The use of DNA sequence data in plant systematics has brought us closer than ever to formulating well-founded hypotheses about phylogenetic relationships, and phylogenetic research keeps on revealing that plant genera as traditionally circumscribed often are not monophyletic. Here, we assess the monophyly of all genera of vascular plants found in Germany. Using a survey of the phylogenetic literature, we discuss which classifications would be consistent with the phylogenetic relationships found and could be followed, provided monophyly is accepted as the primary criterion for circumscribing taxa. We indicate whether and which names are available when changes in generic assignment are made (but do not present a comprehensive review of the nomenclatural aspects of such names). Among the 840 genera examined, we identified c. 140 where data quality is sufficiently high to conclude that they are not monophyletic, and an additional c. 20 where monophyly is questionable but where data quality is not yet sufficient to reach convincing conclusions. While it is still fiercely debated how a phylogenetic tree should be translated into a classification, our results could serve as a guide to the likely consequences of systematic research for the taxonomy of the German flora and the floras of neighbouring countries.
Version of record first published online on 24 March 2016 ahead of inclusion in April 2016 issue.
“All a taxonomist's decisions are subject to revision in time” (Davis & Heywood 1973), and indeed even the most cursory comparison of taxa at any rank through the history of plant systematics reveals that their circumscription has changed again and again. For example, as regards Rothmaler's “Exkursionsflora von Deutschland”, used as basis for the present paper, well over 40 genera have been subject to changes in circumscription when comparing the 19th and 20th editions of the Grundband (Jäger 2012). Major reasons for changes in taxon circumscription, as described and discussed in great detail by e.g. Davis & Heywood (1973) and Stuessy (2009), include the discovery of new species, availability of new data (characters), new approaches in data analysis, and, often related to the preceding point, changes in concepts of classification. At any point in time an author suggesting change of taxon circumscription will have believed to provide something “better”, where “better” had to be evaluated against the aim and purpose of the classification.
Post-Linnaean plant systematists (and indeed some systematists before Linnaeus) increasingly aimed at producing a “natural system” in which a priori selection of characters used for classification was replaced by the simultaneous evaluation of many characters (Davis & Heywood 1973; Stuessy 2009). With the publication of Darwin's (1859) “Origin of Species”, introducing the concept of evolution, “natural” obtained a new meaning, and “natural” taxa were interpreted as groups of common ancestry. Although “Post-Darwinian systems have differed little in content, though they have differed in arrangement, from those of the later pre-Darwinian taxonomists” (Davis & Heywood 1973), “After Darwin, virtually all comprehensive systems of classification of plants were avowedly phylogenetic” (Stuessy 2009).
We are far from having DNA sequences of all species, and probably even farther from resolving all relationships among species and higher level lineages. However, most plant systematists (hopefully) will agree that the use of DNA sequence data in plant systematics has brought us closer than ever to formulating solid hypotheses about phylogenetic relationships, which could serve as basis for classification. Perhaps ironically, exactly how to translate a phylogenetic tree into a classification has resulted in fierce debates. Probably the majority of authors will argue that the branching pattern of a phylogenetic tree should be the primary criterion for classification and that only monophyletic taxa (consisting of ALL descendants of one common ancestor) should be accepted. Some nevertheless maintain that nonmonophyletic taxa should also be accepted in order for classification to depict not only patterns of phylogenetic relationship, i.e. tree topology, but also degree of (phenotypic) divergence. (Most of these latter authors will describe the taxa they think of as paraphyletic; however, as is evident from Fig. 1, assessments of taxa as either paraphyletic or as polyphyletic based on tree topology alone are alternative ways to read a phylogenetic tree.) It is not our aim here to summarize or add to that discussion. For that, the reader is referred to a recent review by Schmidt-Lebuhn (2012), a proponent for recognizing only monophyletic taxa, and a response to that review by Stuessy & Hörandl (2014), opponents to that view. It is also not the aim of this paper to provide a general review of changes of genus concepts through time, which have been reviewed and discussed repeatedly (e.g. Humphreys & Linder 2009 and references therein).
Instead, our aims are: (1) Taking the generic circumscriptions of the 19th edition of the Grundband of Rothmaler (Jäger 2005) as starting point, to identify genera which are not or not unequivocally monophyletic. Such conclusions are based on a thorough survey of the phylogenetic literature. We make an effort to assess the quality of published phylogenies in terms of taxon sampling, DNA regions analysed and support for relationships identified. This sometimes results in the conclusion that a given genus may or may not be monophyletic, but that the data available are too preliminary for drawing taxonomic conclusions. Genera not included in our compilation below either are monophyletic or have not been investigated in detail sufficient to draw conclusions on the matter. Some of the changes that we discuss have already been incorporated in the most recent, 20th edition of the Grundband of Rothmaler (Jäger 2011), and a proportion of those have been justified by Jäger (2012). We nevertheless base our discussion on the previous, 19th edition (Jäger 2005) in order to explain the phylogenetic basis for all of these recent changes. (2) We will discuss, with reference to discussions in the literature, which classifications would be consistent with the phylogenetic relationships found and could be followed provided monophyly is accepted as the primary criterion for circumscribing genera (and taxa in general). In general, these options are either to expand genera in order to include former satellites (based, as it turned out, on single or few autapomorphic characters), or to split genera into smaller generic entities. Recent trends with respect to these two strategies have been discussed by Humphreys & Linder (2009). Where easily available, we indicate whether and which names could be used when changes in generic assignment are made. We do not, however, present a comprehensive review of the nomenclatural aspects of such names.
It is our aim to convince the users of Floras, who want to name plant species for very different reasons and perhaps more often than not are rather reluctant to accept new names, that the name changes discussed here reflect the progress of systematic botany and should be considered just as the results of other branches of biology or of any other science are considered when based on solid evidence.
We follow family circumscriptions and the linear order of families as found in Jäger (2011) and will not discuss these further. For more information on angiosperm family circumscriptions the reader is referred to Stevens (2001 onwards) and APG III (2009).
As already indicated above, several changes of the generic circumscriptions used in the 19th edition of Rothmaler (Jäger 2005) have been made in the 20th edition of that work (Jäger 2011), and some of these changes have been discussed and justified by Jäger (2012). Similarly, some changes resulting from novel phylogenetic work have been implemented by Buttler & Hand (2008a, 2008b, 2011, 2013) and Hand & Buttler (2009, 2012, 2014) in their “Liste der Gefäßpflanzen Deutschlands”. Work similar to that presented here has been presented for other European Floras (British Isles: Stace 2010; Italy: Banfi & al. 2005, 2011; Nordic countries: http://euphrasia.nu/checklista/ and http://www.slu.se/en/collaborative-centres-and-projects/dyntaxa/) and some Floras of other parts of the world have explicitly accepted only presumably monophyletic taxa (e.g. Baldwin & al. 2012, 2015).
In the end, it is left to the authors of Floras to decide which principles and strategies they follow when circumscribing genera (and other taxa). However, as pointed out at various times in history (for references see Humphreys & Linder 2009), judgement should always be based on global and not regional knowledge. For example, it may appear shocking from a German perspective that Anagallis L. (incl. Centunculus L.), Glaux L. and Trientalis L. are all nested in a highly paraphyletic Lysimachia L. and should be included in the latter genus (Manns & Anderberg 2009; see below for details), but from a global perspective it is rather less so. Anderberg & al. (2007) pointed out similarities between, e.g. Anagallis arvensis L. and Lysimachia nemorum L., between Trientalis and the North American Lysimachia subg. Seleucia Bigelow and the South American Lysimachia subg. Theopyxis (Grisebach) J. D. Ray, and, while acknowledging their morphological distinctness, between Glaux and Lysimachia mauritiana Lam.
In the following, we describe and discuss the results of our literature survey. We looked at ALL genera contained in Jäger (2005) but present results only for genera that are not or not unequivocally monophyletic. The work presented here is work in progress. Any comment on what we have written is highly welcome and will help us in future updates of this paper.
Lycophytes and ferns
Lycopodiaceae (L.-B. Zhang)
Diphasiastrum was resolved as embedded within Lycopodium L. (Wikström & Kenrick 1997, 2001). Thus recognition of Diphasiastrum Holub as a separate genus would make Lycopodium paraphyletic, and Zhang & Iwatsuki (2013) suggested inclusion of Diphasiastrum in Lycopodium. However, this finding was based on plastid rbcL data and limited sampling only, and a final decision should await better sampling and use of additional DNA sequences.
Hymenophyllaceae (L.-B. Zhang)
A new classification of the family recognized only nine genera (Ebihara & al. 2006), and Trichomanes speciosum Willd. should now be known as Vandenboschia speciosa (Willd.) G. Kunkel. Trichomanes L. in a new circumscription is a mainly neotropical genus with a few species in continental Africa, Madagascar and the Indian Ocean (Ebihara & al. 2006) and was resolved as sister to Vandenboschia Copel. based on plastid rbcL data (Ebihara & al. 2007).
Aspleniaceae (L.-B. Zhang)
Plastid data resolved the family into two well-supported clades, Asplenium L. and Hymenasplenium Hayata (van den Heede & al. 2003; Schneider & al. 2004), which have different chromosome base numbers as well as distinct root characters (Murakami 1995; Schneider 1996). All other small segregate genera are nested within Asplenium (van den Heede & al. 2003; Schneider & al. 2004). Thus, synonymization of Ceterach Willd. and Phyllitis Hill with Asplenium is advocated (e.g. Smith & al. 2006; Lin & Vianne 2013). Consequently, P. scolopendrium (L.) Newman should be A. scolopendrium L. and C. officinarum Willd. should be A. ceterach L.
Thelypteridaceae (L.-B. Zhang)
Lastrea Bory was resolved as part of Oreopteris Holub based on plastid markers (He & Zhang 2012), and L. limbo sperma (All.) Ching should now be known as O. limbosperma (All.) Holub.
Hydrocharitaceae (J. W. Kadereit)
A phylogenetic analysis of the family based on nuclear, plastid and mitochondrial DNA sequences (Chen & al. 2012) provides some evidence that Egeria Planch, may not be monophyletic when Elodea Michx. is treated as a distinct genus. As only two of five species of Elodea were included in that study, and support in the relevant part of the tree is not entirely convincing, treatment of the two genera as separate is acceptable for the time being. If combined, as has been done in the past (for discussion see Les & al. 2006), Elodea would be the name to be used.
Zosteraceae (J. W. Kadereit)
The finding that Heterozostera tasmanica (M. Martens ex Asch.) Hartog is deeply nested in Zostera L. (Les & al. 1997; Les & al. 2002; Kato & al. 2003; Tanaka & al. 2003) opens the option to maintain Zostera including Heterozostera (Setch.) Hartog as one genus, or to divide this group into two or three genera. In both latter options Z. marina L. would remain in Zostera and Z. noltii Hornem. would have to be combined in Nanozostera Toml. & Posl. as N. noltii (Hornem.) Toml. & Posl. Subdivision into three genera has been advocated and justified with morphological distinctness in inflorescence and vegetative characters by Tomlinson & Posluzny (2001), and maintainance of Zostera as one genus has been recommended by Les & al. (2002).
Potamogetonaceae (J. W. Kadereit)
It has been shown that species of Potamogeton subg. Coleogeton (Rchb.) Raunk. constitute a monophyletic lineage which is well-supported sister to the remainder of Potamogeton L. (Lindqvist & al. 2006). As this lineage is morphologically well characterized, as well summarized by Preston (2005; but see also Wiegleb & Kaplan 1998), it could be separated at generic rank as Stuckenia Börner, as argued by Lindqvist & al. (2006) and other authors (Les & Haynes 1996; Holub 1997; Haynes & al. 1998; Kaplan 2008), or could be maintained within Potamogeton as argued by Wiegleb & Kaplan (1998). If treated as a distinct genus, P. pectinatus L. should be known as S. pectinata (L.) Börner and P. filiformis Pers. as S. filiformis (Pers.) Börner.
Dioscoreaceae (J. W. Kadereit)
Tamus L. is clearly nested in Dioscorea L. (Caddick & al. 2002a, 2002b; Wilkin & al. 2005). In consequence, T. communis L. should be treated as D. communis (L.) Caddick & Wilkin. An alternative option, to split Dioscorea into many smaller genera, as suggested by Huber (1998), was discussed but rejected by Caddick & al. (2002b).
Liliaceae (J. W. Kadereit)
According to studies by Peterson & al. (2008; see also Peterson & al. 2004) and Zarrei & al. (2009), all based on a broad species sample and both nuclear and plastid sequences, a non-monophyletic Lloydia Rchb. is nested in Gagea Salisb. If Lloydia should be included in Gagea, as suggested by Peruzzi & al. (2008) and Zarrei & al. (2011), a name for L. serotina (L.) Rchb. in Gagea would be available (G. serotina (L.) Ker Gawl.).
Orchidaceae (M. Kropf)
Initiated by molecular phylogenetic studies by Pridgeon & al. (1997) and Bateman & al. (1997), European orchids, and especially the genus Orchis L. s.l., have become a prime example for recent rearrangements in generic delimitations (Stace 2010). Although subsequent phylogenetic studies (cf. Cozzolino & al. 1998, 2001; Aceto & al. 1999; Bateman 2001; Pridgeon & al. 2001; Bateman & al. 2003) generated support for (most of) these rearrangements (but almost exclusively based on ITS sequence variation only), most remained subject to fierce debates in the orchid community (cf. Wucherpfennig 1999, 2002, 2005; Bateman 2001, 2009, 2012a, 2012b; Buttler 2001; Devos & al. 2006; Kretzschmar & al. 2007; Tyteca & Klein 2008, 2009; Scopece & al. 2010; Paulus 2012; Tyteca & al. 2012). Possible and partially implemented rearrangements (cf. Jäger 2012) include either splitting of polyphyletic genera into smaller genera (e.g. Orchis s.l.), or inclusion of genera, either with several species (e.g. Nigritella Rich. in Gymnadenia R. Br.) or monospecific (e.g. Aceras anthropophorum (L.) R. Br. in Orchis s.str.), in (otherwise) paraphyletic genera in order to obtain monophyletic entities.
Phylogenetic studies placed the (previously) monospecific Aceras anthropophorum close to Orchis italica Poir. (Pridgeon & al. 1997; Bateman & al. 2003). This close relationship at the base of the Orchis s.str. group (see below) was not only supported by the original ITS sequence data, but also by seed ornamentation patterns (Gamarra & al. 2012), hybridization patterns (Klein 1989, 2004; Scopece & al. 2007), and the nuclear OrcLFY, OrcPI, OrcP2 loci (Montieri & al. 2004; Cantone & al. 2009, 2011), although support by the mitochondrial coxl marker (Inda & al. 2010a) and chloroplast rpl16 intron data (Inda & al. 2012) was ambiguous due to low resolution. Therefore, Bateman (2012a: 111–114) noted that the “most obviously problematic taxa are the readily recognized anthropomorphic species Orchis (Aceras) anthropophora (L.) All. and O. italica … shown as the two earliest-diverging species, making the anthropomorphic species paraphyletic relative to a monophyletic nonanthropomorphic group”. Given the absence of a final solution to the question which taxon is indeed basally branching in Orchis s.str. (i.e. a sister group relationship between A. anthropophorum and Orchis s.str. is still possible; see Pridgeon & al. 1997; Bateman & al. 2003; see also Jacquemyn & al. 2011), and the still debated future treatment of Orchis s.str. in general (see below), one could also retain A. anthropophorum as the only species of Aceras and the only Orchis-like species without a spur. On the other hand, the inclusion of Aceras in Orchis s.str. is one of the most widely accepted changes of controversial generic circumscriptions in European orchids (see Bateman 2009: Tab. 1).
ITS phylogenies implied inclusion of Coeloglossum viride (L.) Hartm. in an otherwise paraphyletic Dactylorhiza Necker ex Nevski (as Dactylorhiza viridis (L.) R. M. Bateman & al.; see Bateman & al. 1997, 2003; Pillon & al. 2007). Further molecular markers, especially chalcone synthase variation (Inda & al. 2010b; see also Inda & al. 2010a, 2012), supported this inclusion because C. viride was found nested in Dactylorhiza. However, evidence against its inclusion exists, and a combined ITS and ETS phylogenetic tree resolved C. viride as sister to Dactylorhiza (Devos & al. 2006). The latter authors also compiled morphological characters differing between the two groups (Devos & al. 2006: Table 1; see also Wucherpfennig 1999). Most strikingly, C. viride has a nectariferous spur (van der Pijl & Dodson 1966), whereas Dactylorhiza has food-deceptive flowers. As Coeloglossum Hartm. is the earlier name, a proposal to conserve Dactylorhiza over Coeloglossum was needed (Cribb & Chase 2001).
As a consequence of studies uncovering the (morphological) heterogeneity and the phylogenetic intermingling of different infrageneric species groups of the closely related genera Liparis Rich. and Malaxis Sol. ex Sw., the monospecific Hammarbya paludosa (L.) Kuntze, certainly a close relative of these two genera (although not sampled in the respective studies; e.g. Cameron 2005), “has often been included in a broadly defined genus Malaxis” (Pridgeon & al. 2005: 464–465). There are a number of unique features characterizing H. paludosa (e.g. incumbent anthers (Szlachetko & Margońska 2002), vegetative reproduction by bulbils at the leaf margin), which, however, have been doubted to be sufficient for differentiation at the generic level given the high variation in Malaxis s.l. (Wucherpfennig 2005). Independently, and referring to recent (but still unpublished) phylogenetic analyses by G. A. Salazar, Pridgeon & al. (2005: 464–465) stated that H. paludosa “does not lie in the main Malaxis clade (Salazar, pers. comm.) but rather is sister to a large clade that includes both Malaxis s.str. and Liparis s.str.” Until comprehensive phylogenetic evidence on relationships among Malaxis and Liparis becomes available (see below), H. paludosa presently can be maintained in a monospecific genus.
The monospecific Chamorchis alpina (L.) Rich. and the dispecific Traunsteinera Rchb., represented by the widespread T. globosa (L.) Rchb. in Germany, form an independent, well-supported clade (Cozzolino & al. 2001; Bateman & al. 2003). This surprising result refutes the originally hypothesized sister group relationship between the latter taxon and Orchis s.str. (Pridgeon & al. 1997) and induced Pridgeon & al. (2005: 228) to state: “However, the two morphologically distinct genera are sufficiently similar in ITS sequences to be potentially viewed as congeneric”. If treated as congeneric, Chamorchis represents the older genus name (cf. Alrich & Higgins 2011).
In (still unpublished) molecular phylogenetic analyses by Bateman and colleagues, Neottia nidus-avis (L.) Rich. is nested within a paraphyletic Listera R. Br. as sister to L. ovata (L.) R. Br. (illustrated by Pridgeon & al. 2005: 492). Already Chase & al. (2003) had considered Listera species as photosynthetic members of Neottia Guett. (without presenting the respective phylogenetic analysis, except for a “summary”, i.e. Fig. 1 on p. 73) and suggested that the two genera should be combined (see also Tesitelová & al. 2015, where Neottia (n = 2) is nested within Listera (n = 3) based on ITS, 18S and trnL(UAA) intron data). A close relationship between the two genera has long been documented (e.g. Dressler 1990), and a combined generic treatment as Neottia, which is the older name, has already been published by Szlachetko (1995). In this treatment, however, Neottia ovata Bluff & Fingerh. is placed in N. subg. Listera (R. Br.) Szlach. The third species of a newly circumscribed Neottia native in Germany is N. cordata (L.) Rich. (L. cordata (L.) R. Br.).
Pridgeon & al. (1997) “took the controversial step of sinking the morphologically distinct Nigritella back into synonymy with Gymnadenia s.str., which would otherwise have been paraphyletic.” (Pridgeon & al. 2001: 229). These authors stressed that “despite superficial differences in flower form and resupination, Nigritella shares several morphological characters with Gymnadenia: palmatedigitate tubers; two lateral, lobe-like stigmas; and two pollinia each with a caudicle…” (Pridgeon & al. 2001: 298). However, other authors, especially Wucherpfennig (1999, 2002), advocated maintaining Nigritella as a genus based on at least ten (“superficial”) morphological characters, but also based on allozyme data (Hedrén & al. 2000). It was noted that a study of character evolution across Orchidinae clearly showed that Nigritella is a morphologically derived lineage (Wucherpfennig 2002) arguing for keeping the genus Nigritella even within a paraphyletic Gymnadenia. However, in a more recent molecular analysis of ITS and rpl16 intron sequences, Pillon & al. (2006) documented a sister group relationship between their Nigritella (n = 2) and Gymnadenia (n = 5) samples. This illustrates that molecular phylogenetic relationships obtained obviously depend on taxon sampling, type of data analyses performed and outgroup selection (see Pillon & al. 2007). In consequence, Nigritella can still be recognized as a morphologically well-defined genus, until more comprehensive analyses are available.
Finally, species of Orchis s.l. were placed in at least three major and only distantly related groups based on ITS data (Bateman & al. 1997, 2003; Pridgeon & al. 1997). These three groups in principle correspond to hybridization patterns (Klein 1989, 2004; Scopece & al. 2007). As regards the first group, the formerly monospecific Neotinea Rchb. f. was expanded by Pridgeon & al. (1997) and Bateman & al. (1997) to encompass the “… small-flowered, essentially trilobed-lipped species of the ustulata-group that were formerly included in Orchis s.l. These could in theory have been treated as a genus separate from the more narrowly delimited original concept of Neotinea, given the relatively long molecular branch and distinct vegetative markings of N. maculata…” (Pridgeon & al. 2001: 228). Relevant for the German flora, the combinations N. ustulata (L.) R. M. Bateman & al. (O. ustulata L.) and N. tridentata (Scop.) R. M. Bateman & al. (O. tridentata Scop.) were provided (Bateman & al. 1997). However, the small flowers of N. maculata (Desf.) Stearn are different from the ustulata-group by producing nectar (Pridgeon & al. 2001; Duffy & al. 2009), and by being 100 % autogamous (Duffy & al. 2009), while the species of the deceptive ustulata-group depend on pollinatormediated outcrossing. This would provide arguments for treating N. maculata as an independent genus. If this approach is taken, the names Odontorchis ustulata (L.) D. Tyteca & E. Klein and Odontorchis tridentata (L.) D. Tyteca & E. Klein are available (Tyteca & Klein 2008).
