Solanum L. is one of the ten most species-rich genera of flowering plants (Frodin 2004). With approximately 1400 species (J. Bennett and S. Knapp, unpubl.) occurring on all temperate and tropical continents, the genus occupies an incredibly wide range of habitats and habits, but the highest diversity of both groups and species occurs in circum-Amazonian tropical South America. Solanum is recognised by its usually pentamerous flowers with fused sepals and petals, stellate to pentagonal corollas, and stamens with short filaments and anthers opening by terminal pores. The genus was one of Linneaus’s (1753) larger, with 23 species mostly described from European or African material. The French botanist Michel-Félix Dunal included 235 species in his thesis (Dunal 1813), mostly the result of extensive European exploration of the Americas. By 1816, when Dunal revised this work (Dunal 1816), this number had risen to 321; many of these additional taxa were based on specimens collected by Alexander von Humboldt and Aimé Bonpland in tropical South America (see Knapp 2007b). The last time the genus was monographed in its entirety was in Candolle’s Prodromus (Dunal 1852) which included 900 Solanum species. Subsequent work on the taxonomy of Solanum has largely been limited to rearrangements of infrageneric taxa, or to the species-level revisions of smaller groups within the genus (see references in Table 1) and floristic treatments. The combination of large numbers of species with relatively poorly circumscribed groups within the genus has meant that Solanum taxonomy has proceeded in a piecemeal fashion until relatively recently. A project funded by the United States National Science Foundation’s Planetary Biodiversity Inventory program begun in 2004 sought to redress this situation by attempting to accelerate species-level taxonomic work across the genus as a whole and at the same time providing a robust phylogenetic framework for this taxonomy. Work by participants of the ‘PBI Solanum’ project (see http://www.solanaceaesource.org ) will result in a modern monographic treatment of the entire genus available on-line. This treatment is part of this collaborative effort.

Dunal (1813, 1816) divided the genus into two major groups, Inermia (unarmed solanums) and Aculeata (armed solanums), based on presence or absence of prickles. He renamed these at the sectional level (but illegitimately, as he cited groups he had previously named as sections, e.g., Pteroidea Dunal [Dunal 1816] in synonymy) “Pachystemonum”, for species with stout anthers and no prickles, and “Leptostemonum”, for species with tapering anthers, usually with stellate hairs and usually possessing prickles (Dunal 1852); these divisions were essentially the same as his Inermia and Aculeata. In Candolle’s Prodromus, Dunal (1852) erected a classification of subsections and ambiguous grades to reflect morphology, mostly of leaf division and inflorescence position. Dunal (1852) maintained as separate the genera Lycopersicon Mill. and Cyphomandra Sendtn., both now subsumed in a monophyletic Solanum (see Bohs 2005; Peralta et al. 2008). The German botanist Georg Bitter, who worked in Berlin, Bremen and Göttingen in the years between the two World Wars, published extensively in Solanum (see complete bibliography in Weber 1928) but never constructed a classification of the entire genus. In the mid-20^th century, Seithe (1962) compiled Bitter’s voluminous work on Solanum classification in a comprehensive system based largely on hair types; she recognised two major divisions, “chorus subgenerum Solanum (L.) Seithe”, the simple and branched hair solanums, and “chorus subgenerum Stellatipilum Seithe”, the stellate haired solanums. Danert (1967, 1970) used his own research on growth form and branching patterns in Solanaceae to propose a new subgeneric classification system for Solanum; he largely kept the same groupings as Seithe (1962) but recognised them at different ranks. D’Arcy (1972) assembled all the subgeneric names in Solanum and provided lectotypes for those that remained untypified. He divided the genus (in what he called a “provisional conspectus”) into seven subgenera (Solanum L., Archaesolanum Marzell, Bassovia (Aubl.) Bitter, Brevantherum (Seithe) D’Arcy (=Minon Raf.), Lyciosolanum Bitter, Potatoe (G.Don) D’Arcy and Leptostemonum Bitter [cited by D’Arcy as Leptostemonum (Dunal) Bitter]) and 52 sections based on combinations of characters he felt were more realistic than those used in the past. His subgenus Leptostemonum is equivalent to that of previous workers, but his divisions of the “non-spiny solanums” were quite different. Like Dunal (1852), but not Seithe (1962), D’Arcy excluded Lycopersicon and Cyphomandra from Solanum. Nee (1999) used only the New World species of Solanum to construct a simplified system in which he recognised 3 subgenera – Bassovia (comprised of the species previously recognised as Cyphomandra and their relatives, in which he included section Pteroidea, see Knapp and Helgason 1997), Solanum (the rest of the “non-spiny” solanums) and Leptostemonum (the “spiny” solanums). He recognised 21 sections with numerous subsections, and listed component taxa in each. Since the largest species diversity of Solanum occurs in the New World, Nee’s (1999) system has been very useful for focusing in-depth taxonomic studies on smaller, putatively monophyletic groups. Child and Lester (2001) provided a synopsis of infrageneric classification of Solanum, based largely on the work of Bitter and Seithe (1962). They did not indicate component species of any of their recognised sections, which comprise fewer species than those of Nee (1999).

