Phylogeny and staminal evolution of Salvia (Lamiaceae, Nepetoideae) in East Asia

Guo-Xiong Hu; Atsuko Takano; Bryan T Drew; En-De Liu; Douglas E Soltis; Pamela S Soltis; Hua Peng; Chun-Lei Xiang

doi:10.1093/aob/mcy104

. 2018 Jun 25;122(4):649–668. doi: 10.1093/aob/mcy104

Phylogeny and staminal evolution of Salvia (Lamiaceae, Nepetoideae) in East Asia

Guo-Xiong Hu ^1,², Atsuko Takano ³, Bryan T Drew ⁴, En-De Liu ¹, Douglas E Soltis ^5,⁶, Pamela S Soltis ⁵, Hua Peng ^1,^✉, Chun-Lei Xiang ^1,^✉

PMCID: PMC6153483 PMID: 29945172

Abstract

Background and Aims

Salvia is the largest genus within Lamiaceae, with about 980 species currently recognized. East Asia, with approx. 100 species, is one of the three major biodiversity centres of Salvia. However, relationships within this lineage remain unclear, and the staminal lever mechanism, which may represent a key innovation within the genus, has been understudied. By using six genetic markers and nearly comprehensive taxon sampling, this study attempts to elucidate relationships and examine evolutionary trends of staminal development within the East Asia (EA) Salvia clade.

Methods

Ninety-one taxa of EA Salvia were sampled and 34 taxa representing all other major lineages of Salvia were included for analysis. Two nuclear [internal transcribed spacer (ITS) and external transcribed spacer (ETS)] and four chloroplast (psbA–trnH, ycf1–rps15, trnL–trnF and rbcL) DNA markers were used for phylogenetic analysis employing maximum parsimony (MP), maximum likelihood (ML) and BEAST, with the latter also used to estimate divergence times.

Key Results

All Salvia species native to East Asia form a clade, and eight major subclades (A–G) were recognized. Subclade A, comprising two limestone endemics (S. sonchifolia and S. petrophila), is sister to the remainder of EA Salvia. Six distinct stamen types were observed within the EA clade. Stamen type A, with two fully fertile posterior thecae, only occurs in S. sonchifolia and may represent the ancestral stamen type within EA Salvia. Divergence time estimates showed that the crown of EA Salvia began to diversify approx. 17.4 million years ago.

Conclusions

This study supports the adoption of a broadly defined Salvia and treats EA Salvia as a subgenus, Glutinaria, recognizing eight sections within this subgenus. Stamen type A is ostensibly plesiomorphic within EA Salvia, and the other five types may have been derived from it. Staminal morphology has evolved in parallel within the EA Salvia, and staminal structure alone is inadequate to delimit infrageneric categories.

Keywords: Salvia, phylogeny, staminal evolution, stamen movement, Mentheae, Salvia sonchifolia, Salvia plebeia, subg. Glutinaria, sect. Sobiso

INTRODUCTION

Salvia, with about 980 species and a nearly cosmopolitan distribution, is the largest genus in the angiosperm family Lamiaceae (Alziar, 1988–1993; Walker et al., 2004; Wei et al., 2015; Drew et al., 2017; Will and Claßen-Bockhoff, 2017). The genus has undergone major species radiations in Mesoamerica/South America (at least 500 spp.), south-western Asia and the Mediterranean region (approx. 250 spp.), and Eastern Asia (approx. 100 spp.) (Alziar, 1988–1993; Walker and Sytsma, 2007). The genus is utilized throughout its range for medicinal purposes, and many species are of economic importance. For instance, Salvia miltiorrhiza (‘Danshen’ in Chinese), endemic to China, is a traditional Chinese medicine that is widely used to treat cardiovascular and cerebrovascular diseases and hyperlipidaemia (Wang, 2010). Salvia hispanica, an important Mesoamerican staple food and medicinal plant in pre-Columbian times, is now available commercially worldwide as a ‘superfood’ (Ali et al., 2012). Additionally, at least 150 species are widely marketed in the horticultural trade (Clebsch, 2008).

Salvia has traditionally been distinguished from other genera of Lamiaceae by an unusual morphological character in which two fertile stamens are separated by a significantly elongated connective tissue. Based on calyx, corolla and stamen morphology, Bentham (1832–1836) first established an infrageneric classification for Salvia. His treatment placed the 266 species known at the time into 14 sections. In subsequent updates, he classified 406 Salvia species into 12 sections (Bentham, 1848), and eventually these sections were organized into four subgenera (Bentham, 1876), thus forming the first comprehensive subgeneric classification of Salvia. Subsequently, Briquet (1897) provided an updated global synopsis of the genus, recognizing eight subgenera and 17 sections. These two comprehensive classifications were subsequently modified by various authors (e.g. Stibal, 1934; Epling, 1938, 1939; Pobedimova, 1954; Wu, 1977; Murata and Yamazaki, 1993).

Due to tremendous diversity in habit, floral morphology and staminal morphology across Salvia, infrageneric boundaries within the genus have been notoriously difficult to define (Bentham, 1876; Briquet, 1897; Stibal, 1934; Pobedimova, 1954; Wu, 1977; Murata and Yamazaki, 1993; Drew et al., 2017; Will and Claßen-Bockhoff, 2017). To avoid troublesome issues resulting from subgeneric circumscriptions, some researchers adopted ‘species-groups’ or sections (often monotypic) rather than explicitly defining subgenera (e.g. Epling, 1939; Hedge, 1974, 1982a, b).

Recent molecular phylogenetic studies, however, have demonstrated that traditionally defined Salvia is non-monophyletic, as five genera (Dorystaechas, Meriandra, Perovskia, Rosmarinus and Zhumeria), most lacking an elongated connective as in traditionally defined Salvia, are embedded within it (Walker et al., 2004, 2015; Walker and Sytsma, 2007; Drew and Sytsma, 2011, 2012; Takano and Okada, 2011; Jenks et al., 2012; Li et al., 2013; Will and Claßen-Bockhoff, 2014, 2017; Drew et al., 2017; Fragoso-Martínez et al., 2018). Two competing ideas regarding circumscription of the genus have recently been proposed. One option is to name the five embedded genera as Salvia and maintain Salvia in a broad (although slightly expanded) sense (González-Gallegos, 2015; Drew et al., 2017). The other is to break up Salvia into several smaller genera (Will et al., 2015; Will and Claßen-Bockhoff, 2017) and maintain the names of the five embedded genera. Will et al. (2015), advocating the latter option, transferred 14 Salvia species (distributed from south-west Asia to northern Africa) to the resurrected genus Pleudia. Later, based on expanded phylogenetic sampling of Mediterranean Salvia, Will and Claßen-Bockhoff (2017) suggested splitting Salvia into six genera [i.e. Salvia sensu stricto (s.s.). Lasemia, Ramona, Glutinaria, Pleudia and Polakia] and retaining the generic status of the five embedded genera. Will and Claßen-Bockhoff (2017) did not provide formal taxonomic treatments, but offered suggestions for future nomenclatural revisions. Conversely, based on phylogenetic, taxonomic, morphological and practical considerations, Drew et al. (2017) treated the five embedded genera as subgenera within Salvia (subg. Dorystaechas, subg. Meriandra, subg. Perovskia, subg. Rosmarinus and subg. Zhumeria) to maintain a broadly defined Salvia, and provided nomenclatural revisions for the 15 species belonging to the five embedded genera. Based upon the taxonomic treatment suggested by Will and Claßen-Bockhoff (2017), about 750 Salvia species would be transferred to the resurrected genera Glutinaria, Lasemia, Ramona, Pleudia and Polakia. Consequent nomenclatural changes would lead to confusion in other subjects such as horticulture, ecology and phytochemistry. Furthermore, the boundaries between the genera advocated by Will and Claßen-Bockhoff (2017) are not morphologically distinct, which would lead to ongoing taxonomic confusion. Inclusion of the five embedded genera in a broadly defined Salvia is phylogenetically legitimate and will probably be accepted by botanists as well as workers in other disciplines. Therefore, we prefer to maintain a broadly defined Salvia following Drew et al. (2017).

The East Asian (EA) radiation of Salvia comprises 82 species native (72 endemic) to China, 12 species native (nine endemic) to Japan and three species native (one endemic) to the Korean Peninsula (Murata and Yamazaki, 1993; Li and Hedge, 1994; Lee, 2004; Hu et al., 2013, 2014, 2017; Takano et al., 2014; Hu and Peng, 2015; Xiang, 2016; Xiang et al., 2016a). The EA Salvia are highly diverse in terms of root, leaf, calyx, corolla and staminal morphology and habitat (Fig. 1). Additionally, all EA Salvia are herbaceous, contrasting with the other two major centres of diversity where shrubs are common. Based on staminal morphology, EA Salvia have been classified into three subgenera: subg. Salvia, subg. Sclarea and subg. Allagospadonopsis (Wu, 1977; Murata and Yamazaki, 1993).

Fig. 1. — Morphological diversity of EA *Salvia*. 1–4, Root morphology (1, *S. castanea*; 2, *S. miltiorrhiza*; 3, *S. plectranthoides*; 4, *S. cavaleriei* var. *erythrophylla*). 5–10, Leaf morphology (5, *S. sonchifolia*; 6, *S. luteistriata*; 7, *S. wardii*; 8, *S. prionitis*; 9, *S. bowleyana*; 10, *S. japonica*). 11–13, Bract morphology (11, *S. scapiformis*; 12, *S. trijuga*; 13, *S. atropurpurea*). 14–17, Calyx morphology (14, *S. scapiformis*; 15, *S. sonchifolia*; 16, *S. substolonifera*; 17, *S. hylocharis*). 18–24, Corolla diversity (18, *S. sonchifolia*; 19, *S. miltiorrhiza*; 20, S. honania; 21, *S. japonica*; 22, *S. liguliloba*; 23, *S. campanulata*; 24, *S. prattii*). 25–34, Stamen diversity (25, *S. sonchifolia*; 26, *S. luteistriata*; 27, *S. plebeia*; 28 and 29, *S. plectranthoides*; 30, *S. petrophila*; 31, *S. bowleyana*; 32, *S. honania*; 33, *S. cavaleriei*; 34, *S. scapiformis*). Photographs: 12–13 by E. D. Liu, 24 by Y. P. Chen, 29 by X. X. Zhu, others by G. X. Hu.

Early molecular phylogenetic studies, focusing on Salvia as a whole (Walker et al., 2004; Walker and Sytsma, 2007), indicated that EA Salvia may represent an independent lineage, but these studies had sparse sampling within EA Salvia. Subsequently, two phylogenetic studies focusing on EA Salvia have been conducted. Using three DNA makers [rbcL, trnL–trnF and an internal transcribed spacer (ITS)], Takano and Okada (2011) first reported that the 11 species of Salvia from Japan are monophyletic. Later, based upon four DNA markers (psbA–trnH, rbcL, matK and ITS), Li et al. (2013) inferred phylogenetic relationships of 37 Chinese and four Japanese Salvia species and found that the Chinese (except for S. deserta, widespread in Xinjiang Province of China, Russia, Kyrgyzstan and Kazakhstan) and Japanese species of Salvia formed a clade. Unfortunately, due to limited sampling and misidentification of a few key species (this remark is based on results presented here), the phylogenetic backbone of EA Salvia was not resolved.

Additionally, Li et al. (2013) suggested discordance between phylogenetic results and traditional classifications, although morphological evidence was not provided. Recently, Will and Claßen-Bockhoff (2017), based on ITS sequences, recognized three clades of EA Salvia. Most of the the 46 EA Salvia ITS sequences used by Will and Claßen-Bockhoff (2017) were from GenBank. However, based on phylogenetic analyses presented here, we found that some ITS sequences (e.g. sequences of S. plectranthoides 247, S. chienii DQ123828# and S. evansiana FJ593405#) used in Will and Claßen-Bockhoff (2017) were questionable, which may have resulted from species misindentification and/or GenBank uploading errors.

Traditionally, Salvia has been defined largely by a lever-like staminal feature that is formed by elongate connectives and filaments (Himmelbaur and Stibal, 1932–1934; Claßen-Bockhoff et al., 2004a). Staminal morphology in Salvia is highly diverse, and up to 11 distinct stamen types have been described within traditionally defined Salvia (Himmelbaur and Stibal, 1932–1934; Claßen-Bockhoff et al., 2003, 2004a, b; Walker and Sytsma, 2007; Will and Claßen-Bockhoff, 2014; Walker et al., 2015). Himmelbaur and Stibal (1932–1934) first hypothesized parallel evolution of the lever mechanism within Salvia, and this has been repeatedly corroborated by molecular phylogenetic studies (Walker et al., 2004, 2015; Walker and Sytsma, 2007; Will and Claßen-Bockhoff, 2014; Drew et al., 2017).

Wu (1977) recognized three distinct stamen types for EA Salvia species: (1) connectives ± curved, upper arms longer than or equal to lower arms, posterior thecae fertile, fused; (2) connectives ± straight, not curved, posterior thecae sterile, fused; and (3) connectives ± straight, not curved, posterior thecae sterile, separated. Based on the staminal morphology alone, Wu (1977) classified EA Salvia into three subgenera (subg. Salvia, Sclarea and Allagospadonopsis). Based on floral morphology and Wu’s (1977) treatment, Huang et al. (2014) hypothesized an evolutionary trend of stamen types for EA Salvia, from a ‘short-lever type’ (subg. Salvia), to a ‘long-lever type’ (subg. Sclarea), to a ‘degraded-lever type’ (subg. Allagospadonopsis). However, while preparing this paper, we observed additional stamen types within EA Salvia and found that the stamen types defined by Wu (1977) are too general to describe stamen morphology accurately within EA Salvia [see Fig. 1 (25–34)]. Additionally, the three subgenera circumscribed by Wu (1977) have been demonstrated to be non-monophyletic. The evolutionary trajectory of stamen types for EA Salvia therefore needs to be re-evaluated. Additionally, it remains unclear whether parallel evolution of staminal morphology has occurred within EA Salvia, as sampling within EA Salvia was limited in previous phylogenetic research regarding staminal evolution (Walker and Sytsma, 2007; Will and Claßen-Bockhoff, 2014).

In this study, based on the most comprehensive geographic, taxonomic and genetic sampling to date, we reconstruct the phylogeny of EA Salvia and clarify inter-relationships within the group. Furthermore, we present a detailed summary of staminal morphology within EA Salvia and elucidate evolutionary trends in staminal morphology in the context of a phylogenetic framework. Finally, based on phylogenetic and morphological considerations, we provide an updated taxonomic treatment for EA Salvia.

MATERIALS AND METHODS

Nomenclature and taxon sampling

Names of Salvia subgenera and major clades follow Drew et al. (2017). A total of 91 taxa, representing 78 species, ten varieties and three forms from China, Japan and the Korean Peninsula, were sampled for this study. Except for the monotypic sect. Aethiopis and ser. Piasezkianae, our sampling represents all subgenera, sections and series sensuWu (1977) and Murata and Yamazaki (1993). In addition, 33 species of Salvia from other major lineages of Salvia (subg. Calosphace, Audibertia, Dorystaechas, Meriandra, Perovskia, Rosmarinus, Zhumeria, the ‘S. aegyptiaca clade’ and the ‘S. officinalis clade’) were sampled. Melissa and Lepechinia from subtribe Salviinae were also sampled. Horminum and Hedeoma were selected as outgroups based on Drew and Sytsma (2012). In total, our data set included 172 accessions, of which 145 were newly generated for this study (see Supplementary Data Appendix). Although sequences of many Chinese taxa are available from GenBank, we sampled and identified all Chinese taxa used in this study, and all sequences of Chinese taxa used here were independently produced to ensure independent results and accuracy.

DNA extraction, amplification and sequencing

Total genomic DNA was extracted from silica gel-dried leaf materials or herbarium specimens (S. brachyloma, S. filicifolia and S. potaninii) using the modified cetyltrimethylammoniun bromide (CTAB) method of Doyle and Doyle (1987). Four chloroplast DNA markers (psbA–trnH, ycf1–rps15, trnL–trnF and rbcL) and two nuclear regions [the ITS and the external transcribed sequence (ETS)] were selected for phylogenetic analyses. Primers used in this study are listed in Table 1.

Table 1.