The second fairly well-supported clade encompasses all species of former Orchis that have 2n = 36 (or 2n = 32 in the case of O. papilionacea L.) chromosomes as well as Anacamptis pyramidalis (L.) Rich. (Pridgeon & al. 1997). Pridgeon & al. (2001: 255) stated that while “A. pyramidalis is distinctive… The other members of this newly circumscribed genus Anacamptis Rich. are difficult to distinguish morphologically from Orchis s.str., but their flowering stems bear cauline sheathing leaves.” Members of Anacamptis in this new circumscription in the German flora are A. coriophora (L.) R. M. Bateman & al. (O. coriophora L.), A. morio (L.) R. M. Bateman & al. (O. morio L.) and A. palustris (Jacq.) R. M. Bateman & al. (O. palustris Jacq.).
The remaining Orchis s.l. taxa should then, following Pridgeon & al. (1997) and Bateman & al. (1997), be treated as Orchis s.str. comprising an anthropomorphic species group (with flowers shaped like “little men”, i.e. sepals and petals forming a compact head and the labellum showing “arms” and “legs”; e.g. O. militaris L., the type of Orchis) plus Aceras R. Br. (see above) and a non-anthropomorphic group (e.g. O. mascula (L.) L.). However, suggestions have been put forward to split Orchis s.l. further (Tyteca & Klein 2008, 2009), partly ignoring problems of paraphyly (criticized, e.g. by Scopece & al. 2010; Bateman 2012a). However, the two supported species groups within Orchis s.str. (Bateman & al. 2003) could be treated as O. subg. Orchis (i.e. O. militaris, O. purpurea Huds. and O. simia Lam.) and O. subg. Masculae H. Kretzschmar & al. (i.e. O. mascula, O. pallens L. and O. spitzelii Saut. ex W. D. J. Koch; Kretzschmar & al. 2007). Tyteca & al. (2012) compiled morphological and pollinator assemblage data for these two groups and concluded that all their information as well as molecular (Bateman & al. 2003) and seed micromorphology data (Gamarra & al. 2012) are in favour of a separation at the generic level, i.e. as Orchis and Androrchis D. Tyteca & E. Klein (Tyteca & al. 2012; see also Tyteca & Klein 2008 for respective names, i.e. Androrchis mascula (L.) D. Tyteca & E. Klein, A. pallens (L.) D. Tyteca & E. Klein and A. spitzelii (Saut. ex W. D. J. Koch) D. Tyteca & E. Klein).
Several orchid genera have been shown not to be monophyletic: Liparis and Malaxis, both comprising about 250 species (Mabberley 2008), are to some extent intermingled (Cameron 2005); Platanthera Rich. should include Piperia Rydb. (Bateman & al. 2003; already implemented there); and Herminium L. is phylogenetically intermingled with Peristylus Blume or Habenaria Willd. (Douzery & al. 1999; Bateman & al. 2003). However, irrespective of exact phylogenetic relationships, which are not yet completely resolved, the nomenclature of the species occurring in the German flora will not be influenced if their respective monophyletic clades are preserved at the generic level, as Liparis loeselii (L.) Rich., Platanthera bifolia (L.) Rich. and Herminium monorchis (L.) R. Br. are the types of the respective genus names (Alrich & Higgins 2011), and Platanthera montana (F. W. Schmidt) Rchb. f. (P. chlorantha Cust. ex Rchb.), the second native species of this genus, is definitely closely related to the type, P. bifolia (Bateman & al. 2003). However, Malaxis monophyllos (L.) Sw. might be affected by future changes: a BLAST search of recently published matK barcodes of this species (Kim & al. 2014; Xiang & al. 2014) revealed higher DNA sequence similarity to a group of Liparis species around the type, L. loeselii, than to the Malaxis species group around the type, M. spicata Sw. (cf. Cameron 2005). On the other hand, this critical point in the systematics of Malaxideae could alternatively be solved by choosing a wide genus concept. In this case, Malaxis would be an older name than Liparis (and Hammarbya; see above).
In summary, one major problem with respect to several recently suggested changes in generic circumscription in European orchids is that new molecular phylogenetic hypotheses often are based on only one molecular marker (i.e. ITS; Bateman & al. 2003). Other molecular markers often resulted in limited phylogenetic resolution given the probably young age of several European orchid lineages (cf. Inda & al. 2010a, 2010b, 2012). Although sometimes combined evidence of ITS plus cpDNA variation seems to improve results (e.g. Pillon & al. 2006), it does not in other cases, indicating the dominance of the ITS information (e.g. Inda & al. 2012). Moreover, it is striking that the overall taxon sampling, some 20 years after the first molecular phylogenetic publications, is still incomplete. Also, multiple samples of single taxa have rarely been included. In consequence, many molecular phylogenetic relationships have still not been solved satisfactorily, and some nomenclatural changes accordingly are premature, giving rise to frequent debate.
Amaryllidaceae (J. W. Kadereit)
Using a broad sample of Galanthus L. and Leucojum L., Lledó & al. (2004) reported that the former genus is deeply nested in the latter. In order to maintain these two genera, the authors recommend to recognize the additional genus Acis Salisb. for large parts of Leucojum. Generic allocation of G. nivalis L., L. aestivum L. and L. vernum L. would remain unaffected if this approach would be taken.
Cyperaceae (B. Gehrke)
Carex L. has been found to be paraphyletic and to include all other members of the Cariceae, i.e. Cymophyllus Mack., Kobresia Willd., Schoenoxiphium Nees and Uncinia Pers. (Roalson & al. 2001; Starr & al. 2004). The results of the molecular phylogenetic work are unambiguous. Retaining Kobresia would lead to the necessity of describing a myriad of morphologically indistinguishable smaller genera and would also mean that Kobresia would have to be either greatly extended to include many species of Carex subg. Psyllophora (Degl.) Peterm. (= Primocarex Kük.) or that Kobresia (and Uncinia) would have to be split into various smaller lineages. Combination of all names of Cymophyllus, Kobresia, Schoenoxiphium and Uncinia in Carex are currently underway (Global Carex Group 2015). The names Carex myosuroides Vill. for Kobresia myosuroides (Vill.) Fiori and Carex simpliciuscula Wahlenb. for K. simpliciuscula (Wahlenb.) Mack. should be used.
Eleogiton (L.) Link was recently discovered to be nested in Isolepis R. Br. (Muasya & al. 2001). Isolepis was thought to be characterized by having one or more pseudolateral spikelets and an erect culm, but the nodding culm of the single terminal spikelet, believed to characterize Eleogiton, is now known to have evolved from within Isolepis (Muasya & al. 2001). Eleogiton fluitans (L.) Link was therefore recently changed to I. fluitans (L.) R. Br.
Recent studies suggest that Schoenoplectus mucronatus (L.) Palla and S. supinus (L.) Palla are not part of Schoenoplectus (Rchb.) Palla, but belong to Schoenoplectiella Lye, a cosmopolitan group, which is most closely related to the African Pseudoschoenus (C. B. Clarke) Oteng-Yeb. (Shiels & al. 2014). Schoenoplectiella differs morphologically from Schoenoplectus by having an unbranched inflorescence (Jung & Choi 2010), whereas Schoenoplectus has a pseudo-lateral branched inflorescence. Both genera have culm-like primary bracts opposed to the inflorescence with leafy bracts in Scirpus L. (Jung & Choi 2010). If recognition of Schoenoplectiella as suggested by Lye (2003) should be accepted, both S. mucronatus and S. supinus must be excluded from Schoenoplectus as Schoenoplectiella mucronata (L.) J. Jung & H. K. Choi and Schoenoplectiella supina (L.) Lye. However, final decisions must await a better understanding of relationships between Pseudoschoenus and Schoenoplectiella.
Poaceae (M. Röser)
A number of molecular phylogenetic studies employing nuclear and chloroplast DNA markers have shown that Festuca L. s.l. is a large paraphyletic group that encompasses Lolium L., Micropyrum (Gaudin) Link, Vulpia C. C. Gmelin and a number of further genera (Torrecilla & Catalán 2002; Catalán & al. 2004, 2007; Torrecilla & al. 2004; Inda & al. 2008). Lolium is nested within a more ancestral broad-leaved clade, whereas Micropyrum and Vulpia belong to the presumably more recently derived fine-leaved Festuca lineages. Vulpia additionally appears to be polyphyletic and encompasses separate diploid and tetraploid/hexaploid lineages, which are not sufficiently understood to date. Because of several uncertainties concerning limited sampling of intermediate taxa and missing representation of several Festuca groups, Catalán & al. (2007) argued for maintenance of Lolium, Micropyrum and Vulpia. This would require no name changes for taxa of the German flora. Micropyrum and Vulpia were included in Festuca by Soreng & al. (2015), but Lolium was kept separate and considered congeneric with Schedonorus P. Beauv. (syn. F. subg. Schedonorus (P. Beauv.) Peterm.), which was segregated from Festuca.
Polyploidy and hybridization play an important role in the evolution of Sesleria Scop. and allies. Preliminary data from Amplified Fragment Length Polymorphisms (AFLPs) and plastid DNA (trnL-ndhF) sequences support the recognition of Oreochloa Link as a separate genus (with only O. disticha (Wulfen) Link represented in the German flora) and underline that Psilathera ovata (Hoppe) Deyl diverges from the remainder of Sesleria (Lakušić 2013). Further study including a more comprehensive taxon sampling is needed to clarify whether the monospecific Psilathera Link (only P. ovata (Hoppe) Deyl in the German flora) can be maintained or should be merged with Sesleria as was done in Jäger (2011) and by Lazarević & al. (2015).
Delimitation of genera allied with Helictotrichon Besser ex Schult. & Schult. f. is a long-term matter of debate. Molecular phylogenetic studies using different chloroplast DNA and nuclear ITS sequences of a sufficiently broad sample of relevant taxa suggest to acknowledge three genera occurring in the German flora, namely Avenula (Dumort.) Dumort., Helictochloa Romero Zarco and Helictotrichon s.str. (Döring & al. 2007; Quintanar & al. 2007; Schneider & al. 2009; Röser & al. unpubl. data). Avenula is represented by A. pubescens (Huds.) Dumort., Helictochloa by H. pratensis (L.) Romero Zarco and H. versicolor (Vill.) Romero Zarco and Helictotrichon s.str. by H. parlatorei (Woods) Pilg. The description of the new genus Helictochloa, type designations and transfer of species to Helictochloa have been published by Romero Zarco (2011).
The distinctiveness of Anthoxanthum L. and Hierochloe R. Br. has repeatedly been questioned due to the occurrence of seemingly intermediate species in Africa and SE Asia. Following Schouten & Veldkamp (1985), the two genera have been merged by some authors (Wu & Phillips 2006; Allred & Barkworth 2007; Kellogg 2015; Soreng & al. 2015). The study by Pimentel & al. (2013), using AFLPs, chloroplast and nuclear DNA sequences, suggests that the intermediate taxa originated by ancient hybridization between the two genera. The question as to whether Anthoxanthum and Hierochloe should be kept separate or amalgamated in a single genus thus remains unanswered.
Ranunculaceae (E. Welk)
Traditionally Aconitum L., Consolida (DC.) Gray and Delphinium L. (and Aconitella Spach, see Soják 1969) were grouped in tribe Delphinieae. Molecular phylogenetic research revealed three Delphinium species (D. subg. Staphisagria J. Hill) to form the sister clade to all other Delphinieae (Jabbour & Renner 2011a; 2011b), and Consolida incl. Aconitella to be nested in Delphinium excl. D. subg. Staphisagria. The position of D. subg. Staphisagria is supported by biochemical, karyological and morphological characters. Furthermore, Wang & al. (2013) found a sister position of the Chinese Aconitum gymnandrum Maxim. to Delphinium (sensu Jabbour & Renner 2012). In order to arrive at monophyletic Aconitum and Delphinium, name changes are required. Of these, inclusion of Consolida (and Aconitella) into Delphinium (Jabbour & Renner 2012) is relevant for the German flora. In consequence, C. ajacis (L.) Schur, C. hispanica (Costa) Greuter & Burdet and C. regalis Gray should be listed as D. ajacis L., D. hispanicum Costa and D. consolida L., respectively.
Based on molecular phylogenetic analyses, Bittkau & Comes (2009) found Garidella L. to be clearly monophyletic while Nigella L., its sister group, was not well supported as monophyletic. This may imply future inclusion of Garidella in Nigella, which, however, would not affect naming of the German species of Nigella.
Combined analyses of DNA sequence data, biochemical data and morphology by Compton & al. (1998) suggested to include Cimicifuga Wernisch and Souliea Franch. in Actaea L. (also Compton & Culham 2002; Gao & al. 2008). However, it has also been argued to keep the genera separate based on the fleshy fruits of Actaea (e.g. Wang & al. 1997; Lee & Park 2004). Actaea and Cimicifuga can also be distinguished using seed morphology and seed anatomical features (Ghimire & al. 2015). If Cimicifuga and Souliea should be included in Actaea based on phylogenetic relationships, German Actaea will not be affected because Actaea L. is the oldest genus name.
Hoot & al. (1994) suggested that Hepatica Mill., Knowltonia Salisb. and Pulsatilla Mill. should be included in Anemone L. s.l. (cf. Ehrendorfer & Samuel 2001; Schuettpelz & al. 2002). However, Pfosser & al. (2011) argued that these genera could also be retained because of unsuitable outgroup selection (Clematis L.) in Hoot & al. (1994) and Schuettpelz & al. (2002). Using Ranunculus ficaria L. as outgroup in their study, a position of Clematis within Anemone s.l. became probable. The sister-group relationship of species of A. subg. Anemonidium (Spach) Juz. (A. subsect. Anemonidium Spach, A. subsect. Himalayicae (Ulbr.) Tamura, A. subsect. Keiskea Tamura and A. subsect. Omalocarpus (DC.) Tamura) to Hepatica found in all studies renders Anemone paraphyletic in relation to the embedded Hepatica and Pulsatilla. Similar to combined karyological and molecular phylogenetic analyses by Mlinarec & al. (2012), Hoot & al. (2012) found, again with Clematis as outgroup, that A. subg. Anemonidium contains Anemonastrum Holub and Hepatica, while Pulsatilla is positioned within A. subg. Anemone. Accordingly, they suggested to incorporate Hepatica in Anemone as A. sect. Hepatica (Mill.) Spreng. or A. subg. Hepatica (Mill.) Peterm. For Pulsatilla they suggested inclusion in Anemone as A. sect. Pulsatilloides DC. or A. subg. Pulsatilloides (DC.) Juz. An alternative solution might be splitting Anemone into at least two genera corresponding to the x = 7/8 divergence seen in Anemoninae. At the moment, it seems best to wait for further analyses before combining the large number of taxa affected. However, from the results of all studies cited it seems inevitable for Anemonastrum Holub to be subsumed in Anemone again. The resulting combination is Anemone narcissiflora L.
Caltha L. has been divided into two sections: the monophyletic C. sect. Psychrophila (DC.) Bercht. & J. Presl in the S hemisphere and the paraphyletic C. sect. Caltha in the N hemisphere (Schuettpelz & Hoot 2004). Based on a broader sampling, Cheng & Xie (2014) showed that Thacla Spach (Caltha natans Pall.) diverged first in the genus, and that the other species fall into two monophyletic clades, i.e. Caltha s. str. and Psychrophila. Thus, it would be possible to raise Psychrophila to genus rank, but this would inevitably require C. natans to be raised to Thacla. Any decision here will not affect the name of C. palustris L.
A number of molecular phylogenetic studies revealed that Ranunculus L. in a wide sense is polyphyletic (Lehnebach & al. 2007; Hoot & al. 2008; Wang & al. 2009). Although the entire tribe Ranunculeae could be recognized as a very broadly circumscribed Ranunculus, this would result in a morphologically highly heterogeneous group. The morphological and geographical independence of Ficaria Huds. and Ceratocephala Moench is comparable to that of Myosurus L. It thus seems to be justified to follow Emadzade & al. (2010) who proposed to recognize Ceratocephala, Ficaria and Myosurus (plus several other small genera) as separate genera, but to include Batrachium (DC.) Gray and Aphanostemma A. St.-Hil., sometimes recognized as separate genera in the past, in a then monophyletic Ranunculus.
Berberidaceae (J. W. Kadereit)
Monophyletic Berberis L. with simple leaves clearly is nested in a paraphyletic grade of Mahonia Nutt. with compound leaves (Kim & al. 2004; Adhikari & al. 2015), a pattern of relationship already postulated by Ahrendt (1961). As the two genera are very similar to each other in many respects (for discussion see Adhikari & al. 2012), and the different lineages of Mahonia would be difficult to justify at generic rank, they probably are best treated as one genus, Berberis, as was done by these authors. Mahonia aquifolium (Pursh) Nutt. had originally been described as B. aquifolium Pursh.
Papaveraceae (J. W. Kadereit)
Papaver L. is part of a group of four genera distributed almost entirely in the Old World (Schwarzbach & Kadereit 1995). The other three genera are Meconopsis Vig., Roemeria Medik, and Stylomecon G. Taylor. Subdivision into these four genera is based largely on capsule morphology. Various analyses of these four genera (Kadereit & al. 1997; Carolan & al. 2006; Kadereit & al. 2011; Xiao 2013; Liu & al. 2014) revealed that patterns of relationship cut across traditional generic delimitations (see also Fig. 1). First, three subgroups of Papaver, i.e. (1) Papaver s.str. (all sections except P. sect. Argemonidium Spach, P. sect. Californica Kadereit, P. sect. Horrida Elkan and P. sect. Meconella Spach), (2) P. californicum A. Gray (P. sect. Californica) and (3) P. aculeatum Thunb. (P. sect. Horrida) form a clade together with Meconopsis cambrica (L.) Vig. and Stylomecon heterophylla (Benth.) G. Taylor. Second, P. sect. Argemonidium is most closely related to Roemeria. Third, P. sect. Meconella Spach is most closely related to one of three subgroups of Meconopsis. While this pattern of relationships allows several classifications, the following option has partly been followed (Kadereit & Baldwin 2011; Kadereit & al. 2011). A newly circumscribed Papaver should contain Meconopsis cambrica, Papaver s.str., P. aculeatum, P. californicum and Stylomecon heterophylla. Of the species found in Germany, P. confine Jord., P. dubium L., P. lecoqii Lamotte and P. rhoeas L. would remain in Papaver. Meconopsis cambrica was originally described as P. cambricum L., and the name P. heterophyllum (Benth.) Greene is available for Stylomecon heterophylla. Papaver sect. Argemonidium, represented by P. argemone and P. hybridum in the German flora, should be united with Roemeria, with which it shares sepal and pollen characters (Kadereit & al. 1997). The combination R. argemone (L.) C. Morales & al. is available for P. argemone.
Although Papaver alpinum L. as part of P. sect. Meconella should clearly be excluded from Papaver, the exact relationships of P. sect. Meconella to Himalayan Meconopsis are not sufficiently clear yet to suggest a formal name. However, it seems to be sister clade to a newly circumscribed Meconopsis (excl. Cathcartia Hook. f. and M. cambrica) and probably is best treated as a distinct genus.
Crassulaceae (J. T. Klein)
Sedum L. has repeatedly been shown to be highly polyphyletic (van Ham & al. 1994; van Ham & 't Hart 1998; Mort & al. 2001; Mayuzumi & Ohba 2004; Gontcharova & al. 2006; Carrillo-Reyes & al. 2009). In the most recent phylogenetic analysis of Crassulaceae based on combined nuclear ITS and chloroplast DNA (Klein & Kadereit in prep.), the 20 species of Sedum found in Germany fall into several lineages.
(1) Sedum rosea (L.) Scop. represents a lineage that is often accepted as the genus Rhodiola L., with c. 60 spp. mostly found in C and E Asia, in which S. rosea should be known as R. rosea L.
(2) Sedum spurium M. Bieb. represents a lineage that is often accepted as the genus Phedimus Raf., with c. 20 spp. mostly found in SW to E Asia, in which S. spurium should be known as P. spurius (M. Bieb.) 't Hart. Phedimus and Rhodiola are sister to each other and could be combined in one genus. However, among other morphological differences, Phedimus spp. have hermaphrodite flowers, whereas most Rhodiola spp. have unisexual flowers.
(3) Sedum maximum (L.) Hoffm., S. telephium L. and S. vulgare (Haw.) Link represent a lineage that is often accepted as the genus Hylotelephium H. Ohba, with c. 30 spp. distributed mainly in C and E Asia, and should be known as H. maximum (L.) Holub, H. telephium (L.) H. Ohba and H. vulgare (Haw.) Holub, respectively. Hylotelephium is closely related to the C to E Asian genera Meterostachys Nakai, Orostachys Fisch. (non-monophyletic, see below) and Sinocrassula A. Berger.
(4) Sedum forsterianum Sm., S. ochroleucum Chaix and S. rupestre L. represent a lineage that should be accepted as the genus Petrosedum Grulich, as was done, e.g., by Thiede & Eggli (2007). Petrosedum is closely related to a small group of SW Asian Sedum spp. that has not yet been excluded from Sedum (S. ser. Nana 't Hart & Alpinar).
The remaining species, including the type, Sedum acre L., fall into a large clade of the family that contains a large number of other genera (see below). If the species discussed above were to remain in a monophyletic Sedum, essentially two thirds of the family would have to be included in that genus. Accordingly, segregation of three of the above four genera, i.e. Rhodiola, Phedimus and Petrosedum, is likely to be stable irrespective of future name changes in other parts of the family. As regards Hylotelephium, future name changes are conceivable because relationships between this genus and Meterostachys, Orostachys and Sinocrassula are not yet fully resolved.
The large clade of the family containing the type consists of two subclades, known as the Leucosedum-clade and the Acre-clade (van Ham & 't Hart 1998), respectively.