DNA sequence data has revolutionised hypotheses of angiosperm relationships (APG III 2009) and Solanum is no exception. Based on cladistic analyses of DNA sequence data Solanum can be divided into 13 well-supported monophyletic clades (Bohs 2005; Weese and Bohs 2007; see Table 1), the largest of which is the group commonly known as the spiny solanums (the Leptostemonum clade), which is largely coincident with Dunal’s (1852) “Leptostemonum”. The largest non-spiny solanum clades (see Table 1) are the Geminata clade, the Potato clade, the Brevantherum clade and the Morelloid/Dulcamaroid clade; this last comprises all the herbaceous species previously placed in section Solanum (see Weese and Bohs 2007) and allied groups, plus the species previously placed in sections Jasminosolanum Bitter, Dulcamara (Moench) Dumort., Lysiphellos Bitter, Nitidum A.Child and Andropedas Rusby plus a variety of other taxa (see Bohs 2005 for discussion). Most current taxonomic work in Solanum is being undertaken in this cladistic framework underpinned by molecular data and concomitantly, more taxa are being added to molecular analyses to test its stability and robustness (e.g., Stern et al. 2011). As more taxa have been added the thirteen clades have generally been supported, but many species have not yet been included in molecular analyeses and their relationships remain to be tested.

The Dulcamaroid clade as recognised by Bohs (2005) and treated here is comprised of elements from several previously recognised subgenera and sections of Solanum. Species of the group are morphologically quite variable (see Morphology below); this in part accounts for these species’ chequered classification history and extensive synonymy. The common European woody nightshade or bittersweet, Solanum dulcamara L., has a long history of use in local medicine (Dunal 1813) in Eurasia (see species treatment of Solanum dulcamara). It was the only member of this group described by Linnaeus (1753), who recognised its extreme variability but did not explicitly name any of these variants, although he did explicitly name variants of other highly variable taxa. He did differentiate an African polynomial taken from Dillenius (1732) as “ß”. This polynomial refers to the South African species Solanum africanum Mill. (long known as Solanum quadrangulare Thunb.), provisionally a member of the African Non-Spiny clade (sensu Bohs 2005), further underlining the morphological similarities between the Dulcamaroid clade as here recognised and the unrelated “African Non-Spiny” (ANS) clade (see below).