List of primers used in this study (F represents a forward primer and R represents a reverse primer)

DNA markers	Sequences (5′–3′)⁾	References
ITS	ITS5 (F): GGAAGTAAAAGTCGTAACAAGG	White et al. (1990)
	ITS4 (R): TCCTCCGCTTATTGATATGC	White et al. (1990)
	ITSA (F): GGAAGGAGAAGTCGTAACAAGG	Blattner (1999)
	ITSB (R): CTTTTCCTCCGCTTATTGATATG	Blattner (1999)
ETS	ETS-bdf1 (F): GTGAGTGGTGKTTGGCGYGT	Drew and Sytsma (2011)
	18S-IGS(R): GAGACAAGCATATGACTACTGGCAGGATCAACCAG	Baldwin and Markos (1998)
	ETS-B (F): ATAGAGCGCGTGAGTGGTG	Beardsley and Olmstead (2002)
	18S-E (R): GCAGGATCAACCAGGTAGCA	Baldwin and Markos (1998)
psbA–trnH	psbAF (F): GTTATGCATGAACGTAATGCTC	Sang et al. (1997)
	trnHR (R): CGCGCATGGTGGATTCACAAATC	Sang et al. (1997)
rbcL	Z1F (F): ATGTCACCACAAACAGAAACTAAAGCAAGT	Soltis et al. (1992)
	Z1351R(R): CTTCACAAGCAGCAGCTAGTTCAGGACTCC	Soltis et al. (1992)
trnL–trnF	trn-c (F): CGAAATCGGTAGACGCTACG	Taberlet et al. (1991)
	trn-f (R): ATTTGAACTGGTGACACGAG	Taberlet et al. (1991)
ycf1–rps15	ycf1 5711f (F): CTTGTATGRATCGTTATTGKTTTG	Drew and Sytsma (2011)
	ycf1 rps15r (R): CAATTYCAAATGTGAAGTAAGTCTCC	Drew and Sytsma (2011)

Open in a new tab

The standard 25 μL PCR mixtures contained 1 μL of each primer (10 μm, Sangon Biotechnology, Shanghai, China), 2.5 μL of 10× reaction buffer (Mg²⁺ free), 2.5 μL of dNTP mixture, 1.5 μL of MgCl₂, 0.3 μL of Taq polymerase (2.5 U μL^–1, Tiangen Biotech, Beijing, China), 2 μL of unquantified template DNA, 1 μL of bovine serum albumin (BSA, 20 mg ml^–1) and deionized water added to achieve a final volume of 25 μL. Amplification for all six markers was performed as follows: an initial denaturation at 94 °C for 4 min, followed by 35 cycles of 30 s denaturation (94 °C), 90 s annealing (50 °C) and 2.5 min extension (72 °C), ending with a final extension at 72 °C for 7 min.

Amplification products were checked on 1 % TAE agarose gels and purified using the QIAquick PCR purification kit (BioTeke, Beijing, China) following the manufacturer’s instructions. Sequencing reactions were performed with the dideoxy chain termination method running on an ABI-PRISM-3730 automated sequencer (Sangon Biotechnology). Sequencing primers for DNA markers were the same as for the PCR primers.

Sequence alignment and phylogenetic analyses

Sequences were checked and assembled using Sequencher v.4.1.4 (Gene Codes, Ann Arbor, MI, USA). Alignments were initially performed using MUSCLE (Edgar, 2004) as implemented in MEGA v.6.0 (Tamura et al., 2013) and then manually adjusted using PhyDE v.0.9971 (Müller et al., 2010). Two separate matrices [combined nuclear ribosomal DNA (nrDNA) and combined chloroplast DNA (cpDNA)] were used for phylogenetic analyses using the following approaches: maximum parsimony (MP), maximum likelihood (ML) and Bayesian inference (BEAST). To assess incongruence between the combined nrDNA and combined cpDNA, we compared the resultant trees from the two genetic compartments for supported incongruence, and also performed the incongruence length difference (ILD) test (Farris et al., 1994) as implemented in PAUP* version 4.0b10 (Swofford, 2003). While Yoder et al. (2001) suggested that the ILD test was not useful in testing data partition compatibility, Hipp et al. (2004) demonstrated this argument to be unconvincing and pointed out that the ILD test has value as a tool for assessing data partition congruence.

The MP analyses were performed using PAUP* v.4.0b10 (Swofford, 2003) with the following settings: heuristic search option, tree-bisection-reconnection (TBR) branch swapping with 1000 random sequence addition replicates and ten trees saved per replicate. All characters were unordered and equally weighted; gaps were treated as missing data. A strict consensus tree was summarized from all of the most parsimonious trees retained. Bootstrap support values were calculated from 1000 rapid bootstrap replicates, with each comprising ten random sequence addition replicates, with only one tree saved per replicate.

Partitioned ML analyses were performed using RAxML-HPC2 on XSEDE v.8.1.11 (Stamatakis, 2014) as implemented on the CIPRES computer cluster (http://www.phylo.org/) (Miller et al., 2010). The GTRCAT model was used for analyses and bootstrapping; bootstrap iterations (–#/–N) was set to 1000, and other parameters followed default settings. All trees were visualized using TreeGraph v.2 (Stöver and Müller, 2010).

Divergence time estimation

Divergence time analyses were performed using BEAST v.1.8.3 (Drummond et al., 2012) as implemented on the CIPRES computer cluster (http://www.phylo.org/) (Miller et al., 2010). Dates were estimated using a lognormal relaxed molecular clock and the Yule model of speciation. Models of nucleotide evolution were evaluated with jModelTest2 (Darriba et al., 2012) using the Akaike information criterion (AIC).

Although divergence times of the Salvia crown have been investigated previously (e.g. Drew and Sytsma, 2012; Drew et al., 2017), sampling within EA Salvia was limited, and consequently the divergence times of the main clades of EA Salvia are not clear. Here, based on broad sampling of EA Salvia, we employed a two-step scheme to estimate divergence times within Salvia independently. For the initial step, we used the Lamiaceae-wide data set from Drew and Sytsma (2012) and used two constraint strategies. (1) We used the minimum (48.3) and maximum (71.9) Lamiaceae crown dates from Yao et al. (2016) as an age constraint for the Lamiaceae crown. This constraint was applied using a truncated normal prior (with the aforementioned minimum and maximum dates), a mean of 60.1 and an s.d. of 15. This approach approximates a uniform distribution, but allows for a slightly higher prior probability in the centre of the curve relative to the edges. (2) We combined the above-described and the constraint strategy used by Drew and Sytsma (2012).

In the second step, we used the 95 % height probability distribution (HPD) from the Lamiaceae-wide analyses as a constraint for the root of the tree and for the crown of Salvia. We constrained the root of the tree using a truncated normal distribution, with a lower limit of 26.6 and an upper limit of 46.3, a mean of 36.45 and an s.d. of 13. We constrained the crown of Salvia with a uniform distribution that had a lower age of 18.5 and an upper age of 34.1.

For both Lamiaceae-wide analyses we conducted two separate runs of 60 million generations and saved samples every 5000 generations. After assessing results in Tracer v1.6 (Rambaut et al., 2014), we discarded the first 6 million generations as burn-in. For the Salvia data set analyses, we conducted two separate runs of 100 million generations, with samples saved every 5000 generations. After assessing the results in Tracer, we discarded the first 10 % of the trees as burn-in. The log files were checked for convergence using Tracer. In both steps of our analyses, all ESS (explained sum of squares) values were well over 200; trees from separate runs were combined with LogCombiner v 1.8.4 and summarized with Tree Annotator v. 1.8.4 (both included in the BEAST package), and the chronogram was visualized using FigTree v. 1.4.2 (Rambaut, 2014).

Stamen morphology

Staminal morphology of most EA Salvia was observed based on fresh specimens collected in the field, with a few observations based on herbarium specimens (Salvia brachyloma, S. filicifolia and S. potaninii). Stamen type was summarized based on observation results. To better understand staminal morphology of EA Salvia, we provide stamen type schematics. In terms of morphology and size, there is no major differences among species with the same stamen type, and therefore each type is illustrated referring to a single species.

Walker and Sytsma (2007) listed 14 stamen types within Salvia and named them using Latin upper case letters ranging from A to N, in which only one stamen type (N) was described from EA Salvia. When Will and Claßen-Bockhoff (2014) showed stamen diversity of African Salvia, the letters A, B and C were used, but they represented different stamen types as compared with Walker and Sytsma (2007). Here, we used six upper case letters (A–F) to distinguish six distinct stamen types of EA Salvia. The naming system is independent of the already mentioned systems.

RESULTS

Alignment and phylogenetic reconstruction

After inspecting the cpDNA and nrDNA trees from our various analyses, it was apparent that these genomes do not exhibit the same gene tree histories within EA Salvia. There were myriad well-supported differences between the two data sets, with the nrDNA phylogeny more in accordance with morphology (see the Discussion). Additionally, ILD P-values were <0.01, a threshold that can indicate significant incongruence between data sets (Cunningham, 1997). Thus, we did not perform analyses on a combined cpDNA and nrDNA data set.

Nuclear DNA analysis.

After removing ambiguously aligned sites, the aligned length of the combined nuclear data set included 1109 bp (ITS, 672 bp; ETS, 437 bp), of which 494 bp (44.5 %) were parsimony informative. Apart from collapsed or weakly supported nodes, MP, ML and BEAST trees generated similar topologies. Therefore, only the BEAST tree is shown, with posterior probabilities (PP) and ML bootstrap (MLBS) values given above branches, and MP bootstrap (MPBS) values below branches (Fig. 2).

The monophyly of subtribe Salviinae was well supported (Fig. 2; PP = 1.00, MLBS = 85 %, MPBS = 99 %; all values are reported in this order below), in which Melissa was sister to Lepechinia, and these together were sister to Salvia. Within Salvia, subg. Perovskia was sister to subg. Rosmarinus, and this clade (subg. Perovskia + subg. Rosmarinus) was then sister to the ‘Salvia officinalis clade’ (0.85, –, –); subg. Zhumeria and the S. aegyptiaca clade formed a clade (1.00, 100 %, 91 %); subg. Dorystaechas was sister to subg. Meriandra, and this clade was sister to subg. Audibertia + subg. Calosphace (1.00, 99 %, 83 %).

All EA Salvia formed a well-supported clade (EA clade: 1.00, 100 %, 100 %). Salvia glutinosa, a widely distributed species ranging from western Asia to Europe, was embedded in the EA clade. Within the EA clade, eight subclades (G1–G8) were recognized: (G1) subclade Sonchifoliae (1.00, 100 %, 100 %), including S. sonchifolia and S. petrophila, which are sister to the rest of the EA clade; (G2) subclade Notiosphace (1.00, 100 %, 100 %) containing only the widespread and enigmatic S. plebeia; (G3) subclade Substoloniferae (1.00, 100 %, 99 %) including S. trijuga and S. substolonifera; (G4) subclade Glutinaria (1.00, 97 %, 93 %) consisting of two species distributed from Europe to western China (S. glutinosa and S. nubicola) and another four species endemic to Japan and Taiwan Island (S. koyamae, S. glabrescens, S. sakuensis and S. nipponica); (G5) subclade Annuae (1.00, 95 %, 90 %) comprising three morphologically similar species (S. roborowskii, S. umbratica and S. tricuspis); (G6) subclade Eurysphace (1.00, 82 %, 78 %) including 33 species of subg. Salvia sensuWu (1977); (G7) subclade Drymosphace (0.99, –, 62 %), consisting of ten species of subg. Sclarea sensuWu (1977). Subclade G7 can be further divided into two groups: (1) the Salvia miltiorrhiza group (0.94, 72 %, 88 %), comprising S. miltiorrhiza and its morphologically similar species (S. sinica, S, bowleyana, S. paramiltiorrhiza and S. dabieshanensis) as well as two other morphologically unique species (S. honania and S. meiliensis); and (2) Salvia yunnanensis, S. plectranthoides and S. nanchuanensis, three morphologically similar species which might represent another group, the ‘S. plectranthoides group’ (0.51, –, –). The final subclade (G8), subclade Sobiso (1.00, 92 %, 78 %), consists of 21 species of subg. Allagospadonopsis and subg. Sclarea sensuWu (1977) and includes two major lineages: (1) the S. lutescens group (1.00, 87 %, 66 %), comprising six species (S. hayatana endemic to Taiwan Island and the other five endemic to Japan); and (2) the S. chinensis group (1.00, 94 %, 88 %), including 15 species (with the exception of S. japonica, the other 14 species are endemic to Japan or China).

Chloroplast DNA analysis.

After removing ambiguous sites, the aligned length of the combined cpDNA data set was 3204 bp (psbA–trnH, 461 bp; rbcL, 1239 bp; ycf1–rps15, 657 bp; and trnL–trnF, 847 bp), of which 384 bp (12 %) were parsimony informative. Apart from collapsed or weakly supported nodes, MP, ML and BEAST trees generated similar topologies. Therefore, only the BEAST tree is shown, with PP and MLBS values given above branches and MPBS values below branches (Supplementary Data Fig. S1).

The monophyly of subtribe Salviinae was supported in BEAST and ML analyses (Supplementary Data Fig. S1; 1.00, 71 %, –); Melissa was sister to Lepechinia, and these together were sister to Salvia. Within Salvia, the Salvia officinalis clade was sister to subg. Rosmarinus (0.56, 84 %, 51 %), and this clade was then sister to subg. Perovskia (1.00, 85 %, 89 %); subg. Dorystaechas, subg. Audibertia and subg. Calosphace formed a clade, which was then sister to subg. Meriandra (1.00, 93 %, 98 %); subg. Zhumeria and the S. aegyptiaca clade formed a clade (1.00, 100 %, 100 %). Monophyly of the EA clade was again supported in cpDNA analyses (1.00, 100 %, 99 %). The cpDNA tree recovered the subclades G1, G2 and G4 of the nrDNA tree, but failed to recover the other five subclades (G3, G5, G6, G7 and G8). In the cpDNA tree, subclade G1 was sister to the rest of the EA clade, as in the nrDNA tree. Within this subclade, however, the monophyly of S. petrophila was not supported, as the accession from Guangxi (S. petrophila 1) was sister to S. sonchifolia instead of grouping with another accession from Guizhou (S. petrophila 2). Subclade G2 was recovered in the cpDNA tree (1.00, 100 %, 100 %). The sister relationship between S. trijuga and S. substolonifera (i.e. subclade G3) was not recovered in the cpDNA tree, but the monophyly of each species was supported (1.00, 100 %, 100 %). Subclade G4 was supported (1.00, 99 %, 98 %), and this clade had S. glutinosa as sister to the remaining five species. Subclade G5 (S. roborowskii, S. tricuspis and S. umbratica) was not recovered as S. umbratica did not form a clade with the two former species. Subclade G6 was split into two lineages, and taxa of subclade G5 were embedded in one of the lineages. Taxa of subclades G7 and G8 nested together and formed a clade (1.00, 92 %, 62 %).

Staminal morphology

To better elucidate staminal morphology within EA Salvia, we provided schematics illustrating EA Salvia staminal diversity rather than photographs. Six distinct stamen types (A–F) were observed from EA Salvia (Fig. 3). Stamen type A was only found in Salvia sonchifolia. In stamen type A [Figs 1 (25) and 3], the filaments are clearly longer than the connectives, with arms sub-equal, the anterior thecae fused, the posterior thecae similar to the anterior thecae in size and fertility, both thecae fertile, fused and in line with the lower arms distinguished from those with posterior thecae that are vertical to the lower arms [see Fig. 1 (25–34)]. Stamen type B is common in subclades G2, G3, G5 and G6. In stamen type B [Figs 1 (26, 27) and 3), the filaments are slightly shorter than or sub-equal to the connectives, the arms sub-equal, the anterior thecae connivent or slightly separated, the posterior thecae developed but clearly smaller than the upper thecae, and the posterior thecae fused with sparse (type B1) to rarely no pollen (type B2) present. Salvia plebeia (G2) and S. trijuga (G3) lack pollen (type B2), but the other taxa within the four subclades possess some pollen on the posterior thecae (type B1). Stamen type C was common in subclades G1, G4 and G7, and was variable. In stamen type C [Figs 1 (28–31) and 3], the filaments are clearly shorter than the connectives, the upper arms clearly longer than the lower arms, the anterior usually connivent but easily separated, and the posterior thecae poorly developed (type C1) or obviously reduced (types C2 and C3), and fused with no pollen present. Stamen type C1 [Figs 1 (28, 29) and 3], with obvious posterior thecae, was found in all taxa of subclade G4 and the S. plectranthoides group of subclade G7. Stamen type C2 [Figs 1 (30) and 3], with two obsolete and outwardly reflexed posterior thecae, was only found in S. petrophila of subclade G1. Stamen type C3 [Figs 1 (31) and 3], with inflated lower arms and extremely reduced posterior thecae, was found in species of the S. miltiorrhiza group of subclade G7, with the exception of S. honania and S. meiliensis. Stamen type D was only present in two morphologically similar species (S. honania and S. meiliensis). The structure of the type D stamen [Figs 1 (32) and 3] is similar to that of type C, but the type D has the smallest posterior thecae of all EA Salvia, extremely divaricate and opposite anterior thecae, and the lower arms are not inflated. Stamen type E was only observed in two species (S. cavaleriei and S. prionitis) of the S. chinensis group of subclade G7. In stamen type E [Figs 1 (33) and 3], the filaments are clearly shorter than the connectives, the upper arms clearly longer than the lower arms, the anterior thecae connivent or separated, the posterior thecae slightly developed and fused, with no pollen present. Stamen type F [Figs 1 (34) and 3], restricted to subclade Sobiso, is similar to type E, but its posterior thecae are completely lost and separated, and the anterior thecae are connivent (F1) or extremely divaricate (F2). Stamen type F1 was observed for most species of the S. chinensis group, and type F2 was found in the S. lutescens group.

Fig. 3. — Morphological diversity and hypothetical evolutionary trends of stamen types of EA *Salvia*. Parallel evolution occurs in types B, C and F. Type A is illustrated referring to *S. sonchifolia*, B1 to *S. campanulata*, B2 to *S. trijuga*, C1 to *S. plectranthoides*, C2 to *S. petrophila*, C3 to *S. bowleyana*, D to *S. honania*, E to *S. cavaleriei* var. *simplicifolia*, F1 to *S. scapiformis*, and F2 to *S. omerocalyx*.