The Leucosedum-clade, which also includes Dudleya Britton & Rose, Mucizonia A. Berger, Pistorinia DC., Prometheum (A. Berger) H. Ohba, Rosularia Stapf, Telmissa Fenzl and Sedella Britton & Rose, contains seven species of German Sedum, i.e. S. album L., S. atratum L., S. cepaea L., S. dasyphyllum L., S. hispanicum L., S. rubens L. and S. villosum L. These seven species are scattered across a number of subclades, which partly contain one or more of the genera listed above.
The remaining five species of German Sedum, i.e. S. acre, S. alpestre Vill., S. annuum L., S. sexangulare L. and S. sarmentosum Bunge fall into the Acre-clade, which also includes Cremnophila Rose, Echeveria DC., Graptopetalum Rose, Lenophyllum Rose, Pachyphytum Link, Klotsch & Otto, Thompsonella Britton & Rose and Villadia Rose. In this Acre-clade, S. acre is supported sister to all remaining taxa, and S. alpestre, S. annuum, S. sarmentosum and S. sexangulare again are scattered across a number of subclades.
In view of the relationships described above, several potential options exist for a monophyletic Sedum. (1) Sedum could be treated as monospecific with only its type, S. acre. Of course, species that have not been sampled yet may fall into this clade. (2) The entire Acre-clade could be treated as Sedum. This, however, would imply inclusion of Cremnophila, Echeveria, Graptopetalum, Lenophyllum, Pachyphytum, Thompsonella and Villadia. Sedum in such circumscription would contain c. 500 species. (3) The Acre-clade and the Leucosedum-clade (with c. 160 species) could be combined into Sedum, which would require additional inclusion of Dudleya, Mucizonia, Pistorinia, Prometheum, Rosularia, Telmissa and Sedella.
Whereas recognition of a monospecific Sedum (option 1) would require description of a large number of genera for former species of that genus, options 2 and 3 would require combination in one genus of morphologically very different genera that are geographically widely distributed. Of these three options, option 1 appears best to us, although the new genera that will have to be described partly may not be easy to differentiate morphologically or geographically. However, as Sedum has not yet been completely sampled, and many relationships within the Acre- and Leucosedum-clades are not supported, I recommend to retain all Sedum species of the Acre- and Leucosedum-clades in Sedum until relationships are understood better. However, such Sedum clearly is not monophyletic.
Saxifragaceae (J. W. Kadereit)
The non-monophyly of Saxifraga L., first shown by Soltis & al. (1993), has been confirmed in several studies (for discussion see Fernández Prieto & al. 2013; Tkach & al. 2015). In particular, a group of 70–90 species from North America and Eurasia is only very distantly related to the remainder of Saxifraga and has to be treated as the genus Micranthes Haw. The one species affected in the German flora is Saxifraga stellaris L., which should be treated as M. stellaris (L.) Galasso & al. Following Soltis (2007), the two genera are clearly distinct in pollen and ovule characters.
Linaceae (J. W. Kadereit)
A broadly sampled phylogeny of Linum L. and relatives by McDill & al. (2009) showed that the South American Cliococca Bab., the North American Hesperolinon (A. Gray) Small and Sclerolinon C. M. Rogers and the Eurasian Radiola Hill (with only R. linoides Roth) are nested within Linum. McDill & al. (2009) proposed to return these four genera to Linum, in which they have been classified before. Radiola linoides should then be L. radiola L.
Euphorbiaceae (J. W. Kadereit)
Following Webster (2014; see discussion of literature there), Euphorbia L. is best treated as one large genus with >2000 spp. as the four major clades recognized (Chamaesyce, Esula, Euphorbia and Rhizanthium), partly treated as subgenera (see, e.g., Bruyns & al. 2006; Zimmermann & al. 2010; Horn & al. 2012; Yang & al. 2012), cannot be defined morphologically. Accordingly, segregation of subclades, e.g. the Chamaesyce Clade (Webster 2014; E. subg. Chamaesyce Raf., e.g. Yang & al. 2012), at genus rank would result in a paraphyletic Euphorbia. In consequence, all species of Euphorbia growing in Germany should be retained in that genus.
Fabaceae (C. M. Ritz)
The circumscription of the genera Cytisus Desf. (40 spp.) and Genista L. (90 spp.) has been subject to long-standing discussions. The first published molecular phylogenies based on plastid (rbcL) and ITS data revealed two well-supported lineages, Cytisus and Genista, each containing numerous segregate taxa of uncertain position (Käss & Wink 1995, 1997). Reviewing the published phylogenies, Cristofolini and Troia (2006) proposed a new sectional classification of Cytisus. Since raising all monophyletic entities within Cytisus s.l. to generic rank would lead to an impracticably high number of small and often monospecific genera, the authors advocated inclusion of Chamaecytisus Link. (30 spp.), Lembrotropis Griseb. (monospecific) and Sarothamnus Wimm. (five spp.) in Cytisus. Molecular studies did indeed not separate Chamaecytisus from Cytisus s.str. (Käss & Wink 1995; Cubas & al. 2002; Pardo & al. 2004), and species with an intermediate morphology blur the boundaries between the two genera (Cristofolini & Conte 2002). Lembotropis nigricans (L.) Griseb. (Cytisus nigricans L.) is morphologically very distinct by its elongate racemes, calyx shape and naviculate hairs, but is phylogenetically nested within Cytisus (Käss & Wink 1995, 1997). Sarothamnus scoparius (L.) W. D. J. Koch (C. scoparius (L.) Link), which is widespread in Europe, is part of C. sect. Spartiopsis Dumort. with four more species distributed in the Iberian Peninsula (Cristofolini & Troia 2006).
Molecular phylogenies based on plastid and ITS markers support the monophyly of three subgenera of Genista, but the segregate genera Genistella Ortega (Genista sagittalis L. / Genistella sagittalis (L.) Gams) and Ulex L. (20 spp.; U. europaeus L. in Germany) are nested in the Genista clade (Pardo & al. 2004). However, a comprehensive revision of the complex is still missing.
The neophytic Amorpha fruticosa L. represents a poorly understood polyploid complex within the North American genus Amorpha L. (16 spp.). The monophyly of the genus is questionable: it is supported by plastid sequences, while phylogenies based on nuclear genes suggest its paraphyly because the clade also contains the North American shrubs Errazurizia rotundata (Wooton) Barneby and Parryella filifolia Torr. & A. Gray (McMahon & Hufford 2004, 2005; Straub & Doyle 2014). However, Amorpha is a Linnaean genus, and accordingly the name of the introduced A. fruticosa will remain unchanged if the above-named species are included in Amorpha.
Planted as ornamentals in Europe, Wisteria Nutt. contains four to seven deciduous lianas distributed in E Asia and North America. Phylogenetic reconstructions based on plastid and nuclear genes suggest the inclusion of the evergreen lianas Afgekia Craib and Callerya Endl. in Wisteria (Li & al. 2014). Since Wisteria is the oldest name, the names of the cultivated species in Germany will be not affected.
Coronilla L. (nine spp.) and Securigera DC. (13 spp.) each represent monophyletic entities in a highly supported clade that is sister to Hippocrepis L. according to an ITS-based phylogeny (Sokoloff & al. 2007). However, detailed analyses based on other genetic markers so far are missing. Based on present knowledge, two options, either adopting a large Coronilla s.l. including Securigera (Sokoloff 2003) or recognizing two genera (Lassen 1989), are equally possible. In the first case the name C. varia L. and in the second case S. varia (L.) Lassen should be used.
The most comprehensive study of Anthyllis L., based on plastid and nuclear markers, support its monophyly when Hymenocarpus Rchb. is included in Anthyllis and the Mediterranean genera Dorycnopsis Boiss. (two spp.) and the monospecific Tripodion Medik, are segregated (Degtjareva & al. 2012). Contradicting results were reported in an ITS-based phylogenetic study with a relatively small taxon sampling (Nanni & al. 2004). This study placed two annual species of Anthyllis, which were clearly part of Anthyllis in the study by Degtjareva & al. (2012), together with Tripodion near Lotus L. However, this result is questionable because resolution and taxon sampling were much lower than in the study by Degtjareva & al. (2012). In any case, the name of the German A. vulneraria L. would not be affected. Lotus (190 spp.) in its traditional circumscription is polyphyletic and divided into an Old World clade and several New World lineages (Allan & Porter 2000; Allan & al. 2003). The latter have now been recognized as four genera (Arambarri & al. 2005). Studies focusing on the highly supported Old World clade revealed that the segregate genera Tetragonolobus Scop. (five spp.) and Dorycnium Mill. (ten spp.) should be returned to Lotus (Degtjareva & al. 2006; Degtjareva & al. 2008), a result already suggested by morphological studies (Polhill 1981). However, the phylogenies published so far rely on nuclear ribosomal DNA only. Since incongruencies between markers are a common phenomenon in Fabaceae, additional genetic data are required. When Dorycnium and Tetragonolobus are included in Lotus, the names L. germanicus (Gremli) Peruzzi (D. germanicum (Gremli) Rikli), L. herbaceus (Vill.) Jauzein (D. herbaceum Vill.), L. hirsutus L. (D. hirsutum (L.) Ser.) and L. maritimus L. (T. maritimus (L.) Roth) should be used.
All phylogenies based on plastid and nuclear markers published so far suggest a close relationship between Calophaca Fisch., Caragana L. and the monospecific Asian Halimodendron DC. (Sanderson & Wojciechowski 1996; Wojciechowski & al. 2000; Zhang & al. 2009; Duan & al. 2015). The morphologically distinct Calophaca and Halimodendron are probably nested within Caragana, but statistical support for this was low and more research is needed (Zhang & al. 2009). In any case, the name of the introduced Caragana arborescens Lam. will not be affected by any changes in generic circumscriptions because Caragana is the oldest genus name.
The monophyly of Hedysarum L., containing c. 180 spp. distributed in the N hemisphere, still remains questionable. Two N African species have been excluded from Hedysarum and recognized as Greuteria Amirahmadi & Kaz. Osaloo (Amirahmadi & al. 2014). According to plastid phylogenies, Hedysarum (including the monospecific genus Sartoria Boiss. from Turkey) is monophyletic (Amirahmadi & al. 2014; Duan & al. 2015). The close relationship of Hedysarum and Sartoria has also been corroborated by biochemical analyses (Arslan & Ertuğrul 2010). In contrast, trees based on ITS sequences separated H. sect. Hedysarum (containing the type, H. alpinum L.) from H. sect. Multicaulia (Boiss.) B. Fedtsch. and H. sect. Stracheya (Benth.) B. H. Choi. The latter two were sister to a clade comprising Onobrychis Mill. (Amirahmadi & al. 2014; Duan & al. 2015). Further studies including sequences of nuclear low-copy genes are needed to unravel the reasons for these incongruencies. If non-monophyly of Hedysarum should obtain further support, either all species of Onobrychis and some other smaller genera should be transferred to a very large Hedysarum, or Hedysarum should be split into several genera. In the latter case the name H. hedysaroides (L.) Schinz & Thell. would remain unchanged because this species is closely related to the type of the genus name.
Similar results were obtained for Onobrychis Mill. Plastid phylogenies supported Onobrychis as a monophyletic entity but ITS phylogenies failed to do so (Amirahmadi & al. 2014; Duan & al. 2015).
All published phylogenies revealed a close relationship between Trigonella L. (60 spp.) and Melilotus Mill. (20 spp.), which is supported by morphology (e.g. incised margin of stipules, notched apex of standard, smooth surface of seed coat). Most reconstructions based on either plastid, ITS or nuclear low-copy genes revealed Trigonella as paraphyletic in relation to Melilotus (Bena 2001; Steele & Wojciechowski 2003; Steele & al. 2010; Dangi & al. 2015). In contrast, a combined analysis of ITS and plastid data showed well-supported monophyly of both genera (Dangi & al. 2015). However, taxon sampling in both genera has not been sufficiently exhaustive to solve this problem. The so-called medicagoid species of Trigonella (23 spp.) distributed in the Mediterranean area share a complex explosive tripping mechanism of pollination with Medicago (Small & al. 1987). In support of this, nuclear ribosomal sequences corroborate the inclusion of these species in Medicago (Bena 2001).
A recent comprehensive study of tribe Vicieae based on plastid and ITS sequences revealed that neither Vicia L. (140 spp.) nor Lathyrus L. (160 spp.) are monophyletic in their current delimitation (Schaefer & al. 2012). Comparable results were also obtained by earlier studies based on matK sequences of a small number of species (Steele & Wojciechowski 2003; Wojciechowski & al. 2004). Lathyrus is paraphyletic in relation to two monophyletic groups: the Caucasian Vavilovia Fed. (two spp.) and Pisum L. (three spp.; Smykal & al. 2011; Schaefer & al. 2012). Vicia appears to be paraphyletic because annual species of V. sect. Ervum (L.) Taub. (e.g. V. tetrasperma (L.) Schreb.) and V. sect. Ervilia (Link.) W. Koch (including V. sect. Ervoides (Godr.) Kupicha and Trigonellopsis Rech. f. and V. hirsuta (L.) Gray) were sister to Lathyrus s.l. and the remaining species of Vicia including Lens Mill, (four spp.; Schaefer & al. 2012). Schaefer & al. (2012) recommended the inclusion of Pisum and Vavilovia in Lathyrus. Vicia could be then recognized as a monophyletic entity by including Lens and re-transferring V. articulata Hornem., V. ervilia (L.) Willd., V. hirsuta (L.) Gray, V. parviflora Cav., V. sylvatica L. and V. tetrasperma (L.) Schreber to either Ervilia Link or Ervum L.
Polygalaceae (J. W. Kadereit)
Several studies (Eriksen 1993; Persson 2001; Forest & al. 2007; Abbott 2011), of which the study by Abbott (2011), although not including full results, used a large sample and both nuclear and plastid sequences, have shown that the large genus Polygala L. is highly polyphyletic. In consequence, several segregate genera of groups formerly included in Polygala have been proposed (for summary see Pastore 2012). Of the species of Polygala known in Germany, P. chamaebuxus L. should be removed from Polygala. According to Abbott (2011), this species is part of a lineage which should be called Polygaloides Haller and be treated as P. chamaebuxus (L.) O. Schwarz. Although not all other German species of the genus were sampled in any of the published phylogenies, their close relationship to each other has been documented (Lack 1995) and it seems safe to assume that they all will remain in Polygala.
Rosaceae (B. Gehrke)
Many genera of the Potentilleae, such as Comarum L., Dasiphora Raf., Duchesnea Sm. and even Fragaria L., have at some point been included in Potentilla L. (Mabberley 2002). However, recent molecular phylogenetic work clearly showed that Fragariinae and Potentillinae are distinct lineages. Based on molecular work authors tend to recognize Potentillinae as comprising only two genera. These are (1) Potentilla s.str. excluding P. fruticosa L., P. palustris (L.) Scop, and P. rupestris L. (see below) and including, amongst others, Duchesnea indica (Andrews) Teschem (as P. indica (Andrews) Th. Wolf), which is deeply nested in Potentilla s.str., and (2) Argentina Hill (Feng & al. 2015), a mostly Asian group, including P. anserina L. (as A. anserina (L.) Rydb.). The separation of Argentina s.l. and Potentilla s.str., which are sister lineages, is based on differences in the insertion of the styles, with Potentilla s.str. having subterminal styles, whereas Argentina has lateral ones (Dobes & Paule 2010; Sojak 2010; Feng & al. 2015). However, considering the relationship between these two genera, it would also be possible not to recognize Argentina as a separate genus and use the name Potentilla for all species of the Potentillinae (Eriksson & al. 2015).
The other monophyletic subtribe in the Potentilleae, the Fragariinae, has its highest species diversity in Asia and includes numerous smaller lineages as well as Alchemilla L., Fragaria and Sibbaldia L. Well nested in Fragariinae and more closely related to Fragaria than to Alchemilla or even Potentilla are P. fruticosa and P. rupestris. These should be treated as Dasiphora fruticosa (L.) Rydb., a monospecific genus, and Drymocallis rupestris (L.) Soják. Drymocallis Soják is a small genus confined to the N hemisphere. Alternatively, Fragaria could be extended to include Dasiphora and Drymocallis, amongst some other Asian groups, but the genus then would no longer be united by its characteristic fleshy receptacle. Leaving D. fruticosa and D. rupestris in Potentilla would necessitate including Alchemilla, Fragaria and Sibbaldia in Potentilla as well, which is obviously not desirable. Most authors therefore seem to prefer to recognize Dasiphora and Drymocallis as genera separate from Fragaria.
Alchemilla forms a clade with Aphanes L. and the mainly South American Lachemilla Rydb., easily recognizable by the lack of petals and the presence of only four calyx and epicalyx lobes (Notov & Kusnetzova 2004; Gehrke & al. 2008). Molecular phylogenetic work revealed the existence of a fourth, previously unknown clade with Alchemilla species from Africa (Gehrke & al. 2008). Aphanes is clearly nested among Alchemilla, Lachemilla and African Alchemilla (Gehrke & al. 2008). As there are no obvious morphological features to separate the African clade of Alchemilla from the European clade, and the entire clade is readily recognizable by floral morphology despite differences in life cycle, size and leaf morphology, I would like to recommend to include Aphanes in Alchemilla leading to reusing the names Alchemilla arvensis (L.) Scop. for Aphanes arvensis L. and Alchemilla microcarpa Boiss & Reut. for Aphanes inexspectata W. Lippert. Irrespective of this, Alchemilla, Aphanes, and Lachemilla in their traditional circumscriptions differ in habit and some details of floral morphology. Whereas Alchemilla and Lachemilla species are perennial and usually have four introrse stamens inserted at the outer side of the discus (Alchemilla) or 2(-4) extrorse stamens inserted at the inner side of the discus (Lachemilla), Aphanes species are annual or short-lived and have only a single extrorse stamen at the inner side of the discus.
Potentilla palustris is most closely related to Alchemilla as circumscribed above according to chloroplast data and to Sibbaldia using nuclear data. Unless included in either of these two genera, which is not desirable from a morphological point of view, it should be reinstated as Comarum palustre L. It seems that especially the Asian species of Sibbaldia require more work (Eriksson & al. 2015), but it is most likely that S. procumbens L. can retain its name.
Molecular phylogenetic work in combination with morphological character optimization has shown that Rosaceae contain only three major lineages (Potter & al. 2007): Dryadoideae, Rosoideae and Spiraeoideae. The last includes the formerly recognized Amygdaloideae, Maloideae, Prunoideae as well as Pyrinae. Evolution of derived fruit types (pome, drupe, achene) has been shown to be more complex than traditionally hypothesized (Morgan & al. 1994; Potter & al. 2002; Potter & al. 2007).
In the newly defined Spiraeoideae, the most prominent result of molecular phylogenetic work is the recognition that the species of Sorbus L. fall into two major clades. As part of the first major clade, Sorbus s.str., which is closely related to Pyrus L., should include only pinnate-leaved species (Campbell & al. 2007; Potter & al. 2007; Lo & Donoghue 2012). In this clade, S. domestica L. should be placed in the monospecific genus Cormus Spach according to Lo & Donoghue (2012) because this species is sister to a clade formed by Sorbus s.str. and Micromeles Decne. according to chloroplast data, with a weakly supported incongruent placement of Mi- cromeles as sister to Aria (Pers.) Host (see below) according to nuclear ITS sequences. If this approach were taken, the only species remaining in Sorbus found in Germany would be the type, S. aucuparia L. Chloroplast and combined chloroplast and nuclear data suggest that Sorbus species with simple leaves are not closely related to Sorbus s.str., but are a subclade of the second major clade also including Cydonia Mill., Malus Mill. and others. In this subclade of simple-leaved Sorbus species, Lo & Donoghue (2012) suggested to recognize the monospecific genera Aria (with S. aria (L.) Crantz apparently as A. nivea Host), Chamaemespilus Medik. (with S. chamaemespilus (L.) Crantz as C. alpina (Mill.) K. R. Robertson & J. B. Phipps) and Torminalis Medik. (with S. torminalis (L.) Crantz as T. clusii (M. Roem.) K. R. Robertson & J. B. Phipps). However, Chamaemespilus and Torminalis form a well-supported clade together with Aria and could also be included in Aria (Li & al. 2012a; Lo & Donoghue 2012; Sennikov 2014). Lo & Donoghue (2012) pointed out that the former inclusion of Aria and satellite genera in Sorbus reflects the finding that numerous apomictic microspecies in Europe and W Asia are of apparent hybrid origin involving species of Aria (incl. Torminalis) and Sorbus s.str. (Aas & al. 1994; Nelson-Jones & al. 2002). Maintainance of Sorbus as one genus would require sinking Cotoneaster Medik., Crataegus L., Malus Mill. and many other genera in Pyrus L. (Sennikov 2014), which is evidently even less desirable.
Rhamnaceae (J. W. Kadereit)
Phylogenetic studies in Rhamnaceae, focusing on Frangula Mill. and Rhamnus L., suggested that Frangula and Rhamnus are distinct genera, and that Rhamnus is best divided into Rhamnus s.str., the Old World genus Oreoherzogia W. Vent and the New World genus Ventia Hauenschild (Hauenschild & al. 2016). Of the German species of Rhamnus, R. pumila Turra falls into Oreoherzogia, in which it should be known as O. pumila (Turra) W. Vent. Following Hauenschild & al. (2016), Rhamnus s.str. and Oreoherzogia can be distinguished by the number of lateral leaf vein pairs (3–5 in Rhamnus, 6–20 in Oreoherzogia) and by the position of a seed furrow (lateral-medial in Rhamnus, dorso-medial in Oreoherzogia).
Urticaceae (J. W. Kadereit)
Evidence has been presented that the generic circumscription of Parietaria L. in relation to Gesnouinia Gaudich. and Soleirolia Gaudich. may require modification (Wu & al. 2013). However, no sufficiently well-sampled phylogeny is available yet to tackle this problem.
