The EDIT Platform for Cybertaxonomy (Ciardelli et al. 2009; Berendsohn 2010; BGBM 2011+, 2016+, Luther et al. 2019) (https://cybertaxonomy.eu/) was used to manage the taxonomic data, and to present them online as well as in text document format. The platform unites a set of open-source software tools that have been developed over the past 25 years at the Botanic Garden and Botanical Museum Berlin. This platform aims at including all aspects of taxonomic treatments and the workflows to create and maintain them (Borsch et al. 2015; Kilian et al. 2015; Henning et al. 2018). It provides the database system, editing tools, the online portal, and publication pipelines for this treatment. All sources of information can be cited for almost any item in the database, so that the information is transparent and appropriately credited.

The database component is structured according to the Common Data Model (CDM), a fully standard-based object-oriented data-model covering in detail the entire scope of taxonomic data (Müller et al. 2017). The principal software tool used in the context of the work presented here was the Taxonomic Editor, an operating-system independent frontend used to input and edit the taxonomic data in the online CDM database.

The appended taxonomic backbone was generated from an output of the EDIT platform using the functionalities of MS Access and MS Word processing software (Berendsohn et al. 2018). The contribution to the World Flora Online taxonomic backbone will be accomplished by submitting a WFO-DwCA (Darwin Core Archive) file generated from the EDIT platform.

The terminology, editorial approaches, and the handling of source citations in the EDIT Platform are applied here as described in Berendsohn et al. (2018), Korotkova et al. (2021) and The Caryophyllales Network (2024+; https://caryophyllales.org/Editorial).

A list of Plumbaginaceae names was received from the World Flora Online (WFO 2017). This list was based on The Plant List 1.1 (TPL 2013) and included 2925 names with a unique WFO-identifier. The list was uploaded to the Tropicos (1991+) name matching service to obtain publication details, author and literature abbreviations were standardized and the names then matched with and imported into the EDIT Platform, resulting in a dataset comprising 2990 names. The first author then preliminarily placed the imported names not yet classified as either taxon names or synonyms. Afterwards, the names from the World Checklist of Vascular Plants (WCVP, obtained from Kew in December 2019) were matched with the records already present in the database and further names covered in the Euro+Med PlantBase were added manually. Newly published names or names that were missing from other databases were also entered manually.

Botanical literature, both in print and online, online databases, phylogenetic studies, monographs, regional or species group treatments and checklists as well as personal taxonomic knowledge of the authors were used to evaluate the taxon concepts at species level.

The taxonomic backbone includes a core part with accepted names and their synonyms. Names excluded from the core checklist were assigned to the following categories: “Unplaced taxa” currently contains only 2 invalidly published hybrid designations that were described by Pignatti and used in later publications. “Unplaced generic subdivisions” contains names of sections and subsections that we refrained from classifying awaiting further evidence from phylogenetic studies. “Names of verified uncertain application” lists names that probably will never be placed. The categories “Invalid horticultural names and combinations” and “Excluded designations” list designations that have been in use but which we did not want to include in the synonymy (e.g. erroneous author citations). “Excluded names” contains names outside the Plumbaginaceae that were erroneously part of the original WFO backbone. In contrast, the “unresolved names” offers a provisional category for practical reasons to accommodate names for which the correct application or status has yet to be determined. To classify the unresolved names in the correct place, further literature and/or herbarium revisions are required but this investigation exceeds the scope of this study. The circumscription of taxa is always indicated by a secundum “sec.” reference (Berendsohn 1995), a reference that indicates the circumscription of a taxon and its distinction from other taxa. The sec. references are either literature references, or original work done here, and then are referred as this publication. The “syn. sec.” reference of the synonyms refers to a reference stating the synonymy to the accepted name or to one of its synonyms. The sec. reference of the names that are excluded from the core taxonomic backbone is normally the source of the name, i.e., the dataset from where the name has been imported.

The authors collaborated both by exchange of corrections in the formatted output produced by the EDIT Platform and by using a preliminary password-protected online portal that gave direct access to the CDM database.

The Euro+Med Plant Base (2006+) as a continental-level checklist was used as primary source for taxon concepts at species level for many of the European, Mediterranean and North African Plumbaginaceae species (Plumbaginaceae treatment by Domina 2011+). The taxonomic treatment by Erben (1993) in the Flora Iberica was further considered for the species from the Iberian Peninsula and the Balearic Islands. Different literature was reviewed for the Italian species (e.g., Brullo 1988; Arrigoni 2015; Peruzzi et al. 2015; Brullo and Guarino 2017). A checklist of vascular plants of Greece (Dimopoulos et al. 2013, 2016) as well as the treatment by Brullo and Erben (2016) were the main source for species from Greece. Sell (2018) is used as the principal reference for the species of Great Britain.

The African species were treated based on the African Plant Database (http://africanplantdatabase.ch), floras and other literature (e.g., Lobin et al. 1995; Mucina and Hammer 2019).

Taxa from SW and Central Asia and Russia were treated according to relevant literature (e.g., Linczevski 1952; Rechinger and Schiman-Czeika 1974; Bokhari and Edmondson 1982) but also based on the personal knowledge of the first author on taxa from the Caucasus and Middle East.

The accepted species from SE Asia and China are based on literature and the online version of Flora of China (Peng and Kamelin 1996) (http://www.efloras.org).

The taxon concept of American taxa at species level were adopted from the Vascular Plants of the Americas online database (Ulloa Ulloa et al. 2018+) (https://www.tropicos.org/Project/VPA), Flora of North America (Morin 1993) (http://floranorthamerica.org/Plumbaginaceae) and the treatments by Luteyn (1976, 1990).

The majority of the Statice species were transferred to Armeria, Limonium or other Plumbaginaceae genera by Kuntze (1891). His treatment is the main source of nomenclatural information for Statice in our database together with other literature.

A total of 198 different literature references or online databases were used as secundum reference or in notes for Plumbaginaceae.