Divergence time estimation

The chronogram inferred from the nrDNA (ITS and ETS) data set is shown in Fig. 4 and Supplementary Data Fig. S2. Divergence times presented here are consistent with previous studies (Drew and Sytsma, 2012; Drew et al., 2017). The crown age of subtribe Salviinae is estimated to be 30.15 Ma, with the 95 % HPD 23.44–37.22 Ma. The Salvia crown arose approx. 27.79 Ma (95 % HPD: 22.25–34.10 Ma). The crown of the Salvia officinalis clade + subg. Rosmarinus and subg. Perovskia clade arose approx. 23.54 Ma (95 % HPD: 16.62–30.60 Ma). The crown of the S. aegyptiaca + subg. Zhumeria clade began to diversify approx. 17.07 Ma (95 % HPD: 11.45–23.05 Ma). The crown of the subg. Calosphace + subg. Audibertia clade and subg. Meriandra + subg. Dorystaechas clade arose approx. 20.27 Ma (95 % HPD: 14.91–25.73 Ma), and the crown of subg. Glutinaria (EA Salvia clade) began to diversify approx. 17.40 Ma (95 % HPD: 12.37–23.11 Ma). The results of the divergence times for major lineages of EA Salvia are summarized in Table 2.

Table 2.

Divergence time estimates, based on BEAST analyses of nrDNA, for major clades within EA Salvia

Node	Age estimated in this study (Ma)
	Mean	95 % HPD
Subg. Glutinaria	17.40	12.37–23.11
Sect. Sonchifoliae	3.00	1.06–5.38
Sect. Notiosphace	1.92	0.78–3.35
Sect. Substoloniferae	5.38	3.00–8.16
Sect. Glutinaria	4..88	2.85–6.07
Sect. Annuae	3.49	1.72–5.56
Sect. Eurysphace	4.91	2.89–7.13
Sect. Drymosphace	4.77	2.93–6.75
Sect. Sobiso	4.89	3.00–6.88
S. lutescens group	4.02	2.35–5.81
S. chinensis group	2.89	1.70–4.91

Open in a new tab

DISCUSSION

Based on nearly comprehensive taxon sampling, we corroborated that EA Salvia are monophyletic (both nrDNA and cpDNA analyses) and recognized eight major lineages based on the nrDNA data set. Additionally, we provide a detailed description of staminal diversity within EA Salvia. Based upon phylogenetic results and morphological evidence, we maintain a broadly defined Salvia and treat the EA Salvia as a new subgenus, Glutinaria, and recognize eight sections within this subgenus.

Possible causes of incongruence between nuclear and plastid phylogenies

Although the cpDNA phylogeny strongly supported EA Salvia as monophyletic, resolution within the cpDNA tree was quite low as compared with the nrDNA tree, and topologies inferred from these two data sets displayed obvious and widespread discordance (Fig. 2; Supplementary Data Fig. S1). Similar discordances between genomes have been noted in clades elsewhere in Lamiaceae, and ancient hybridization with chloroplast capture has been hypothesized to have contributed to the discordance (Albaladejo et al., 2005; Drew and Sytsma, 2013; Xiang et al., 2013; Drew et al., 2014; Deng et al., 2015; Walker et al., 2015). In this study, subclades G7 and G8 formed two distinct lineages in the nrDNA tree (Fig. 2), but taxa from these two lineages were mixed together in the cpDNA phylogeny, forming a well-supported clade (Supplementary Data Fig. S1). Additionally, previous chromosome research indicated that all taxa of G7 and G8 are diploid and have the same number of chromosomes (Funamoto et al., 2000; Zhao et al., 2006; Wang et al., 2009; Hu et al., 2016). Therefore, ancient hybridization with chloroplast capture is likely to be responsible for the discordance in the placement of these two lineages. Another case probably involving chloroplast capture is that of S. petrophila and S. sonchifolia. In the nrDNA tree, two accessions of S. petrophila formed a clade. In the cpDNA tree, however, S. petrophila was non-monophyletic, with one accession sister to S. sonchifolia (Supplementary Data Fig. S1). The discordance may have been caused by the accession of S. petrophila from Guangxi ‘capturing’ the chloroplast genome of S. sonchifolia.

Rapid speciation events have occurred throughout the tree of life (Enard and Paabo, 2004). Incomplete lineage sorting, often occurring in taxa associated with rapid radiations, may significantly influence phylogenetic relationships (Enard and Paabo, 2004; Pollard et al., 2006). In EA Salvia, the core subclades (G6, G7 and G8) diversified recently (Fig. 4), and interspecific relationships within the three subclades remain unresolved. Recent rapid radiation may have occurred in these groups, and incomplete lineage sorting may be a possible cause resulting from incongruence between the nuclear and plastid data sets. Additionally, the relative lack of informative sites within the cpDNA matrix (12 % vs. 45 % of nrDNA) may also contribute to the conflict. Phylogenomic approaches based on next-generation sequencing (NGS) data could address this issue.

Phylogeny of East Asian Salvia

Primarily on the basis of staminal morphology, EA Salvia were previously classified into three subgenera: subg. Salvia, Sclarea and Allagospadonopsis (Wu, 1977; Murata and Yamazaki, 1993). However, our phylogenetic analyses do not support these traditionally defined subgenera. In the nrDNA tree, taxa of subg. Salvia are spread across four lineages (G1, G4, G5 and G6), taxa of subg. Sclarea across six lineages (G1, G2, G3, G4, G7 and G8) and taxa of subg. Allagospadonopsis across two lineages (G3 and G8) (see Fig. 2).

Compared with the nrDNA tree, the cpDNA tree with limited resolution does not accurately reflect morphological relationships (Fig. 2; Supplementary Data Fig. S1). As nrDNA trees have been shown to reflect relationships based on morphological characters more accurately, the following discussion of phylogeny within EA Salvia mainly refers to the nuclear topology, and naming of lineages follows the nrDNA tree.

Subclade Sonchifoliae (G1)

Both nrDNA and cpDNA trees indicate that the recently described Salvia petrophila (Hu et al., 2014) and S. sonchifolia form a clade that is sister to the rest of EA Salvia. It was unexpected that these taxa would be phylogenetically closely related, because the species are quite different in terms of flower morphology, but they are very similar in vegetative features (see below).

Salvia sonchifolia was described by Wu (1976) based on specimens collected from limestone mountain areas of south-eastern Yunnan. We recently found three new populations in western Guangxi, adjacent to south-eastern Yunnan (Hu, 2015). Wu (1977) placed this species into subg. Salvia based upon its curved connectives, sub-equal arms and fused posterior thecae. However, it can be easily distinguished from the rest of subg. Salvia due to its oblong rosettes of basal leaves, fully fertile posterior thecae and long corolla tubes, and was therefore placed into a monotypic series (ser. Sonchifoliae) by Wu (1977). Salvia petrophila, a newly described species (Hu et al., 2014), occurs only on moist limestone cliffs in two adjacent national nature reserves (Maolan National Nature Reserve in southern Guizhou and Mulun National Nature Reserve in northern Guangxi) in south-west China.

However, prior to flowering, it is difficult to distinguish S. sonchifolia from S. petrophila, as they have similar leaves and habitats. In fact, specimens of S. petrophila were first collected in 1984 from the Maolan National Nature Reserve (Chen 2635, HGAS!) and were misidentified as S. sonchifolia because the collections were in bud. In flower, however, S. petrophila can be easily distinguished from S. sonchifolia by its falcate (vs. sub-circular) and relatively long upper corolla lips (1.2–1.5 cm vs. 0.6–0.7 cm), longer connectives (15–18 mm vs. 1.4–1.7 mm), unequal arms (upper arms twice as long as the lower vs. sub-equal arms), aborted posterior thecae lacking pollen (vs. fully fertile posterior thecae) and exserted styles (vs. included styles). Based on clear differences in flower morphology, Hu et al. (2014) did not regard them as closely related species. Instead, Hu et al. (2014) placed S. petrophila in sect. Drymosphace of subg. Sclarea sensuWu (1977), because its flower morphology is similar to that of S. miltiorrhiza. In the nrDNA analyses, S. petrophila is sister to S. sonchifolia. In the cpDNA analyses, however, S. sonchifolia is sister to one accession of S. petrophila, then these together are sister to another accession of S. petrophila (Supplementary Data Fig. S1). The discordance between nrDNA and cpDNA may be partially caused by the accession of S. petrophila from Guangxi ‘capturing’ the chloroplast genome of S. sonchifolia. Possible synapomorphies for subclade G1 include thickened fleshy roots, sub-succulent, oblong and basal leaves, and limestone habitats.

Subclade Notiosphace (G2)

This monotypic subclade was represented by five individuals of Salvia plebeia in our study. Thunberg (1784) first described S. plebeia (from Japan) as Ocimum virgatum. Subsequently, Brown (1810), presumably unaware of Thunberg’s species delimitation, established S. plebeia based on collections from Australia but without type designation. Recently, Sales et al. (2010) made a typification of the name. The phylogenetic position of S. plebeia has long puzzled taxonomists. Bentham (1832–1836) was the first to place S. plebeia, along with S. aegyptiaca, in sect. Notiosphace, but he was unsure about the placement of the latter. In his subsequent study (Bentham, 1848), a number of species considered to be allied to S. aegyptiaca (including S. japonica and S. chinensis from East Asia) were added into sect. Notiosphace. However, Briquet (1897) excluded S. japonica and S. chinensis from sect. Notiosphace in his classification of Salvia. Based on morphological differences in habit, calyces, corollas and stamens, Stibal (1935) considered S. plebeia to be unrelated to S. japonica and its allies. Instead, he argued that his newly described species, S. substolonifera, was related to S. plebeia by virtue of sharing an annual habit, campanulate calyces and small corollas (0.4–0.6 cm long) (Stibal, 1934). Based on morphology, Sales et al. (2010) concluded that S. plebeia is a phenetically isolated species, with affinities for species from Asia. In Flora Reipublicae Popularis Sinicae (Wu, 1977), although S. plebeia is considered to be a member of subg. Sclarea, the species is placed in the monotypic sect. Notiosphace based on an annual or biennial habit, simple leaves and small corollas, and subequal connective arms (Fig. 5). Although previous molecular phylogenetic studies showed that S. plebeia was in the S. glutinosa clade (EA clade), consensus on its phylogenetic position has not been reached (Takano and Okada, 2011; Li et al., 2013; Will and Claßen-Bockhoff 2017), leaving its phylogenetic position unclear. Here, our molecular phylogenetic analyses suggest that S. plebeia is an independent lineage in EA Salvia, worthy of a monotypic sectional delimitation.

Fig. 5. — Morphology of *S. plebeia*. (A) Plant; (B) root; (C) leaf, adaxially; (D) inflorescence; (E) corolla, showing stamen morphology; (F) bract. Photographs by G. X. Hu.

Salvia plebeia has perhaps the widest native geographic distribution of any species within the genus (Sales et al., 2010). It is distributed from Iran and Afghanistan in the west to Japan in the east, from far eastern Russia in the north to Australia in the south (Sales et al., 2010). In Asia, S. plebeia usually grows in disturbed habitats. As a weed, it is the most widespread sage in China, where only Qinghai, Xizang and Xinjiang lack records (Wei et al., 2015). However, S. plebeia is usually found in natural habitats in Australia. There is no consensus on its origin and dispersal. Froissart (2007) argued that S. plebeia originated in Asia, and its colonization in Australia was the result of a recent human introduction. However, because it occurs in non-weedy habitats in Australia, Sales et al. (2010) posited (but did not advocate) another scenario in which S. plebeia had an Australian origin and subsequently arrived in Asia via long-range dispersal. At any rate, the fact that S. plebeia is morphologically distinct, geographically widespread and the only (ostensibly) native species of Salvia in Australia is noteworthy. Our results clearly demonstrate that S. plebeia is a distinct lineage within EA Salvia, but further explanation as to its origin and current distribution pattern need to be evaluated by an in-depth phylogeographic study.

Subclade Substoloniferae (G3)

Only two endemic Chinese species, S. substolonifera and S. trijuga, were included in this lineage. Since they were described, no one has hypothesized that S. substolonifera and S. trijuga are closely related species. Geographically, they have markedly disjunct distributions and distinctly different habitats. The distribution of S. substolonifera ranges from eastern China to eastern Sichuan and north-eastern Yunnan, belonging to the Sino-Japanese distribution pattern (Wu, 1979, 1991), and it grows in riparian areas, moist rocky crevices and mesic forests. Salvia trijuga, however, is usually found in dry habitats in north-western Yunnan, south-western Sichuan and south-eastern Xizang, and has a Sino-Himalayan distribution pattern (Wu, 1979, 1991).

Morphologically, S. substolonifera is an annual ascending or sub-prostrate herb, with small flowers (calyx 3–4 mm, corolla 5–7 mm), but S. trijuga is an erect perennial herb, with larger flowers (calyx 10–11 mm, corolla approx. 3 cm). Taxonomically, S. substolonifera and S. trijuga have been included in different subgenera. Stibal (1934) placed S. substolonifera and S. plebeia in sect. Notiosphace (sensu subg. Eusalvia Stib.), as they both possess an annual habit, a campanulate calyx and a small corolla (0.4–0.6 cm). Wu (1977), however, placed S. substolonifera into subg. Allagospadonopsis based on the separated sterile posterior thecae. On the basis of falcate upper corolla lips and fused sterile posterior thecae, both Stibal (1934) and Wu (1977) treated S. trijuga as a member of sect. Drymosphace of subg. Sclarea. Our phylogenetic analyses, however, showed that S. substolonifera and S. trijuga are sister species, forming a distinct clade (Fig. 2). While both S. substolonifera and S. trijuga were sampled in the study of Li et al. (2013), these two species did not group together. In the study of Li et al. (2013), S. substolonifera grouped with taxa of subg. Allagospadonopsis, and S. trijuga was sister to S. pauciflora E. Peter, an illegitimate name now replaced by S. wuana C.L. Xiang (Xiang et al., 2016b). A recent phylogenetic study by Will and Claßen-Bockhoff (2017) showed that these two species form a lineage together with S. evansiana. In our study, two accessions of S. substolonifera and three accessions of S. trijuga were sampled from different localities, and all analyses strongly supported them as monophyletic. The different results in the studies by Li et al. (2013) and Will and Claßen-Bockhoff (2017) may have been a result of species misidentification. Indeed, some key morphological characters also support S. trijuga and S. substolonifera as closely related species. For example, these two species can be easily distinguished from the rest of EA Salvia by bearing unique truncate apices of the upper calyx lips [Fig. 1 (16)]. Additionally, except for size, they share similar corolla (± galeate upper corolla lips) and stamen morphology (arcuate connectives and sub-equal arms). In the field, we observed fused posterior thecae in S. substolonifera. As the fused posterior thecae in Salvia species can be easily (and unknowingly) split when dissecting flowers, the ‘free stamens’ of S. substolonifera described in previous studies may not represent its real status (Stibal, 1934; Wu, 1977; Li and Hedge, 1994). Possible synapomorphies for this clade may include ternate and simple leaves, campanulate calyces, truncate upper calyx lips and sub-equal arms.

Subclade Glutinaria (G4)

In both the nrDNA and cpDNA trees, S. glutinosa, S. nubicola, S. koyamae, S. glabrescens, S. sakuensis and S. nipponica formed an independent lineage with high support values. The floral morphology of these species is similar to that of taxa of the S. miltiorrhiza group (sensu subg. Sclarea), including tubular–campanulate calyces, strongly falcate upper corolla lips, obviously exserted styles and unequal connective arms. However, foliar morphology (simple leaves) and flowering time (August to October) indicate that taxa of this subclade resemble taxa of subg. Salvia, which is characterized by simple leaves, arcuate connectives, sub-equal arms and fused posterior thecae, and bloom usually from August to October. Based on the falcate upper corolla lips and unequal connective arms, Wu (1977) placed S. glutinosa and S. nubicola (two species distributed from western China to Europe) into subg. Sclarea. On the basis of simple leaves and larger corollas (2–3 cm long) (vs. the corolla length of other taxa from Japan ranging from 0.4 to 1.2 cm), Murata and Yamazaki (1993) included S. koyamae, S. glabrescens, S. sakuensis and S. nipponica (four species endemic to Japan and Taiwan Island) in subg. Salvia. In addition to forming a clade in our phylogenetic analyses, these six species are also morphologically similar, indicating that the six species may belong neither to subg. Salvia nor to subg. Sclarea sensuMurata and Yamazaki (1993) and Wu (1977). Therefore, both morphological and molecular evidence suggests that these six species are a distinct lineage. Simple leaves, tubular–campanulate calyces, falcate upper corolla lips, unequal connective arms and fused deformed posterior thecae may be synapomorphies for this clade.

Although we did not obtain DNA sequences for this study, Salvia chanryoenica, a species endemic to the Korean Peninsula, should also be included in this lineage based on leaf and flower morphology (Lee, 2004). Species in this clade display a noteworthy distribution pattern in that S. glutinosa and S. nubicola are distributed from the Himalayan region to Europe, while the other five species are endemic to Japan, the Korean Peninsula or Taiwan Island. Speciation and dispersal patterns within this lineage require additional study.