Myricaceae (J. W. Kadereit)
Myrica L. has been found to be diphyletic by Huguet & al. (2005). Following these authors (for discussion of nomenclature see their paper), M. gale L. lectotypifies the genus name Myrica, and M. pensylvanica Mirb. should be treated as Morella pensylvanica (Mirb.) Kartesz.
Onagraceae (C. M. Ritz)
Heterogeneity of Epilobium L. in stamen characters had already been noticed by Linnaeus. Several sections are recognized in the genus, of which only E. sect. Chamaenerion Ség. and E. sect. Epilobium grow in Germany. While the former has alternate leaves, weakly zygomorphic flowers with only a very short hypanthium, almost entire petals, recurved stamens of almost equal length, a recurved style and pollen in monads (type: E. angustifolium L.), E. sect. Epilobium has opposite leaves, actinomorphic flowers with a distinct hypanthium, emarginate petals, erect stamens of different length, an erect style and pollen in tetrads (lectotype: E. hirsutum L.). All phylogenetic analyses of the family, partly using a broad taxon sampling and both nuclear and plastid sequences (Baum & al. 1994; Levin & al. 2003, 2004) invariably demonstrated that E. sect. Chamaenerion is sister to the remainder of the genus. Considering this pattern of relationship, it is both possible to treat E. sect. Chamaenerion at generic rank on account of its morphological distinctness, as was done in most North American Floras, or to include it in Epilobium. If treated as a distinct genus, this would affect classification of E. angustifolium, E. dodonaei Will. and E.fleischeri Hochst. The name Chamaenerion has long been discussed controversially. Chamaenerion Ség. instead of Chamaenerion Hill or Chamerion Raf. has to be used according to Sennikov (2011).
As shown in the well-sampled phylogeny of Onagraceae by Levin & al. (2004), Oenothera L. is only monophyletic when Calylophus Spach, Gaura L. and Stenosiphon Spach are included, as was done by Wagner & al. (2007).
Lythraceae (J. W. Kadereit)
A phylogenetic analysis of Lythraceae including several species of Lythrum L. and Peplis portula L. (Morris 2007) clearly showed that Peplis L. is deeply nested in Lythrum and should, as already done by Webb (1967), be treated as L. portula (L.) D. A. Webb.
Malvaceae (J. W. Kadereit)
A well-sampled phylogenetic analysis of Alcea L., Althaea L., Lavatera L. and Malva L. using nuclear and plastid sequences by Escobar García & al. (2009) revealed that, probably with the exception of Alcea, these genera are not monophyletic. This had been shown before for Lavatera and Malva by Ray (1995). The two species of Althaea found in Germany fall into two only distantly related clades, with Althaea hirsuta L. as representative of one clade more closely related to Malva / Lavatera and Althaea officinalis L. as representative of the second clade more closely related to Alcea. Species of Malva fall into three separate clades, of which the one containing M. alcea L. and M. moschata L. is more closely related to one of two clades of Lavatera that contains L. thuringiaca L. than to a second clade of Malva with M. verticillata L., M. sylvestris L. and M. neglecta Wallr. As is evident, these patterns of relationship require taxonomic changes. Escobar García & al. (2009) did not present a new classification of this “Malva alliance”, but both Banfi & al. (2005, 2011) and Stace (2010) suggested to recognize an enlarged Malva containing Lavatera and Althaea hirsuta and relatives.
Resedaceae (J. W. Kadereit)
A phylogenetic analysis of a broad sample of Resedaceae using nuclear and plastid sequences by Martín-Bravo & al. (2007) demonstrated that Reseda L. is paraphyletic in relation to the genera Ochradenus, Oligomeris and Randonia. This group of genera consists of two major lineages, and the four species of Reseda found in Germany fall into both. Reseda alba L. and R. luteola L. fall into two different subclades of the lineage that also contains Oligomeris, whereas R. lutea L. and R. odorata L. fall into two different subclades of the lineage that also contains Ochradenus and Randonia. Although the authors argued that Ochradenus and Randonia should be recognized at generic rank, they do not propose subdivision of Reseda into smaller genera. If this should eventually be proposed, the name Reseda would have to be applied to a clade containing R. lutea, the type of the genus name.
Brassicaceae (M. A. Koch)
Brassicaceae, currently recognized to contain 325 genera in 51 tribes (Al-Shehbaz 2012; Koch & al. 2012; Kiefer & al. 2014), show high levels of homoplasy in almost every morphological character used in the circumscription of tribes and genera in the past. Consequently, reliable systematic concepts often have to be obtained from molecular data, and many changes of tribal and generic circumscriptions have become necessary.
Based on molecular data, Erophila DC. is nested in Draba L. (Jordon-Thaden & al. 2010) and should be included in that genus, and E. verna (L.) Chevall. should be known as D. verna L. If recognized at species rank, E. praecox (Stev.) DC. and E. spathulata Lang should be D. praecox (Stev.) and D. spathulata (Lang) Sadler, respectively.
Several species of a formerly widely defined Arabis L. have to be transferred to other genera: A. glabra (L.) Bernh. has to be treated as Turritis glabra L., A. pauciflora Garcke as Fourraea alpina (L.) Greuter & Burdet and A. turrita L. as Pseudoturritis turrita (L.) Al-Shehbaz (Koch & al. 1999, 2000, 2001; Karl & Koch 2014). None of these three genera groups in tribe Arabideae any longer (Koch & al. 2007; Couvreur & al. 2010). Even after these changes, Arabis is still a paraphyletic taxon. Since A. alpina L. is the type of the genus name, all remaining Arabis species might be transferred to newly introduced genera in the future.
Cardaminopsis Hayek is the sister group of Arabidopsis thaliana (L.) Heynh. (Koch & Matschinger 2007; Hohmann & al. 2014), and it has been widely accepted to include Cardaminopsis in Arabidopsis Heinh. The German species of Cardaminopsis will be A. arenosa (L.) Lawalrée, A. halleri (L.) O'Kane & Al-Shehbaz and A. lyrata subsp. petraea (L.) O'Kane & Al-Shehbaz (Al-Shehbaz 2012; Kiefer & al. 2014).
A new classification of Thlaspi L. was proposed four decades ago (Meyer 1973, 1979), recognizing the genera Microthlaspi F. K. Mey., Noccaea Moench and Thlaspi for species of Thlaspi s.l. in the German flora. This concept has been confirmed by a series of molecular studies (e.g. Mummenhoff & al. 1997a, 1997b; Koch & Mummenhoff 2001). Microthlaspi and Noccaea do not group in tribe Thlaspideae, but are members of tribe Coluteocarpeae (Koch & German 2013). For the German flora, T. caerulescens J. Presl & C. Presl, T. cepaeifolia (Wulfen) Koch and T. montanum L. were transferred to Noccaea and should be recognized as N. caerulescens (J. Presl & C. Presl) F. K. Mey., N. cepaeifolia (Wulfen) Rchb. and N. montana (L.) L. K. Mey., respectively. Thlaspi perfoliatum, with its two morphologically slightly differentiated cytotypes (T. erraticum Jord. and T. improperum Jord.; Koch & Bernhardt 2004), has to be included in Microthlaspi as M. perfoliatum (L.) F. K. Mey. It has also been proposed to combine most genera of tribe Coluteocarpeae in a broadly defined Noccaea (Al-Shehbaz 2012). However, since comprehensive molecular analyses of the entire tribe with its more than 125 species (Koch & German 2013) are lacking, this concept should not be followed at the moment.
Considering the German flora, Alyssum saxatile L. has been shown to be member of a clade including various species of Aurinia Desv., which is sister to Bornmuellera Hausskn. and Clypeola L. (Cecchi & al. 2010; Resetnik & al. 2013). Consequently, A. saxatile is best treated as Aurinia saxatilis (L.) Desv. All other Alyssum species in Germany belong to a then monophyletic Alyssum.
Integration of Dentaria L. in Cardamine L. (Carlsen & al. 2009) and of Coronopus Mill. in Lepidium L. (Al-Shehbaz & al. 2002; Mummenhoff & al. 2008) is strongly supported and both are nested in the respective genera in molecular analyses. The four Dentaria species of the German flora should be known as Cardamine bulbifera (L.) Crantz, C. enneaphyllos (L.) Crantz, C. heptaphyllos (Vill.) O. E. Schulz and C. pentaphyllos (L.) Crantz. Coronopus didymus (L.) Sm. and C. squamatus (Forrsk.) Asch. are now best treated as Lepidium didymum L. and L. coronopus (L.) Al-Shehbaz, respectively.
Pritzelago Kuntze and Hymenolobus Nutt. of tribe Erysimeae are best included in Hornungia Bernh. These three genera form a well-supported clade (Mummenhoff & al. 2001; Kropf & al. 2003), and it has been demonstrated that there is no single character that reliably distinguishes the three genera (Al-Shehbaz & Appel 1997). Consequently, the following names should be used: Hornungia alpina (L.) O. Appel, H. petraea (L.) Rchb. and H. procumbens (L.) Hayek.
Maximum-likelihood trees derived from ITS1 and ITS2 sequences available from BrassiBase (Koch & al. 2012; Kiefer & al. 2014; http://brassibase.cos.uni-heidelberg.de/ clearly show that Cheiranthus cheiri L. is nested in Erysimum L., where it should be called E. cheiri (L.) Crantz. Hirschfeldia Moench of tribe Brassiceae consists of one species only: H. incana (L.) Lagr.-Fossat is most closely related to Erucastrum C. Presl. (including its type, E. virgatum C. Presl.; Arias & al. 2014). However, since Erucastrum as currently treated is a polyphyletic genus, and various other Erucastrum species might be transferred to different genera in future (Arias & Pires 2012), it seems best to keep Hirschfeldia separate until various phylogenetic hypotheses have been tested in more detail.
Santalaceae (J. W. Kadereit)
Thesium L. was found to be monophyletic only when the genera Austroamericum Hendrych and Thesidium Sonder are included (Moore & al. 2010). Discussing the options of either sinking these two genera into Thesium or maintaining them, requiring splitting of Thesium in its traditional circumscription into several smaller genera, Moore & al. (2010) preferred the former option for morphological reasons.
Polygonaceae (K. Wesche)
In Rumex L. two monophyletic subgenera can be distinguished: R. subg. Acetosa (Mill.) Rech. f. and R. subg. Rumex. This is the approach currently chosen in most C European floras, although it is possible (but not mandatory) to raise these subgenera to generic rank (Hejný & Slavik 1990). According to molecular analyses, R. subg. Acetosa includes the sometimes separately treated R. subg. Acetosella (Meisn.) Rech. f. (Schuster & al. 2015). This is supported by shared morphological characters, e.g. the presence of hastate leaves.
The taxonomy of Polygonum L. has posed particular challenges. The traditional broad concept had survived two centuries in spite of repeated criticism including calls to split the genus into up to nine sections, which commenced as early as 1856 (Meisner 1856). Based on morphological evidence, Haraldson (1978) reinforced these earlier proposals for splitting Polygonum, which have since been confirmed by studies of both plastid and nuclear DNA markers (Lamb Frye & Kron 2003; Galasso & al. 2009; Schuster & al. 2015). Polygonum s.l. clearly is polyphyletic and should be split into several genera, partly even belonging to different tribes. Some details, however, are still controversial, given that new molecular studies continue to differ from preceding ones, and no final conclusions have been reached. Accordingly, all inferences remain somewhat tentative.
Species of tribe Polygoneae have outer tepals with one primary vein and include a range of life forms. Polygonum s.str. is characterized by a distinct pollen morphology and by outer tepals that do not develop large appendages in fruit (Schuster & al. 2011a). In Germany it comprises few, mainly ruderal species (P. aviculare L. agg. - including P. arenastrum Borean, P. oxyspermum Ledeb. and P. raii Bab., the latter sometimes treated as a subspecies of P. oxyspermum). In our context, these species are distinct by having essentially solitary or at the most approximate flowers in axillary glomerules and a silvery ochrea. Genetic studies support the monophyly of Polygonum L. s.str. (Schuster & al. 2015).
A clade related to Polygonum L. s.str. contains the genera Reynoutria Houtt. and Fallopia Adans. Their taxonomy is notorious for frequent changes and their treatment is inconsistent among C European Floras (Fischer & al. 2008; Jäger 2011; Tison & de Foucoult 2014). Fallopia in its traditional circumscription contains mostly lianas, while Reynoutria includes extremely tall herbs that are invasive in many regions. Both taxa share the presence of extrafloral nectaries and have wings on the floral bracts. Viable intergeneric hybrids are known, and polyploidy and extreme morphological variability add to the taxonomic difficulties. In consequence, Reynoutria has often been included in a broader Fallopia s.l., where it was treated as a section. Uncertainty about the treatment of the two genera pertains, although molecular approaches have used both chloroplast and nuclear markers for a very good taxonomic coverage. These studies support the monophyly of each of the two genera (Schuster & al. 2011b, 2015). The S hemisphere Muehlenbeckia Meisn., however, has been identified as closely related (Haraldson 1978), and recent molecular studies implied that it is indeed sister to Fallopia (Schuster & al. 2011b, 2015). This is in line with the fact that both Fallopia and Muehlenbeckia share a base chromosome number of 10 (11 in Reynoutria) and contain flavones (absent in Reynoutria). Including Reynoutria but not Muehlenbeckia in a broadly circumscribed Fallopia would thus result in a polyphyletic group. In view of this, keeping Fallopia, Muehlenbeckia and Reynoutria as separate genera currently is the best — but not necessarily final — solution.
The second large tribe relevant for relationships of Polygonum s.l. in Germany are the Persicarieae, which are monophyletic and morphologically distinct by the presence of three veins arising from the base of the tepals, of nectaries and of non-dilated stamen filaments (Lamb Frye & Kron 2003; Kim & Donoghue 2008; Sanchez & Kron 2008). The tribe includes Aconogonon (Meisn.) Rchb., Bistorta Mill, and Persicaria (L.) Mill. Persicaria is characterized by spicate or capitate panicles, a usually entire but often ciliate or pectinate ochrea, and has 4–8 stamens and 4 or 5 tepals. All recent treatments agree that it is monophyletic and should be excluded from tribe Polygoneae (Kim & Donoghue 2008; Fan & al. 2013). Thus, the following combinations should be used for the German species: Persicaria amphibia (L.) Delarbre, P. hydropiper (L.) Delarbre, P. lapathifolia (L.) Delarbre, P. maculosa Gray, P. minor (Huds.) Opiz, P. mitis (Schrank) Assenov and P. pensylvanica (L.) M. Gómez. Except for P. maculosa (formerly Polygonum persicaria L.), epithets could be directly adopted from former names in Polygonum. Although the highly variable P. amphibia is a morphologically distinct taxon within Persicaria (Kim & Donoghue 2008), there is no need to raise it to genus level (Galasso & al. 2009).
Bistorta Mill, is morphologically distinct (with a rosette of basal leaves and usually only one terminal, spicate panicle), and both chloroplast and nuclear data imply that it is monophyletic within Persicarieae (Kim & Donoghue 2008; Fan & al. 2013). Molecular approaches, however, are not fully consistent with respect to its exact relationships to Aconogonon and Koenigia L. Nonetheless, most current Floras and also molecular studies (Galasso & al. 2009; Sanchez & al. 2011; Schuster & al. 2011a) accept its generic rank. The German species thus have to be named B. officinalis Delarbre and B. vivipara (L.) Delarbre (formerly Polygonum bistorta L. and P. viviparum L., respectively).
The taxonomy of Aconogonon is particularly complicated. Species in this group have been placed in Persicaria, Polygonum or Rubrivena M. Krái (the last for A. polystachyum (Meisn.) Small as the only species of Aconogonon s.l. occurring in Germany). Recent molecular studies implied that Aconogonon species are distinct from Bistorta and Persicaria, but also revealed their close relationship with the mostly boreal and polar Koenigia (Galasso & al. 2009; Sanchez & al. 2009). Aconogonon and Koenigia have broadly similar pollen, and the two genera cannot easily be separated by morphological characters. Studies based on cpDNA have suggested that Koenigia in its traditional circumscription may be nested between Aconogonon and Rubrivena (Sanchez & al. 2011). The so-far most comprehensive study covering many taxa and employing both cpDNA and nuclear markers (Schuster & al. 2015) confirms this close relationship and finds one large clade that comprises all analysed species of Aconogonon and Koenigia (and Rubrivena). While most Koenigia species form a distinct group, some (but not all!) accessions of K. delicatula (Meisn.) H. Hara are sister group to a clade comprising other Aconogonon and Koenigia species (incl. A. polystachyum). This implies that Koenigia in its traditional sense is not monophyletic. Relationships of Aconogonon are even more puzzling, with a number of polyploid Aconogonon species being more closely related to Koenigia than to other members of the genus. Moreover, different accessions of some Aconogonon species appear on very different branches in the Aconogonon / Koenigia clade. Details of the evolution of this group clearly are not fully understood, and thus Schuster & al. (2015) advocate the fusion of all taxa in one large genus. They propose to unite them under the name Koenigia, which was chosen for priority reasons. These authors also draw the necessary taxonomic consequences and provide the new combination K. polystachya (Meisn.) T. M. Schust. & Reveal.
Though using a somewhat smaller species set, Fan & al. (2013) also presented a comprehensive molecular study, which confirmed the odd position of K. delicatula (plus one Aconogonon species). In their analysis, A. polystachyum is nested within other Aconogonon species, which jointly form the sister clade to the core Koenigia species. Fan & al. (2013) also discussed the possibility to adopt a broad concept of Koenigia. However, they acknowledged that merging the larger Aconogonon in the smaller Koenigia is somewhat impractical and also remarked on the apparently different chromosome base numbers in the two groups. They advocated keeping the two genera independent and placing the odd K. delicatula in a new monospecific genus, for which no valid name is available yet. This would also be supported by some of its morphological features that are transient to Persicaria. Splitting the whole complex into several, partly new genera indeed is an alternative solution to the problem implied by the tree of Schuster & al. (2015), but would presumably result in the formation of many small genera such as Rubrivena. Given that details of the evolution of Acononogon / Koenigia remain unclear, I opt for an intermediate position. The special position of Aconogonon and Koenigia in Persicarieae is undebated, but instead of drawing far-reaching taxonomic consequences, I rather acknowledge the level of uncertainty by keeping Aconogonon as a separate genus for the time being. In line with Fan & al. (2013), I regard evidence for a separate genus Rubrivena as questionable and maintain the established name A. polystachyum for the taxon occurring in C Europe.
Caryophyllaceae (M. S. Dillenberger)
Regarding generic delimitations in the Caryophyllaceae, Greenberg & Donoghue (2011) stated: “none of the eight largest genera (Arenaria, Cerastium, Dianthus, Gypsophila, Minuartia, Paronychia, Silene, Stellaria) appear to be strictly monophyletic”. For some genera taxonomic adjustments have already been made (e.g. Dillenberger & Kadereit 2014), but not for all. All taxonomic changes that were recently made for taxa in the German flora, or that need to be made in the future, are related to these eight genera.
There are several problems concerning the monophyly of Cerastium L. and Stellaria L. Cerastium is an almost cosmopolitan genus with about 100 species. Stellaria is cosmopolitan, too, and contains c. 120 species (Mabberley 2008). Both genera have emarginate to deeply lobed petals, but this character is shared with other genera, e.g. Myosoton Moench (Bittrich 1993). Myosoton is a monospecific genus, with M. aquaticum (L.) Moench as its only species. This species has recently (Jäger 2011; Seybold 2011) been treated as part of Stellaria, as S. aquatica L. This is congruent with the findings of Greenberg & Donoghue (2011), where S. aquatica is nested with good support in a clade of several Stellaria species, including S. media (L.) Vill. but not the type of Stellaria, S. holostea L., and is closely related to S. bungeana Fenzl. Unfortunately, Stellaria does not become monophyletic by including Myosoton. With good support, S. holostea is sister to a clade containing the largest part of Stellaria, but also Cerastium, Holosteum L. and Moenchia Ehrh. Furthermore, Cerastium is not monophyletic since a well-supported clade of two species, C. cerastoides (L.) Britton and C. dubium (Bastard) Guépin, is sister to Holosteum. This position is poorly supported, but Moenchia is sister to the rest of Cerastium with good support, making it impossible to retain the two Cerastium species in Cerastium without including at least Moenchia. To amend these various violations of monophyly there are at least two possible solutions:
(1) The first solution is to merge Cerastium, Holosteum, Moenchia and Stellaria (including Myosoton) in one large genus with c. 230 species. Which name among those with equal priority (i.e. Cerastium, Holosteum and Stellaria) is correct for this genus needs further investigation. This genus combines most species with deeply lobed petals, but also some species with entire or emarginate petals.
(2) The second solution is to change generic circumscriptions and to describe new genera. On the basis of the phylogeny of Greenberg & Donoghue (2011), it is clear that Stellaria needs to be split into different genera. Stellaria retains only S. holostea and probably closely related species that were not included in the phylogeny of Greenberg & Donoghue (2011). The largest number of Stellaria species have to be transferred to a new genus. This new genus contains all former German Stellaria species except S. holostea. This genus is then sister to a clade containing Cerastium, Holosteum and Moenchia. Moenchia can be retained unmodified and is sister to Cerastium. Cerastium contains all German species with four or more styles. The two species with three styles that are sister to Holosteum, i.e. C. cerastoides and C. dubium, are best included in Holosteum, which also has three styles, or those two species (and maybe other Cerastium species from other regions with three styles) should be treated as a new genus. Both solutions require a large number of taxonomic changes and a decision between them cannot be easily made. However, changes in the circumscription of the above genera are inevitable.
In Gypsophila L. and relatives of interest (i.e. Dianthus L., Petrorhagia (Ser.) Link and Vaccaria Wolf.) two issues need to be discussed. The first is the treatment of Vaccaria. Vaccaria is a monospecific genus containing only V. hispanica (Mill.) Rauschert. This species is native to Eurasia, especially the Mediterranean region, but has become naturalized in large parts of the world (S Africa, Australia and North and South America). The phylogeny of Greenberg & Donoghue (2011) unambiguously placed Vaccaria within Gypsophila. It differs from Gypsophila mainly by its calyx wings. The position in the phylogeny allows two alternative solutions.