Below the genus, we included the subgenera and sections of Limonium that had been revised in recent studies (e.g., Malekmohammadi et al. 2017a; Koutroumpa et al. 2018). A large group of Limonium species forming a well-supported clade were mostly not circumscribed at the sectional level due to the low internal phylogenetic resolution (Koutroumpa et al. 2018) and the difficulty in identifying diagnostic morphological characters that would be required to characterize sections. These species are currently classified under the non-formal “Mediterranean lineage” until further molecular and morphological data will allow their assignment to sections.

Though sections have been also described for Armeria, Acantholimon and Goniolimon (Boissier 1848; Bunge 1872; Linczevski 1952; Sauvage and Vindt 1952), only a small part of their species have been assigned into them. Furthermore, phylogenetic studies have shown that these sections are non-monophyletic (e.g., Fuertes Aguilar and Nieto Feliner 2003 for Armeria, and Moharrek et al. 2017 for Acantholimon). Therefore, sectional classifications of these genera were not included in the core checklist but were assigned to the “Unplaced generic subdivisions” category. Finally, aggregates, complexes and species groups that represent informal taxonomic units are not included in the checklist, but are mentioned in notes.

When necessary, all references to official herbaria of the type information follow the acronyms in Thiers (2024+).

We utilised Koutroumpa et al.’s (2018, 2021) ITS dataset (nrDNA) of the largest Plumbaginaceae phylogeny and added recently generated GenBank sequences for the two Aegialitis species to test the sister relationship of Aegialitideae. A Bayesian approach was employed using MrBayes v.3.2.7 (Ronquist et al. 2012), following the methodology described in Koutroumpa et al. (2018). The results from the nrDNA dataset were compare to previous phylogenetic inferences that relied only on chloroplast data for the genus (rbcL, trnL-F and matK; Lledó et al. 2001, 2005a; Koutroumpa et al. 2018, 2021). The new phylogenetic insights regarding the position of Aegialitis inform the subfamilial division of Plumbaginaceae.

The taxonomic backbone, encompassing the taxa in the family and their synonyms as the core checklist, along with the lists of names and taxa not included in the core, is provided in the Suppl. material 1 as a static treatment. This list offers a snapshot reflecting the current state of knowledge. An online version is accessible through the Caryophyllales portal (https://caryophyllales.org). The taxonomic backbone comprises scientific names (both taxonomically accepted names and synonyms), author names standardized according to IPNI (2000+), and standardized nomenclatural citations. The URL of the protologue is provided where available, preferably connecting to the specific page of the protologue (e.g., links available online through the Biodiversity Heritage Library, BHL 2005+), otherwise a link to the entire publication is provided. Additional information, such as distribution area, common names, types, and the source of nomenclatural status, will be available for some taxa in the online portal. Discussion notes provide details on decisions regarding the status of a taxon, classifications into groups or aggregates by certain authors, and additional data like hybrid parents.

The taxonomic backbone is divided into a core part, encompassing all accepted taxa and their synonyms, and lists of names that could not be resolved or are excluded from the core part (“names of verified uncertain application”, “unresolved names”, “unplaced taxa”, “invalid horticultural names and combinations”, “excluded designations”, “excluded names non Plumbaginaceae” and “unplaced generic subdivisions”).

The Plumbaginaceae, as presented here, comprises 3 tribes, 26 accepted genera, 2 subgenera, 17 sections, 2 subsections, 1,179 accepted species, 105 subspecies, 79 varieties, 3 forms and 37 nothotaxa. The core checklist in the taxonomic backbone assigns 2,782 synonyms to accepted names, whereas 30 homotypic synonyms are found in non-core sections. A total of 4,446 scientific name records for Plumbaginaceae are included, incorporating 94 invalid designations and 70 illegitimate names). Table 1 shows the core database statistics including the number of taxa and synonyms assigned to each accepted genus.

Through a review of both old and recent literature, numerous hitherto unresolved names, could be placed. Some names require further revision, notably 45 names mostly from Statice. Five Limonium names are of verified uncertain application. A separate list contains 46 horticultural designations, including 19 synonyms, identified as nomina nuda. Twenty-eight names from the original WFO data set were excluded as they do not belong to the Plumbaginaceae. Most of these names are Phlox names from the Polemoniaceae family, relocated to a genus named Armeria in that family (with reference to Linnaeus 1737) by Kuntze (1891), with the argument that Armeria Willd. was an invalid synonym of Statice. The rest belong to other taxa, Aegialitis Trin. and Plumbago esquirolii H.Lév., members of the Poaceae and Linaceae families, respectively.

New combinations for three Limonium names are implemented in the taxonomic novelties section of this paper, together with the new combinations Limonium sect. Jovibarba and Dyerophytum socotranum.

In the ITS Bayesian tree, representatives of Aegialitis (Aegialitideae) form a well-supported clade (posterior probability [pp] = 1; Suppl. material 2) sister to the genera of Plumbagineae (Fig. 3; Suppl. material 2). However, the sister relationship between Plumbagineae and Aegialitideae received low support (pp = 0.63; Suppl. material 2). Plumbagineae and Limonieae are reciprocally monophyletic with the highest support (pp = 1; Suppl. material 2).

Figure 3. 

Topological incongruence in the sister relationships of the three tribes in Plumbaginaceae, using Polygonaceae as outgroup A plastid cpDNA tree (rbcL, trnL-F, matK; Lledó et al. 2001, 2005a; Koutroumpa et al. 2018) B nrDNA tree, ITS (see Suppl. material 2), and 353 low copy nuclear loci (Baker et al. 2022).

Molecular phylogenetic studies have shown that Plumbaginaceae are well supported as monophyletic and sister to Polygonaceae (e.g., Lledó et al. 1998; Cuénoud et al. 2002; Yang et al. 2015; Yao et al. 2019; Baker et al. 2022). Several studies provided insights into the phylogenetic relationships within Plumbaginaceae. The main ones include Lledó et al. (1998, 2001, 2005a), Fuertes Aguilar and Nieto Feliner (2003), Akhani et al. (2013), Moharrek et al. (2017), Malekmohammadi et al. (2017a) and Koutroumpa et al. (2018).