Subclade Annuae (G5)

This subclade comprises three species (Salvia roborowskii, S. umbratica and S. tricuspis). The distribution of these three species ranges from northern China to south-western China, with S. roborowskii extending to Bhutan and Nepal. Morphologically, these species are very similar, with the main diagnostic characters separating the species being corolla colour and length. Zhu et al. (2011) described a new speceis (S. chuanxiensis) from western Sichuan, China, but it was synonymized with S. tricuspis by Xiang et al. (2016a). Wu (1977) placed the three species in subsect. Annuae based on annual or biennial habits, many branched stems and cauline hastate–sagittate leaves. Based on staminal morphology, taxa of this lineage appear to be related to subclade G6, which have arcuate connectives, sub-equal arms and fused posterior thecae with scant pollen, but their hastate–sagittate leaves resemble those of some taxa from subclade G4. Therefore, morphological data are consistent with phylogenetic evidence suggesting that these three species are an independent lineage.

One noteworthy finding is that the corolla tube length of these three species seems to be negatively correlated with elevation. Salvia umbratica has the longest corolla (2.3–2.8 cm) and lowest elevation (600–2000 m), while S. roborowskii has the shortest corolla (1–1.3 cm) and occurs at the highest elevation (2500–3700 m). Corolla length (2.1–2.3 cm) and elevation (1400–3000 m) of S. tricuspis are intermediate between the former two species. In the specimen studies and field surveys, we found morphological characters of some populations intermediate between S. tricuspis and S. roborowskii, and these two species overlap somewhat in distribution. Thus, we speculate that natural hybridization may occur between these two species.

Subclade Eurysphace (G6)

Taxa within this lineage mirror subsect. Perennes of subg. Salvia sensuWu (1977). In Wu’s (1977) classification, he divided subg. Salvia into two sections. Section Eusphace only included S. officinalis (introduced from Europe). Our present and previous molecular analyses (Walker et al., 2004; Walker and Sytsma, 2007; Li et al., 2013; Will and Claßen-Bockhoff, 2014, 2017; Drew et al., 2017) indicate that S. officinalis and EA Salvia reside in two distinct clades. Section Eurysphace was divided into two subsections, subsect. Annuae and Perennes. Subsection Annuae comprises three annual or biennial species (S. roborowskii, S. tricuspis and S. umbratica), and both molecular and morphological evidence support subsect. Annuae as an independent lineage (see Subclade Annuae above). Subsection Perennes includes 42 species, of which 35 were sampled in the present study. Within this subsection, except for S. sonchifolia (associated with limestone in south-eastern Yunnan and western Guangxi) and S. nubicola (distributed from eastern Afghanistan to western Xizang, China), all other species are distributed in the Hengduan Mountains and adjacent areas, with a clear Sino-Himalayan distribution pattern (Wu, 1979, 1991). Our molecular phylogenetic analyses indicate that 33 of the 35 species sampled from subsect. Perennes group into subclade G6, with S. sonchifolia and S. nubicola embedded into G1 and G4, respectively. While we failed to obtain DNA sequences from S. alatipetiolata, S. dolichantha, S. himmelbaurii, S. mekongensis, S. schizocalyx, S. schizochila, and S. luteistriata, these seven species of subsect. Perennes should also be included in G6 based on their morphology and Sino-Himalayan distribution pattern. Within subclade G6, interspecific relationships remain unresolved, indicating a potential recent rapid radiation associated with the uplift of the Qinghai–Tibet Plateau (QTP). Possible synapomorphies for this lineage include: perennial herbs, simple leaves, oval–round bracts, campanulate calyces, relatively large corollas (length >1.5 cm), arcuate connectives with sub-equal arms, posterior thecae poorly developed but clearly reduced relative to the anterior thecae, and the posterior thecae fused with sparse pollen grains.

Subclade Drymosphace (G7)

All species of this lineage correspond to sect. Drymosphace of subg. Sclarea sensuWu (1977). In Wu’s (1977) treatment, he divided sect. Drymosphace into three series. Series Miltiorrhizae comprises eight species (S. miltiorrhiza, S. bowleyana, S. sinica, S. trijuga, S. yunnanensis, S. nubicola, S. cavaleriei and S. prionitis); series Honaniae is monotypic; and series Plectranthoidites includes three species (S. plectranthoides, S. nanchuanensis and S. breviconnectivata). Based on floral similarities, the subsequently described S. meiliensis (Su et al., 1984) should be included in ser. Honaniae, and S. dabieshanensis, S. vasta and S. paramiltiorrhiza (Li and Hedge, 1994) should be placed in ser. Miltiorrhizae. In this study, except for S. vasta and S. breviconnectivata, all other species of sect. Drymosphace sensuWu (1977) were sampled. However, the monophyly of sect. Drymosphace was not supported in this study. Phylogenetic and morphological evidence indicates that S. trijuga and S. nubicola are members of subclades G3 and G4, respectively (see Subclade Substoloniferae and Subclade Glutinaria above), and that the newly described S. petrophila arose via an early split in the EA Salvia clade (see Subclade Sonchifoliae above). Additionally, our phylogenetic analyses showed that S. cavaleriei and S. prionitis were embedded in subclade G8. Wu (1977) placed S. cavaleriei and S. prionitis in ser. Miltiorrhizae based on unequal connective arms and fused deformed sterile posterior thecae. However, except for deformed sterile posterior thecae, other morphological characters of these two species implicate them as members of subclade G8. For example, the fibril roots and small corollas (length usually <1 cm) observed in these two species are present in all taxa from subclade G8.

Although subclade G7 is well supported, interspecific relationships remain unresolved. Within the Salvia miltiorrhiza group, S. honania and S. meiliensis are two unique species, with clearly exserted stamens adhering laterally to the corolla wall and styles and two opposite anterior thecae [Figs 1 (28) and 6H, I). According to specimen studies and field observations, we found that there are no diagnostic characters between the two species, and the recently described S. meiliensis appears to be conspecific with S. honania. Pollen morphology also supports their similarity in that both have sub-oblate pollen, wide muri and large secondary lumina (C. L. Xiang, Kunming Institute of Botany, CAS, China, unpubl. res.). However, molecular phylogenetic results do not follow morphology, as accessions of these two species do not group together. Also, S. miltiorrhiza and its morphological allies did not group together, but S. miltiorrhiza, S. bowleyana, S. sinica, S. dabieshanensis, S. paramiltiorrhiza and S. vasta are morphologically similar species. The main diagnostic characters for them are corolla colour, leaf surface trichomes and annulate corolla tubes. However, field investigations found that the corolla colour of these species varies along a continuum (Fig. 6A–G), and leaf trichomes are also variable. Given our current knowledge, it may be more appropriate to regard S. miltiorrhiza and its allies as a species complex. At any rate, the identities of these species require further study.

Fig. 6. — Variation of corolla colour for *S. miltiorrhiza* group of sect. *Drymosphac*e *sensu*Wu (1977). (A, B) *S. miltiorrhiza*; (C, D) *S. bowleyana*; (E, F) *S. sinica*; (G) *S. dabieshanensis*. (H) *S. honania*; (I) *S. meiliensis*. Photographs by G. X. Hu.

Within subclade G7, S. plectranthoides, S. nanchuanensis and S. yunnanensis form a weakly supported clade (S. plectranthoides group), sister to the S. miltiorrhiza group. In terms of corolla morphology, S. yunnanensis is similar to S. miltiorrhiza and its allies, and can be readily distinguished from S. plectranthoides and S. nanchuanensis by funnelform corolla tubes (vs. tubular corolla tubes) and falcate upper corolla lips (vs. straight upper corolla lips). However, in the present study, S. yunnanensis grouped with one accession of S. plectranthoides (from Guizhou, south-western China), instead of being embedded within the S. miltiorrhiza group. Salvia plectranthoides, S. nanchuanensis and S. yunnanensis share similar staminal morphology (type C1), providing additional possible support for their relationship and distinguishing them from taxa in the S. miltiorrhiza group (types C3 and D; see Fig. 3). While the S. plectranthoides group needs to be confirmed by further molecular studies, this study found both morphological and phylogenetic evidence indicating that the three species and the S. miltiorrhiza group form a distinct lineage that is sister to subclade G8. Possible synapomorphies for subclade G7 include robust taproots, pinnate leaves, relatively long corollas (length >2 cm) and fused deformed posterior thecae.

Salvia breviconnectivata was described by Sun (1976) from Lufeng Village, Zhushan Town, Lunan (current name: Yiliang), Kunming, Yunnan, China and was subsequently placed in ser. Plectranthoidites, together with S. plectranthoides and S. nanchuanensis by Wu (1977). Unfortunately, we could not find either specimens or living plants of S. breviconnectivata in the field despite repeated efforts (only S. plectranthoides was found at the type locality). Based on the original descriptions, except for floral morphology, there are no diagnostic morphological differences separating S. breviconnectivata from S. plectranthoides. This may be why Wu (1977) placed S. breviconnectivata in ser. Plectranthoidites. However, within ser. Plectranthoidites, the species can be readily distinguished from the other two species by having shorter corolla tubes (0.8 cm vs. 1.4–2.5 cm), sub-equal connective arms (vs. upper arms clearly longer than lower arms) and separated posterior thecae (vs. fused posterior thecae). The small corolla and separated posterior thecae resemble taxa of subclade G8. However, S. breviconnectivata lacks fibril roots, one of the putative synapomorphies for subclade G8 (see Subclade Sobiso below). If the species was established based on a population concept, its morphological characters are so unusual in EA Salvia that it may have an enigmatic phylogenetic position. Instead, if based on a single specimen, it is probably an abnormal individual of S. plectranthoides rather than representing an independent species. Due to a lack of morphological and molecular evidence, here we tentatively regard S. breviconnectivata as a doubtful species.

Subclade Sobiso (G8)

In total, 23 EA Salvia species are placed in subg. Allagospadonopsis (Wu, 1977; Su et al., 1984; Murata and Yamazaki, 1993; Li and Hedge, 1994; Takano et al., 2014; Hu and Peng, 2015). Except for S. japonica, broadly distributed in China, Japan and the Korean peninsula, all other species are endemic either to Japan (six species) or to China (16 species). Salvia weihaiensis was first described by Wu and Li (1977) based on collections from Weihai, Shandong, in eastern China and was considered to be a Chinese endemic species. Although Wu (1977) placed this species in subg. Allagospadonopsis, it can be readily distinguished from the remainder of EA Salvia as characterized by its robust taproots, broadly ovate bracts and three-aristate upper calyx lips. Finally, Hu and Peng (2015) concluded that S. weihaiensis is a synonym of S. verbenaca (a species widely distributed in Europe) and speculated that its occurrence in China may have resulted from an accidental nursery escape.

Based on the taxa sampled from Japan and Taiwan Island, Takano and Okada (2011) indicated that subg. Allagospadonopsis was monophyletic. However, the study of Li et al. (2013) showed that subg. Allagospadonopsis was non-monophyletic, with intercalation of S. cavaleriei, S. prionitis and S. plectranthoides from subg. Sclarea. Here we sampled all species of subg. Allagospadonopsis except for S. fragarioides, S. adoxoides and S. piasezkii, three narrowly distributed species endemic to China. Our phylogenetic results showed that S. substolonifera, belonging to subg. Allagospadonopsis, formed an independent lineage together with S. trijuga, and the rest of this subgenus formed a well-supported clade together with S. cavaleriei and S. prionitis sensu subg. Sclarea (Wu, 1977). Therefore, this study confirms that subg. Allagospadonopsis sensuWu (1977) and Murata and Yamazaki (1993) is non-monophyletic. Additionally, no accessions of S. plectranthoides were embedded in subclade Sobiso in our nrDNA trees, and the unusual phylogenetic positions of S. plectranthoides in previous studies may have resulted from species misidentification (Li et al., 2013; Will and Claßen-Bockhoff, 2017).

Within subclade Sobiso, two lineages were recognized. The Salvia chinensis group consisted of 16 species, including one Japanese endemic (S. pygmaea), 14 Chinese endemics and the widely distributed S. japonica (Japan, China and the Korean peninsula). While DNA sequences of S. fragarioides, S. adoxoides and S. piasezkii were unavailable, these three Chinese endemics should also most probably be placed into this group based on geographic distribution and flower morphology.

Within this group, we observed a distinct stamen movement phenomenon in all sampled taxa: the upper connective arms cling close to the upper corolla lips at early anthesis and then bend downward gradually until anterior fertile thecae reach the middle lobe of the lower corolla lips (Fig. 7A, B). The stamen movement was also mentioned by Huang et al. (2014), in which they argued that all taxa of subg. Allagospadonopsis and a few taxa of subg. Sclarea with small flowers share this similarity. However, we observed that S. substolonifera is the only species of subg. Allagospadonopsis without this type of stamen movement, and our phylogenetic results demonstrate that S. substolonifera is a member of subclade G3. Within subg. Sclarea, S. cavaleriei, S. prionitis and S. plebeia are small-flowered taxa (corolla length <1 cm). Among these three species, the first two species with the stamen movement phenomenon are embedded in the S. chinensis group (G8), while S. plebeia, without this phenomenon, represents the distinct subclade Notiosphace (G2).

Fig. 7. — Stamen and style movement in sect. Sobiso. (A, B) Stamen movement of *S. scapiformis* (A) and *S. cavaleriei* var. *simplicifolia* (B; the arrows point to anterior thecae); (C) style movement of *S. omerocalyx*. 1 = early anthesis, 2 = later anthesis, and the red ovals indicate styles. Photographs: (A) and (B) by G. X. Hu, (C) by A. Takano.

The S. lutescens group is another lineage of subclade G8. With the exception of S. hayatana, endemic to Taiwan Island, the five other species are Japanese endemics. Within this group, most species (S. isensis, S. lutescens, S. akiensis and S. omerocalyx) display stylar (as opposed to staminal) movement during anthesis. When anthesis begins, the style clings to the upper corolla lip, and the stigma does not open. Later, the style moves downward slowly, and the stigma begins to open (becomes bilobed). However, the long exserted stamens retain their position throughout anthesis (Fig. 7C). In this group, neither stamens nor styles of S. ranzaniana change position during anthesis. Whether S. hayatana is dynamic in terms of staminal and/or stylar position during anthesis needs to be evaluated further. Stamen and style movement could be regarded as a diagnostic character to differentiate these two groups. In subclade G8, all taxa have fibril roots, lanceolate–linear bracts, small tubular calyces (4–7 mm), small corollas (length usually 5–10 mm, rarely up to 18 mm) and completely reduced posterior thecae (type F). These characters could be regarded as possible synapomorphies of subclade Sobiso. The crown of subclade G8 began to diversify approx. 4.89 Ma. Of the two lineages, the crown of the S. lutescens group diversified earlier than that of the S. chinensis group (4.02 vs. 2.89 Ma), and interspecific relationships within the S. lutescens group were relatively clear (see Fig. 2). In contrast, interspecific relationships of the S. chinensis group remain unresolved, and a recent rapid radiation may have occurred in this group.

Staminal evolution of EA Salvia

Himmelbaur and Stibal (1932–1934) first hypothesized an evolutionary trend in the staminal morphology in Salvia, proceeding from a curved connective bearing two fertile thecae to a situation where the lower connective became sterile and was modified in various ways. This evolutionary trend was observed in subg. Calosphace, subg. Audibertia and the S. officinalis clade, and stamens with fertile posterior thecae have been demonstrated to be plesiomorphic (Claßen-Bockhoff et al., 2004a; Walker and Sytsma, 2007; Will and Claßen-Bockhoff, 2014; Walker et al., 2015).

In this study, we recognize six stamen types for EA Salvia (types A–F; see Fig. 3). Stamen type A is only found in Salvia sonchifolia and is distinct from the other five types by bearing fully fertile posterior thecae. In both the nrDNA and cpDNA trees, S. sonchifolia was one of the two first diverging lineages of EA Salvia, and divergence time analyses showed that this lineage diverged approx. 17.4 Ma (Fig. 4; Supplementary Data Fig. S2). Therefore, phylogenetic and morphological evidence both indicate stamen type A as the ancestral type in EA Salvia, and that the other five types may have been derived from it. Stamen types A and B both have curved connectives and sub-equal upper and lower arms. The differences between them are that the posterior thecae of type B are obviously smaller, vertical to the lower arms, and produce little (B1) or no (B2) pollen. Therefore, we speculate that type B is derived from type A. All of the other four stamen types (C–F) have very unequal arms (the upper arms clearly longer than the lower arms). Posterior thecae within these types gradually degrade (C–E) until total abortion (F) and do not produce pollen. Stamen type F1 resembles type E in that both of them have non-parallel connectives, forming a large gap between the two connectives. Differences between the two types are that stamen type F1 loses posterior thecae completely, with separated lower arms. Both stamen types occur in the S. chinensis group, and stamen type F1 may have evolved from stamen type E. Stamen type F2, limited to the S. lutescens group, also loses posterior thecae completely, and therefore the stamen type may also have evolved from stamen type E. Within EA Salvia, fusion of the posterior thecae restricts access to nectar and forms the lever mechanism (type A–E), which occurs in all subclades (G1–G8) of EA Salvia. Complete reduction of the posterior thecae makes species lose the lever mechanism (type F), which only occurs in subclade G8. By studying the pollination mechanism of S. liguliloba, a Chinese endemic species of subclade G8 with stamen type F, Huang et al. (2015) indicated that the corolla tube of species with stamen type F becomes shorter and narrower, and pollinators are not required to enter the tube to access nectar; this was hypothesized to be an energy-saving and specialized pollination pattern. Given the above, staminal evolutionary trends within EA Salvia may present such a scenario: starting from curved connectives with equal (or sub-equal) arms with two fully fertile and fused posterior thecae (Fig. 3, type A), the posterior thecae become smaller and produce little or no pollen (Fig. 3, type B); subsequently, connectives elongate to make the upper arms obviously longer than the lower arms, and posterior thecae gradually degrade, producing no pollen (Fig. 3, types C, D and E); finally, posterior thecae are completely reduced and the lower arms are separated, resulting in a loss of the lever mechanism (Fig. 3, type F). The evolutionary trend is similar to the previous hypotheses (Himmelbaur and Stibal 1932–1934; Claßen-Bockhoff et al., 2004a), supporting the independent origins of the lever mechanism in Salvia.

