are observed.
Euphorbia planiticola showed an overall reticulate form with prominent ridges developed on the surface (Figs. 1A, B, 2D). The thin ridges on the seed surface are arranged in layers, forming a reticulate pattern (Fig. 1A).
DISCUSSION
Testing the monophyly of five Australian species of sect. Eremophyton
Although species traditionally classified as sect. Eremophyton (= subg. Eremophyton) within genus Euphorbia have recently been shown to be not a monophyletic group by molecular phylogenetic studies (Park and Jansen, 2007), Australian sect. Eremophyton, consisting of five endemic species, is considered to be a natural taxon derived from a common ancestor with basic chromosomal number x = 7 (Hassall, 1977). The above hypothesis that this taxon is a natural taxon is plausible because the Australian species formed a cluster in another study based on phenetic analysis (Hassall, 1976).
Despite the many phylogenetic studies on Euphorbia species using molecular data (Zimmermann et al., 2010; Yang et al., 2012; Dorsey et al., 2013), the species relationships or monophyly of the Australian Eremophyton species have not been clearly identified. The main reason for this result is that no study included all five species of Australian Eremophyton, and each analysis included different species combinations, which inevitably limits the ability to accurately identify the relationships among species. Among these, an interesting study that could verify the monophyly of Australian Eremophyton was the study by Dorsey et al. (2013), in which E. tannensis and E. platinicola, which were presumed to be diploids, were included in the subg. Chamaesyce clade, while polyploid E. boöphthona and E. stevenii were included in the subg. Euphorbia clade, suggesting that they are not a monophyletic group (Fig. 3A).
Phylogenetic analyses using different molecular markers have also shown inconsistent results, so the phylogenetic relationships of the five Australian species have not been clearly proposed to date. Trees generated using ITS + ndhF + matK (Dorsey et al., 2013, Fig. 3A; Yang et al., 2012, Fig. 3C) support the sister-group relation between E. tannensis and E. planiticola while the relationship based on ITS+trnL-trnF data (Zimmermann et al., 2010) placed E. boöphthona as the most closely related species of E. tannensis (Fig. 3B).
Although it is not a phylogenetic tree, the relationship analysis using genetic distance obtained through allozyme analysis revealed that E. tannensis and E. parvicaruncula are the most closely related species within the Australian sect. Eremophyton (Fig. 3D).
Within subg. Eremophyton, seed morphology has been used as a criterion for dividing lower taxa such as sections, and in particular, the presence or absence of a caruncle, the shape of the caruncle, and the presence or absence of horizontal constriction on the seed surface have been used as important characters (Gilbert et al., 1993). Nevertheless, these seed characters are highly diverse within the Australian sect. Eremophyton, which consists of only five species, raising the suspicion that these species may not belong to a single natural taxon. Therefore, considering the seed heterogeneity within the Australian sect. Eremophyton, it can be said that it supports the polyphyletic origin suggested by Dorsey et al. (2013).
Considering the seed characters resulting from this study, E. tannensis and E. parvicaruncula, which share the characteristics of a quadrangular shape and a tuberculate seed surface, can be considered to be very closely related. It consists of the results of allozyme analysis (Park, 2022), while it cannot support the previous interpretation that E. planiticola is closely related to E. tannensis (Yang et al., 2012; Dorsey et al., 2013). In the case of E. planiticola, the seed shape is oval and has a reticulate form with unique ridges developed on the surface, showing significant differences from E. tannensis.
Re-examining the hypothesis of E. stevenii as one of the ancestors of octoploid E. boöphthona
Hassall (1976, 1977) studied the chromosomes of E. boöphthona and found that this species was an octoploid with 2n = 56. However, since tetraploids were not found in Australia, he could not clearly state which species was involved in the speciation of E. boöphthona. However, it was assumed that polyploidization served as a key driving force enabling adaptive radiation of Australian species of sect. Eremophyton.
Recently, a genetic study was conducted to find the ancestor of E. boöphthona using allozyme analysis, and the results of this study revealed that E. boöphthona and E. stevenii share the same alleles in many genetic loci, and based on this, it suggested that E. stevenii, an hexaploid species (2n = 42), may be an ancestor of E. boöphthona (Park, 2022). The fixed heterozygote bending pattern and unique alleles present in E. boöphthona have led to the hypothesis that this species speciated through an allopolyploidization process, with one ancestor involved in the origin of this species being E. stevenii and the other ancestor being unknown or possibly extinct. A recent phylogenetic study (Dorsey et al., 2013) (Fig. 3A) showed that E. boöphthona and E. stevenii do not form a clade with the existing species belonging to the Australian section Eremophyton, but rather shows a close relationship with species belonging to subg. Euphorbia, suggesting that they may be independently evolved. Based on this phylogeny, Dorsey et al. (2013) newly designated sect. Pacificae, which includes 11 species from subg. Euphorbia including E. boöphthona, E. stevenii, E. haeleeleana, E. plumerioides and E. sarcostemmoides. In particular, E. plumerioides and E. sarcostemmoides are Australian species and are likely to be unknown ancestors of E. boöphthona based on their geographical relatedness.
On the base of seed characters observed in this study, E. boöphthona, having a mamillate form with papillae developed along the slits, is unique characteristics, but shared the papillate surface patterns with E. stevenii. Our results on seed morphology support previous genetic studies (Park, 2022) suggesting that E. stevenii is one parent of E. boöphthona, and that the other parent may be extinct or unknown species from sect. Pacificae that shares some of the characteristics of E. boöphthona, such as carunculate and quadrangular seeds.
This study clearly demonstrated that seed morphology can provide important evidence for testing interspecific relationships and identifying ancestral species.
PDF Links
PubReader
ePub Link
Full text via DOI
Download Citation 
Print 