The family was divided into two well-supported clades assigned to subfamilies Plumbaginoideae and Limonioideae (= Staticoideae) (Lledó et al. 1998, 2001, 2005a; Malekmohammadi et al. 2017a; Koutroumpa et al. 2018). Plumbaginoideae comprised the tribe Plumbagineae, whereas Limonioideae was further divided into the tribes Limonieae and Aegialitideae, with the monogeneric Aegialitideae (genus Aegialitis) sister to the Limonieae clade with high support according to molecular phylogenies employing the chloroplast markers rbcL, trnL-F and matK (Lledó et al. 2001, 2005a; Koutroumpa et al. 2018, 2021). However, in a recent phylogenomic study for the angiosperm tree of life using the 353 nuclear bait set, Aegialitis was recovered sister to Plumbagineae clade, comprising Plumbago and Ceratostigma Bunge, with highest support (Baker et al. 2022). These results challenge the subfamilial classification of Aegialitis contradicting previous molecular studies that used two or three chloroplast markers for the genus (Lledó et al. 2001, 2005a; Koutroumpa et al. 2018, 2021). In order to explore whether there is an incongruence between chloroplast and nuclear data or the position of Aegialitis was affected by the very limited taxon sampling in Baker et al.’s (2022) phylogenomic study (only 11 genera of Plumbaginaceae), we inferred an ITS phylogeny adding Aegialitis to Koutroumpa et al.’s (2018, 2021) dataset of the largest Plumbaginaceae phylogeny. Our results confirmed Baker et al.’s (2022) topology, placing Aegialitis sister to the Plumbagineae genera (Fig. 3; Suppl. material 2), yet with low support, showing a conflict between nuclear and chloroplast genomes regarding the placement of Aegialitis. These results further indicate that either incomplete lineage sorting or reticulate evolution may have been implicated in the emergence of this lineage. Aegialitis was regarded as the “most primitive” and aberrant genus of Plumbaginaceae (Prakash and Lim 1995). It exhibits several autapomorphies, namely fleshy corolla, basifixed anthers, elongated fruit (capsule) with spongy mesocarp and seed two or three times longer than the calyx (e.g., Kubitzki 1993; Lledó et al. 2001). Aegialitis also exhibits intermediate features between the tribes Limonieae and Plumbagineae. Specifically, it has similar vegetative and chemical features to Limonieae (Boissier 1848; Maury 1886; Harborne 1967; Hanson et al. 1994; Lledó et al. 2001), but the same breeding system (‘Plumbago‐type’ pollen and monomorphic stigma) and similar anatomical characters to Plumbagineae (Maury 1886; Weber-El Ghobary 1984; Lledó et al. 2001). Taken together, in the present taxonomic treatment, we accept the classification of Plumbaginaceae into three distinct and monophyletic tribes: Aegialitideae, Limonieae and Plumbagineae. We abstain from dividing the family into the two subfamilies Limonioideae and Plumbaginoideae due to the incongruent placement of Aegialitis observed in phylogenetic studies. Our decision is also informed by the need for additional investigations including analyses of complete chloroplast and nuclear genomes, coupled with detailed morphological studies.

Limonieae currently comprises 21 genera, namely Acantholimon, Armeria, Bakerolimon Lincz., Bamiania Lincz., Bukiniczia Lincz., Cephalorhizum Popov & Korovin, Ceratolimon M.B.Crespo & Lledó, Chaetolimon (Bunge) Lincz., Dictyolimon Rech.f., Ghaznianthus Lincz., Goniolimon Boiss., Limoniastrum Fabr., Limonium, Limoniopsis Lincz., Myriolimon Lledó, Erben & M.B.Crespo, Muellerolimon Lincz., Neogontscharovia Lincz., Popoviolimon Lincz., Psylliostachys (Jaub. & Spach) Nevski, Saharanthus M.B.Crespo & Lledó and Vassilczenkoa Lincz. These genera constitute five well-supported subclades in Limonieae: i) Limonium, ii) Ceratolimon-Limoniastrum, iii) Armeria-Psylliostachys, iv) Bakerolimon-Muellerolimon-Myriolimon-Saharanthus, and v) Goniolimon-Acantholimon s.l. with Acantholimon s.l. comprising the small genera Bamiania, Bukiniczia, Chaetolimon, Cephalorhizum, Dictyolimon, Gladiolimon, Popoviolimon and Vassilczenkoa (Malekmohammadi et al. 2017a; Moharrek et al. 2017; Koutroumpa et al. 2018). The sister relationships between these subclades remained largely unresolved (Koutroumpa et al. 2018). However, in the recent angiosperm phylogeny by Baker et al. (2022), the authors sampled representatives of eight Limonieae genera belonging to four out of the five aforementioned subclades (except iv) and found Ceratolimon-Limoniastrum sister to a clade comprising Goniolimon-Acantholimon s.l., Limonium and Psylliostachys-Armeria, with Limonium sister to Psylliostachys-Armeria subclade. All these relationships were highly supported, yet further sampling of genera under a phylogenomic approach is essential to draw clear conclusions regarding sister relationships within Limonieae. In addition, three small genera (Ghaznianthus, Limoniopsis and Neogontscharovia) have not been sampled yet in a phylogenetic framework.