(1) Vaccaria remains a monospecific genus that is sister to the largest part of Gypsophila. Therefore at least G. takhtadzhanii Schischk. ex Ikonn, has to be excluded from Gypsophila because it is sister to Vaccaria and the rest of Gypsophila.
(2) Vaccaria hispanica is included in Gypsophila as G. vaccaria (L.) Sm.
I prefer the second solution for different reasons. Vaccaria is quite similar to Gypsophila and its inclusion in that genus will not require large changes in the circumscription of Gypsophila. The other point is that it is difficult to justify splitting Gypsophila into different genera only to retain Vaccaria as an independent genus. As Greenberg & Donoghue (2011) included only few of the 150 Gypsophila species in their phylogeny, I cannot foresee to what extent an independent Vaccaria would affect subdivision of Gypsophila.
The second issue concerns the paraphyly of Petrorhagia in relation to Dianthus, and the position of Gypsophila muralis L. and several other Gypsophila species from outside Germany as sister to Dianthus and Petrorhagia instead of being part of the rest of Gypsophila. Petrorhagia is a genus with 33 species distributed from the Canary Islands across the Mediterranean region to Kashmir (Mabberley 2008). Although the phylogeny of Greenberg & Donoghue (2011) contains only three species of Petrorhagia, it unambiguously shows that the genus is paraphyletic. Two solutions seem possible:
(1) Dianthus, Petrorhagia and Gypsophila muralis (and some more Gypsophila species from outside Germany) are included in a more broadly circumscribed Dianthus.
(2) Petrorhagia is split into at least two genera, and G. muralis is transferred into a new, probably monospecific genus. Regarding the other Gypsophila species in this group from outside Germany, this solution would require establishing additional small to monospecific genera for those Gypsophila species. Dianthus, Petrorhagia and the Gypsophila species of this clade show some morphological variation. It is difficult to decide whether this variation is sufficient to justify splitting Petrorhagia into different genera that can be distinguished from each other and from Dianthus and the small genera containing former Gypsophila species, or whether all species of this clade are sufficiently alike to be merged into one genus, i.e. Dianthus. Linnaeus (1753a) described the type of Petrorhagia, P. saxifraga (L.) Link, as D. saxifragus L., P. prolifera (L.) P. W. Ball & Heywood as D. prolifer L., but no name for G. muralis is available in Dianthus.
Minuartia L. (sensu McNeill 1962) comprises about 175 species that are distributed in the N hemisphere. It was delimited from most other genera of Caryophyllaceae by a combination of three styles and three capsule valves. Molecular phylogenies revealed that the genus consists of ten independent lineages (Fior & al. 2006; Harbaugh & al. 2010; Greenberg & Donoghue 2011; Dillenberger & Kadereit 2014), each of which is closest relative of another genus or group of genera. According to Dillenberger & Kadereit (2014) the genus is best divided into 11 genera instead of including other genera in Minuartia. The ten lineages were divided into 11 genera because in one case there was no morphological or karyological character or combination of characters to define this clade as one genus. Therefore two subclades with more uniform morphologies were described as genera. Including other genera in Minuartia would have affected most genera of subfam. Alsinoideae or subfam. Alsinoideae and subfam. Caryophylloideae. In consequence, several species of Minuartia in the C European flora need to be treated as part of other genera. Minuartia species transferred to other genera are: Cherleria sedoides L. (M. sedoides (L.) Hiern), Facchinia cherlerioides (Sieber) Dillenb. & Kadereit (M. cherlerioides (Sieber) Bech.), present in the German flora only with F. cherlerioides subsp. aretioides (Port, ex J. Gay) Dillenb. & Kadereit, F. rupestris (Scop.) Dillenb. & Kadereit (M. rupestris (Scop.) Schinz & Thell.), Sabulina austriaca (Jacq.) Rchb. (M. austriaca (Jacq.) Hayek), S. stricta (Sw.) Rchb. (M. stricta (Sw.) Hiem), S. tenuifolia (L.) Rchb. (M. hybrida (Vill.) Schischk.), S. verna (L.) Rchb. (M. verna (L.) Hiem) and S. viscosa (Schreb.) Rchb. (M. viscosa (Schreb.) Schinz & Thell.). The only two species in Germany that remain in Minuartia are M. rubra (Scop.) McNeill and M. setacea (Thuill.) Hayek.
Silene L. contains c. 700 species that are restricted to the N hemisphere (Mabberley 2008). Although the genus is large, there exist only small problems with its monophyly. One point concerns Lychnis L., which contains c. 20 species distributed in N-temperate and arctic regions (Bittrich 1993). Its treatment as separate from Silene L. has repeatedly been regarded as doubtful (see Oxelman & Lidén 1995). Lychnis has usually five styles and five capsule teeth, whereas Silene has three or five styles and six or ten capsule teeth. Even the most recent phylogeny of the Caryophyllaceae could not unambiguously determine the position of Lychnis (Greenberg & Donoghue 2011). In that study Silene seems to be paraphyletic in relation to Lychnis. However, this position is not well supported, and a change of position is possible. For the moment, the species of Lychnis in the German flora, i.e. L. coronaria (L.) Desr. and L. flos-cuculi L., should be maintained, but future inclusion in Silene, as S. coronaria (L.) Clairv. and S. flos-cuculi (L.) Clairv., seems likely.
The second problem is related to Cucubalus baccifer L. Although Silene is not sufficiently well supported, the position of C. baccifer seems to be clearly within Silene (Greenberg & Donoghue 2011). Therefore it seems advisable to treat this species as S. baccifera (L.) Roth.
Several problems hinge on the acceptance of Heliosperma Rchb. and other smaller genera. When accepting Heliosperma, several smaller genera need to be recognized in order to keep Silene monophyletic. One of these genera is Atocion Adans. Based on a molecular phylogeny, Lidén & al. (2001) excluded five species, including S. armeria L. and S. rupestris L., from Silene and included them in Atocion. These results were verified with a large sample of Silene and related genera by Greenberg & Donoghue (2011) and should have taxonomic consequences. Atocion is sister to Viscaria Bernh, and the names for the two species are A. armeria (L.) Raf. and A. rupestre (L.) Oxelman. An inclusion of Atocion in Silene would also affect Eudianthe Rchb., Heliosperma and Viscaria and is therefore not desirable. Atocion is glabrous, has elliptic or oblanceolate leaves, a regular dichasium, and flowers with entire or emarginate petals and three stigmas (Lidén & al. 2001). Silene species with the same character combination of hairiness, inflorescence type and stigma number have lower leaves that are spathulate and petals that are usually lobed. Furthermore, these Silene species have anastomosing calyx veins, but Atocion has non-anastomosing veins (Lidén & al. 2001).
Irrespective of the inclusion of Lychnis in Silene or its separate treatment, L. viscaria L. is not part of either of these two genera. The species clearly belongs to a wellsupported clade that is sister to Atocion (Greenberg & Donoghue 2011). The correct genus name for the species of this clade is Viscaria Bernh., and L. viscaria should be known as V. vulgaris Bernh. Viscaria vulgaris is the type of Viscaria.
Another problem is related to Silene pusilla Waldst. & Kit., which is nested in the well-supported Heliosperma (Rchb.) Rchb. The inclusion of S. pusilla in Heliosperma as H. pusillum (Waldst. & Kit.) Rchb. is justified and necessary.
Chenopodiaceae (G. Kadereit)
Chenopodium L. in its traditional wide circumscription, comprising c. 150 spp. worldwide, has been shown to be highly polyphyletic with Chenopodium lineages spread all over the phylogeny of subfam. Chenopodioideae (Kadereit & al. 2010; Fuentes-Bazán & al. 2012a, 2012b). According to Fuentes-Bazán & al. (2012a, 2012b), species of Chenopodium belong to six different genera: Blitum L., Chenopodiastrum S. Fuentes & al., Chenopodium L. s.str., Dysphania R. Br., Lipandra Moq. and Oxybasis Kar. & Kir. Although the sampling for the molecular analyses was far from complete, the polyphyly of Chenopodium seems well supported and future studies will reveal where unsampled species belong. Twenty of the 23 species of former Chenopodium occurring in the German flora were included in the molecular studies by Fuentes-Bazán & al. (2012a, 2012b), and these are distributed among all six genera. Blitum is represented by three species: B. bonus-henricus (L.) Rchb. (C. bonus-henricus L.), B. capitatum L. (C. capitatum (L.) Aschers.) and B. virgatum L. (C. foliosum Aschers.). Chenopodiastrum is represented by Chenopodiastrum hybridum (L.) S. Fuentes & al. (Chenopodium hybridum L.) and Chenopodiastrum murale (L.) S. Fuentes & al. (Chenopodium murale L.). Species with glandular hairs and an aromatic odour clearly need to be classified in Dysphania, which is only distantly related to core Chenopodium. In the German flora these are D. ambrosioides (L.) Mosyakin & Clemants (C. ambrosioides L.), D. botrys (L.) Mosyakin & Clemants (C. botrys L.), D. pumilio (R. Br.) Mosyakin & Clemants (C. pumilio R. Br.) and D. schraderiana (Schult) Mosyakin & Clemants (C. schraderianum Schult). Lipandra is represented by L. polysperma (L.) S. Fuentes & al. (C. polyspermum L.) and Oxybasis by O. chenopodioides (L.) S. Fuentes & al. (C. botryodes Sm.), O. glauca (L.) S. Fuentes & al. (C. glaucum L.), O. rubra (L.) S. Fuentes & al. (C. rubrum L.) and O. urbica (L.) S. Fuentes & al. (C. urbicum L.). Of the remaining species present in the German flora, Chenopodium album L., C. berlandieri Moq., C. ficifolium Sm., C. opulifolium Schrader ex Koch & Ziz, C. patericola Rydb. and C. vulvaria L. belong to Chenopodium s.str. Chenopodium hircinum Schrader, C. strictum Roth and C. suecicum Murr have not yet been included in molecular analyses. Chenopodium aristatum L. (Dysphania aristata (L.) Mosyakin & Clemants) is a neophyte in the German flora and should be treated as Teloxys aristata (L.) Moq. This monospecific genus is closely related to Cycloloma Moq., Dysphania and Suckleya A. Gray (Kadereit & al. 2010; Fuentes-Bazán & al. 2012a).
Halimione Aellen is well-supported sister group of the large genus Atriplex L., from which it can be distinguished by unique seed and fruit characters (Kadereit & al. 2010). Inclusion of Halimione into Atriplex as proposed in Sukhorukov (2006) is possible, but not recommended by the present author (G. Kadereit).
Bassia All. and Kochia Roth were both found to be polyphyletic in molecular studies (Kadereit & Freitag 2011; Kadereit & al. 2014). Most species of Kochia including the two species present in Germany, K. laniflora (S. G. Gmelin) Borbás and K. scoparia (L.) Schrader, have been included in Bassia, and the remaining species were classified in two new genera, Eokochia Freitag & G. Kadereit and Grubovia Freitag & G. Kadereit. Other species of Bassia (B. dasyphylla Kuntze, B. hirsuta (L.) Kuntze and B. sedoides (Schrad.) Asch.) needed to be transferred to new genera (Grubovia dasyphylla (Fisch. & C. A. Mey.) Freitag & G. Kadereit, Spirobassia hirsuta (L.) Freitag & G. Kadereit and Sedobassia sedoides (Schrad.) Freitag & G. Kadereit) in order to define monophyletic genera in Camphorosmeae (Kadereit & Freitag 2011). Of these new genera only Spirobassia (S. hirsuta) occurs in Germany.
Salsola L. is a large and highly polyphyletic genus (Akhani & al. 2007). Unfortunately there is disagreement among experts concerning the typification of Salsola. Mosyakin & al. (2014) proposed a conserved type, S. kali L., while Akhani & al. (2014) argued in favour of the current type, S. soda L. If S. soda is accepted as type of Salsola, S. kali has to be included in Kali Mill., as Kali soda Moench (Akhani & al. 2007).
Nyctaginaceae (J. W. Kadereit)
As shown by Levin (2000), Oxybaphus Willd. is clearly part of Mirabilis L. where it should have the rank of section. Accordingly, O. nyctagineus (Michx.) Sweet should be treated as M. nyctaginea (Michx.) MacMill.
Hydrangeaceae (J. W. Kadereit)
Although Philadelphus L. appears to be paraphyletic in relation to the monospecific Carpenteria Torr. (Guo & al. 2013), classification of P. coronarius L. as a Philadelphus would not be affected as P. coronarius is the type of the genus name. Philadelphus inodorus L. falls into the same clade as P. coronarius.
Primulaceae (J. W. Kadereit)
Mast & al. (2001) demonstrated that Cortusa L. is deeply nested in Primula L. Accordingly, it should be treated as P. matthioli (L.) V. A. Richt.
As summarized by Manns & Anderberg (2009), several studies using either nuclear, plastid or both nuclear and plastid sequences have shown that a non-monophyletic Anagallis L. (incl. Centunculus L.), Glaux L. and Trientalis L. (as well as the non-C-European genera Asterolinon Hoffsgg. & Link and Pelletiera A. St. Hil.) are all nested in a highly paraphyletic Lysimachia L. Based on a careful consideration of morphological variation in this group of genera, and facing the choice between including all in Lysimachia or splitting Lysimachia in such a way that at least some of the above genera can be maintained, Manns & Anderberg (2009) argue: “It is, however, difficult to establish morphological characters to distinguish between different subgroups within Lysimachia and the morphological distinctiveness of these subgroups is not very high. Furthermore, the characters used to recognize Lysimachia are also present in Anagallis and to large extent also in Asterolinon, Pelletiera and Trientalis. Consequently, proposal of new genera for some Lysimachia (e.g. L. nemorum L. and L. serpyllifolia Schreb.), or transfer of L. nemorum and allied taxa to Anagallis would inevitably result in poorly diagnosed genera. Choosing among alternatives, we find it better to merge the smaller segregate genera with Lysimachia, rather than splitting Lysimachia further.” Through earlier work and the work by Banfi & al. (2005) and Manns & Anderberg (2009) combinations are available for C European Anagallis (plus Centunculus), Glaux and Trientalis as species of Lysimachia. These would be L. arvensis (L.) U. Manns & Anderb. (Anagallis arvensis L.), L. europaea (L.) U. Manns & Anderb. (Trientalis europaea L.), L.foemina (Mill.) U. Manns & Anderb. (A. foemina Mill.), L. maritima (L.) Galasso & al. (Glaux maritima L.) and L. tenella L. (A. tenella (L.) L.).
Ericaceae (M. D. Pirie)
Three genera have been recently re-delimited to make them monophyletic. The first is Kalmia L., which becomes monophyletic only after inclusion of Loiseleuria Desv. (Gillespie & Kron 2013). Accordingly, Loiseleuria procumbens (L.) Desv. should be known as Kalmia procumbens (L.) Gift & al. ex Galasso & al.
The second is Rhododendron L., with c. 850 species, which should include Ledum L. based on morphological evidence by Kron & Judd (1990) and molecular evidence by, e.g., Goetsch & al. (2005). In Germany, the native L. palustre L. should be known as R. tomentosum Harmaja and the introduced L. groenlandicum Oeder as R. groenlandicum (Oeder) Kron & Judd.
The third is Monotropa L., which is replaced by Hypopitys Hill, in Jäger (2011). Evidence from nuclear encoded markers suggests that the type of Monotropa, M. uniflora L., and that of Hypopitys, H. monotropa Crantz (M. hypopitys L.), are more closely related to other monotropoid genera than to each other (Bidartondo & Bruns 2001). Species delimitation within Hypopitys is controversial, but resolution of the precise number and delimitation of species (including H. hypophegea G. Don in Germany) across its broad geographic range seems unlikely to further affect generic boundaries.
Problems in generic delimitation remain in Vaccinium L. A number of different genera are apparently nested between its c. 450 species, with no evidence to suggest that the type, V uliginosum L., is closely related to any of the other species of the German flora, and clear indication that V myrtillus L. is more closely related to species elsewhere (Powell & Kron 2002). As the specialists are apparently not in favour of expanding the circumscription of Vaccinium it is likely that name changes will yet be required, but the current phylogenetic hypothesis is insufficiently resolved and sampled to offer a solution.
Rubiaceae (F. Ehrendorfer)
Since more than 20 years ago, DNA-analytical phylogenetic studies on the critical tribe Rubieae (Rubiaceae) have become available (e.g. Ehrendorfer & al. 1994; Manen & al. 1994; Natali & al. 1995, 1996; Soza & Olmstead 2010a, 2010b; and particularly Ehrendorfer & Barfuss 2014: Fig. 1 & 2, with clades and their reference numbers). These studies have made it increasingly clear that the traditional genera Asperula L. and Galium L., both well represented in the flora of Germany (Jäger 2005), are polyphyletic in their present circumscriptions. Monophyly was documented only for Cruciata Mill., Rubia L. and Sherardia L. In order to achieve monophyly for Asperula and Galium, one would have to lump all these genera (and several others except Rubia) into a giant Galium s.latiss. with about 900 species worldwide and a very complex infrageneric classification.
If a more narrow generic concept for C European Rubieae is preferred, Asperula would have to be restricted to its type, the annual A. arvensis L., and its perennial sister taxon A. taurina L. (clade V-B). The large A. sect. Cynanchicae (DC.) Boiss. (with A. cynanchica L. and A. neilreichii Beck), centred in the Mediterranean area, is more closely related to Sherardia (both in clade V-A) than to Asperula s.str. and might also deserve separate generic status. This also applies to A. tinctoria L., a member of the traditional A. sect. Glabella Griseb. (clade V-C) with a disjunct Eurasian distribution. Also into clade V-C falls Galium sect. Aparinoides (Jord.) Gren., a sube lade of limnic habitats with a worldwide distribution, typified by G. palustre L., a well-known element of the European flora. The morphological distinctness and deviating chromosome base number x = 12 (otherwise mostly x = 11 in Rubieae) also suggest generic separation of this subclade.
It was no surprise to find two Galium species (G. boreale L. and G. rotundifolium L.) from G. sect. Platygalium (DC.) W. D. J. Koch in the same clade (V-D) as the generally recognized genera Cruciata and Valantia L.: they all are characterized by whorls of two leaves and only two additional leaf-like stipules. This and the relevant DNA data could justify the transfer of G. boreale and G. rotundifolium to a separate genus, corresponding to G. sect. Platygalium s.latiss. (also including the former European genus Trichogalium Fourr., the American genus Relbunium (Endl.) Benth. & Hook, and probably also the monotypie Microphysa Schrenk from C Asia) with a worldwide distribution and up to 230 other, clearly related former Galium and Relbunium species centred in E Asia and the Americas.
The majority of the remaining C European Galium species (24 in Germany; Jäger 2005) always have leaves and leaf-like stipules in whorls of more than four (and up to 12). They are clearly verified as members of a worldwide “monophylum” that corresponds to clade VI and the genus Galium s.str. with about 350 predominantly Old World species. The relationships of its species in Germany correspond quite well with the following more or less DNA-supported taxonomic sections: G. sect. Aparine W. D. J. Koch (G. aparine L. and G. spurium L.), G. sect. Aspera (DC.) W. D. J. Koch, syn.: G. sect. Microgalium Griseb. (G. parisiense L.), G. sect. Galium (G. album Mill., G. aristatum L., G. glaucum L., G. intermedium Schultes [G. schulte sii Vest], G. lucidum All., G. mollugo L., G. xpomeranicum Retz., G. sylvaticum L., G. truniacum (Ronn.) Ronn. and G. verum L.), G. sect. Hylaea (Griseb.) Ehrend. (G. odoratum (L.) Scop.), G. sect. Kolgyda Dumort. (G. tricomutum Dandy and G. verrucosum Huds.), G. sect. Leptogalium (G. anisophyllon Vill., G. megalospermum All, G. noricum Ehrend., G. pumilum Murray, G. saxatile L., G. sterneri Ehrend, and G. valdepilosum H. Braun) and G. sect. Trachygalium K. Schum. (G. uliginosum L.).
A more detailed presentation of our current knowledge concerning relationships within tribe Rubieae in C Europe can be found in Kästner & Ehrendorfer (in press). Before one can begin to execute the possible and DNA-supported taxonomic and nomenclatural changes within the Rubieae discussed above, further critical research appears obligatory.
Gentianaceae (J. W. Kadereit)
Several phylogenetic studies of Gentianaceae-Swertiinae (Chassot & al. 2001; von Hagen & Kadereit 2001, 2002) have shown that generic circumscriptions in this group require substantial revision. Thus, it is evident that Gentianella ciliata (L.) Borkh. and G. tenella (Rottb.) Börner are only distantly related to Gentianella s.str., and should be treated as Gentianopsis ciliata (L.) Ma and Comastoma tenellum (Rottb.) Toyok., respectively. Even after exclusion of these (and related) species, Gentianella is polyphyletic, as is Swertia L. If this eventually should result in the recognition of several smaller genera, the generic identity of the remaining German species of Gentianella would remain unaffected as they fall into the same clade as the type of the genus name, G. campestris (L.) Börner. As S. perennis L. is the type of Swertia, recognition of segregate genera will not affect the generic identity of S. perennis. For descriptions and discussion of genera see Struwe & al. (2002). Inclusion of Comastoma (Wettst.) Toyok. and Lomatogonium A. Braun in Gentianella, as suggested by Banfi & al. (2005), who in consequence provided a combination for L. carinthiacum (Wulfen) Rchb. in Gentianella, is not justified by the data available unless a much larger number of lineages, including several lineages of Swertia, are included in Gentianella.
Oleaceae (J. W. Kadereit)
As first suspected by Wallander & Albert (2000) on the basis of plastid sequences, a monophyletic Ligustrum L. was found deeply nested in a paraphyletic Syringa L. using nuclear sequences (Li & al. 2002). In consequence, inclusion of Ligustrum in Syringa may have to be considered once stronger evidence for such relationship is available. Interestingly, one species of Ligustrum, L. sempervirens (Franch.) Lingelsh., sometimes classified as a separate genus, is intermediate in fruit morphology between Syringa (capsules) and Ligustrum (berries or drupes) by having berries that become leathery and eventually dehisce.