It is clear based on our analyses that there is parallel evolution of staminal morphology within EA Salvia. For instance, stamen type C occurs mainly in subclade G7, but it is also observed in all six species of subclade G4 and S. petrophila of subclade G1. Likewise, stamen type B is found in four distinct subclades (G2, G3, G4 and G6). EA Salvia have traditionally been placed into three subgenera on the basis of staminal morphology alone (Wu, 1977; Murata and Yamazaki, 1993), but the three subgenera are not supported by our molecular analyses. Parallel evolution of stamen types may be largely responsible for the previous inaccurate infrageneric classifications of EA Salvia. Parallel evolution of staminal morphology has also been noted among other clades of Salvia (Himmelbaur and Stibal, 1932–1934; Claßen-Bockhoff et al., 2004; Walker and Sytsma, 2007; Will and Claßen-Bockhoff, 2014). Therefore, stamen structure should not be the only diagnostic character for delimiting infrageneric (or generic) categories.

Diversification of East Asian Salvia

We estimate that Salvia diverged from other Salviinae during the early Oligocene (30.15 Ma; Fig. 4, node 1) and then began to diversify during the middle Oligocene (27.79 Ma; Fig. 4, node 2). The EA Salvia diverged from other Salvia in the early Oligocene (25.32 Ma; Fig. 4, node 6) and extant EA Salvia began to diversify during the mid-Miocene (17.4 Ma; Fig. 4, node 7). The divergence time estimates of Salvia presented here are mostly consistent with previous studies (Drew and Sytsma, 2012; Walker et al., 2015; Drew et al., 2017). However, the divergence estimate of EA Salvia in Drew et al. (2017) is younger than that estimated here (approx. 12 Ma vs. 17.4 Ma). Most probably, the failure of Drew et al. (2017) to include the early diverging EA Salvia subclades G1 and G2 led to the difference in divergence times.

The East Asian flora is a major biodiversity hotspot, and is often divided into two sub-kingdoms: the Sino-Japanese forest sub-kingdom and the Sino-Himalayan forest sub-kingdom (Wu, 1979, 1991). The Sino-Japanese flora includes most paleoendemic taxa while the Sino-Himalayan bears more neoendemic taxa. Therefore, the former has traditionally been considered to be older than the latter (Wu and Wu, 1996; Li and Li, 1997). However, based on molecular and fossil data, Chen et al. (2018) argue that the two floras probably have a similar age. Our study supports the findings of Chen et al. (2018). Geographically, EA Salvia has typical Sino-Japanese and Sino-Himalayan distribution patterns, in which species of the G6 clade are mainly distributed in the Hengduan Mountains and adjacent regions (Sino-Himalayan), while species within the G8 lineage are mainly found in sub-tropical China (Central/South/East), the Korean Peninsula and the Japanese Archipelago (Sino-Japanese). Divergence time analyses estimated that subclades G6 and G8 both began to diversify since the early Pliocene (Fig. 4: nodes G6 and G8).

The uplift of the QTP and the initiation of the East Asia monsoon around the early Miocene greatly influenced the East Asian flora (Yin and Harrison, 2000; Decelles et al., 2007; Chen et al., 2018). The initiation of EA Salvia diversification approx. 17.4 Ma may have been spurred by these geological and/or climatic events. Subclade G6 is a well-supported subclade in the nrDNA tree, but interspecific relationships are mostly unresolved. As all species of this subclade are distributed in the Hengduan Mountains and adjacent regions, a rapid radiation may have occurred in this lineage in association with Pliocene/Pleistocene QTP uplift and Pleistocene glaciation events. Indeed, several taxa in this region (e.g. Isodon, Saussurea, Aconitum and Gentiana) are hypothesized to have experienced rapid radiations in association with QTP uplift episodes since the late Miocene (Yu et al., 2014; Favre et al., 2016). As a consequence of QTP uplift events, environmental heterogeneity was increased in association with increasingly high mountains and deep valleys, which probably triggered bursts of speciation (Yu et al., 2014).

Stamen and style movement in EA Salvia

Stamen and style movement have been reported in many angiosperm plants, and their hypothesized adaptive significance includes avoidance of self-pollination, promotion of outcrossing, delayed autonomous self-pollination and reduction in intrafloral male–female interference (Ruan and Teixeira da Silva, 2011). A lever-like staminal mechanism, which features pollinator-induced movement of the stamens, was the main trait that traditionally defined Salvia. In this study, we observed active staminal and stylar movements that were not pollinator induced, with both movements being irreversible compared with such movement induced by pollinators.

Staminal movement has been considered to be a key factor affecting male reproductive success, and can directly determine the contact frequency and precision of anther/pollen with pollinators (Schlindwein and Wittmann, 1997; Taylor et al., 2006). To date, four main types of stamen movement have been described: (1) stimulated movement is found in Cactaceae (Schlindwein and Wittmann, 1997), Berberidaceae (Lechowski and Białczyk, 1992) and Ericaceae (Nagy et al., 1999); (2) simultaneous and slow movement is found in Calycanthaceae (Azuma et al., 2005; Du et al., 2012); (3) quick and explosive movement is found in Morus alba and Cornus canadensis (Taylor et al., 2006; Whitaker et al., 2007); and (4) cascade (successive) movement is found in Nasa macrothyrsa and Ruta graveolens (Weigend et al., 2010; Ren and Tang, 2012). In EA Salvia, stamen movement corresponds to the simultaneous and slow type, and seems to be restricted to the S. chinensis group of subclade Sobiso (G8). At the beginning of anthesis, both the styles and the upper connective arms cling to the upper corolla lips, with the styles behind the upper connective arms and anterior thecae (Fig. 7A, B). The configuration prevents the style from receiving pollen from pollinators or self-pollinating. When the upper connective arms move downward to the middle lobes of lower corolla lips, the styles remain fixed in position or move downwards slightly. This separation ensures that the style of one flower will receive pollen from a different flower. Although we did not test stylar receptivity at different flowering stages, we hypothesized that taxa of G8 are protandrous based on positional changes at full anthesis. If protogynous, the styles will logically waste some opportunities to receive pollen because of blocking from the upper arms and anterior thecae at the onset of anthesis. This type of stamen movement is also observed in Chimonanthus praecox (Azuma et al., 2005; Du et al., 2012). In contrast, Chimonanthus praecox is protogynous, in which stamens with immature pollen first recurve outward (the stigmas are receptive), then the stamens gradually become upright and ultimately enclose the carpels. Except for promotion of outcrossing and avoidance of self-pollination, as suggested by previous studies (Lloyd and Yates, 1982; Barrett, 2002), stamen movement of Salvia can separate male and female functions well spatially (herkogamy) and temporally (dichogamy), which will reduce sexual interference between female and male function within a flower.

In contrast to stamen movement, we observed obvious style movement in most species of the S. lutescens group of subclade G8, in which the style moves downward slowly and the stigma lobes cluster together first then bifurcate. The obvious style movements were also observed in another two morphologically similar species (S. honania and S. meiliensis) of subclade G7. Based on the available data (personal field photos), we speculate that all EA Salvia except few for taxa of the S. chinensis group of subclade G8 may have an apparent or cryptic style movement phenomenon. Similar movement (although not described) has been observed in S. hierosolymitana, which is native to the eastern Mediterranean and belongs to the S. officinalis clade. Salvia hierosolymitana has been demonstrated to be protandrous, which allows it to reduce the risk of geitonogamy and promote outcrossing (Leshem et al., 2011). Based on this observation, all EA Salvia may be protandrous and its adaptive significance needs further study.

Taxonomic treatment

Based on staminal morphology, EA Salvia have been placed in three subgenera: subg. Salvia, subg. Sclarea and subg. Allagospadonopsis (Wu, 1977; Murata and Yamazaki, 1993). However, our molecular phylogenetic results do not support these delimitations. Based on their phylogenetic results, Will and Claßen-Bockhoff (2017) suggested treating EA Salvia either as three distinct genera or as three sections of a new genus. In this study, we recognized eight distinct lineages that should be given equal taxonomic weight. Following the philosophy of Will and Claßen-Bockhoff (2017), this would mean that EA Salvia should be treated as either eight genera or eight sections of a single genus. Treating EA Salvia as eight separate genera would be confusing to say the least; furthermore, it seems untenable to treat EA Salvia as a single genus because we were unable to find any single morphological character that distinguishes EA Salvia from Salvia in the other centres of diversity, particularly in south-western Asia and the Mediterranean region. Therefore, following the suggestion of Drew et al. (2017), and based on our molecular analysis and morphological investigation, we formally treat the EA Salvia clade as a subgenus, including eight sections, below. Morphologically, all species of EA Salvia are herbaceous and have the same basic chromosome number, x = 8 (Gill, 1971; Yang et al., 2004; Zhao et al., 2006; Wang et al., 2009; Hu et al., 2016). These two characters distinguish this group and could be regarded as diagnostic characters of the EA Salvia clade.

Salvia subg. Glutinaria (Raf.) G.X. Hu, C.L. Xiang & B.T. Drew, comb. & stat. nov.

Basionym: Glutinaria Raf. in Fl. Tellur. 3: 93. 1836.

Type: Salvia glutinosa L. in Sp. Pl. 1: 26. 1753.

Herbs perennial, rarely biennial or annual. Leaves simple or pinnately compound. Verticillasters two- to many flowered. Calyx tubular to campanulate, two-lipped; upper lip entire or three-mucronate, rarely truncate or three-dentate; lower lip two-toothed. Corolla two-lipped; tube straight or curved, annulate or not; upper lip straight or falcate; lower lip three-lobed, the middle lobe largest. Stamens two; connectives elongated, usually articulating with the filament; anterior thecae fertile, connivent or separated; posterior thecae developed, reduced or completely lost, sterile, rarely fully fertile (S. sonchifolia), fused or separated. Style two-lobed. Nutlets triquetrous, ovoid or oblong, glabrous.

Sect. Sonchifoliae (C.Y. Wu) G.X. Hu, C.L. Xiang & H. Peng, stat. nov. ≡ Ser. Sonchifoliae C.Y. Wu in Fl. Reipubl. Popularis Sin. 66: 581. 1977 – Type: Salvia sonchifolia C.Y. Wu. in Fl. Yunnan. 1: 679. 1977.
Sect. Notiosphace Benth. in Labiat. Gen. Spec. 309. 1833, p.p. – Type: Salvia plebeia R. Br. in Prodr. Fl. Nov. Holland. 501. 1810.
Sect. Substoloniferae (C.Y. Wu) C.L. Xiang & B.T. Drew stat. nov. ≡ Ser. Substoloniferae C.Y. Wu in Fl. Reipubl. Popularis Sin. 66: 583. 1977 – Type: Salvia substolonifera E. Peter. in Acta Horti Gothob. 9: 138. 1934.
Sect. Glutinaria – Type: Salvia glutinosa L. in Sp. Pl. 1: 26. 1753.
Sect. Annuae (C.Y. Wu) C.L. Xiang & H. Peng stat. nov. ≡ Subsect. Annuae C.Y. Wu in Fl. Reipubl. Popularis Sin. 66: 581. 1977 – Type: Salvia roborowskii Maxim. in Bull. Acad. Imp. Sci. Saint-Pétersbourg xxvii: 527. 1881.
Sect. Eurysphace Stib. in Act. Hort. Gothob. 9: 105, 112. 1934, p.p. – Type: Salvia przewalskii Maxim. in Bull. Acad. Imp. Sci. Saint-Pétersbourg xxvii: 527. 1881.
Sect. Drymosphace Benth. in Labiat. Gen. Spec. 195, 218. 1833, p.p. – Type: Salvia miltiorrhiza Bunge in Enum. Pl. China Bor. 50. 1833.
Sect. Sobiso (Raf.) G.X. Hu, A. Takano & B.T. Drew, comb. & stat. nov. ≡ Sobiso Raf. in Fl. Tellur. 3: 94. 1837. – Type: Salvia japonica Thunb. in Syst. Veg., ed. 14. 72. 1784.

SUPPLEMENTARY DATA

Supplementary data are available online at https://academic.oup.com/aob and consist of the following. Figure S1: cladogram based on BEAST analysis of the combined cpDNA (psbA–trnH, ycf1–rps15, trnL–trnF and rbcL) matrix. Figure S2: divergence time estimation of Salviinae based on the nrDNA matrix (ITS and ETS). Appendix: voucher information and GenBank accession numbers for taxa used in this study.

Supplementary Figures Legends

Click here for additional data file.^{(13.9KB, docx)}

Supplementary Appendix

Click here for additional data file.^{(55.3KB, docx)}

Supplementary Figure 1

Click here for additional data file.^{(3.2MB, jpeg)}

Supplementary Figure 2

Click here for additional data file.^{(4.9MB, jpeg)}

ACKNOWLEDGEMENTS

We thank Mr Fei Zhao, Mr Wei Fang, Mr Fazhi Shangguan, Miss Li Chen, Dr Hongjin Dong, Dr Yaping Chen, Dr Ying Tang and Dr Zhikun Wu for help with field work; Dr Zehuan Wang, Dr Wenzhang Ma, Dr Zhuo Zhou, Dr Zhiwei Wang, Dr Xinxin Zhu, Dr Jie Cai, Dr Ting Zhang, Dr Yousheng Chen, Dr Zhixi Fu, Dr Yi Tong, Dr Yan Liu, Dr Weibin Xu, Dr Longfei Fu, Dr Renbin Zhu, Dr Xiaojun Shi, Mr Sheng Huang, Mr Hui Tang and Mr Sirong Yi for collecting plant materials; Mr Xiaoyu Wang for his line drawing of stamen types; the curators of KUN and PE herbaria for access to herbarium materials; and Dr Petra Wester and three anonymous reviewers for helpful comments to improve the manuscript. This work was supported by the Natural Science Foundation of Yunnan Province (2015FB169); the National Natural Science Foundation of China (31370229, 31600164); the Japan Society for the Promotion of Science [26440227]; the united fund of the Natural Science Foundation of Guizhou Province and Guizhou University (LH [2017]7278); and Construction Program of Biology First-class Discipline in Guizhou (GNYL [2017] 009).