Morphological data corroborate some of the inferred sister relationships of the genera within the five subclades. Specifically, Ceratolimon and Limoniastrum have stamen filaments adnate to the corolla up to the apex of the corolla tube, which is a synapomorphy within Plumbaginaceae (Lledó et al. 2000). Armeria and Psylliostachys share a unique calyx trait in which the rib‐like tissue fuses at the base of the calyx limb and is absent along the calyx tube (Crespo and Lledó 2000; Lledó et al. 2001). Muellerolimon and Bakerolimon (Malekmohammadi et al. 2017a; Koutroumpa et al. 2021) share distinctive pollen morphology and shrub habit with articulate, almost leafless stems (Baker 1953), whereas similar stem morphology is present in Myriolimon belonging to the same subclade (Lledó et al. 2003; Lledó et al. 2005b). The majority of small genera in Limonieae are found in the Goniolimon-Acantholimon s.l. subclade and several of them have been previously segregated from these two genera. The monospecific genus Ikonnikovia Lincz., previously segregated from Goniolimon by Linczevski (1952), was found nested in Goniolimon and synonymised by Koutroumpa et al. (2018) on the basis of molecular and morphological data (i.e., styles free from the base, papillose or hairy in the lower part and capitate stigmata distinguish Goniolimon including Ikonnikovia, from the rest of Plumbaginaceae; Boissier 1848; Siebert and Voss 1896). The well-supported Acantholimon s.l. clade comprise representatives of Acantholimon s.s. placed in two subclades with small genera branching in-between them (Moharrek et al. 2017; Koutroumpa et al. 2018). One of the two subclades (clade B sensu Moharrek et al. 2017) is highly supported with the oligospecific genera Dictyolimon and Bukiniczia forming a monophyletic group sister to Acantholimon s.s. species. The other subclade (clade A sensu Moharrek et al. 2017) is not highly supported and consists of the oligospecific genera Vassilczenkoa, Chaetolimon, Popoviolimon, Cephalorhizum and Bamiania, with the latter three forming a monophyletic group sister to Acantholimon s.s. that includes Gladiolimon (Moharrek et al. 2017). The sister relationships between the lineages Vassilczenkoa-Chaetolimon, Popoviolimon-Cephalorhizum-Bamiania and Acantholimon s.s. are not well resolved (Moharrek et al. 2017). Considering the phylogenetic results of Moharrek et al. (2017), Beshko et al. (2019) changed the circumscription of Acantholimon for Flora of Uzbekistan to include Chaetolimon, Vassilczenkoa and Cephalorhizum, and provided recombinations for their species under Acantholimon. Although a wider circumscription for Acantholimon including the smaller genera could avoid naming a non-monophyletic assemblage, the absence of morphological diagnostic characters for Acantholimon s.l., the unresolved relationships between Acantholimon s.s. and some of the smaller genera, the non-comprehensive taxon sampling, and the few (two or three) molecular markers used in the phylogenetic studies hinder a formal revision in the circumscription of the genus. Therefore, in this taxonomic backbone, we adopt a more conservative approach by keeping Bamiania, Bukiniczia, Chaetolimon, Cephalorhizum, Dictyolimon, Popoviolimon and Vassilczenkoa separate from Acantholimon pending further molecular data and a detailed morphological analysis. However, we merge the previously segregated monospecific genus Gladiolimon following Rechinger and Schiman-Czeika (1974) back into Acantholimon as it is found nested into Acantholimon s.s., at a shallow phylogenetic node, in a well-supported clade sister to two species of Acantholimon sect. Acmostegia Bunge, with which it shares the morphological traits that were used for its segregation (Moharrek et al. 2017). Finally, in the Limonium clade of Limonieae, Afrolimon Lincz. and Eremolimon Lincz., two genera previously separated from Limonium, were found nested in the genus (Lledó et al. 2005a; Akhani et al. 2013; Malekmohammadi et al. 2017a) and were formally synonymised by Malekmohammadi et al. (2017a), and Akhani et al. (2013), respectively.

The accepted name or synonym status and number of genera in Limonieae varied in different studies. Kubitzki (1993) accepted 22 genera in Limonieae including Acantholimon, Armeria, Bakerolimon, Bamiania, Bukiniczia, Chaetolimon, Cephalorhizum, Dictyolimon, Ghaznianthus, Gladiolimon, Goniolimon, Ikonnikovia, Limoniastrum, Limoniopsis, Limonium, Muellerolimon, Neogontscharovia, Popoviolimon, Psylliostachys, Vassilczenkoa, as well as Afrolimon and Eremolimon. The three genera Ceratolimon, Myriolimon and Saharanthus were described after the publication of Kubitzki (1993). Ceratolimon and Saharanthus were described by Crespo and Lledó (2000) based on phylogenetic results of Lledó et al. (2000). Lledó et al. (2005b) proposed the new name Myriolimon to replace their illegitimate Myriolepis (Boiss.) Lledó, Erben & M.B.Crespo, a combination that they had published before (Lledó et al. 2003), but that was considered homonymous with the earlier Myrialepis Becc. (Arecaceae) by the Committee for Spermatophyta and thus ratified at the XVII International Botanical Congress held in Vienna in July 2005 (Brummitt 2005). Molecular support for separation of Myriolimon was argued by Lledó et al. (2005a).

Hernández-Ledesma et al. (2015) accepted 24 genera in this tribe including the genera accepted by Kubitzki (1993) and the three described genera at that time, whereas Eremolimon was considered nested in Limonium. The 21 accepted genera in the current taxonomic backbone differ from Kubitzki (1993) with Ikonnikovia and Afrolimon being synonyms of Goniolimon (Koutroumpa et al. 2018) and Limonium (Malekmohammadi et al. 2017a), respectively. Gladiolimon is merged here in Acantholimon following Rechinger and Schiman-Czeika (1974) and the latest phylogenetic studies (Moharrek et al. 2017; Koutroumpa et al. 2018). Genera that are well supported as monophyletic are Armeria (e.g., Lledó et al. 1998, 2005a; Moharrek et al. 2017; Koutroumpa et al. 2018), Ceratolimon (e.g., Lledó et al. 2000, 2005a; Koutroumpa et al. 2018), Dictyolimon (Moharrek et al. 2017), Goniolimon (e.g., Moharrek et al. 2017; Koutroumpa et al. 2018), Limoniastrum (e.g., Lledó et al. 2000, 2005a; Koutroumpa et al. 2018), Limonium (e.g., Lledó et al. 2005a; Malekmohammadi et al. 2017a; Koutroumpa et al. 2018), Myriolimon (Malekmohammadi et al. 2017a) and Psylliostachys (Moharrek et al. 2017; Koutroumpa et al. 2018), Acantholimon is non-monophyletic (e.g., Moharrek et al. 2014, 2017; Koutroumpa et al. 2018). Monophyly has not been tested yet for Bakerolimon, Cephalorhizum and Chaetolimon as only one species per genus is sampled in available phylogenetic studies (e.g., Moharrek et al. 2017; Koutroumpa et al. 2018). Limoniopsis, Ghaznianthus and Neogontscharovia have not been sampled yet phylogenetically, and Bamiania, Bukiniczia, Muellerolimon, Popoviolimon, Saharanthus and Vassilczenkoa are monospecific genera. Below we discuss the infrageneric classifications and give some examples of recent studies on species delimitation within the large genera of Limonieae.