Dunal (1813) reviewed the medicinal history of Solanum dulcamara in his thesis, and included it in his Inermia in a group characterised as “Foliis lobatis….” together with Solanum lyratum Thunb. and Solanum triquetrum Cav. In another group – “Foliis integerrimus…” he included an additional eight species now included in the Dulcamaroid clade as defined here; Solanum dichotomum Lour. (=Solanum lyratum) was put into a subgroup with lateral inflorescences, while the rest were said to have terminal inflorescences. Dunal had only seen herbarium specimens or live plants of three of these taxa (Solanum dulcamara, Solanum triquetrum and Solanum pubigerum Dunal) and relied on published descriptions for the rest. The greatly increased number of species (19 names recognised as members of the Dulcamaroid clade as treated here) included in the updated synopsis (Dunal 1816) were scattered over three groups based on leaf division. Solanum seaforthianum Andrews was part of Dunal’s (1816) group containing members of section Pteroidea (sensu Knapp and Helagson 1997, part of the Potato clade, Bohs 2005), Solanum cyrrhosum Dunal (=Solanum seaforthianum) was included in a group with Solanum dulcamara and Solanum lyratum; and Solanum nitidum Ruiz & Pav. (together with other members of the Solanum nitidum species group sensu Knapp 1989) with various members of the Geminata clade (sensu Knapp 2002a, 2008a) such as Solanum sessile Ruiz & Pav. and Solanum oblongifolium Dunal. Dunal’s division of Solanum into groups was quite detailed, but definition of the groups themselves was not completely parallel (see Knapp 1983 for a discussion of sectional nomenclature in Solanum). It is clear that the species now recognised here as members of the Dulcamaroid clade were in different groups due to their highly variable leaf morphology that can be pinnatifid to simple on a single stem (see below). Dunal’s (1852) Prodromus treatment included many more species of the Dulcamaroid clade as defined here; he mostly treated these taxa as members of his subsections Dulcamara (in the groups Dulcamara and Subdulcamara) and Micranthes (in the group Oppositifolia * Indubitaria). He included Solanum viscosissimum Sendtn. in his section Tuberarium (group Potatoe with the potatoes and tomatoes), based on its deeply pinnatifid leaves. Difficulty in grouping these taxa which today we recognise as closely related is perhaps understandable as label data from 19^th century herbarium specimens is sparse and usually does not include any mention of habit. Dependence upon leaf morphology and inflorescence position to determine relationships in Dunal’s systems would easily be foiled by the extremes of variability in leaf shape and inflorescence size found in the Dulcamaroid clade.

Seithe (1962) examined 36 species (representing 24 of the species recognised here and 12 species relegated to synonymy here) of the Dulcamaroid clade in her analyses of hair types in the genus. Of these, 11 were classified only as members of subgenus Solanum without assignment to a section; Seithe was unsure as to the affinities of these taxa. The rest of the species were included in her sections Dulcamara (e.g., Solanum dulcamara, Solanum lyratum), Lysiphellos (e.g., Solanum decorticans Sendtn. [=Solanum inodorum Vell.]), Jasminosolanum (e.g., Solanum seaforthianum, Solanum jasminoides Paxt. [=Solanum laxum Spreng.]), Anthoresis (e.g., Solanum aureum Dunal, Solanum storkii C.V.Morton & Standl.) and Pseudocapsicum (e.g., Solanum triquetrum). Those taxa included in Anthoresis are all the members of the Solanum nitidum species group (sensu Knapp 1989) and were grouped together based on the possession of highly branched trichomes. Danert (1970) recognised the same set of sections (renaming section Anthoresis as “section Holophyllum Walp.”), but did not list any of the component species in his system, so understanding his concepts of specific relationships is difficult. He did suggest a close relationship between sections Dulcamara, Jasminosolanum and Aculeigerum Seithe (equivalent in part to the Wendlandii/Allophyllum clade of Bohs 2005) and the vining taxa from Africa (recognised as sections Afrosolanum Bitter, Lemurisolanum Bitter, Benderianum Bitter, Macronesiotes Bitter, and Quadrangulare Bitter). He pointed out the extreme morphological similarity between members of his section Holophyllum and section Brevantherum Seithe; species in both these groups have dichasial branching, but they differ in their trichome morphology. Danert (1970), despite his emphasis on architecture in classification, put more weight on the trichome characters and agreed with Seithe (1962) that his sections “Holophyllum” (the same as Seithe’s Anthoresis) and Brevantherum, while similar morphologically, were not close evolutionarily.

D’Arcy (1972) included most of the vining dulcamaroid species in his subgenus Potatoe (G.Don) D’Arcy, along with the potatoes (section Petota Dumort.) and section Regmandra (Dunal) Ugent, in sections Dulcamara and Jasminosolanum (typified with Solanum dulcamara and Solanum jasminoides [=Solanum laxum] respectively). This placement was based on their possession of a scandent habit, pinnate leaves and articulation of pedicels above the base (D’Arcy 1972). Species such as Solanum pulverulentum Pers. (=Solanum cutervanum Zahlbr.) were included in his subgenus Brevantherum (=subgenus Minon) in section Holophylla (G.Don) Walp. due to their possession of complex branched trichomes (D’Arcy 1972). Solanum decorticans (=Solanum inodorum) was segregated as section Lysiphellos of subgenus Solanum. He included all of the African vining species (which all have simple leaves) in subgenus Solanum as Bitter’s original sections (sensu Seithe 1962). D’Arcy’s separation of the members of the Dulcamaroid clade in two different subgenera was the first explicit hypothesis as to the affinities of these taxa on a higher level.