LITERATURE CITED

Albaladejo RG, Aguilar JF, Aparicio A, Feliner GN. 2005. Contrasting nuclear–plastidial phylogenetic patterns in the recently diverged Iberian Phlomis crinita and P. lychnitis lineages (Lamiaceae). Taxon 54: 987–998. [Google Scholar]
Ali NM, Yeap SK, Ho WY, Beh BK, Tan SW, Tan SG. 2012. The promising future of chia, Salvia hispanica L. Journal of Biomedicine and Biotechnology 2012: 171956. doi: 10.1155/2012/171956. [DOI] [PMC free article] [PubMed] [Google Scholar]
Alziar G. 1988–1993. Catalogue synonymique des Salvia L. du monde (Lamiaceae). I–VI. Biocosme Mesogéen 5: 87–136, 6: 79–115, 163–204, 7: 59–109, 9: 413–497, 10: 33–117. [Google Scholar]
Azuma H, Toyota M, Asakawa Y. 2005. Floral scent chemistry and stamen movement of Chimonanthus praecox (L.) Link (Calycanthaceae). Acta Phytotaxonomica et Geobotanica 56: 197–201. [Google Scholar]
Baldwin BG, Markos S. 1998. Phylogenetic utility of the external transcribed spacer (ETS) of 18S–26S rDNA: congruence of ETS and ITS trees of Calycadenia (Compositae). Molecular Phylogenetics and Evolution 10: 449–463. [DOI] [PubMed] [Google Scholar]
Barrett SCH. 2002. Sexual interference of the floral kinds. Heredity 88: 154–159. [DOI] [PubMed] [Google Scholar]
Beardsley PM, Olmstead RG. 2002. Redefining Phrymaceae: the placement of Mimulus, tribe Mimuleae, and Phryma. American Journal of Botany 89: 1093–1102. [DOI] [PubMed] [Google Scholar]
Bentham G. 1832–1836. Labiatarum Genera et Species. London: Ridgeway. [Google Scholar]
Bentham G. 1848. Labiatae. In: De Candolle A, ed. Prodromus systematis naturalis regni vegetabilis, Vol. 12 Treuttel et Wurtz: Paris, 27–603. [Google Scholar]
BenthamG 1876. Labiatae. In: Bentham G, Hooker JD, eds. Genera Plantarum, Vol. 2 London, UK: Reeve and Co, 1160–1196. [Google Scholar]
Blattner F. 1999. Direct amplification of the entire ITS region from poorly preserved plant material using recombinant PCR. Biotechniques 27: 1180–1186. [DOI] [PubMed] [Google Scholar]
Briquet J. 1897. Salvia. In: Engler A, Prantl K, eds. Die naturlichen Pflanzenfamilien nebst ihrer Gattungen und wichtigeren Arten, Vol. IV Leipzig: Wilhelm Engelmann, 183–287. [Google Scholar]
Brown R. 1810. Prodromus Florae Novae Hollandiae, et Insulae van Diemen. London. [Google Scholar]
Chen YS, Deng T, Zhou Z, Sun H. 2018. Is the East Asian flora ancient or not?National Science Review (in press). [Google Scholar]
Claßen-Bockhoff R, Wester P, Tweraser E. 2003. The staminal lever mechanism in Salvia L. (Lamiaceae) – a review. Plant Biology 5: 33–41. [Google Scholar]
Claßen-Bockhoff R, Crone M, Baikova E. 2004a. Stamen development in Salvia: homology reinvestigated. International Journal of Plant Science 165: 475–498. [Google Scholar]
Claßen-Bockhoff R, Speck T, Tweraser E, Wester P, Thimm S, Reith M. 2004b. The staminal lever mechanism in Salvia L. (Lamiaceae): a key innovation for adaptive radiation?Organisms, Diversity & Evolution 4: 189–205. [Google Scholar]
Clebsch B. 2008. The new book of salvias: sages for every garden. Portland, ME: Timber Press. [Google Scholar]
Cunningham CW. 1997. Can three incongruence tests predict when data should be combined?Molecular Biology and Evolution 14: 733–740. [DOI] [PubMed] [Google Scholar]
Darriba D, Taboada GL, Doallo R, Posada D. 2012. jModelTest 2: more models, new heuristics and parallel computing. Nature Methods 9: 772. [DOI] [PMC free article] [PubMed] [Google Scholar]
Decelles PG, Quade J, Kapp P, Fan M, Dettman DL, Ding L. 2007. High and dry in central Tibet during the Late Oligocene. Earth & Planetary Science Letters 253: 389–401. [Google Scholar]
Deng T, Nie ZL, Drew BT, et al. 2015. Does the arcto-tertiary biogeographic hypothesis explain the disjunct distribution of northern hemisphere herbaceous plants? The case of Meehania (Lamiaceae). PLoS One 10: e0117171. doi: 10.1371/journal.pone.0117171. [DOI] [PMC free article] [PubMed] [Google Scholar]
Doyle JJ, Doyle JL. 1987. A rapid DNA isolation procedure for small quantities of fresh leaf tissue. Phytochemical Bulletin 19: 11–15. [Google Scholar]
Drew BT, Sytsma KJ. 2011. Testing the monophyly and placement of Lepechinia in the tribe Mentheae (Lamiaceae). Systematic Botany 36: 1038–1049. [Google Scholar]
Drew BT, Sytsma KJ. 2012. Phylogenetics, biogeography, and staminal evolution in the tribe Mentheae (Lamiaceae). American Journal of Botany 99: 933–953. [DOI] [PubMed] [Google Scholar]
Drew BT, Sytsma KJ. 2013. The South American radiation of Lepechinia (Lamiaceae): phylogenetics, divergence times and evolution of dioecy. Botanical Journal of the Linnean Society 171: 171–190. [Google Scholar]
Drew BT, Cacho NI, Sytsma KJ. 2014. The transfer of two rare monotypic genera, Neoeplingia and Chaunostoma, to Lepechinia (Lamiaceae), and notes on their conservation. Taxon 63: 831–842. [Google Scholar]
Drew BT, González-Gallegos JG, Xiang CL, et al. 2017. Salvia united: the greatest good for the greatest number. Taxon 66: 133–145. [Google Scholar]
Drummond AJ, Suchard MA, Xie D, Rambaut A. 2012. Bayesian phylogenetics with BEAUti and the BEAST 1.7. Molecular Biology and Evolution 29: 1969–1973. [DOI] [PMC free article] [PubMed] [Google Scholar]
Du W, Qin KZ, Wang XF. 2012. The mechanism of stamen movement in Chimonanthus praecox (Calycanthaceae): differential cell growth rates on the adaxial and abaxial surfaces of filaments after flower opening. Plant Systematics and Evolution 298: 561–567. [Google Scholar]
Edgar R. 2004. MUSCLE: multiple sequence alignment with high accuracy and high throughput. Nucleic Acids Research 32: 1792–1797. [DOI] [PMC free article] [PubMed] [Google Scholar]
Enard W, Paabo S. 2004. Comparative primate genomics. Annual Review of Genomics and Human Genetics 5: 351–378. [DOI] [PubMed] [Google Scholar]
Epling C. 1938. The California salvias. A review of Salvia, section Audibertia. Annals of the Missouri Botanical Garden 25: 95–188. [Google Scholar]
Epling C. 1939. A revision of Salvia, subgenus Calosphace. Beihefte Feddes Repertorium Specierum Novarum Regni Vegetabilis 110: 1–383. [Google Scholar]
Farris JS, Källersjö M, Kluge AG, Bult C. 1994. Testing significance of incongruence. Cladistics 10: 315–319. [Google Scholar]
Favre A, Michalak I, Chen C, Wang J, Pringle JS, Matuszak S, et al. 2016. Out-of-Tibet: the spatio-temporal evolution of Gentiana (Gentianaceae). Journal of Biogeography 43: 1967–1978. [Google Scholar]
Fragoso-Martínez I, Martínez-Gordillo M, Salazar GA, et al. 2018. Phylogeny of the Neotropical sages (Salvia subg. Calosphace; Lamiaceae) and insights into pollinator and area shifts. Plant Systematics and Evolution 304: 43–55. [Google Scholar]
Froissart C. 2007. Salvia plebeia, les grandes voyages d’une petite plante. Mémoires Académie d’Orleans 6: 55–65. [Google Scholar]
Funamoto T, Zushi M, Harana T, Nakamura T. 2000. Comparative karyomorphology of the Japanese species of Salvia L. (Lamiaceae). Journal of Phytogeography & Taxonomy 48: 11–18. [Google Scholar]
Gill LS. 1971. Chromosome studies in Salvia (Labiatae): West Himalayan species. Experientia 27: 596–598. [Google Scholar]
González-Gallegos JG. 2015. Two new Salvia species (Lamiaceae) from the Sierra Madre Occidental, Durango, Mexico. Systematic Botany 40: 1093–1101. [Google Scholar]
Hedge IC. 1974. A revision of Salvia in Africa and the Canary Islands. Notes from the Royal Botanic Garden, Edinburgh 33: 1–121. [Google Scholar]
Hedge IC. 1982a. Salvia. In: Davis PH, ed. Flora of Turkey and the East Aegean Islands, Vol. 7 Edinburgh: Edinburgh University Press, 400–461. [Google Scholar]
Hedge IC. 1982b. Salvia. In: Rechinger KH, ed. Flora Iranica, Vol. 150 Graz: Akademische Druck und Verlags-Anstalt, 403–476. [Google Scholar]
Himmelbaur W, Stibal E. 1932–1934. Entwicklungsrichtungen in der Blutenregion der Gattung Salvia L. I–III. Biologia Generalis 8: 449–474; 9: 129,–150; 10: 17–48. [Google Scholar]
Hipp AL, Hall JC, Sytsma KJ. 2004. Congruence versus phylogenetic accuracy: revisiting the incongruence length difference test. Systematic Biology 53: 81–89. [DOI] [PubMed] [Google Scholar]
Hu GX. 2015. Phylogeny of Salvia (Lamiaceae) in East Asia. PhD Thesis, Kunming Institute of Botany, Chinese Academy of Sciences, Kunming, China. [Google Scholar]
Hu GX, Peng H. 2015. Identity of Salvia weihaiensis (Lamiaceae) from China. Phytotaxa 202: 298–300. [Google Scholar]
Hu GX, Xiang CL, Liu ED. 2013. Invasion status and risk assessment for Salvia tiliifolia, a recently recognised introduction to China. Weed Research 53: 355–361. [Google Scholar]
Hu GX, Liu Y, Xu WB, Liu ED. 2014. Salvia petrophila sp. nov. (Lamiaceae) from north Guangxi and south Guizhou, China. Nordic Journal of Botany 32: 190–195. [Google Scholar]
Hu GX, Xiang CL, Liu ED, Dong HJ, Funamoto T. 2016. Karyotypic study of eighteen taxa of Salvia (Lamiaceae) from China. Caryologia 69: 50–57. [Google Scholar]
Hu GX, Liu ED, Zhang T, Cai J, Xiang CL. 2017. Salvia luteistriata (Lamiaceae), a new species from northeastern Sichuan, China. Phytotaxa 314: 123–128. [Google Scholar]
Huang YB, Wei YK, Ge BJ, Wang Q. 2014. Pollination mechanisms of genus Salvia (Lamiaceae) in East Asia (China). Acta Ecologica Sinica 34: 2282–2289. [Google Scholar]
Huang YB, Wei YK, Wang Q, Xiao YE, Yue XY. 2015. Floral morphology and pollination mechanism of Salvia liguliloba, a narrow endemic species with degraded lever-like stamens. Chinese Journal of Plant Ecology 39: 753–761. [Google Scholar]
Jenks AA, Walker JB, Kim SC. 2012. Phylogeny of New World Salvia subgenus Calosphace (Lamiaceae) based on cpDNA (psbA–trnH) and nrDNA (ITS) sequence data. Journal of Plant Research 126: 483–496. [DOI] [PubMed] [Google Scholar]
Lechowski Z, Białczyk J. 1992. Effect of external calcium on the control of stamen movement in Berberis ulgaris L. Biologia Plantarum 34: 121–130. [Google Scholar]
Lee YN. 2004. Flora of Korea. Seoul: Kyohak Publishing Co., Ltd. [Google Scholar]
Leshem L, Keasar T, Shmida A. 2011. Female-biased nectar production in the protandrous, hermaphroditic shrub Salvia hierosolymitana (Lamiaceae). Australian Journal of Botany 59: 18–25. [Google Scholar]
Li HW, Hedge IC. 1994. Salvia. In: Wu CY, Raven PH, Hong DY, eds. Flora of China, Vol. 13 Beijing and St. Louis: Science Press and Missouri Botanical Garden Press, 196–224. [Google Scholar]
Li HW, Li J. 1997. The Tanaka–Kaiyong line: an important floristic line for the study of the flora of East Asia. Annals of the Missouri Botanical Garden 84: 888–892. [Google Scholar]
Li QQ, Li MH, Yuan QJ, Cui ZH, Huang LQ, Xiao PG. 2013. Phylogenetic relationships of Salvia (Lamiaceae) in China: evidence from DNA sequence datasets. Journal of Systematics and Evolution 51: 184–195. [Google Scholar]
Lloyd DG, Yates JMA. 1982. Intrasexual selection and the segregation of pollen and stigma in hermaphrodite plants, exemplified by Wahlenbergia albomarginata (Campanulaceae). Evolution 36: 903–913. [DOI] [PubMed] [Google Scholar]
Miller MA, Pfeiffer W, Schwartz T. 2010. Creating the CIPRES Science Gateway for inference of large phylogenetic trees. Gateway Computing Environments Workshop (GCE), 2010, IEEE. [Google Scholar]
Müller J, Müller K, Quandt D, Neinhuis C. 2010. PhyDE – phylogenetic data editor, version 0.995. Available from: www.phyde.de. [Google Scholar]
Murata G, Yamazaki T. 1993. Salvia. In: Iwatsuki K, Yamazaki T, Boufford D, Ohba H, eds. Flora of Japan, Vol IIIa Tokyo: Kodansha, 302–307. [Google Scholar]
Nagy ES, Strong L, Galloway LF. 1999. Contribution of delayed autogamous selfing to reproductive success in Mountain Laurel, Kalmia latifolia (Ericaceae). American Midland Naturalist 142: 39–46. [Google Scholar]
Pobedimova EG. 1954. Salvia. In: Shishkin BK, ed. Flora of the U.S.S.R., Vol. 21 Moscow: Izdatel’stvo Akademii Nauk SSR, 154–255. [Google Scholar]
Pollard DA, Iyer VN, Moses AM, Eisen MB. 2006. Widespread discordance of gene trees with species tree in Drosophila: evidence for incomplete lineage sorting. PLoS Genetics 2: e173. [DOI] [PMC free article] [PubMed] [Google Scholar]
Rambaut A. 2014. FigTree, v.1.4.2. http://tree.bio.ed.ac.uk/software/figtree/. [Google Scholar]
Rambaut A, Suchard MA, Drummond AJ. 2014. Tracer v1.6. http://beast.bio.ed.ac.uk/Tracer. [Google Scholar]
Ren MX, Tang JY. 2012. Up and down: stamen movements in Ruta graveolens (Rutaceae) enhance both outcrossing and delayed selfing. Annals of Botany 110: 1017–1025. [DOI] [PMC free article] [PubMed] [Google Scholar]
Ruan CJ, Teixeira da Silva JA. 2011. Adaptive significance of floral movement. Critical Reviews in Plant Sciences 30: 293–328. [Google Scholar]
Sales F, Hedge I, Christie F. 2010. Salvia plebeia R. Br.: taxonomy, phytogeography, autogamy and myxospermy. Pakistan Journal of Botany 42: 99–110. [Google Scholar]
Sang T, Crawford D, Stuessy T. 1997. Chloroplast DNA phylogeny, reticulate evolution, and biogeography of Paeonia (Paeoniaceae). American Journal of Botany 84: 1120–1136. [PubMed] [Google Scholar]
Schlindwein C, Wittmann D. 1997. Stamen movements in flowers of Opuntia (Cactaceae) favour oligolectic pollinators. Plant Systematics and Evolution 204: 179–193. [Google Scholar]
Soltis PS, Soltis DE, Smiley CJ. 1992. An rbcL sequence from a Miocene Taxodium (bald cypress). Proceedings of the National Academy of Sciences, USA 89: 449–451. [DOI] [PMC free article] [PubMed] [Google Scholar]
Stamatakis A. 2014. RAxML version 8: a tool for phylogenetic analysis and post-analysis of large phylogenies. Bioinformatics 30: 1312–1313. [DOI] [PMC free article] [PubMed] [Google Scholar]
Stibal E. 1934. A revision of Chinese and East Burmese species of Salvia. Acta Horti Gotoburgensis 9: 101–165. [Google Scholar]
Stibal E. 1935. Revision der Gruppe der Salvia japonica. Acta Horti Gotoburgensis 10: 55–69. [Google Scholar]
Stöver BC, Müller KF. 2010. TreeGraph 2: combining and visualizing evidence from different phylogenetic analyses. BMC Bioinformatics 11: 7. doi: 10.1186/1471-2105-11-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
Su SW, Shen ZA, He JQ. 1984. New species of the genus Salvia from Anhui. Acta Botanica Yunnanica 6: 55–62. [Google Scholar]
Sun XC. 1976. Salvia breviconnectivata. In: Wu CY, ed. Flora Yunnanica, Vol. 1 Beijing: Science Press, 686–687. [Google Scholar]
Swofford DL. 2003. PAUP*. Phylogenetic analysis using parsimony (* and other methods), version 4.0b10. Sunderland, MA: Sinauer Associates. [Google Scholar]
Taberlet P, Gielly L, Pautou G, Bouvet J. 1991. Universal primers for amplification of three non-coding regions of chloroplast DNA. Plant Molecular Biology 17: 1105–1109. [DOI] [PubMed] [Google Scholar]
Takano A, Okada H. 2011. Phylogenetic relationships among subgenera, species, and varieties of Japanese Salvia L. (Lamiaceae). Journal of Plant Research 124: 245–252. [DOI] [PubMed] [Google Scholar]
Takano A, Sera T, Kurosaki N. 2014. A new species of Salvia (Lamiaceae) from Chugoku District, Japan, Salvia akiensis sp. nov. Acta Phytotaxonomica et Geobotanica 65: 99–104. [Google Scholar]
Tamura K, Stecher G, Peterson D, Filipski A, Kumar S. 2013. MEGA6: molecular evolutionary genetics analysis version 6.0. Molecular Biology and Evolution 30: 2725–2729. [DOI] [PMC free article] [PubMed] [Google Scholar]
Taylor PE, Card G, House J, Dickinson MH, Flagan RC. 2006. High-speed pollen release in the white mulberry tree, Morus alba L. Sexual Plant Reproduction 19: 19–24. [Google Scholar]
Thunberg CP. 1784. Salvia japonica. In: Murray JA, ed. Systema Vegetabilium, 14th edn Göttingen: Dietrich, 72. [Google Scholar]
Walker JB, Sytsma KJ. 2007. Staminal evolution in the genus Salvia (Lamiaceae): molecular phylogenetic evidence for multiple origins of the staminal lever. Annals of Botany 100: 375–391. [DOI] [PMC free article] [PubMed] [Google Scholar]
Walker JB, Sytsma KJ, Treutlein J, Wink M. 2004. Salvia (Lamiaceae) is not monophyletic: implications for the systematics, radiation, and ecological specializations of Salvia and tribe Mentheae. American Journal of Botany 91: 1115–1125. [DOI] [PubMed] [Google Scholar]
Walker JB, Drew BT, Sytsma KJ. 2015. Unravelling species relationships and diversification within the iconic California Floristic Province sages (Salvia subgenus Audibertia, Lamiaceae). Systematic Botany 40: 826–844. [Google Scholar]
Wang BQ. 2010. Salvia miltiorrhiza: chemical and pharmacological review of a medicinal plant. Journal of Medicinal Plant Research 4: 2813–2820. [Google Scholar]
Wang M, Zhang L, Ding CB, et al. 2009. Meiotic observations of eight taxa in the genus Salvia. Caryologia 62: 334–340. [Google Scholar]
Wei YK, Wang Q, Huang YB. 2015. Species diversity and distribution of Salvia (Lamiaceae). Biodiversity Science 23: 3–10. [Google Scholar]
Weigend M, Ackermann M, Henning T. 2010. Reloading the revolver-male fitness as a simple explanation for complex reward partitioning in Nasa macrothyrsa (Loasaceae, Cornales). Biological Journal of the Linnean Society 100: 124–131. [Google Scholar]
Whitaker DL, Webster LA, Edwards J. 2007. The biomechanics of Cornus canadensis stamens are ideal for catapulting pollen vertically. Functional Ecology 21: 219–225. [Google Scholar]
White TJ, Bruns T, Lee S, Taylor J. 1990. Amplification and direct sequencing of fungal ribosomal RNA genes for phylogenetics. In: Innis M, Gelfand D, Sninsky J, White T, eds. PCR protocols: a guide to methods and applications. San Diego: Academic Press, 315–322. [Google Scholar]
Will M, Claßen-Bockhoff R. 2014. Why Africa matters: evolution of Old World Salvia (Lamiaceae) in Africa. Annals of Botany 114: 61–83. [DOI] [PMC free article] [PubMed] [Google Scholar]
Will M, Claßen-Bockhoff R. 2017. Time to split Salvia s.l. (Lamiaceae) – new insights from Old World Salvia phylogeny. Molecular Phylogenetics and Evolution 114: 33–58. [DOI] [PubMed] [Google Scholar]
Will M, Schmalz N, Claßen-Bockhoff R. 2015. Towards a new classification of Salvia s.l.: (re) establishing the genus Pleudia Raf. Turkish Journal of Botany 39: 693–707. [Google Scholar]
Wu CY. 1976. Salvia sonchifolia. In: Wu CY, ed. Flora Yunnanica, Vol. 1 Beijing: Science Press, 679–681. [Google Scholar]
Wu CY. 1977. Salvia. In: Wu CY, ed. Flora Reipublicae Popularis Sinicae, Vol. 66 Beijing: Science Press, 70–196. [Google Scholar]
Wu CY. 1979. The regionalization of Chinese flora. Acta Botanica Yunnanica 1: 1–22. [Google Scholar]
Wu CY. 1991. The areal-types of Chinese genera of seed plants. Acta Botanica Yunnanica S4: 1–139. [Google Scholar]
Wu CY, Li HW. 1977. Salvia weihaiensis. In: Wu CY, ed. Flora Reipublicae Popularis Sinicae, Vol. 66 Beijing: Science Press, 185–193. [Google Scholar]
Wu CY, Wu SG. 1996. A proposal for a new floristic kingdom (realm): the E. Asiatic Kingdom, its delineation and characteristics. In: Zhang AL and Wu SG eds. Proceedings of the First International Symposium on Floristic Characteristics and Diversity of East Asian Plants. Beijing: Chinese Higher Education Press, 3–24. [Google Scholar]
Xiang CL. 2016. Lamiaceae. In: Xiang CL, Liu QX, Peng H, eds. Species catalogue of China, plants, Vol. 1(IX) Beijing: Sciences Press, 1–85. [Google Scholar]
Xiang CL, Zhang Q, Scheen AC, Cantino PD, Funamoto T, Peng H. 2013. Molecular phylogenetics of Chelonopsis (Lamiaceae: Gomphostemmateae) as inferred from nuclear and plastid DNA and morphology. Taxon 62: 375–386. [Google Scholar]
Xiang CL, Hu GX, Peng H. 2016a. New combinations and new synonyms in Lamiaceae from China. Biodiversity Science 24: 719–722. [Google Scholar]
Xiang CL, Hu GX, Peng H. 2016b. Salvia wuana (Lamiaceae), a new name for S. pauciflora E. Peter. Phytotaxa 255: 99–100. [Google Scholar]
Yang ZY, Gong X, Pan YZ. 2004. Cytological study of six Salvia species (Lamiaceae) from Hengduanshan Mountains region of China. Caryologia 57: 360–366. [Google Scholar]
Yao G, Drew BT, Yi TS, Yan HF, Yuan YM, Ge XJ. 2016. Phylogenetic relationships, character evolution and biogeographic diversification of Pogostemon s.l. (Lamiaceae). Molecular Phylogenetics and Evolution 98: 184–200. [DOI] [PubMed] [Google Scholar]
Yin A, Harrison TM. 2000. Geologic evolution of the Himalayan–Tibetan orogen. Annual Review of Earth & Planetary Sciences 28: 211–280. [Google Scholar]
Yoder AD, Irwin JA, Payseur BA. 2001. Failure of the ILD to determine data combinability for slow loris phylogeny. Systematic Biology 50: 408–424. [PubMed] [Google Scholar]
Yu XQ, Maki M, Drew BT, et al. 2014. Phylogeny and historical biogeography of Isodon (Lamiaceae): rapid radiation in south-west China and Miocene overland dispersal into Africa. Molecular Phylogenetics and Evolution 77: 183–194. [DOI] [PubMed] [Google Scholar]
Zhao HX, Zhang L, Fang X, Yang RW, Ding CB, Zhou YH. 2006. Studies on chromosome numbers of Salvia miltiorrhiza, S. flava and S. evansiana. China Journal of Chinese Materia Medica 31: 1847–1849. [PubMed] [Google Scholar]
Zhu ZY, Min BQ, Wang QL. 2011. Taxa Nova Salviorum labiatarum. Bulletin of Botanical Research 31: 1–3. [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementary Figures Legends