Limonium

It is the largest and most diverse genus of the family Plumbaginaceae distributed worldwide (Kubitzki 1993) with c. 70% of its species being endemic in the Mediterranean area (Koutroumpa et al. 2018). The monophyly of Limonium is confirmed by multiple molecular phylogenetic studies (Lledó et al. 1998, 2005a; Malekmohammadi et al. 2017a; Koutroumpa et al. 2018). Limonium contains two well-supported monophyletic subgenera L. subg. Limonium and L. subg. Pterocladus (Spach) H.Arnaud (Lledó et al. 2005a; Malekmohammadi et al. 2017a; Koutroumpa et al. 2018). Twenty-five species are here classified under L. subg. Pterocladus and the rest are assigned to Limonium subg. Limonium.

Boissier (1848) provided the first infrageneric treatment for Limonium (under Statice) recognizing 13 sections and 10 subsections, which were mostly transferred to Limonium by Sauvage and Vindt (1952). Since Boissier (1848), some of the subsections were raised to sectional rank (e.g., Statice sect. Limonium subsect. Sarcophyllae raised to L. sect. Sarcophylla by Linczevski 1952), several new sections were described (e.g., L. sect. Limoniodendron Sventenius (1960), L. sect. Nephrophyllum Rech.f. by Rechinger and Schiman-Czeika 1974), whereas some of Boissier’s sections were segregated from Limonium as independent genera (namely Dictyolimon by Rechinger and Schiman-Czeika (1974), Psylliostachys by Nevski (1937) and Myriolimon by Lledó et al. (2003, 2005b)). Sectional classification has been updated recently following molecular phylogenetic studies (Malekmohammadi et al. 2017a; Koutroumpa et al. 2018). The new Limonium sect. Iranolimon M.Malekm., Akhani & Borsch was described by Malekmohammadi et al. (2017a) to accommodate species of an Irano-Turanian subclade previously classified under L. sect. Sarcophylla (Boiss.) Lincz. The latter section was originally described based on the woody habit of its species, which turned out to have convergently evolved in two unrelated lineages (Malekmohammadi et al. 2017a). Also, the new Limonium sect. Circinaria (Boiss.) M.Malekm. was validated to include species previously assigned to Afrolimon that were found nested within Limonium (Lledó et al. 2005a; Malekmohammadi et al. 2017a). Koutroumpa et al. (2018) described the new monospecific section L. sect. Tenuiramosa Koutr. (L. anthericoides (Schltr.) R.A.Dyer) which is sister to L. sect. Pterocladus (Spach) Bokhari and both constitute L. subg. Pterocladus. Furthermore, Koutroumpa et al. (2018) amended L. sect. Limonium, L. sect. Nephrophyllum and L. sect. Sarcophylla, and published new combinations for L. sect. Pruinosa (Batt.) Koutr. and L. sect. Pterocladus subsect. Nobilia (Boiss.) Koutr.

The extensive sampling of Mediterranean endemics of Limonium in Koutroumpa et al.’s (2018) study of Plumbaginaceae revealed that they all belong to a large, well-supported internal clade, namely the “Mediterranean lineage”. Nevertheless, species relationships within the lineage remained largely unresolved. Only few species of the “Mediterranean lineage” were assigned to four morphologically well-defined sections (i.e., L. sect. Polyathrion (Boiss.) Sauvage & Vindt, L. sect. Pruinosa, L. sect. Siphonantha (Boiss.) Sauvage & Vindt and L. sect. Schizhymenium (Boiss.) Sauvage & Vindt), two of which were represented by multiple species in the phylogeny and recovered as monophyletic (Koutroumpa et al. 2018). A sectional classification for the remaining species within this lineage at the moment is difficult. Species of this lineage have diversified very recently (mostly during the Pleistocene; Lledó et al. 2005a; Koutroumpa et al. 2021) and to resolve their phylogenetic relationships, many molecular characters will be required. In addition, combined effects of polyploidy, apomixis and hybridization may have blurred species limits and make the identification of diagnostic morphological characters a difficult task. Therefore, all these species are provisionally assigned to the phylogenetically well-defined “Mediterranean lineage”. Apart from causing taxonomic complexity, polyploidy, apomixis and hybridization have been considered as the main factors for promoting speciation of Limonium in the Mediterranean region (Ingrouille 1984; Kubitzki 1993; Palacios et al. 2000; Lledó et al. 2005a). Indeed, Koutroumpa et al. (2021) found a significant shift in diversification rates for the “Mediterranean lineage” and showed that the turbulent geological history and climatic oscillations in the Mediterranean, combined with the significant role of apomixis triggered species radiation in Limonium.