Plantaginaceae (D. C. Albach)
A hundred years ago, Veronica L. included all Scrophulariaceae with a tetramerous flower and short corolla tube, two stamens and a flattened capsule. In that circumscription the genus included approximately 300 species. Subsequent authors treated more and more groups of distinct species as separate genera, such as Hebe Juss. mainly from Australasia, Pseudolysimachion Opiz from Eurasia (V. longifolia L. and V. spicata L. in the German flora) and Veronicastrum Farw. from E Asia and E North America. The first DNA-based phylogenetic analyses (e.g. Albach & Chase 2001; Wag staff & al. 2002; Albach & al. 2004a) supported the separation of some genera (Paederota L. and Veronicastrum), but demonstrated that most genera split off in the 19th and 20th centuries are nested in a lineage that should be recognized as a monophyletic Veronica. These results caused a sometimes heated discussion on whether autapomorphies need to be considered as important as symplesiomorphies (e.g. Brummitt 2006). However, subsequent analyses added support to the molecular results and demonstrated that autapomorphies of these segregate genera are not as clear as sometimes believed, and that morphological transitions between Veronica and groups considered distinctive commonly exist. For example, such transitional species between Pseudolysimachion and Veronica occur in E Asia and Japan (Albach 2008). Thus, based on molecular and morphological arguments, these analyses suggest inclusion of these genera in Veronica rather than further splitting (Albach & al. 2004b; Garnock-Jones 2007). In C Europe, reintegration of Australasian Hebe and relatives and North American Synthyris Benth. will be of interest mainly to horticulturists, but reintegration of Pseudolysimachion, the species of Veronica with dense, spicate inflorescences, reverses a split adopted by many European Floras since the 1960s (Holub & Pouzar 1967). All European species of Pseudolysimachion were originally described as species of Veronica. Therefore, only taxonomic changes at the intraspecific level were necessary (Albach 2008).
Lamiaceae (M. S. Dillenberger)
Ballota L. contains c. 30 species that occur in Europe, the Mediterranean area, W Asia and, with one species, S Africa (Mabberley 2008). Several species of Ballota were included in a phylogenetic analysis of subfam. Lamioideae (Bendiksby & al. 2011b). This phylogeny unambiguously showed that Ballota is not monophyletic. The type of Ballota, B. nigra L. (the only species of the genus in Germany), is well-supported sister to Marrubium L., represented in the German flora by M. peregrinum L. and M. vulgare L. The other Ballota species are sister to this B. nigra—Marrubium clade. Only two other Ballota species, B. frutescens (L.) Woods and B. integrifolia Benth., form a separate clade that is sister to the former clade and a clade containing species of Moluccella L., Otostegia Benth. and Sulaimania Hedge & Rech. f. There are two solutions to obtain a monophyletic Ballota. The first is to merge all species of this clade (i.e. Ballota, Marrubium, Moluccella, Otostegia and Sulaimania) in one genus. The second solution is to exclude B. frutescens and B. integrifolia from Ballota and to combine the rest of Ballota including B. nigra and Marrubium in one genus. In order to avoid creation of one very large and heterogeneous genus, it seems reasonable to take the second approach. Since both genera were described by Linnaeus (1753b), it remains unclear at this point which genus name should be used. Marrubium contains 40 species (Mabberley 2008), so that a comparable number of new combinations would be needed when using either name.
The treatment and circumscription of Clinopodium L. is very different in different Floras of Germany (e.g. Jäger & Werner 2005; Jäger 2011; Seybold 2011). Clinopodium in its broad circumscription, including Acinos Mill., Bancroftia Billb., Calamintha Mill., New World Micromeria Benth. and Satureja L. contains c. 100 species and is almost cosmopolitan (Mabberley 2008). Seybold (2011) included Acinos and Calamintha, but not Satureja, in Clinopodium, and Jäger & Werner (2005) treated Acinos, Calamintha, Clinopodium and Satureja as separate genera. A molecular phylogeny of subtribe Menthinae illustrates the whole dimension of the problem (Bräuchler & al. 2010). In this phylogeny, Clinopodium is highly polyphyletic and numerous genera are nested among different Clinopodium clades. The species of Acinos form a well-supported clade together with Ziziphora L., a genus of c. 20 species distributed from the Mediterranean area to C Asia, Afghanistan and Himalaya (Mabberley 2008). In this clade, Acinos and Ziziphora are not supported as monophyletic. Calamintha species are in a well-supported clade with the type and other species of Clinopodium. Another genus that causes problems with respect to the monophyly of Clinopodium is Monarda L., a small genus of c. 16 mostly North American species (Mabberley 2008) occurring in Germany with one introduced species, M. didyma L. (Jäger & Werner 2005). The large number of genera, species and clades makes several solutions possible. For the German species only two solutions need to be discussed. The first is to include all species of Acinos, Calamintha, Clinopodium and Monarda in one genus, together with the whole or parts of Acanthomintha (A. Gray) A. Gray, Blephilia Raf., Bystropogon L′Hér., Conradina A. Gray, Cuminia Colla, Cunila D. Royen ex L., Cyclotrichium Mandenova & Schengelia, Dicerandra Benth., Glechon Spreng., Hedeoma Pers., Hesperozygis Epling, Hoehnea Epling, Killickia Bräuchler & al., Mentha L., New World Micromeria Benth., Minthostachys (Benth.) Spach, Monardella Benth., Obtegomeria Doroszenko & P. D. Cantino, Piloblephis Raf., Poliomintha A. Gray, Pycnanthemum Michx., Rhododon Epling, Stachydeoma (Benth.) Small and Ziziphora. Alternatively, Clinopodium can be split into clades which could be treated as morphologically recognizable genera. In view of substantial morphological variation of the lineages concerned it is not meaningful to merge so many genera only to prevent Clinopodium from being split. Although it is not clear how exactly Clinopodium will be split in the future, the impact of this approach on German species can easily be seen. Since Calamintha is very closely related to the type of Clinopodium, and this relationship is well supported, there is no other solution than to transfer Calamintha to Clinopodium. The Calamintha species in Germany, C. menthifolia Host and C. nepeta (L.) Savi, will have to become known as Clinopodium menthifolium (Host) Stace and Clinopodium nepeta (L.) Kuntze. No species name in Clinopodium is available for the hybrid taxon Calamintha xfoliosa Opiz; at subspecies level Clinopodium nepeta nothosubsp. subisidoratum (Borbás) Govaerts has been used. It is not possible to treat the species of Acinos as part of Clinopodium without including in Clinopodium all genera listed above. Although the relationships between Acinos and Ziziphora are not fully resolved, it seems necessary to combine these two genera in one genus. Ziziphora has priority over Acinos, and the species of Acinos accordingly need new names in Ziziphora. These are not yet available. The only German Clinopodium species, C. vulgare L., is the type of the genus name and will therefore most likely not be affected by any changes of generic circumscriptions. The only genus of this group that seems to be unproblematic is Satureja. This genus, together with Gontscharovia Boriss., is part of a polytomy with the Clinopodium-clade (the numerous genera listed above) and a clade of Old World Micromeria (Bräuchler & al. 2010). Even if Satureja is paraphyletic in relation to Gontscharovia, Satureja has priority over Gontscharovia and no taxonomic changes will be necessary in the German flora.
A long-discussed problem is the correct placement and naming of species belonging to Galeobdolon Adans. / Lamiastrum Heist, ex Fabr. (Dandy 1967; Holub 1970; Rauschert 1974; Mennema 1989; Krawczyk & al. 2013). Choice of genus name is a nomenclatural problem, which will not be dicussed here. In recent Floras of or covering Germany, either both names were used: Galeobdolon (Jäger & Werner 2005; Jäger 2011) and Lamiastrum (Heywood & Richardson 1972; Seybold 2009), or the species of Galeobdolon / Lamiastrum were included in Lamium L. (Seybold 2011). Molecular phylogenies of subfam. Lamioideae (Bendiksby & al. 2011b) and of Lamium (including species of Galeobdolon / Lamiastrum; Bendiksby & al. 2011a) clearly showed that a well-supported clade of species of Galeobdolon / Lamiastrum is sister to a well-supported Lamium. Accordingly, both inclusion of Galeodolon / Lamiastrum in Lamium and treatment as two distinct genera would result in monophyletic genera. When included in Lamium, G. argentatum Smejkal, G. flavidum (F. Herm.) Holub, G. luteum Huds. and G. montanum (Pers.) Pers. ex Rchb. should be L. argentatum (Smejkal) Henker ex G. H. Loos, L. flavidum F. Herm., L. galeobdolon (L.) L. and L. montanum (Pers.) Hoffm. ex Kabath, respectively. The treatment of these four taxa at species level has been questioned. When treated as subspecies of Lamium galeobdolon (e.g. by Bendiksby & al. 2011a), the names to be used would be L. galeobdolon subsp. argentatum (Smejkal) J. Duvign., L. galeobdolon subsp. flavidum (F. Herm.) Á. Löve & D. Löve, L. galeobdolon subsp. galeobdolon and L. galeobdolon subsp. montanum (Pers.) Hayek, respectively.
Majorana Mill, and Origanum L. are two genera containing commonly used spices. Origanum is distributed in Eurasia and contains c. 38 species (Mabberley 2008). Majorana hortensis Moench was first described as O. majorana L. A phylogenetic analysis by Katsiotis & al. (2009) showed that M. hortensis is nested among other species of Origanum, so that recognition of M. hortensis would make Origanum paraphyletic. Therefore the inclusion of M. hortensis in Origanum, as O. majorana, is appropriate.
Salvia L. in its traditional circumscription is a large genus of 800–900 tropical to temperate species (Mabberley 2008). Recent molecular studies in the genus clearly showed that Salvia is highly polyphyletic (Walker & Sytsma 2007; Will & Claßen-Bockhoff 2014) and will have to be split into several genera (Will & Claßen-Bockhoff 2014; M. Will pers. comm.). The only alternative would be to include several smaller genera in Salvia, e.g. Rosmarinus L., which would inflate this large genus even more. The German flora is largely unaffected by these changes. According to the different phylogenies available, only one species of the German flora, S. glutinosa L., will have to be transferred to a new genus.
Stachys L. is a large genus of c. 450 species distributed in temperate and warm regions of the world, including tropical mountains but excluding Australasia (Mabberley 2008). Molecular phylogenies have shown that Stachys is highly polyphyletic, with many different genera nested among different Stachys clades (Bendiksby & al. 2011b; Salmaki & al. 2013). In Germany eight species and one hybrid taxon of Stachys can be found. These fall into four larger Stachys clades (Salmaki & al. 2013). Only S. arvensis L. and S. palustris L. fall into the clade containing the type of Stachys, S. sylvatica L. This clade also contains Haplostachys Hillebr., Phyllostegia Benth., Stenogyne Benth. and Suzukia Kudô. Sideritis L. is one of those genera nested among different Stachys clades (Bendiksby & al. 2011b; Salmaki & al. 2013). This genus of c. 140 N-temperate species of the Old World and Macaronesia (Mabberley 2008) occurs in Germany with only one species, S. montana L. Relationships among the genera listed above and several others are complicated and not fully resolved. At this point two solutions seem possible. One is to include all species of a clade called Eurystachys Salmaki & Bendiksby (including Stachys and Sideritis; Salmaki & al. 2013) in one genus. The other is to split Stachys into a large number of smaller genera. Both solutions are problematic. The first would result in a large genus that is morphologically heterogeneous and, according to Salmaki & al. (2013), c. 194 new combinations would have to be made. The second solution would allow maintaining morphologically distinct genera. However, it would require dividing Stachys into several genera that would be difficult to delimit (Salmaki & al. 2013). In this second approach a similarly high number of combinations would be necessary. The first solution would allow keeping all German species of Stachys in Stachys, which, however, would also have to include Sideritis. In the second solution, Stachys alpina L., S. annua L., S. byzantina K. Koch, S. germanica L. and S. recta L. most likely would need to be excluded from Stachys. The relationships of Sideritis are unresolved. It therefore remains unclear whether Sideritis montana would need a new name when opting for the second solution. A further difficulty of the second solution is a high level of incongruence between the nuclear and plastid data sets analysed (Salmaki & al. 2013). Future changes in this group are clearly necessary. They will affect large numbers of species on a global scale, but only few species of the German flora.
Orobanchaceae (D. C. Albach)
There has been some debate about the monophyly of Orobanche L., and some publications re-used the name Phelipanche Pomel, introduced for some morphologically deviant species more commonly treated as O. sect. Trionychon Wallr. (in Germany O. arenaria Borkh., O. purpurea Jacq. and O. ramosa L.; see lead 1 in the key to Orobanche in Jäger 2011). The group differs from the type section not only in flower morphology but also in seed ultrastructure (Plaza & al. 2004) and pollen morphology (Abu Sbaih & al. 1994). Whereas first cpDNA-based phylogenetic analyses suggested O. ramosa to be nested in the rest of Orobanche (Young & al. 1999; Manen & al. 2004), subsequent analyses using ITS (Schneeweiss & al. 2004) revealed a biphyletic Orobanche with O. sect. Trionychon and New World species of the genus forming a clade and O. sect. Orobanche sister to Diphelypaea Nicolson. More detailed analyses of cpDNA sequences demonstrated that the nested position of O. sect. Trionychon is due to horizontal gene transfer (Park & al. 2007). However, analyses of another nuclear marker (PhyA; Bennett & Mathews 2006) as well as cpDNA analyses removing introgressed sequences (Park & al. 2008) agree on phylogenetic relationships with Orobanche being monophyletic and O. sect. Trionychon and New World species being sister to O. sect. Orobanche. Thus, no taxonomic changes will be necessary.
Linderniaceae (D. C. Albach)
Only two species of Linderniaceae occur in Germany and are commonly still recognized under Lindernia L., one being the type of the genus name, L. procumbens (Krock.) Borbás. Lindernia dubia (L.) Pennell, though, has been demonstrated to be more closely related to Micranthemum Michx. than to Lindernia (Fischer & al. 2013). However, no generic realignment has been proposed so far. Lindernia dubia had been recognized as separate from Lindernia before under the names Gratiola dubia L. or Ilysanthes riparia Raf., but was included in Lindernia by Pennell (1935). Ilysanthes Raf. had been separated from Lindernia based on the reduction of the androecium to two stamens, which Pennell (1935) did not consider stable enough to merit generic rank. Fischer & al. (2013) seemed to favour inclusion of Micranthemum in Lindernia. However, since Micranthemum also has only two stamens and occurs sympatrically with L. dubia in North America, combining L. dubia in Micranthemum remains a possibility.
Convolvulaceae (J. W. Kadereit)
A monophyletic Calystegia R. Br. is clearly nested within Convolvulus L. (Stefanović & al. 2002; Carine & al. 2004; Williams & al. 2014) and should be classified in Convolvulus following Stefanović & al. (2002). Combinations are available for most German species of Calystegia, and C. pulchra Brummitt & Heywood should be Convolvulus dubius J. L. Gilbert, C. sepium (L.) R. Br. should be Convolvulus sepium L., C. silvatica (Kit.) Griseb. should be Convolvulus silvaticus Kit. and C. soldanella (L.) Roem. & Schult, should be Convolvulus soldanella L.
Solanaceae (J. W. Kadereit)
Lycopersicon Mill, is clearly nested in Solanum L. (Spooner & al. 1993) and should be treated in that genus. Accordingly, the tomato should be called S. lycopersicum L.
Whitson & Manos (2005) demonstrated that the two species of Physalis L. listed for Germany, P. alkekengi L. and P peruviana L., fall into two distantly related clades of Physalinae. The authors argued: “To correct the paraphyly of Physalis, nomenclatural changes are required. Options include restricting the name Physalis to P. alkekengi, the type, and renaming the 75+ species of New World Physalis, or broadening the circumscription of Physalis by uniting the majority of the Physalinae into a single genus. However, the least taxonomically disruptive approach for dealing with this problem is to re-typify Physalis using a Linnaean species that is a member of the morphologically typical Rydbergis clade, such as P. pubescens. The atypical species could then be recognized as four small genera (for P. carpenteri, P. alkekengi, P. microphysa, and subgenus Physalodendron), which would produce a morphologically homogeneous Physalis. A proposal to re-typify Physalis is currently in progress.” This proposal has been made by Whitson (2011), and conservation of Physalis L. with conserved type has been recommended (Applequist 2012). If accepted, P. alkekengi should be known as Alkekengi officinarum Moench.
Boraginaceae (M. Weigend)
Generic limits in the large family Boraginaceae (1500— 1600 spp.) are highly problematic and numerous re-alignments of generic limits are required, but few of these problems concern the German flora. The genus Omphalodes Moench is represented only by two species in Germany: O. scorpioides (Haenke) Schrank and O. verna Moench. Omphalodes scorpioides has been shown to be more closely related to Mertensia Roth than to the typical representatives of Omphalodes (Weigend & al. 2013), and is now accommodated in the monospecific genus Memoremea A. Otero & al. as Memoremea scorpioides (Haenke) A. Otero & al. (Otero & al. 2014). This is clearly supported by molecular data, but also by gross differences in habit and its aberrant fruit morphology (circular wing of the nutlet forming a hollow ring, not a flat appendage). Omphalodes in the narrowest sense is restricted to those perennial, rhizomatous herbs which are closely related to O. verna, the type of the genus name. This group ranges from N Spain to N Iran. Other groups from Asia and the Americas previously assigned to Omphalodes either have already been segregated from the genus (Otero & al. 2014) or will likely be removed to other genera.
The genus Buglossoides Moench is also represented by only two species in Germany. Buglossoides arvensis (L.) I. M. Johnst., an annual weed with tiny white flowers and four triangular-ovate, verrucose nutlets and B. purpurocaerulea (L.) I. M. Johnst., a perennial herb with large, blue, hypocrateriform flowers and single, smooth, spherical nutlets. They represent the C European representatives of two highly natural and monophyletic species groups, which are retrieved as sister groups in molecular studies (Weigend & al. 2009; Cecchi & al. 2014). These species groups have recently been segregated into two different, easily distinguished genera (Cecchi & al. 2014): Buglossoides s.str., essentially comprising the two species B. arvensis and B. incrassata and largely restricted to the circum-Mediterranean region and Europe (and introduced as weeds elsewhere), and Aegonychon Gray with a total of three species, one narrow S Italian endemic and the widespread A. purpurocaeruleum (L.) Holub, in W Eurasia as sister to the morphologically barely distinguishable Japanese endemic A. zollingeri (A. DC.) Holub (Cecchi & al. 2014). The clear morphological differences between these two groups justify their separation into two well-defined genera, but phylogenetic data would equally permit a broader delimitation of Buglossoides, including Aegonychon.
The genera Eritrichium Schrad. ex Gaudin, Hackelia Opiz and Lappula Moench have a confused taxonomic history, but Hackelia was finally segregated from Lappula by Johnston (1923). The only C European species of Hackelia and Lappula and the types of those names, H. deflexa (Wahlenb.) Opiz and L. squarrosa (Retz.) Dumort., have recently often been treated as belonging to a single genus, i.e. Lappula. Recent molecular studies retrieved these two species in widely separate clades in tribe Eritrichieae, together with the bulk of the species currently assigned to the respective genera. There is therefore both morphological (Johnston 1923) and molecular (Weigend & al. 2013) evidence supporting the recognition of the two genera. The exact limits between Eritrichium and Hackelia and Lappula still require additional work, with several extra-European segregate genera apparently nested in them, and some species incorrectly placed. This, however, does not concern the German or European flora.
The delimitation of Anchusa L., characterized by radially symmetrical flowers, from Lycopsis L. with curved, slightly zygomorphic flowers, has been contentious in the past. Morphological differences are small but striking, and the segregate Lycopsis is currently not generally recognized. Hilger & al. (2004) advocated the subdivision of Anchusa into several smaller genera, including the separation of Lycopsis. However, their molecular data failed to retrieve the two species of Lycopsis as monophyletic, and there was no statistical support for Anchusa excluding Lycopsis. Generic limits in Anchusa s.1. clearly require more work, and it seems more sensible at this stage to recognize a single, more widely defined genus Anchusa until much better data are available.
Two other genera represented in Germany will likely be subject to re-definition in the near future, without affecting the taxonomy of German species: Both species of Cynoglossum L. were retrieved in the core-clade of Cynoglossum s.1. (Weigend & al. 2013) and certainly will remain part of a redefined Cynoglossum. However, Cynoglossum likely will have to include a whole range of W Eurasian segregate genera (Hilger & al. 2015). Similarly, Heliotropium europaeum L. is the type of Heliotropium L. and therefore will not be affected by name change, irrespective of how the limits of Heliotropium, with the large genus Tournefortia L. deeply nested in it (Luebert & al. 2011), will ultimately be redefined.
Apiaceae (K. Spalik)
Hacquetia DC. is nested within Sanicula L. (Valiejo-Roman & al. 2002; Calvifio & Downie 2007) and should therefore be sunk into synonymy; for its only species, H. epipactis (Scop.) DC., the name S. epipactis (Scop.) E. H. L. Krause is available.
Apium L. s.l. is polyphyletic and among its European species only the type, A. graveolens L., is retained in the genus; the other true celeries are distributed throughout the S hemisphere (Spalik & al. 2010). For its other European members, the genus Helosciadium W. D. J. Koch has been reinstated (Hardway & al. 2004; Spalik & al. 2009; Ronse & al. 2010) including H. inundatum (L.) W. D. J. Koch (A. inundatum (L.) Rchb. f.), H. nodiflorum (L.) W. D. J. Koch (A. nodiflorum (L.) Lag.) and H. repens (Jacq.) W. D. J. Koch (A. repens (Jacq.) Lag.). The species of Helosciadium are hydrophytes or helophytes and are closely related to the morphologically and ecologically similar Berula W. D. J. Koch and Sium L., members of tribe Oenantheae (Spalik & al. 2014).