Click here for additional data file.^{(13.9KB, docx)}

Supplementary Appendix

Click here for additional data file.^{(55.3KB, docx)}

Supplementary Figure 1

Click here for additional data file.^{(3.2MB, jpeg)}

Supplementary Figure 2

Click here for additional data file.^{(4.9MB, jpeg)}

[CIT0001] Albaladejo RG, Aguilar JF, Aparicio A, Feliner GN. 2005. Contrasting nuclear–plastidial phylogenetic patterns in the recently diverged Iberian Phlomis crinita and P. lychnitis lineages (Lamiaceae). Taxon 54: 987–998. [Google Scholar]

[CIT0002] Ali NM, Yeap SK, Ho WY, Beh BK, Tan SW, Tan SG. 2012. The promising future of chia, Salvia hispanica L. Journal of Biomedicine and Biotechnology 2012: 171956. doi: 10.1155/2012/171956. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CIT0003] Alziar G. 1988–1993. Catalogue synonymique des Salvia L. du monde (Lamiaceae). I–VI. Biocosme Mesogéen 5: 87–136, 6: 79–115, 163–204, 7: 59–109, 9: 413–497, 10: 33–117. [Google Scholar]

[CIT0004] Azuma H, Toyota M, Asakawa Y. 2005. Floral scent chemistry and stamen movement of Chimonanthus praecox (L.) Link (Calycanthaceae). Acta Phytotaxonomica et Geobotanica 56: 197–201. [Google Scholar]

[CIT0005] Baldwin BG, Markos S. 1998. Phylogenetic utility of the external transcribed spacer (ETS) of 18S–26S rDNA: congruence of ETS and ITS trees of Calycadenia (Compositae). Molecular Phylogenetics and Evolution 10: 449–463. [DOI] [PubMed] [Google Scholar]

[CIT0006] Barrett SCH. 2002. Sexual interference of the floral kinds. Heredity 88: 154–159. [DOI] [PubMed] [Google Scholar]

[CIT0007] Beardsley PM, Olmstead RG. 2002. Redefining Phrymaceae: the placement of Mimulus, tribe Mimuleae, and Phryma. American Journal of Botany 89: 1093–1102. [DOI] [PubMed] [Google Scholar]

[CIT0008] Bentham G. 1832–1836. Labiatarum Genera et Species. London: Ridgeway. [Google Scholar]

[CIT0009] Bentham G. 1848. Labiatae. In: De Candolle A, ed. Prodromus systematis naturalis regni vegetabilis, Vol. 12 Treuttel et Wurtz: Paris, 27–603. [Google Scholar]

[CIT0010] BenthamG 1876. Labiatae. In: Bentham G, Hooker JD, eds. Genera Plantarum, Vol. 2 London, UK: Reeve and Co, 1160–1196. [Google Scholar]

[CIT0011] Blattner F. 1999. Direct amplification of the entire ITS region from poorly preserved plant material using recombinant PCR. Biotechniques 27: 1180–1186. [DOI] [PubMed] [Google Scholar]

[CIT0012] Briquet J. 1897. Salvia. In: Engler A, Prantl K, eds. Die naturlichen Pflanzenfamilien nebst ihrer Gattungen und wichtigeren Arten, Vol. IV Leipzig: Wilhelm Engelmann, 183–287. [Google Scholar]

[CIT0013] Brown R. 1810. Prodromus Florae Novae Hollandiae, et Insulae van Diemen. London. [Google Scholar]

[CIT0014] Chen YS, Deng T, Zhou Z, Sun H. 2018. Is the East Asian flora ancient or not?National Science Review (in press). [Google Scholar]

[CIT0015] Claßen-Bockhoff R, Wester P, Tweraser E. 2003. The staminal lever mechanism in Salvia L. (Lamiaceae) – a review. Plant Biology 5: 33–41. [Google Scholar]

[CIT0016] Claßen-Bockhoff R, Crone M, Baikova E. 2004a. Stamen development in Salvia: homology reinvestigated. International Journal of Plant Science 165: 475–498. [Google Scholar]

[CIT0017] Claßen-Bockhoff R, Speck T, Tweraser E, Wester P, Thimm S, Reith M. 2004b. The staminal lever mechanism in Salvia L. (Lamiaceae): a key innovation for adaptive radiation?Organisms, Diversity & Evolution 4: 189–205. [Google Scholar]

[CIT0018] Clebsch B. 2008. The new book of salvias: sages for every garden. Portland, ME: Timber Press. [Google Scholar]

[CIT0019] Cunningham CW. 1997. Can three incongruence tests predict when data should be combined?Molecular Biology and Evolution 14: 733–740. [DOI] [PubMed] [Google Scholar]

[CIT0020] Darriba D, Taboada GL, Doallo R, Posada D. 2012. jModelTest 2: more models, new heuristics and parallel computing. Nature Methods 9: 772. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CIT0021] Decelles PG, Quade J, Kapp P, Fan M, Dettman DL, Ding L. 2007. High and dry in central Tibet during the Late Oligocene. Earth & Planetary Science Letters 253: 389–401. [Google Scholar]

[CIT0022] Deng T, Nie ZL, Drew BT, et al. 2015. Does the arcto-tertiary biogeographic hypothesis explain the disjunct distribution of northern hemisphere herbaceous plants? The case of Meehania (Lamiaceae). PLoS One 10: e0117171. doi: 10.1371/journal.pone.0117171. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CIT0023] Doyle JJ, Doyle JL. 1987. A rapid DNA isolation procedure for small quantities of fresh leaf tissue. Phytochemical Bulletin 19: 11–15. [Google Scholar]

[CIT0024] Drew BT, Sytsma KJ. 2011. Testing the monophyly and placement of Lepechinia in the tribe Mentheae (Lamiaceae). Systematic Botany 36: 1038–1049. [Google Scholar]

[CIT0025] Drew BT, Sytsma KJ. 2012. Phylogenetics, biogeography, and staminal evolution in the tribe Mentheae (Lamiaceae). American Journal of Botany 99: 933–953. [DOI] [PubMed] [Google Scholar]

[CIT0026] Drew BT, Sytsma KJ. 2013. The South American radiation of Lepechinia (Lamiaceae): phylogenetics, divergence times and evolution of dioecy. Botanical Journal of the Linnean Society 171: 171–190. [Google Scholar]

[CIT0027] Drew BT, Cacho NI, Sytsma KJ. 2014. The transfer of two rare monotypic genera, Neoeplingia and Chaunostoma, to Lepechinia (Lamiaceae), and notes on their conservation. Taxon 63: 831–842. [Google Scholar]

[CIT0028] Drew BT, González-Gallegos JG, Xiang CL, et al. 2017. Salvia united: the greatest good for the greatest number. Taxon 66: 133–145. [Google Scholar]

[CIT0029] Drummond AJ, Suchard MA, Xie D, Rambaut A. 2012. Bayesian phylogenetics with BEAUti and the BEAST 1.7. Molecular Biology and Evolution 29: 1969–1973. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CIT0030] Du W, Qin KZ, Wang XF. 2012. The mechanism of stamen movement in Chimonanthus praecox (Calycanthaceae): differential cell growth rates on the adaxial and abaxial surfaces of filaments after flower opening. Plant Systematics and Evolution 298: 561–567. [Google Scholar]

[CIT0031] Edgar R. 2004. MUSCLE: multiple sequence alignment with high accuracy and high throughput. Nucleic Acids Research 32: 1792–1797. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CIT0032] Enard W, Paabo S. 2004. Comparative primate genomics. Annual Review of Genomics and Human Genetics 5: 351–378. [DOI] [PubMed] [Google Scholar]

[CIT0033] Epling C. 1938. The California salvias. A review of Salvia, section Audibertia. Annals of the Missouri Botanical Garden 25: 95–188. [Google Scholar]

[CIT0034] Epling C. 1939. A revision of Salvia, subgenus Calosphace. Beihefte Feddes Repertorium Specierum Novarum Regni Vegetabilis 110: 1–383. [Google Scholar]

[CIT0035] Farris JS, Källersjö M, Kluge AG, Bult C. 1994. Testing significance of incongruence. Cladistics 10: 315–319. [Google Scholar]

[CIT0036] Favre A, Michalak I, Chen C, Wang J, Pringle JS, Matuszak S, et al. 2016. Out-of-Tibet: the spatio-temporal evolution of Gentiana (Gentianaceae). Journal of Biogeography 43: 1967–1978. [Google Scholar]

[CIT0037] Fragoso-Martínez I, Martínez-Gordillo M, Salazar GA, et al. 2018. Phylogeny of the Neotropical sages (Salvia subg. Calosphace; Lamiaceae) and insights into pollinator and area shifts. Plant Systematics and Evolution 304: 43–55. [Google Scholar]

[CIT0038] Froissart C. 2007. Salvia plebeia, les grandes voyages d’une petite plante. Mémoires Académie d’Orleans 6: 55–65. [Google Scholar]

[CIT0039] Funamoto T, Zushi M, Harana T, Nakamura T. 2000. Comparative karyomorphology of the Japanese species of Salvia L. (Lamiaceae). Journal of Phytogeography & Taxonomy 48: 11–18. [Google Scholar]

[CIT0040] Gill LS. 1971. Chromosome studies in Salvia (Labiatae): West Himalayan species. Experientia 27: 596–598. [Google Scholar]

[CIT0041] González-Gallegos JG. 2015. Two new Salvia species (Lamiaceae) from the Sierra Madre Occidental, Durango, Mexico. Systematic Botany 40: 1093–1101. [Google Scholar]

[CIT0042] Hedge IC. 1974. A revision of Salvia in Africa and the Canary Islands. Notes from the Royal Botanic Garden, Edinburgh 33: 1–121. [Google Scholar]

[CIT0043] Hedge IC. 1982a. Salvia. In: Davis PH, ed. Flora of Turkey and the East Aegean Islands, Vol. 7 Edinburgh: Edinburgh University Press, 400–461. [Google Scholar]

[CIT0044] Hedge IC. 1982b. Salvia. In: Rechinger KH, ed. Flora Iranica, Vol. 150 Graz: Akademische Druck und Verlags-Anstalt, 403–476. [Google Scholar]

[CIT0045] Himmelbaur W, Stibal E. 1932–1934. Entwicklungsrichtungen in der Blutenregion der Gattung Salvia L. I–III. Biologia Generalis 8: 449–474; 9: 129,–150; 10: 17–48. [Google Scholar]

[CIT0046] Hipp AL, Hall JC, Sytsma KJ. 2004. Congruence versus phylogenetic accuracy: revisiting the incongruence length difference test. Systematic Biology 53: 81–89. [DOI] [PubMed] [Google Scholar]

[CIT0116] Hu GX. 2015. Phylogeny of Salvia (Lamiaceae) in East Asia. PhD Thesis, Kunming Institute of Botany, Chinese Academy of Sciences, Kunming, China. [Google Scholar]

[CIT0047] Hu GX, Peng H. 2015. Identity of Salvia weihaiensis (Lamiaceae) from China. Phytotaxa 202: 298–300. [Google Scholar]

[CIT0048] Hu GX, Xiang CL, Liu ED. 2013. Invasion status and risk assessment for Salvia tiliifolia, a recently recognised introduction to China. Weed Research 53: 355–361. [Google Scholar]

[CIT0049] Hu GX, Liu Y, Xu WB, Liu ED. 2014. Salvia petrophila sp. nov. (Lamiaceae) from north Guangxi and south Guizhou, China. Nordic Journal of Botany 32: 190–195. [Google Scholar]