We follow the mentioned recent advances in the infrageneric classification of Limonium and assign taxa to major clades corresponding to the two subgenera, and further classify them to one of the 17 accepted sections and two subsections or the “Mediterranean lineage”. We achieved this by combining information from the latest phylogenetic analyses (Malekmohammadi et al. 2017a; Koutroumpa et al. 2018) with other data (e.g., morphology, chromosome counts, geographic distributions) obtained from an extensive literature search for Limonium taxa that were not yet sampled in a phylogenetic framework (see for example table S3 in Koutroumpa et al. 2018). As a result, we can summarize Limonium subg. Pterocladus to comprise L. sect. Tenuiramosa (one species) and L. sect. Pterocladus (24 species) that is further divided into L. subsect. Nobilia (18 species) and L. subsect. Odontolepidea (six species) (Fig. 4). Limonium subg. Limonium is divided into three distinct well-supported clades (B1, B2 and B3 sensu Koutroumpa et al. 2018; Fig. 4), with L. sect. Limoniodendron (one species; clade B1) being sister to a clade comprising mostly non-Mediterranean taxa (clade B2) and the “Mediterranean lineage” (clade B3). Clade B2 includes ten morphologically and phylogenetically well-defined sections (Fig. 4), namely L. sect. Circinaria (eight species), L. sect. Ctenostachys (Boiss.) Sauvage & Vindt (11 species), L. sect. Iranolimon (nine species), L. sect. Jovibarba (Boiss.) M.Malekm. & Koutr. (three species), L. sect. Limonium (25 species), L. sect. Nephrophyllum (16 species), L. sect. Plathymenium (Boiss.) Lincz. (28 species), L. sect. Sarcophylla (12 species), L. sect. Siphonocalyx Lincz. (12 species) and L. sect. Sphaerostachys (Boiss.) Bokhari (four species). The “Mediterranean lineage” (clade B3) comprise four small sections (Fig. 4) L. sect. Polyarthrion (four species), L. sect. Pruinosa (six species), L. sect. Schizhymenium (two species) and L. sect. Siphonantha (four species), whereas 479 species are assigned to this lineage but not classified further into sections or subsections due to the reasons explained above.

Figure 4. 

Phylogenetic relationships of the major clades in Limonium and the corresponding infrageneric units, following Koutroumpa et al. (2018). The size of triangles is proportional to the number of species assigned to the different sections and the “Mediterranean lineage”.

Only a few studies explore species limits and relationships in Limonium at shallow phylogenetic levels. A recent example is the phylogenomic investigation by Pina-Martins et al. (2023) on seven representatives of Limonium sect. Limonium. The authors sampled multiple individuals per species and analyzed more than 10,000 SNPs obtained from genotyping by sequencing (GBS). The large amount of molecular data employed in the study could resolve species relationships that were previously mostly unresolved when only few molecular markers were used (see e.g., Malekmohammadi et al. 2017a, 2017b; Koutroumpa et al. 2018). Pina-Martins et al.’s (2023) phylogeny recovered L. brasiliense (Boiss.) Kuntze, L. californicum (Boiss.) A.Heller, L. carolinianum (Walter) Britton, L. humile Mill. and L. narbonense Mill. as monophyletic with high support. The widespread L. vulgare Mill. formed a large, well-supported clade with L. maritimum Caperta, Cortinhas, A.P.Paes, Guara, Esp.Santo & Erben nested in it. Limonium maritimum, represented by a single population although it is widely distributed along the Portuguese coast (Cortinhas et al. 2015), differed from L. vulgare only by 34 out of 10,000 SNPs. Moreover, the populations of L. vulgare across its distributional range showed high genetic structure based on the phylogenetic and clustering analyses (Pina-Martins et al. 2023). Limonium maritimum was described as separate species based on morphometric data, in which few diagnostic traits had slightly smaller yet largely overlapping size ranges compared to the closely related L. vulgare (Cortinhas et al. 2015). Taken together, the nested phylogenetic position of L. maritimum within L. vulgare and the low genetic and morphometric differentiation of L. maritimum compared to L. vulgare, question the recognition of the former as a separate species. Therefore, we consider L. maritimum as a synonym of L. vulgare.

Iamonico et al. (2022) examined species boundaries in four endemic species along the Tyrrhenian coast and Ponziane Archipelago (central Italy) combining molecular and morphometric data. They analyzed ITS sequences and found the same ribotype in all populations of the four species, except for two populations displaying individuals with dual ribotypes. This suggests a possible hybrid origin, though not addressed in the study. Morphometric analyses revealed that individuals from the two populations with dual ITS ribotypes were differentiated in morphospace, and at species level, L. pandatariae Pignatti was distinguished from L. circaei Pignatti, while L. amynclaeum Pignatti and L. pontium Pignatti occupied a central position relative to the other species, with considerable overlap among individuals. The authors proposed merging all species under L. pontium and recognizing the populations with mixed ITS genotypes as two different subspecies. However, considering the potential hybrid nature of these populations and the limitations of using a single molecular maker for species resolution within the “Mediterranean lineage” (see e.g., Koutroumpa et al. 2018), we suggest maintaining the four previously recognized species separate until further molecular studies provide clarity. Additionally, we propose synonymizing the newly described L. pontium subsp. terracinense Iberite, Iamonico, De Castro & Nicolella under L. amynclaeum.

We used phylogenetic studies for the circumscription of taxa in Limonium, where available (e.g., Akhani et al. 2013; Pina-Martins et al. 2023). As many species are of restricted range, in the absence of such studies it was considered adequate to use morphological circumscriptions from regional floras (e.g., Erben 1993, 2012), regional monographs (e.g., Brullo and Erben 2016, Limonium in Greece; Doğan et al. 2020, Limonium in Turkey), taxonomic revisions of specific species groups (e.g., Bogdanović and Brullo 2015, Limonium cancellatum group), and protologues (Pignatti 1955, 1982) as secundum references for the taxa of Limonium.

The tribe Plumbagineae comprises Ceratostigma, Dyerophytum Kuntze, Plumbagella Spach and Plumbago. Ceratostigma is sister to the rest of the genera (Koutroumpa et al. 2018). Plumbago is polyphyletic (Lledó et al. 2001, 2005a; Koutroumpa et al. 2018) with Plumbago europaea L., the type species of the genus, sister to the monotypic Plumbagella and the tropical/subtropical species of Plumbago sister to Dyerophytum. All phylogenetic relationships within Plumbagineae are highly supported (Koutroumpa et al. 2018). Plumbago and Plumbagella are characterised by glandular calyces, a distinct diagnostic feature for the family (Kubitzki 1993). Although Plumbago forms a non-monophyletic assemblage, a formal revision of its generic circumscription would require a comprehensive taxon sampling in a phylogenetic framework (currently four out of 19 Plumbago species sampled in Koutroumpa et al. 2018) and a revision of the diagnostic characters of the well-supported phylogenetic groups and corresponding genera. Therefore, no changes in generic circumscriptions are advisable for the time being.