Carum L., the type of Careae, includes c. 30 species that in molecular analyses are located in several disparate clades interspersed with species of Chamaesciadium C. A. Mey., Fuernrohria K. Koch and Grammosciadium DC., with only few species closely related to the type of Carum, C. carvi L. (Zakharova & al. 2012). Carum verticillatum (L.) W. D. J. Koch is a very distant relative of its nominative congeners and, therefore, was placed in the reinstated monospecific genus Trocdaris Raf.; its proper name is T. verticillata (L.) Raf. (Zakharova & al. 2012). This species forms an isolated lineage in a clade of hydrophytic umbellifers constituting tribe Oenantheae (Spalik & al. 2014).
The genera Angelica L., Cnidium Cusson, Libanotis Haller ex Zinn, Peucedanum L., Selinum L., Seseli L. and Trinia Hoffm. are part of the taxonomically difficult tribe Selineae (Spalik & al. 2004; Downie & al. 2010). Many of its genera are polyphyletic while at the same time many monophyletic lineages have unnecessarily been split into small segregates. Numerous species have not yet been included in molecular phylogenetic studies, and the generic boundaries remain unclear. Phylogenetic relationships within this tribe were mostly examined using only nuclear ITS sequences that have some limitations. Moreover, the tribe originated relatively recently, c. 12 Mya, and underwent rapid radiation (Banasiak & al. 2013: Appendix S2). In effect, internal branches of the phylogenetic trees obtained from molecular data are short and often poorly supported, precluding unambiguous taxonomic inferences.
Seseli sensu amplo encompasses 100–120 species and is obviously polyphyletic: its species occur in tribes Apieae, Pimpinelleae and Selineae (Downie & al. 2010), and in Selineae they are placed in several clades (Spalik & al. 2004). Seseli hippomarathrum Jacq. together with three other congeners forms a clade that is not most closely related to S. tortuosum L., the type of the genus name; for this group, a restitution of Hippomarathrum G. Gaertn. & al. has been considered (Spalik & al. 2004). Depending on taxonomic sampling and the method of phylogenetic inference, this clade was placed sister to the Seseli clade (Spalik & al. 2004) or sister to Peucedanum s.l. (see Appendix S2 in Banasiak & al. 2013). Detailed molecular and morphological studies are necessary to elucidate the taxonomic status of this group. Upon restitution of Hippomarathrum the name H. pelviforme G. Gaertn. & al. would be available for S. hippomarathrum. Seseli annuum L. has not yet been included in molecular analyses; therefore, its phylogenetic affinities remain unknown.
Libanotis pyrenaica (L.) Bourg eau is closely related to L. montana Crantz, the type of Libanotis, and in Flora iberica (Aedo & Vargas 2003) the former was synonymized with the latter. In molecular analyses, the clade containing these two species is sister to a clade containing the type of Seseli (Spalik & al. 2004; Banasiak & al. 2013). If a broad definition of Seseli is adopted, e.g. based on the Seseli clade in Spalik & al. (2004), then Libanotis should be sunk into Seseli and the species is to be named S. libanotis (L.) W. D. J. Koch.
Ligusticum mutellinoides (Crantz) Vill. (Pachypleurum mutellinoides (Crantz) Holub) is also closely related to the Libanotis—Seseli clade in tribe Selineae, whereas the types of Ligusticum L. and Pachypleurum Ledeb. are placed in the Acronema clade, which deserves rank as a separate tribe (Downie & al. 2010; Banasiak & al. 2013). Depending on the delineation of Libanotis and Seseli, Ligusticum mutellinoides may be included in either of these two genera. Alternatively, Neogaya Meisn. may be reinstated. Its type is N. simplex (L.) Meisn., a taxonomic synonym of L. mutellinoides. In molecular phylogenetic trees, Ligusticum mutellina (L.) Crantz is placed in the Conioselinum chinense clade far from the type of Ligusticum and, therefore, should be excluded from the genus and placed in the reinstated Mutellina Wolf, as M. purpurea (Poir.) Reduron & al. (Valiejo-Roman & al. 2006).
Cnidium dubium (Schkuhr) Schmeil & Fitschen is not most closely related to the type of Cnidium, C. monnieri (L.) Spreng., and should therefore be recognized as Kadenia dubia (Schkuhr) Lavrova & V. N. Tikhom. (Valiejo-Roman & al. 2006).
Trinia is exceptional in Apiaceae due to its dioecious breeding system, and this feature seems to be synapomorphic for the genus. So far, only T. hispida Hoffm. has been included in molecular phylogenetic analyses and it was placed in the Seseli clade very close to the type of Seseli (Spalik & al. 2004). If this placement is confirmed upon extended sampling of species and molecular markers, then either Trinia is to be included into the synonymy of Seseli or the latter is to be restricted to a clade of only a few closest relatives of its type.
Peucedanum sensu amplo includes c. 100–120 species worldwide and is a “dustbin” genus encompassing taxa that do not fit elsewhere. The European species have often been transferred to small segregate genera including Cervaria Wolf, Dichoropetalum Fenzl (= Holandrea Reduron & al.), Imperatoria L., Oreoselinum Mill., Thysselinum Adans. and Xanthoselinum Schur. Of these, however, only Cervaria and Dichoropetalum are unambiguously supported by molecular data because their types are distant relatives of Peucedanum officinale L., the type of the genus name. The remaining segregates form the Peucedanum s.l. clade that can be retained as one genus (Spalik & al. 2004). Therefore, the use of the names Dichoropetalum carvifolia (Vill.) Pimenov & Kljuykov (P. carvifolia Vill.) and Cervaria rivini Gaertn. (P. cervaria (L.) Lapeyr.) is advocated. If a very narrow definition of Peucedanum is adopted, the names Imperatoria ostruthium L. (P. ostruthium (L.) W. D. J. Koch), Oreoselinum nigrum Delarbre (P. oreoselinum (L.) Moench), Thysselinum palustre (L.) Hoffm. (P. palustre (L.) Moench) and Xanthoselinum alsaticum (L.) Schur (P. alsaticum L.) are available for the respective species of Peucedanum.
Molecular data have demonstrated that Laserpitium L. is polyphyletic (Weitzel & al. 2014; Lyskov & al. 2015), and this polyphyly is strongly supported by nrDNA and cpDNA markers (Banasiak & al. in press). Six closely related species including the type, L. gallicum L., as well as L. latifolium L. constitute Laserpitium s.str. Laserpitium siler L. forms an isolated lineage that is not closely related to the type and, therefore, the restitution of the monospecific Siler Crantz has been postulated; the respective name for L. siler is S. montanum Crantz. Laserpitium prutenicum L. is more closely related to Daucus L. than to Laserpitium s.str. and, together with its closest relative, L. hispidum M. Bieb., it deserves to be placed in a new genus, Silphiodaucus (Koso-Pol.) Spalik & al. (Banasiak & al. in press). The respective name for L. prutenicum would be S. prutenicus (L.) Spalik & al.
Dipsacaceae (J. W. Kadereit)
Virga Hill, with V. pilosa (L.) Hill and V. strigosa (Roem. & Schult.) Holub clearly groups in Dipsacus L. (Avino & al. 2009; Carlson & al. 2009) and these two species should be known as Dipsacus pilosus L. and D. strigosus Roem. & Schult., respectively.
Valerianaceae (J. W. Kadereit)
Neither Valerianella Mill, nor Valeriana L. are monophyletic according to Hidalgo & al. (2004) and Bell & Donoghue (2005). However, inclusion of Fedia Gaertn. emend. Moench in Valerianella, and both inclusion of Plectritis (Lindl.) DC. in Valeriana and exclusion of some species of Valeriana could make the two genera monophyletic.
Campanulaceae (N. Kilian)
The two German species of Lobelia L., L. dortmanna L. and L. erinus L., fall into two different clades of a highly paraphyletic Lobelia (Antonelli 2008). If this should result in splitting of Lobelia, an approach considered premature by Lammers (2011), L. erinus would belong to a different genus.
Wahlenbergia Roth has been shown to be polyphyletic (Haberle & al. 2009; Roquet & al. 2009; Prebble & al. 2012; Cupido & al. 2013). Wahlenbergia hederacea (L.) Rchb., the only species present in Germany, is not only misplaced in Wahlenbergia (typified by W. elongata (Willd.) Schrad., a synonym of the S African W. capensis (L.) A. DC.; Lammers 2007) but also in the otherwise monophyletic wahlenbergioid group of genera (Cupido & al. 2013). It appears instead to be a close relative of Feeria Buser and Jasione L. (Prebble & al. 2012; Cupido & al. 2013; Crowl & al. 2014; but not so in Mansion & al. 2012), but its systematic position still needs clarification.
A number of molecular phylogenetic studies of Campanulaceae (Eddie & al. 2003; Park & al. 2006; Roquet & al. 2008, 2009; Borsch & al. 2009; Haberle & al. 2009; Mansion & al. 2012; Crowl & al. 2014) have shown that Campanula L. in its present circumscription is not monophyletic, and that the species of this genus fall into at least four major clades, each containing other genera of the family. Referring, with a view on the German flora, to the analysis based on the most comprehensive sampling by Mansion & al. (2012), which also provides the best resolution so far, the three largest major clades are relevant. These are: (1) the Campanula s.str. clade (Park & al. 2006; Roquet & al. 2008, 2009; Borsch & al. 2009; Mansion & al. 2012), including the type of the genus name, C. latifolia L., and comprising clades 13–17 in Mansion & al. (2012), contains the majority of the Campanula species in Germany (C. alliariifolia Willd., C. alpina Jacq., C. barbata L., C. bononiensis L., C. cervicaria L., C. glomerata L., C. latifolia, C. medium L., C. rapunculoides L., C. sibirica L., C. thyrsoides L. and C. trachelium L., all nested in clade 17). The Campanula s.str. clade also includes the species of the S European Trachelium L., but the different analyses demonstrate that this genus does not constitute a natural group but is found dismembered in clades 13 and 16 in Mansion & al. (2012). (2) The Rapunculus clade (clades 5–12 in Mansion & al. 2012) includes all but one of the remaining species in Germany (C. baumgartenii Becker, C. cochleariifolia Lam., C. rhomboidalis L., C. rotundifolia L. [incl. C. gentilis Kovanda] and C. scheuchzeri Vill. in clade 12; C. patula L. and C. rapunculus L. in clade 9) and also contains (in clade 6) the genera Adenophora Fisch, and Hanabusaya Nakai. (3) The third major clade, which has low support, comprises the well-supported clades 2–4 in Mansion & al. (2012), in which several Campanula lineages (among them the last German member C. persicifolia L. in clade 3) are mixed with Asyneuma Griseb. & Schenk, Legousia Durande and Phyteuma L. as well as with the American genera Githopsis Nutt., Heterocodon Nutt, and Triodanis Raf. Faced with different classificatory options, i.e. (1) treating all clades containing species of Campanula as one genus, (2) limiting Campanula to the Campanula s.str. clade, and (3) splitting Campanula into numerous small genera, an option briefly discussed by Park & al. (2006), Roquet & al. (2008) concluded: “We favor the first option in order to arrive at a generic delimitation that reflects the evolutionary history of Campanula. This approach is more consistent with previous taxonomic work, Campanula has always been very rich in number of species, and it does not seem to us reasonable to divide it ad nauseam. … However, a comprehensive study of the currently recognized genera that fall within Campanula should be conducted before changing their taxonomic status.” If this approach would be taken, all species of Adenophora, Legousia and Phyteuma would have to be treated as Campanula.
Menyanthaceae (J. W. Kadereit)
Although Nymphoides Ség. was found to be non-monophyletic, with one species more closely related to one clade of a non-monophyletic Villarsia Vent, than to the remaining species of Nymphoides (Tippery & al. 2008), N. peltata (S. G. Gmel.) Kuntze will not change name even when combined with Villarsia because Nymphoides is the older name.
Cardueae (A. Susanna & N. Garcia-Jacas)
Extensive molecular analyses in subtribe Centaureinae have demonstrated that Centaurea L., as defined in classic terms, was a polyphyletic assemblage (Susanna & al. 1995; Garcia-Jacas & al. 2001). As regards naming of the two main lineages, problems originated with an inadequate type of the genus name proposed by Britton & Brown (1913), a decision later ratified by Dittrich (1993): C. centaurium L. This species belongs to a group of some 20–25 taxa that are not closest relative of the largest part of the genus. Two alternate solutions were possible for achieving a natural delineation of the two genera that should be recognized: first, to keep the old type and limit the use of Centaurea to this group of species, which would imply renaming more than 200 species in a different genus; second, to conserve a new type belonging to the main group of the genus. This second, more conservative (in terms of botanical nomenclature) option finally prevailed: a new type, C. paniculata L., was proposed by Greuter & al. (2001) and is now the conserved type of Centaurea (Wiersema & al. 2015). The valid name for the genus comprising the smaller group of species is Rhaponticoides Vaill. This change, in Germany, affects only C. ruthenica Lam., which should be known as R. ruthenica (Lam.) M. V. Agab. & Greuter. As for the segregation of C. sect. Cyanus (Mill.) DC. as a separate genus (e.g. Greuter & al. 2001), molecular evidence, although inconclusive, points at a sister relationship of C. sect. Cyanus and C. sect. Centaurea (e.g. Garcia-Jacas & al. 2001). The latest proposal for a classification of the entire genus Centaurea (Hilpold & al. 2014) and the revisions of tribe Cardueae by Susanna and Garcia-Jacas (2007, 2009) do not accept Cyanus Mill, as genetically different from Centaurea.
Cichorieae (N. Kilian)
Lapsana L., together with the equally epappose Mediterranean Rhagadiolus Juss., is nested in Crepis L., as has been shown in nuclear ribosomal (ITS) and chloroplast (matK) DNA marker phylogenies by Enke & Gemeinholzer (2008). To maintain Lapsana, which is monospecific after the well-supported segregation of the E Asian Lapsanastrum Pak & K. Bremer (Pak & Bremer 1995; Deng & al. 2014) and the dispecific Rhagadiolus as separate monophyletic genera, Crepis would have to be split into two morphologically ill-defined entities. This is definitely no practicable solution. If only monophyletic genera should be accepted, merging of both genera with Crepis would be the more appropriate solution, although breaking with a long tradition (no combination of Lapsana in Crepis has been published). The morphological circumscription of Crepis does not, however, preclude the inclusion of Lapsana communis L. (and of Rhagadiolus) if variation is extended to allow for the absence of a pappus. In other subtribes, parallel cases of epappose entities traditionally treated as separate genera are similarly found nested in regularly pappose genera (e.g. Deng & al. 2014).
The members of Hypochaeris L. cluster in two main clades according to the phylogenetic analyses of nuclear ribosomal (ITS) and several chloroplast DNA marker sequences by Samuel & al. (2003, 2006) and Enke & al. (2012). The results, however, are inconclusive as to whether the two clades actually form a sister group and thus to the monophyly of Hypochaeris. Based on these findings, Talavera & al. (2015a) opted for splitting the genus in the forthcoming treatment of Flora iberica, there recognizing the segregates Achyrophorus Adans. (in its narrow sense distributed in the Mediterranean region) and Trommsdorffia Bernh. (with T. maculata (L.) Bernh. [H. maculata L.] and T. uniflora (Vill.) Soják [H. uniflora Vill.] in the German flora), a solution that necessitates recognition of at least a fourth genus for the NW African-South American clade of Hypochaeris s.l.
Leontodon L. in its traditional circumscription is at least diphyletic (Samuel & al. 2006; Enke & al. 2012). Leontodon subg. Leontodon and L. subg. Oporinia (D. Don) Peterm., which both received strong support in molecular phylogenies, are nested in two different major clades of the subtribe. This finding from phylogenetic analyses based on both nuclear ribosomal (ITS) and chloroplast (matK) DNA marker sequences necessitates the recognition of L. subg. Oporinia (including L. autumnalis L., L. helveticus Mérat and L. montanus Lam. in Germany) as a separate genus, Scorzoneroides Moench, with S. autumnalis (L.) Moench as type (Greuter & al. 2006) and S. helvetica (Mérat) Holub and S. montana (Lam.) Holub as additional species in the German flora. The authorship of Scorzoneroides should be attributed to Moench (Meth.: 549. 1794), because an earlier place of publication of that name and other genus names (in a German translation dated 1754–1756 of a pre-Linnaean work by Vaillant) is expected to be added to the list of suppressed works by the next International Botanical Congress (Applequist 2014: 1370). Leontodon s.str., moreover, is paraphyletic with respect to the small, chiefly Mediterranean genus Hedypnois Mill., not present in the flora of Germany (Enke & al. 2012). The nrlTS phylogeny by Samuel & al. (2006) and Enke & al. (2012) also provide initial indication (without statistical support, and not supported by the matK phylogeny) that L. sect. Thrincia (Roth) Benth. (with only L. saxatilis Lam. [Thrincia saxatilis (Lam.) Holub & Moravec] in the German flora) forms a clade not sister to the remainder of Leontodon s.str. Based on these findings, Talavera & al. (2015b) revived the genus Thrincia Roth for this clade.
Picris L. is monophyletic after exclusion of the small Mediterranean-SW Asian genus Helminthotheca Zinn (Samuel & al. 2006; Enke & al. 2012). Its segregation has previously been concluded for morphological reasons by Lack (1975). The only species of the latter genus in Germany is II. echioides (L.) Holub (Picris echioides L.), which also provides the type of the genus name.
Scorzonera L. is polyphyletic in all current circumscriptions according to the initial molecular phylogenetic investigations in the subtribe by Mavrodiev & al. (2004) and Owen & al. (2006), using nuclear ribosomal (ITS and ETS) DNA markers and Amplified Fragment Length Polymorphisms (AFLP) variation, respectively. According to these analyses, the clade of Scorzonera s.str. (including the type of the name, S. humilis L., as well as S. purpurea L.) is sister to a clade comprising Podospermum DC. (of which the only member in the German flora, P. laciniatum (L.) DC. [S. laciniata L.], provides the type of that name). The other members of Scorzonera in its wider circumscription, as far as included in the analyses, are distributed over at least three further clades. Two of them, which form a clade sister to the clade comprising Koelpinia Pall, and the Podospermum and Scorzonera s.str. clades (Owen & al. 2006), each include one species in the German flora: S. austriaca L. and S. hispanica L. The third clade is the “Lasiospora clade” (including S. hirsuta L., the type of Lasiospora Cass.), which is sister to all other lineages of the subtribe but has no representatives in the German flora. Apart from the segregation of Podospermum DC. from Scorzonera s.str., which is supported as an option (but not a necessity), the current state of our knowledge of Scorzonera s.l. is still far too preliminary to draw taxonomic conclusions.
Sonchus L. has turned out to be paraphyletic with respect to various smaller Mediterranean-Macaronesian and Australian-New Zealand segregates as well as to the SE Pacific Ocean island endemics Dendroseris D. Don and Thamnoseris Phil, in a series of molecular phylogenetic analyses based on both nuclear ribosomal and chloroplast DNA markers (Kim & al. 2007 and references therein). The preferred and envisaged taxonomic solution is the broadening of the generic concept for Sonchus and (re)inclusion of all these genera (Mejias & Kim 2012). A splitting approach would inevitably dismember even the four German representatives of the genus, the congenerity of which has never been questioned.
The systematics of the Lactuca alliance, which is represented in the German flora by the genera Cicerbita Wallr., Mycelis Cass, and Lactuca L., has been in lively debate for more than 200 years. The first molecular phylogenetic analyses published (Koopman & al. 1998; Wang & al. 2013) explained the difficulties in arriving at a natural classification with frequent convergent evolution of morphological characters. Consequences for the generic classification of the species in Germany are to be expected, but phylogenetic reconstruction is still in progress and any reclassification would be premature at present.
Prenanthes L. has been redefined completely on the basis of molecular phylogenetics, now being understood as a probably monospecific genus, accommodating the chiefly European P. purpurea L. (Kilian & Gemeinholzer 2007; Kilian & al. 2009; Wang & al. 2013).
The placement of the C and SE European Tolpis staticifolia (All.) Sch. Bip., the only representative of Tolpis L. in the flora of Germany, is not settled yet. Tolpis staticifolia and the S and tropical African T. capensis (L.) Sch. Bip. (plus its close ally T. mbalensis G. V. Pope) have been excluded from that chiefly Mediterranean-Macaronesian genus based on palynological differences (Blackmore & Jarvis 1986) and on the results of a chloroplast ndhF sequence phylogeny by Park & al. (2001), which placed the two species as sister to Taraxacum F. H. Wigg. (T. capensis) and Crepis (T. staticifolia), respectively.
Recent molecular phylogenetic analyses of the Hieracium alliance using nuclear ribosomal, low-copy nuclear and chloroplast DNA markers (Fehrer & al. 2007, 2009; Krak & al. 2013) revealed conflicting topologies between the different gene trees in particular due to both reticulate evolution and incomplete lineage sorting during the rapid evolution of the alliance. Discussing the available evidence, the authors concluded that the nuclear ribosomal DNA gene trees provide the best approximation for the reconstruction of the species tree. Accordingly, Hieracium L. in the wide sense is polyphyletic. Hieracium subg. Pilosella (Hill.) Fr. is sister to the W Mediterranean genus Hispidella Lam., both are sister to H. subg. Hieracium and the American H. subg. Chionoracium Sch. Bip. (= Stenotheca Monnier), the four taxa in turn are sister to the chiefly Mediterranean-Macaronesian genus Andryala L., and, finally, H. intybaceum All., which is restricted to the siliceous Alps, forms the sister group to all of them. The taxonomic consequences already widely drawn are the recognition as separate genera of Hieracium and Pilosella Hill (for taxonomy and new combinations needed see Bräutigam & Greuter 2007; for the authorship of Pilosella the above notes on Scorzoneroides also apply). The further consequence in order to arrive at monophyletic entities is the resurrection of the genus Schlagintweitia Griseb. to accommodate H. intybaceum (as S. intybacea (All.) Griseb.) and its few allies (Gottschlich & Greuter 2007; Greuter & Raab-Straube 2008).