Knapp (1989) studied the members of section Holophylla and recognised that the section was morphologically heterogeneous and almost certainly not monophyletic. She segregated a group of 11 species, the Solanum nitidum species group, and identified the unusual pedicel insertion morphology (see below) as a morphological synapomorphy for the group. Her cladistic treatment of the Solanum nitidum species group based on morphology was one of the first explicitly phylogenetic analyses of an entire clade in Solanum. Solanum pubigerum Dunal and Solanum aligerum Schltdl. were suggested as sister taxa to the Solanum nitidum group, but Knapp (1989) suggested the entire group was more closely related to other woody tree solanums such as the members of section Geminata (G.Don) Walp. (the Geminata clade, see Knapp 2002a, 2008a) rather than to species associated with Solanum dulcamara and its relatives. Child (1998) established subsection Nitidum A.Child based on this monograph (Knapp 1989), but did not suggest further relationships or examine material in detail. Child (1998) also segregated the Californian species Solanum xanti A.Gray (=Solanum umbelliferum Eschsch.) and its relatives as section Californisolanum A.Child and followed D’Arcy (1972) in suggesting these species were members of subgenus Potatoe and related to “dulcamaroid taxa”.

In his treatment Solanum in the New World Nee (1999) included the species of the Dulcamaroid clade in sections Dulcamara (the majority of species included here) and Holophylla (members of the Solanum nitidum species group sensu Knapp 1989). His circumscription of section Holophylla also includes all of the members of the Geminata clade as treated by Knapp (2008). All the vining species (plus other species previously recognised as section Parasolanum Bitter, e.g. Solanum corymbosum Jacq.) were combined into a single section Dulcamara by Nee (1999); included in his circumscription are the African Solanum terminale Forssk. and its relatives, which in molecular analyses form the distinct African Non-Spiny clade (Bohs 2005). Although the two groups are morphologically quite similar, in that both contain species that climb with the aid of petioles (see below), sequence differences thus far clearly differentiate them into two clades. Solanum corymbosum and other members of section Parasolanum group are resolved unambiguously as part of the Morelloid clade in all molecular analyses (Bohs 2005; Weese and Bohs 2007).

Two years after the publication of Nee’s (1999) system for New World taxa, Child and Lester (2001) provided a synopsis of infrageneric taxa in Solanum worldwide. This system is an attempt to bring together information from the scattered publications on Solanum taxonomy at the time, rather than a comprehensive reanalysis of infrageneric classification based on new data. Child and Lester (2001) were inspired by Bitter’s many works (see above) and their system is in large part based on his divisions of the genus. No component species in any groups are listed, and although “type” species are given, these are not validly published. They treat members of the Dulcamaroid clade (as far as one can tell from the designations of “types”) in their subgenus Solanum, section Holophyllum (as subsection Nitidum and the “species group Solanum valdiviense sensu A. Child”) and subgenus Potatoe (as “Dulcamaroid taxa”, sections Dulcamara, Jasminosolanum and Californisolanum). Following D’Arcy (1972), they treated all of the African vining species as members of subgenus Solanum.