The Aegialitideae comprise Aegialitis which is recovered as monophyletic with high support in our phylogenetic analysis in which we included sequences for its two species A. annulata (two accessions) and A. rotundifolia (four accessions) that are also recovered as reciprocally monophyletic with high support (Suppl. material 2). In Plumbagineae, highly supported monophyly is inferred for Ceratostigma (Koutroumpa et al. 2018; Zhao et al. 2023) and Dyerophytum (Koutroumpa et al. 2018). Plumbago is non-monophyletic (Lledó et al. 2001, 2005a; Koutroumpa et al. 2018) and Plumbagella is a monospecific genus, therefore, monophyly at the genus level cannot be tested.

To our knowledge, very few phylogenetic studies have explored species limits within the genera of Plumbagineae. Zhao et al. (2023) sampled whole plastid genomes of multiple individuals from five Ceratostigma species in China in a phylogenomic framework and tested for species limits. The inferred phylogeny resolved the interspecific relationships within Ceratostigma with high support in almost all clades and species’ monophyly was confirmed for four out of the five species, namely C. griffithii C.B.Clarke, C. plumbaginoides Bunge, C. ulicinum Prain and C. willmottianum Stapf. Ceratostigma minus Stapf ex Prain was not monophyletic with its individuals found in two distinct clades corresponding to their geographical ranges. Individuals from Hengduan Mountains were sister to C. plumbaginoides and C. willmottianum and individuals from the Qinghai-Tibet Plateau were sister to C. griffithii. The authors attributed the non-monophyletic clustering of C. minus to genetic divergence promoted by geographical barriers in the two mountainous regions but they did not exclude a potential impact of hybridization and introgression to the observed topology. They also highlighted the necessity of comparing chloroplast genome phylogenies with those from nuclear genomes to further understand the evolutionary relationships between and within species.

Graham (2014) investigated the systematics of Dyerophytum using morphometric and molecular data (ITS and two cpDNA regions). She found D. africanum (Lam.) Kuntze highly supported as monophyletic and sister to a clade comprising D. indicum (Gibbs ex Wight) Kuntze, which was monophyletic in the ITS tree, D. pendulum (Balf.f.) Kuntze, and D. socotranum. The latter taxon was originally described as a variety of D. indicum (under the former generic name Vogelia Lam., i.e., Vogelia indica var. socotrana Balf.f.), but it was later unofficially raised to species level by J.R. Edmondson. This taxon was monophyletic in the ITS tree but formed a polytomy with D. pendulum in the cpDNA tree. In the morphometric analysis, individuals of all four Dyerophytum taxa were largely distinct. Consequently, we uphold all four taxa at the species level in this study and formally propose the new combination and status for Dyerophytum socotranum (see Taxonomic and nomenclatural novelties).

For Plumbagineae and Aegialitideae, we follow floras and e-floras (e.g., Peng and Kamelin 1996, Flora of China) in addition to the recent phylogenetic studies (e.g., Zhao et al. 2023) as secundum references.

Our current understanding of phylogenetic relationships and generic concept in the tribe Limonieae show that Armeria and Limonium are two well differentiated entities (e.g., Koutroumpa et al. 2018). However, the classification of these two genera is still historically linked via the name Statice and the names have been in part intermingled. Since the whole Plumbaginaceae were approached in this study, we also revisited the “Statice problem” and a few names could be resolved, resulting in three new combinations (see taxonomic and nomenclatural novelties below).

Tournefort (1694) considered Limonium and Statice as two distinct genera of sea-lavenders and thrifts, respectively. Linnaeus 1737, in his Genera Plantarum) and Species Plantarum (1753), combined them in the single genus Statice L. and regarded Limonium Tourn. as its synonym. Miller (1754, 1768) used the names Limonium Tourn. for sea-lavenders and Statice Tourn. for thrifts, based on the pre-Linnaean treatments. Willdenow (1809) separated the thrifts in the genus Armeria Willd. and called the sea-lavenders Statice L. (p.p.). This caused confusion in the circumscription of the three genera Armeria, Limonium, and Statice because following Willdenow’s publication, the names Armeria and Statice were both applied to the thrifts by different authors (Lawrence 1940) that led to the rejection of the name Statice as a nomen ambiguum. The name Armeria was conserved for the thrifts and Limonium for sea-lavenders in the International Botanical Congress of Cambridge 1930, following the report of the Editorial Committee for Nomenclature (see Bricquet 1935).

Rejection of Statice in favour of Armeria and Limonium caused part of the taxonomic and nomenclatural complexities in Plumbaginaceae. Only 18 species of Statice were described by Linnaeus (1753, 1762, 1767), 11 of which refer to species of Limonium and seven species are now synonyms of accepted species in Acantholimon, Armeria, Goniolimon, Limoniastrum, Myriolimon and Psylliostachys. Many new species were described under Statice before its rejection that are currently mostly synonyms of Limonium or Armeria but Statice names are spread over the Plumbaginaceae and can be found in synonymy status under 11 out of 26 genera of the family. This resulted in a multitude of required taxonomic recombination that is continuing even in this study. Kuntze (1891) recombined most of the described Statice species to their correct generic name. Despite several other works (Greuter and Raus 1987, 1989; Erben 2012), there are still 31 Statice names that have not yet been recombined or their status remains unresolved (classified under “unresolved names” in the database).

The comprehensive treatment in this study, attempting to completely cover Statice names, to clarify nomenclatural status and to highlight the still unresolved taxonomic and/or nomenclatural questions will facilitate future investigations.

Besides the accepted taxonomic subdivisions above species rank (e.g., subgenus, section, subsection), several informal terms are used by different authors to categorize species or infraspecies of similar morphology (Tutin et al. 1972; Greuter et al. 1984). These informal classifications mostly reflect the taxonomic complexity of species groups in which biological processes, such as apomixis, polyploidy and hybridization, play an important role in their speciation as demonstrated for example in Limonium (Koutroumpa et al. 2021).