Senecioneae (J. W. Kadereit)
Phylogenetic analyses of Senecioneae (Pelser & al. 2002, 2007, 2010) have shown that Senecio L. in its traditional circumscription is not monophyletic but rather both poly and paraphyletic. As regards species in the German flora, it is evident that those species that lack outer involucral bracts, i.e. S. congestus (R. Br.) DC., S. gaudinii Gremli, S. helenites (L.) Schinz & Thell., S. integrifolius (L.) Clairv. and S. rivularis (Waldst. & Kit.) DC., need to be segregated as Tephroseris (Rchb.) Rchb., in which they are known as T. palustris (L.) Rchb. (for S. congestus), T. tenuifolia (Gaudin) Holub (for S. gaudinii), T. helenites (L.) B. Nord, (for S. helenites), T. integrifolia (L.) Holub (for S. integrifolius) and T. crispa (Jacq.) Rchb. (for S. rivularis). Tephroseris is only very distantly related to Senecio s.str. and even belongs to a different subtribe of Senecioneae.
Species related to Senecio jacobaea L. should be segregated as Jacobaea Mill., which again is only distantly related to Senecio s.str. These, besides S. jacobaea (J. vulgaris Gaertn.), include S. abrotanifolius L. (J. abrotanifolia (L.) Moench), S. alpinus (L.) Scop. (J. alpina (L.) Moench), S. aquaticus Hill (J. aquatica (Hill) G. Gaertn. & al.), S. erraticus Bertol. (J. erratica (Bertol.) Fourr.), S. erucifolius L. (J. erucifolia (L.) G. Gaertn. & al.), S. incanus subsp. carniolicus (Willd.) Braun-Blanq. (J. incana subsp. carniolica (Willd.) B. Nord.; for a recent account of the S. carniolicus aggregate see Flatscher & al. 2015), S. paludosus L. (J. paludosa (L.) G. Gaertn. & al.) and S. subalpinus Koch. (J. subalpina (W. D. J. Koch) Pelser & Veldkamp). Combinations in Jacobaea are available for all these species (Pelser & al. 2006).
Endocellion Turcz. ex Herder, containing two species in Asia, is clearly nested in Petasites Mill. (Steffen 2013) and should be treated as part of that genus. This does not affect the generic identity of the Petasites species in Germany.
Gnaphalieae (M. Galbany-Casals)
Phylogenetic analyses and morphological data show that Filago L. is not monophyletic, and that the species involved should now be placed in two separate genera not closely related to each other (Galbany-Casals & al. 2010; Andrés-Sánchez & al. 2011): Logfia Cass, includes L. minima (Sm.) Dumort. (F. minima (Sm.) Pers.) and L. gallica (L.) Coss. & Germ. (F. gallica L.), and Filago includes the rest of the species present in Germany. Filago neglecta (Soyer-Willemet) DC. has been claimed to be of hybrid origin between L. gallica and Gnaphalium uliginosum L. (Holub 1976; Jäger 2011), but this is currently considered highly doubtful (Andrés-Sánchez, pers. comm.). However, it is not clear yet if this rarely collected species belongs to Filago or Logfia.
Bombycilaena erecta (L.) Smoljan. has not been treated in Jäger (2005, 2011), but there exists at least one old record of this species from Germany (Andrés-Sánchez & al. 2014). The genus Bombycilaena (DC.) Smoljan. has been shown to be a lineage separate from Micropus L. and Filago in a molecular phylogeny and is currently considered to include only two species from the Old World (Galbany-Casals & al. 2010; Andrés-Sánchez & al. 2014).
Omalotheca Cass, (sensu Holub 1976) has often been considered a synonym of Gnaphalium L. (e.g. Anderberg 1991; Jäger 2005, 2011). However, a molecular phylogeny (Galbany-Casals & al. 2010) has shown that Gnaphalium s.l. is not monophyletic and that these two genera should be considered separate, given that G. supinum L. — the type of Omalotheca — is not closely related to G. uliginosum — the type of Gnaphalium. Additionally, Blöch & al. (2010) showed that G. hoppeanum W. D. J. Koch, G. norvegium Gunnerus and G. sylvaticum L., three species also present in Germany, form a clade with G. supinum. In conclusion, with regard to the German flora, Gnaphalium should be restricted to G. uliginosum, and the other four species named above should be considered to belong to Omalotheca, as O. hoppeana (W. D. J. Koch) Sch. Bip. & F. W. Schultz, O. norvegica (Gunnerus.) Sch. Bip. & F. W. Schultz, O. supina (L.) DC. and O. sylvatica (L.) Sch. Bip. & F. W. Schultz. Smissen & al. (2011) noted that Gnaphalium s.str. includes diploid species (2n = 14), whereas Omalotheca species are all polyploids, and that the latter genus is part of a large clade of ancient allopolyploid origin, together with, among others, genera such as Antennaria Gaertn, Bombycilaena, Filago, Gamochaeta Wedd., Leontopodium R. Br. ex Cass, and Logfia (Galbany-Casals & al. 2010).
Helichrysum Mill, is not monophyletic. Some Australasian species had already been transferred to other genera for morphological reasons (see Bayer 2001 and Ward & al. 2009 for a review) and later were shown not to be part of the main Helichrysum clade (Galbany-Casals & al. 2004; Ward & al. 2009; Smissen & al. 2011). This affects H. bracteatum (Vent.) Willd., an ornamental species (Jäger 2005), which should be known as Xerochrysum bracteatum (Vent.) Tzvelev (Bayer 2001). Anaphalis DC. and Pseudognaphalium Kirp., two genera of hypothesized allopolyploid origin, are embebbed in the main Helichrysum clade (Galbany-Casals & al. 2014). The need for a generic re-circumscription of these three genera, plus others, was extensively discussed by Galbany-Casals & al. (2014), who recommended maintaining Anaphalis, Helichrysum and Pseudognaphalium as independent genera until more data are available. This affects two taxa present in Germany, A. margaritacea (L.) Benth. & Hook. f., an ornamental but naturalized (Jäger 2011) species native to Asia and North America, and P. luteoalbum (L.) Hilliard & B. L. Burtt. The latter species was treated as Gnaphalium luteoalbum L. in Jäger (2005). At present it remains unclear if this species should be included in Helichrysum or Pseudognaphalium, or if it should be treated as Laphangium Tzvelev as was done in Jäger (2011).
Astereae (C. Oberprieler)
The most comprehensive molecular phylogenetic analysis of tribe Astereae based on nrDNA ITS sequences was made by Brouillet & al. (2009). To a large extent, its results are supportive of the generic delimitation proposed by Greuter (2003) for the Euro+Med plantbase treatment of the tribe and of Nesom & Robinson's (2007) treatment of Astereae in Kubitzki's The families and genera of vascular plants (Kadereit & Jeffrey 2007).
In subtribe Solidagininae, results by Brouillet & al. (2009) confirm that Solidago L. is polyphyletic and that the naturalized S. graminifolia (L.) Salisb. should be transferred to Euthamia (Nutt.) Cass. as E. graminifolia (L.) Nutt. because it belongs to another lineage than the type of Solidago (i.e. S. virgaurea L.). While in subtribe Bellidinae the monophyly of both Bellis L. and Bellium L. was repeatedly found in molecular phylogenetic studies based on nrDNA ITS sequences (Fiz & al. 2002; Brouillet & al. 2009; Fiz-Palacios & Valcarcel 2011), phylogenetic analyses in subtribe Asterinae have led to extensive generic rearrangements due to the consistently demonstrated polyphyly of Aster L. in its classical circumscription. According to nrDNA ITS-based analyses by Brouillet & al. (2009), a more narrowly and more naturally circumscribed genus Aster in Germany would only comprise A. alpinus L. and A. amellus L., while A. linosyris (L.) Bernh. should be transferred to the Eurasian genus Galatella Cass. as G. linosyris (L.) Rchb. f., the halophilic A. tripolium L. to the genus Tripolium Nees as T. pannonicum (Jacq.) Dobrocz., and A. bellidiastrum (L.) Scop. not only to the separate and monospecific genus Bellidiastrum Scop. (as B. michelii Scop.) but also to another subtribe (Bellidinae). The last has also been confirmed by the phylogenetic analyses by Fiz & al. (2002) and Fiz-Palacios & Valcarcel (2011). Finally, molecular phylogenies based on nrDNA ITS alone (Brouillet & al. 2009) or on nrDNA ITS + ETS complemented by the intergenic spacer region trnL-trnF of the chloroplast genome (Li & al. 2012b) support the transfer of the naturalized “New World asters” (i.e. A. laevis L., A. lanceolatus Willd., A. novae-angliae L., A. novi-belgii L., A. parviflorus Nees, A. salignus Willd., A. versicolor Willd.) to the genus Symphyotrichum Nees (subtribe Symphyotrichinae). On the other hand, Li & al. (2012b) found no evidence for a close relationship between Callistephus chinensis (L.) Nees and any other genus of subtribe Asterinae and supported its independent generic status. Finally, in subtribe Conyzinae, it has been repeatedly demonstrated (Noyes 2000; Andrus & al. 2009; Brouillet & al. 2009) that neither Conyza Less. nor Erigeron L. as previously defined are monophyletic; a situation that is best accommodated by merging the two genera into Erigeron, as was already suggested by Greuter (2003). This requires the transfer of C. bonariensis (L.) Cronquist, C. canadensis (L.) Cronquist and C. sumatrensis (Retz.) E. Walker to this more broadly circumscribed genus (as E. bonariensis L., E. canadensis L., and E. sumatrensis Retz., respectively).
Anthemideae (C. Oberprieler)
In the S hemisphere subtribe Cotulinae, phylogenetic analyses by Himmelreich & al. (2012) based on sequence variation of nrDNA ITS and intergenic spacer regions (psbA-trnH, trnC-petN) of the chloroplast genome have shown that Cotula L. is non-monophyletic, even when the Mediterranean C. cinerea Delile is excluded as the independent and monospecific genus Brocchia Vis. (as B. cinerea (Delile) Vis.) following results by Oberprieler (2004a). Being the type of Cotula, sinking of Leptinella Cass. and Soliva Ruiz & Pav. into a broader, then monophyletic genus would not affect the name of C. coronopifolia L., naturalized in the N hemisphere. Of subtribe Artemisiinae, as circumscribed by Oberprieler & al. (2007a, 2009), only Artemisia L. and Leucanthemella Tzvelev are represented in our area. In the case of the former genus, there is a consistent tendency supported by many molecular phylogenetic studies of the last years (e.g. Vallès & al. 2003; Sanz & al. 2008; Pellicer & al. 2010, 2011; Garcia & al. 2011) for lumping the numerous small to large segregate genera (i.e. Crossostephium Less., Filifolium Kitam., Mausolea Poljakov, Neopallasia Poljakov, Picrothamnus Nutt., Seriphidium Fourr., Sphaeromeria Nutt. and Turaniphytum Poljakov) into a broadly defined and monophyletic Artemisia. On the other hand, studies focusing on phylogenetic relationships among the remainder of the Artemisiinae sensu Oberprieler & al. (2007a, 2009) in general and on the generic delimitation of Ajania Poljakov versus Chrysanthemum L. in particular, presented no consistent and well-supported evidence for the affiliation of Leucanthemella Tzvelev to any other genus of the subtribe (Masuda & al. 2009; Zhao & al. 2010). As a consequence, Leucanthemella with its sole European species L. serotina (L.) Tzvelev should be treated as an independent evolutionary unit at genus rank. After inclusion, motivated by molecular phylogenetic studies, of the Mediterranean monospecific Otanthus Hoffmanns. & Link and the equally monospecific Turkish endemic Leucocyclus Boiss. in subtribe Matricariinae (Guo & al. 2004; Oberprieler 2004b; Ehrendorfer & Guo 2005), Achillea L. constitutes a monophyletic genus. Support from a comprehensive molecular phylogenetic analysis for the monophyly of the Eurasian and Mediterranean Matricaria L. with its presently accepted six species (Oberprieler & al. 2007b, 2009) is still missing. However, the transfer of the Aegean M. macrotis Rech. f. to Anthemis L. (as A. macrotis (Rech. f.) Oberpr. & Vogt) based on nrDNA sequence information (Oberprieler & Vogt 2006) and the repeatedly shown support for the generic independence of Matricaria (subtribe Matricariinae) from Tripleurospermum Sch. Bip. (subtribe Anthemidinae; e.g. Oberprieler 2004b, 2005; Oberprieler & al. 2007a) and from Microcephala Pobed. (subtribe Handeliinae; e.g. Oberprieler & al. 2007a; Sonboli & al. 2012) contributed strong evidence for the naturalness of Matricaria in its present circumscription. In subtribe Anthemidinae sensu Oberprieler & al. (2007a, 2009) with its species-rich core genera Anthemis L. and Tanacetum L., considerable efforts have been made to achieve a natural delimitation of genera based on molecular phylogenies. After Oberprieler (2001) had shown, with a limited taxon sample, that Anthemis in its traditional circumscription is paraphyletic, and that the species of A. sect. Cota (J. Gay) Rchb. f., distinct in their achene morphology, should be transferred to the independent genus Cota J. Gay ex Guss. (Greuter & al. 2003), Lo Presti & al. (2010) corroborated this finding based on a comprehensive species sampling (c. 75 % of the described species) and sequence information from both nuclear and plastid markers. With the exclusion of further four species from the Caucasus region (i.e. A. calcarea Sosn., A. fruticulosa M. Bieb., A. marschalliana Willd. and A. trotzkiana Bunge) from Anthemis and their accommodation in the newly described genus Archanthemis Lo Presti & Oberpr., and the abovementioned inclusion of Matricaria macrotis (Oberprieler & Vogt 2006), three natural, morphologically distinct genera were established (Lo Presti & al. 2010; Sonboli & al. 2012). To reflect these phylogenetic findings, Anthemis austriaca Jacq. and A. tinctoria L., hitherto treated as Anthemis in Germany, should be transferred to Cota, as C. austriaca (Jacq.) Sch. Bip. and C. tinctoria (L.) J. Gay.
The natural circumscription of Tanacetum L. remains problematic even after considerable taxon and marker sampling. Based on nrDNA ITS and cpDNA trnH-psbA sequence information, Sonboli & al. (2012) could demonstrate that there is no support for a generic separation of the yellow-rayed or rayless species of Tanacetum from the white- and red-rayed species of Pyrethrum Zinn. On the other hand, even after exclusion of several enigmatic species from Tanacetum based on phylogenetic analyses (i.e. T. annuum L. and T. microphyllum DC. transferred to the newly established Vogtia Oberpr. & Sonboli [Sonboli & al. 2012]; T. paradoxum Bornm. transferred to Artemisia [Sonboli & al. 2011]; T. semenovii Herder transferred to Richtera Kar. & Kir. [Sonboli & Oberprieler 2012]) and the suggested inclusion in Tanacetum of many satellite genera (e.g. Balsamita Mill., Gonospermum Less., Gymnocline Cass., Hemipappus K. Koch, Lugoa DC., Spathipappus Tzvelev and Xylanthemum Tzvelev), support for a monophyletic Tanacetum remains weak and awaits phylogenetic reconstructions based on a broader sampling of molecular markers (Sonboli & al. 2012). For the time being, this argues for the presently used broad generic concept of Tanacetum in Germany.
After having been raised from sectional rank in Tanacetum to an independent genus by Heywood (1975), Leucanthemopsis (Giroux) Heywood was considered to be related to Leucanthemum Mill. by Bremer & Humphries (1993) until molecular phylogenetic studies revealed its even closer relationships with three monospecific genera endemic to the Iberian Peninsula, Castrilanthemum Vogt & Oberpr., Hymenostemma Willk. and Prolongoa Boiss. This resulted in its accommodation in the new subtribe Leucanthemopsidinae (Oberprieler & al. 2007a, 2009). More recently, a multilocus phylogenetic analysis of all species of the subtribe in a coalescent-based species-tree reconstruction clearly demonstrated the well-supported monophyly of Leucanthemopsis (Tomasello & al. 2015).
As already discussed by Vogt (1991) in his revision of Leucanthemum Mill. in the Iberian Peninsula, the genus in its traditional circumscription contained species that are only remotely related to its type, L. vulgare Lam. Accommodation of these divergent species in the independent genera Mauranthemum Vogt & Oberpr. and Rhodanthemum B. H. Wilcox & al. by Vogt & Oberprieler (1995) and Bremer & Humphries (1993), respectively, has led to a well-circumscribed and strongly supported monophyletic Leucanthemum, as was recently corroborated by a multi-locus phylogenetic analysis by Konowalik & al. (2015).
In subtribe Santolininae sensu Oberprieler & al. (2007a, 2009), genus delimitations were studied in molecular phylogenetic analyses by Oberprieler (2002). Based on nrDNA ITS and cpDNA trnL-trnF sequence variation, this study demonstrated the paraphyly of Chamaemelum Mill. relative to the monospecific Cladanthus Cass. Transfer of four W Mediterranean Chamaemelum species to Cladanthus led to two well-supported monophyletic sister genera, with the widely cultivated and sporadically naturalized C. nobile (L.) All. and the W Mediterranean C. fuscatum (Brot.) Vasc. as the only members of Chamaemelum.
Glebionis Cass. with the naturalized G. segetum (L.) Fourr. comprises only two species and is the type genus of the small subtribe Glebionidinae (Oberprieler & al. 2007a, 2009). Phylogenetic relationships within this subtribe were studied by Francisco-Ortega & al. (1997), who found little support for the monophyly of the subtribe and for the genus (sub Chrysanthemum) in a maximum-parsimony analysis based on nrDNA ITS sequence variation. While more recent studies using model-based sequence analysis methods (maximum likelihood) gained strong support for the monophyly of the subtribe (Oberprieler 2005; Oberprieler & al. 2007a), relationships among the genera of Glebionidinae, i.e. the species-rich Argyranthemum Webb (24 spp.), Glebionis (two spp.), and the two monospecific genera Heteranthemis Schott and Ismelia Cass., remain unclear, especially after a recent study based on nrDNA ITS sequence variation by Imamura & al. (2015), who found G. coronaria (L.) Spach nested in a group of Argyranthemum species. If future studies should corroborate the non-monophyly of the four genera of Glebionidinae, and their merging would be necessary to arrive at a monophyletic genus, the oldest genus name for this entity would be Heteranthemis Schott. For the time being, however, retaining the four genera in their present circumscriptions appears preferable due to their morphological and geographical distinctness.
Inuleae (J. W. Kadereit)
Phylogenetic analyses of tribe Inuleae have shown that neither Inula L. nor Pulicaria Gaertn. are monophyletic (Anderberg & al. 2005; Englund & al. 2009), but this has not yet been translated into formal taxonomic changes, although possible taxonomic consequences were discussed by Englund & al. (2009). The species of Inula present in Germany fall into at least four different clades, of which I. graveolens (L.) Desf. is more closely related to Pulicaria and its relatives than to Inula and its relatives and has been treated as Dittrichia Greuter. Maintainance of this genus will depend on future treatment of the various lineages of Pulicaria. If, after exclusion of some lineages as suggested by Englund & al. (2009), a broad concept of Pulicaria is adopted, Dittrichia will have to be included in that genus. If, on the other hand, a narrow concept of Pulicaria is adopted, Dittrichia would remain an independent genus and the two species of Pulicaria present in Germany (P. dysenterica (L.) Bernh. and P. vulgaris Gaertn.) would remain in Pulicaria. Adoption of a broad concept of Inula would require inclusion of Carpesium L. and Telekia Baumg. Adoption of a narrow concept would require distribution of the German species in probably several genera, dependent on treatment, and only I. helenium L., as the type, would remain in Inula.
Helenieae (J. W. Kadereit)
Both Bidens L. and Coreopsis L. have been shown not to be monophyletic (Mort & al. 2008), but this has not yet been translated into taxonomic changes.
Heliantheae (J. W. Kadereit)
Both Ambrosia L. and Iva L. have been found not to be monophyletic (Miao & al. 1995). Ambrosia becomes monophyletic after inclusion of Hymenoclea Torr. & A. Gray, as proposed by Panero (2007), whereas parts of Iva are better accommodated in other genera. This affects the German I. xanthiifolia Nutt., which, according to Panero (2007), should be considered a species of Euphrosyne DC. and called E. xanthiifolia (Nutt.) A. Gray.
Madieae (J. W. Kadereit)
Although Eriophyllum Lag. does not appear to be monophyletic (Baldwin & al. 2002), E. lanatum (Push) Forbes, a naturalized ornamental in Germany, is part of the perennial clade, which also contains the type of the genus name. In consequence, no change of name will be necessary should the genus be split.
Among the 840 genera examined, we identified c. 140 where data quality is sufficiently high to conclude that they are not monophyletic, and an additional c. 20 where monophyly is questionable but where data quality is not yet sufficient to reach convincing conclusions. The resolution of these uncertainties will depend on the expansion of taxon and DNA sequence datasets, and on the interpretation of the results by taxonomic specialists. In many cases recognition of non-monophyly offers the options of either to expand genera in order to include former satellites or to split genera into smaller generic entities. As we do not know which of these options will be adopted in each case, we cannot say how the number of genera recognized in the German flora will be affected. General trends in global plant classification, e.g. towards larger genera based on molecular data (Humphreys & Linder 2009), may or may not be reflected in the consequences for the comparatively small and well-studied German flora. However, the summary presented here clearly indicates that considerable further change is inevitable provided monophyly is accepted as the primary criterion for circumscribing genera (and taxa in general). Although such developments may be met with some dismay by users of Floras, they reflect ongoing progress in our scientific understanding of plant diversity.
M. Galbany-Casals would like to thank Santiago Andrés-Sánchez and Rob Smissen, and J. W. Kadereit would like to thank Arne A. Anderberg, Bruce G. Baldwin, Christopher D. Preston and Clive A. Stace — all for helpful advice. Eckehart J. Jäger and an anonymous reviewer are gratefully acknowledged for helpful comments.