[CIT0050] Hu GX, Xiang CL, Liu ED, Dong HJ, Funamoto T. 2016. Karyotypic study of eighteen taxa of Salvia (Lamiaceae) from China. Caryologia 69: 50–57. [Google Scholar]

[CIT0051] Hu GX, Liu ED, Zhang T, Cai J, Xiang CL. 2017. Salvia luteistriata (Lamiaceae), a new species from northeastern Sichuan, China. Phytotaxa 314: 123–128. [Google Scholar]

[CIT0052] Huang YB, Wei YK, Ge BJ, Wang Q. 2014. Pollination mechanisms of genus Salvia (Lamiaceae) in East Asia (China). Acta Ecologica Sinica 34: 2282–2289. [Google Scholar]

[CIT0053] Huang YB, Wei YK, Wang Q, Xiao YE, Yue XY. 2015. Floral morphology and pollination mechanism of Salvia liguliloba, a narrow endemic species with degraded lever-like stamens. Chinese Journal of Plant Ecology 39: 753–761. [Google Scholar]

[CIT0054] Jenks AA, Walker JB, Kim SC. 2012. Phylogeny of New World Salvia subgenus Calosphace (Lamiaceae) based on cpDNA (psbA–trnH) and nrDNA (ITS) sequence data. Journal of Plant Research 126: 483–496. [DOI] [PubMed] [Google Scholar]

[CIT0055] Lechowski Z, Białczyk J. 1992. Effect of external calcium on the control of stamen movement in Berberis ulgaris L. Biologia Plantarum 34: 121–130. [Google Scholar]

[CIT0056] Lee YN. 2004. Flora of Korea. Seoul: Kyohak Publishing Co., Ltd. [Google Scholar]

[CIT0117] Leshem L, Keasar T, Shmida A. 2011. Female-biased nectar production in the protandrous, hermaphroditic shrub Salvia hierosolymitana (Lamiaceae). Australian Journal of Botany 59: 18–25. [Google Scholar]

[CIT0057] Li HW, Hedge IC. 1994. Salvia. In: Wu CY, Raven PH, Hong DY, eds. Flora of China, Vol. 13 Beijing and St. Louis: Science Press and Missouri Botanical Garden Press, 196–224. [Google Scholar]

[CIT0058] Li HW, Li J. 1997. The Tanaka–Kaiyong line: an important floristic line for the study of the flora of East Asia. Annals of the Missouri Botanical Garden 84: 888–892. [Google Scholar]

[CIT0059] Li QQ, Li MH, Yuan QJ, Cui ZH, Huang LQ, Xiao PG. 2013. Phylogenetic relationships of Salvia (Lamiaceae) in China: evidence from DNA sequence datasets. Journal of Systematics and Evolution 51: 184–195. [Google Scholar]

[CIT0060] Lloyd DG, Yates JMA. 1982. Intrasexual selection and the segregation of pollen and stigma in hermaphrodite plants, exemplified by Wahlenbergia albomarginata (Campanulaceae). Evolution 36: 903–913. [DOI] [PubMed] [Google Scholar]

[CIT0061] Miller MA, Pfeiffer W, Schwartz T. 2010. Creating the CIPRES Science Gateway for inference of large phylogenetic trees. Gateway Computing Environments Workshop (GCE), 2010, IEEE. [Google Scholar]

[CIT0062] Müller J, Müller K, Quandt D, Neinhuis C. 2010. PhyDE – phylogenetic data editor, version 0.995. Available from: www.phyde.de. [Google Scholar]

[CIT0063] Murata G, Yamazaki T. 1993. Salvia. In: Iwatsuki K, Yamazaki T, Boufford D, Ohba H, eds. Flora of Japan, Vol IIIa Tokyo: Kodansha, 302–307. [Google Scholar]

[CIT0064] Nagy ES, Strong L, Galloway LF. 1999. Contribution of delayed autogamous selfing to reproductive success in Mountain Laurel, Kalmia latifolia (Ericaceae). American Midland Naturalist 142: 39–46. [Google Scholar]

[CIT0065] Pobedimova EG. 1954. Salvia. In: Shishkin BK, ed. Flora of the U.S.S.R., Vol. 21 Moscow: Izdatel’stvo Akademii Nauk SSR, 154–255. [Google Scholar]

[CIT0066] Pollard DA, Iyer VN, Moses AM, Eisen MB. 2006. Widespread discordance of gene trees with species tree in Drosophila: evidence for incomplete lineage sorting. PLoS Genetics 2: e173. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CIT0067] Rambaut A. 2014. FigTree, v.1.4.2. http://tree.bio.ed.ac.uk/software/figtree/. [Google Scholar]

[CIT0068] Rambaut A, Suchard MA, Drummond AJ. 2014. Tracer v1.6. http://beast.bio.ed.ac.uk/Tracer. [Google Scholar]

[CIT0069] Ren MX, Tang JY. 2012. Up and down: stamen movements in Ruta graveolens (Rutaceae) enhance both outcrossing and delayed selfing. Annals of Botany 110: 1017–1025. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CIT0070] Ruan CJ, Teixeira da Silva JA. 2011. Adaptive significance of floral movement. Critical Reviews in Plant Sciences 30: 293–328. [Google Scholar]

[CIT0071] Sales F, Hedge I, Christie F. 2010. Salvia plebeia R. Br.: taxonomy, phytogeography, autogamy and myxospermy. Pakistan Journal of Botany 42: 99–110. [Google Scholar]

[CIT0072] Sang T, Crawford D, Stuessy T. 1997. Chloroplast DNA phylogeny, reticulate evolution, and biogeography of Paeonia (Paeoniaceae). American Journal of Botany 84: 1120–1136. [PubMed] [Google Scholar]

[CIT0118] Schlindwein C, Wittmann D. 1997. Stamen movements in flowers of Opuntia (Cactaceae) favour oligolectic pollinators. Plant Systematics and Evolution 204: 179–193. [Google Scholar]

[CIT0073] Soltis PS, Soltis DE, Smiley CJ. 1992. An rbcL sequence from a Miocene Taxodium (bald cypress). Proceedings of the National Academy of Sciences, USA 89: 449–451. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CIT0074] Stamatakis A. 2014. RAxML version 8: a tool for phylogenetic analysis and post-analysis of large phylogenies. Bioinformatics 30: 1312–1313. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CIT0075] Stibal E. 1934. A revision of Chinese and East Burmese species of Salvia. Acta Horti Gotoburgensis 9: 101–165. [Google Scholar]

[CIT0076] Stibal E. 1935. Revision der Gruppe der Salvia japonica. Acta Horti Gotoburgensis 10: 55–69. [Google Scholar]

[CIT0077] Stöver BC, Müller KF. 2010. TreeGraph 2: combining and visualizing evidence from different phylogenetic analyses. BMC Bioinformatics 11: 7. doi: 10.1186/1471-2105-11-7. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CIT0078] Su SW, Shen ZA, He JQ. 1984. New species of the genus Salvia from Anhui. Acta Botanica Yunnanica 6: 55–62. [Google Scholar]

[CIT0079] Sun XC. 1976. Salvia breviconnectivata. In: Wu CY, ed. Flora Yunnanica, Vol. 1 Beijing: Science Press, 686–687. [Google Scholar]

[CIT0080] Swofford DL. 2003. PAUP*. Phylogenetic analysis using parsimony (* and other methods), version 4.0b10. Sunderland, MA: Sinauer Associates. [Google Scholar]

[CIT0081] Taberlet P, Gielly L, Pautou G, Bouvet J. 1991. Universal primers for amplification of three non-coding regions of chloroplast DNA. Plant Molecular Biology 17: 1105–1109. [DOI] [PubMed] [Google Scholar]

[CIT0082] Takano A, Okada H. 2011. Phylogenetic relationships among subgenera, species, and varieties of Japanese Salvia L. (Lamiaceae). Journal of Plant Research 124: 245–252. [DOI] [PubMed] [Google Scholar]

[CIT0083] Takano A, Sera T, Kurosaki N. 2014. A new species of Salvia (Lamiaceae) from Chugoku District, Japan, Salvia akiensis sp. nov. Acta Phytotaxonomica et Geobotanica 65: 99–104. [Google Scholar]

[CIT0084] Tamura K, Stecher G, Peterson D, Filipski A, Kumar S. 2013. MEGA6: molecular evolutionary genetics analysis version 6.0. Molecular Biology and Evolution 30: 2725–2729. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CIT0085] Taylor PE, Card G, House J, Dickinson MH, Flagan RC. 2006. High-speed pollen release in the white mulberry tree, Morus alba L. Sexual Plant Reproduction 19: 19–24. [Google Scholar]

[CIT0086] Thunberg CP. 1784. Salvia japonica. In: Murray JA, ed. Systema Vegetabilium, 14th edn Göttingen: Dietrich, 72. [Google Scholar]

[CIT0087] Walker JB, Sytsma KJ. 2007. Staminal evolution in the genus Salvia (Lamiaceae): molecular phylogenetic evidence for multiple origins of the staminal lever. Annals of Botany 100: 375–391. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CIT0088] Walker JB, Sytsma KJ, Treutlein J, Wink M. 2004. Salvia (Lamiaceae) is not monophyletic: implications for the systematics, radiation, and ecological specializations of Salvia and tribe Mentheae. American Journal of Botany 91: 1115–1125. [DOI] [PubMed] [Google Scholar]

[CIT0089] Walker JB, Drew BT, Sytsma KJ. 2015. Unravelling species relationships and diversification within the iconic California Floristic Province sages (Salvia subgenus Audibertia, Lamiaceae). Systematic Botany 40: 826–844. [Google Scholar]

[CIT0090] Wang BQ. 2010. Salvia miltiorrhiza: chemical and pharmacological review of a medicinal plant. Journal of Medicinal Plant Research 4: 2813–2820. [Google Scholar]

[CIT0091] Wang M, Zhang L, Ding CB, et al. 2009. Meiotic observations of eight taxa in the genus Salvia. Caryologia 62: 334–340. [Google Scholar]

[CIT0092] Wei YK, Wang Q, Huang YB. 2015. Species diversity and distribution of Salvia (Lamiaceae). Biodiversity Science 23: 3–10. [Google Scholar]

[CIT0093] Weigend M, Ackermann M, Henning T. 2010. Reloading the revolver-male fitness as a simple explanation for complex reward partitioning in Nasa macrothyrsa (Loasaceae, Cornales). Biological Journal of the Linnean Society 100: 124–131. [Google Scholar]

[CIT0094] Whitaker DL, Webster LA, Edwards J. 2007. The biomechanics of Cornus canadensis stamens are ideal for catapulting pollen vertically. Functional Ecology 21: 219–225. [Google Scholar]

[CIT0095] White TJ, Bruns T, Lee S, Taylor J. 1990. Amplification and direct sequencing of fungal ribosomal RNA genes for phylogenetics. In: Innis M, Gelfand D, Sninsky J, White T, eds. PCR protocols: a guide to methods and applications. San Diego: Academic Press, 315–322. [Google Scholar]

[CIT0096] Will M, Claßen-Bockhoff R. 2014. Why Africa matters: evolution of Old World Salvia (Lamiaceae) in Africa. Annals of Botany 114: 61–83. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CIT0097] Will M, Claßen-Bockhoff R. 2017. Time to split Salvia s.l. (Lamiaceae) – new insights from Old World Salvia phylogeny. Molecular Phylogenetics and Evolution 114: 33–58. [DOI] [PubMed] [Google Scholar]

[CIT0098] Will M, Schmalz N, Claßen-Bockhoff R. 2015. Towards a new classification of Salvia s.l.: (re) establishing the genus Pleudia Raf. Turkish Journal of Botany 39: 693–707. [Google Scholar]

[CIT0099] Wu CY. 1976. Salvia sonchifolia. In: Wu CY, ed. Flora Yunnanica, Vol. 1 Beijing: Science Press, 679–681. [Google Scholar]

[CIT0100] Wu CY. 1977. Salvia. In: Wu CY, ed. Flora Reipublicae Popularis Sinicae, Vol. 66 Beijing: Science Press, 70–196. [Google Scholar]

[CIT0101] Wu CY. 1979. The regionalization of Chinese flora. Acta Botanica Yunnanica 1: 1–22. [Google Scholar]

[CIT0102] Wu CY. 1991. The areal-types of Chinese genera of seed plants. Acta Botanica Yunnanica S4: 1–139. [Google Scholar]

[CIT0103] Wu CY, Li HW. 1977. Salvia weihaiensis. In: Wu CY, ed. Flora Reipublicae Popularis Sinicae, Vol. 66 Beijing: Science Press, 185–193. [Google Scholar]

[CIT0104] Wu CY, Wu SG. 1996. A proposal for a new floristic kingdom (realm): the E. Asiatic Kingdom, its delineation and characteristics. In: Zhang AL and Wu SG eds. Proceedings of the First International Symposium on Floristic Characteristics and Diversity of East Asian Plants. Beijing: Chinese Higher Education Press, 3–24. [Google Scholar]

[CIT0105] Xiang CL. 2016. Lamiaceae. In: Xiang CL, Liu QX, Peng H, eds. Species catalogue of China, plants, Vol. 1(IX) Beijing: Sciences Press, 1–85. [Google Scholar]

[CIT0106] Xiang CL, Zhang Q, Scheen AC, Cantino PD, Funamoto T, Peng H. 2013. Molecular phylogenetics of Chelonopsis (Lamiaceae: Gomphostemmateae) as inferred from nuclear and plastid DNA and morphology. Taxon 62: 375–386. [Google Scholar]

[CIT0107] Xiang CL, Hu GX, Peng H. 2016a. New combinations and new synonyms in Lamiaceae from China. Biodiversity Science 24: 719–722. [Google Scholar]

[CIT0108] Xiang CL, Hu GX, Peng H. 2016b. Salvia wuana (Lamiaceae), a new name for S. pauciflora E. Peter. Phytotaxa 255: 99–100. [Google Scholar]

[CIT0109] Yang ZY, Gong X, Pan YZ. 2004. Cytological study of six Salvia species (Lamiaceae) from Hengduanshan Mountains region of China. Caryologia 57: 360–366. [Google Scholar]

[CIT0110] Yao G, Drew BT, Yi TS, Yan HF, Yuan YM, Ge XJ. 2016. Phylogenetic relationships, character evolution and biogeographic diversification of Pogostemon s.l. (Lamiaceae). Molecular Phylogenetics and Evolution 98: 184–200. [DOI] [PubMed] [Google Scholar]

[CIT0112] Yin A, Harrison TM. 2000. Geologic evolution of the Himalayan–Tibetan orogen. Annual Review of Earth & Planetary Sciences 28: 211–280. [Google Scholar]

[CIT0113] Yoder AD, Irwin JA, Payseur BA. 2001. Failure of the ILD to determine data combinability for slow loris phylogeny. Systematic Biology 50: 408–424. [PubMed] [Google Scholar]

[CIT0119] Yu XQ, Maki M, Drew BT, et al. 2014. Phylogeny and historical biogeography of Isodon (Lamiaceae): rapid radiation in south-west China and Miocene overland dispersal into Africa. Molecular Phylogenetics and Evolution 77: 183–194. [DOI] [PubMed] [Google Scholar]

[CIT0114] Zhao HX, Zhang L, Fang X, Yang RW, Ding CB, Zhou YH. 2006. Studies on chromosome numbers of Salvia miltiorrhiza, S. flava and S. evansiana. China Journal of Chinese Materia Medica 31: 1847–1849. [PubMed] [Google Scholar]

[CIT0115] Zhu ZY, Min BQ, Wang QL. 2011. Taxa Nova Salviorum labiatarum. Bulletin of Botanical Research 31: 1–3. [Google Scholar]

PERMALINK

Phylogeny and staminal evolution of Salvia (Lamiaceae, Nepetoideae) in East Asia

Guo-Xiong Hu

Atsuko Takano

Bryan T Drew

En-De Liu

Douglas E Soltis

Pamela S Soltis

Hua Peng

Chun-Lei Xiang

Abstract

Background and Aims

Methods

Key Results

Conclusions

INTRODUCTION

Fig. 1.

MATERIALS AND METHODS

Nomenclature and taxon sampling

DNA extraction, amplification and sequencing

Table 1.

Sequence alignment and phylogenetic analyses

Divergence time estimation

Stamen morphology

RESULTS

Alignment and phylogenetic reconstruction

Nuclear DNA analysis.

Fig. 2.

Chloroplast DNA analysis.

Staminal morphology

Fig. 3.

Divergence time estimation

Fig. 4.

Table 2.

DISCUSSION

Possible causes of incongruence between nuclear and plastid phylogenies

Phylogeny of East Asian Salvia

Subclade Sonchifoliae (G1)

Subclade Notiosphace (G2)

Fig. 5.

Subclade Substoloniferae (G3)

Subclade Glutinaria (G4)

Subclade Annuae (G5)

Subclade Eurysphace (G6)

Subclade Drymosphace (G7)

Fig. 6.

Subclade Sobiso (G8)

Fig. 7.

Staminal evolution of EA Salvia

Diversification of East Asian Salvia

Stamen and style movement in EA Salvia

Taxonomic treatment

SUPPLEMENTARY DATA

ACKNOWLEDGEMENTS

LITERATURE CITED

Associated Data

Supplementary Materials

ACTIONS

PERMALINK

RESOURCES

Similar articles

Cited by other articles

Links to NCBI Databases