Greuter et al. (1984) defined the term “aggregate” as an informal grouping for easily confused and morphologically allied and (probably) closely related species, the so-called “segregate species”. The following rule is given for naming aggregates: “an aggregate is designated by the oldest name, in terms of nomenclatural priority, of an included species, but without author citation”, yet this is not a formal rule and is not included in the International Code of Nomenclature for algae, fungi, and plants (Turland et al. 2018). In the Flora Europaea (Tutin et al. 1972), the term “group” was used for the same purpose, i.e., to group similar species that are difficult to distinguish. These groups, and not the individual species, were keyed out in the main identification key. A key to the component species within each group was provided along with descriptions for further study and identification. Furthermore, the term “complex” has also been used in literature, especially for Limonium and Armeria (e.g., Pignatti 1982; Baumbach and Hellwig 2007; Róis et al. 2018), to refer to morphologically related taxa, but its application sometimes lacks a global context.

Aggregates, complexes, and groups have been mostly used for Limonium in Plumbaginaceae due to the highly variable nature of its species, with several of them being polyploid and apomictic. These informal subdivisions contain many sexual and apomictic microspecies with narrow geographical distributions. For example, the Limonium binervosum aggregate is a complex assemblage of nine species and over 40 infraspecies in the British Isles (Ingrouille and Stace 1986) that are mostly raised to species level by Sell (2018) (e.g., L. anglicum (Ingr.) P.D.Sell), whereas Pignatti (1972) considered it as a group with only five species. Limonium binervosum (G.E.Sm.) C.E.Salmon is a widespread, apomictic, and variable species (Domina 2011+) but, has not been studied throughout its entire distribution area (Ingrouille and Stace 1986). Further examples of informal subdivisions in Limonium include the aggregates adopted by Greuter et al. (1989) to classify several Mediterranean species, the complexes recognized by Brullo and Guarino (2017) to classify species occurring in Italy and the groups adopted by Brullo and Erben (2016) for the Greek taxa.

All aggregates, complexes, and groups in Limonium are essentially regional classifications that refer to morphologically similar species and infraspecies in a restricted geographical area without considering the wider distribution of the group. It is therefore unclear which species should belong to a group, complex or aggregate when the range is expanded. In addition, the monophyly of these informal subdivisions in a phylogenetic framework has not yet been explored. Therefore, in our global Plumbaginaceae backbone we refrained from using aggregates, complexes or groups as ranks since they lack nomenclatural and phylogenetic status. A note is given in the database for each taxon that is part of a literature-based aggregate or group. These notes highlight the existing alpha-taxonomic confusion associated with the respective taxa.

This database encompasses a total of 4,301 scientific names in the family Plumbaginaceae, surpassing the count in other online databases: Tropicos (1991+) assigns 2,067 names to Plumbaginaceae, IPNI (2000+) lists 3,200 names, the World Checklist of Vascular Plants (Govaerts 2023) 3,727 names, and the Catalogue of Life Checklist (COL 2024) contains 3,769 names. The Global Biodiversity Information Facility’s (GBIF) backbone largely follows COL (using a 2023 version).

With respect to the taxonomy, the Catalogue of Life (COL 2024), sourced from Hassler (2023) includes two subfamilies, 23 genera, 1,149 species, 100 subspecies, and 30 varieties in the accepted names and 166 names are considered “ambiguous synonyms”, of these 158 are unambiguously placed as synonyms in our treatment and four as misapplied names.

The database will be updated continuously according to newly published results of taxonomic and phylogenetic studies and published online in the Caryophyllales portal (https://caryophyllales.org). Future versions that significantly differ from this one will be published as further stable and citable versions. Adding further information such as distribution, common names, protologue link, type species, morphological description, species keys, molecular data, photographs or link to the photographs, cytological data, conservation status, etc. is a future goal for the Plumbaginaceae database, with the initial priority set to nomenclatural types, protologue links and geographical distribution.

Following publication, this information will be stored in ChecklistBank (Döring et al. 2022) (https://www.checklistbank.org/) and shared with the WFO Plant List, contributing to the development of the WFO Backbone. Regular updates will be consistently incorporated into WFO to keep the information current.

The authors have declared that no competing interests exist.

No ethical statement was reported.

Verein der Freunde des Botanischen Gartens und Botanischen Museums Berlin-Dahlem e.V. Foundation of Women from endangered countries of the Freie Universität Berlin Iran National Science Foundation, INSF, Project number 98005681

MM: taxonomic and nomenclatural investigations based on the available literature, herbarium revisions and field studies, as well as data entry. WGB: preparation of the data for the initial imports and checks, nomenclatural issues, transformation of the EDIT Platform’s generic output to the document provided in Supplementary Material online, statistics and data quality control. MM, KK, NK and WGB: editorial control of the database. MBC, GD, HA, and KK: taxonomic and nomenclatural revisions of the draft checklist. MM, KK and TB: design of the structure of the manuscript. MM and KK: writing of the manuscript with inputs from TB. SvM provided initial software training for MM and contributed to the database and to an early version of the manuscript. All authors contributed to the text and approved its final version.

Maryam Malekmohammadi https://orcid.org/0000-0003-4641-7795

Konstantina Koutroumpa https://orcid.org/0000-0001-9391-874X

Manuel B. Crespo https://orcid.org/0000-0002-3294-5637

Gianniantonio Domina https://orcid.org/0000-0003-4184-398X

Nadja Korotkova https://orcid.org/0000-0003-0106-3209

Hossein Akhani https://orcid.org/0009-0003-8484-973X

Sabine von Mering https://orcid.org/0000-0003-2982-7792

Thomas Borsch https://orcid.org/0000-0002-5724-2786

Walter Berendsohn https://orcid.org/0000-0001-9945-7606

All of the data that support the findings of this study are available in the main text or Supplementary Information.