The unexpected relationship of the Solanum nitidum species group (sensu Knapp 1989) with the vining dulcamaroid solanums was revealed in analyses of DNA sequence data (Bohs and Olmstead 1999; Bohs 2005; Weese and Bohs 2007). The monophyly of the Dulcamaroid clade as recognised here had previously never been suggested. Earlier ideas about the relationships of these groups had always emphasised their habit differences and segregated them clearly, but the presence of a “pedicel sleeve” (sensu Knapp 1989) is a clear synapomorphy uniting these seemingly very different taxa. Knowledge of plants in the field also has helped to find characters uniting these taxa; many of the shrubby taxa are scandent or scrambling, a character only apparent with good field notes or observations. The distant relationship of these taxa and the vining species from Africa and Madagascar (the African Non-Spiny clade sensu Bohs 2005) is somewhat surprising, given that they share the unusual characters of twining petioles and pedicel insertion into a small “sleeve”. Only three species (albeit from three of Bitter’s sections!) from the African Non-Spiny clade have been included in molecular analyses to date and with more in-depth sampling relationships are likely to become clearer. What is clear from analyses using molecular sequence data is that the members of the Dulcamaroid clade are not closely related to the potatoes as had suggested D’Arcy (1972) and Child and Lester (2001). In the analysis using the plastid gene ndhF the sister group to the Dulcamaroid clade is the Morelloid clade, comprising members of section Solanum and relatives (Bohs 2005). A subsequent analysis (Weese and Bohs 2007) using two plastid regions (ndhF and trnT-F) and a nuclear region (waxy, GBSSI or Granule Bound Starch Synthetase) resolved the Dulcamaroid clade (with 93% bootstrap support) as part of a polytomy involving the Morelloid, Normania, Archaesolanum and African Non-Spiny (ANS) clades (see Figure 1 in Weese and Bohs 2007).

Although the structure of relationships within the Dulcamaroid clade is very poorly resolved and the taxa sampled relatively few, a few sister relationships are apparent within the clade. In the ndhF analysis (Bohs 2005) five groups form a polytomy: (Solanum ipomoeoides Chodat & Hassl. [=Solanum uncinellum Lindl.]) + (Solanum calileguae Cabrera+ Solanum jasminoides [=Solanum laxum]) + (Solanum crispum Ruiz & Pav. + Solanum amygdalifolium Steud.) + (Solanum aligerum + Solanum nitidum) + (Solanum dulcamara + (Solanum seaforthianum + Solanum wallacei (A.Gray) Parish)). Weese and Bohs (2007) used fewer species in their three-gene analysis, and recovered Solanum pubigerum + (Solanum nitidum + Solanum aligerum) as sister to a polytomy of (Solanum calileguae + Solanum ipomoeoides [=Solanum uncinellum]) + (Solanum crispum + Solanum amygdalifolium) + (Solanum dulcamara + Solanum seaforthianum). These relationships are consistent with those recovered using only ndhF sequences and support the distinction of a group corresponding to section Holophylla in the narrow sense including Solanum nitidum, Solanum pubigerum and Solanum aligerum, but not Solanum crispum. Relationships amongst the vining dulcamaroid species are less clear, but additional sampling will certainly help with the delimitation of smaller groups within the clade. Most of these relationships are very poorly supported in both analyses, so any sister taxon relationships should be regarded as preliminary.

Habit. Members of the Dulcamaroid clade are all woody plants and vary from shrubs or lax shrubs to vines (see Figure 1). Some species (e.g., Solanum dulcamaroides Poir., Solanum uncinellum) are large canopy lianas, while other vining species are woody only at the base (e.g., Solanum dulcamara, Solanum lyratum), especially in temperate climates. Solanum umbelliferum in California has sprawling herbaceous stems arising from a thick woody rootstock or base. Some species spread by underground stems; Solanum dulcamara is often considered a pest of gardens in Europe for this reason. Many of the species of the group are small shrubs that only become vining as they grow (i.e., Solanum viscosissimum) while others spread over adjacent vegetation, but never truly climb (e.g., Solanum angustifidum Bitter) and others are erect and never twine (e.g., Solanum aligerum, Solanum endoadenium Bitter, Solanum storkii).

Climbing members of the group have petioles that coil around supports (Solanum laxum [as Solanum jasminoides] was characterised as a “leaf-climber” by Darwin 1865) anchoring the ascending stems (Figure 1). This climbing mechanism is not found elsewhere in Solanum except in the African Non-Spiny (ANS) clade, which in molecular phylogenetic analyses is not closely related to the Dulcamaroid clade (Weese and Bohs 2007; T. Särkinen and S. Knapp, unpublished data). Darwin (1865) measured the time taken for petioles to clasp supports and found it very slow compared to other twining plants; once a petiole has clasped a support it accumulates woody tissue and thickens considerably.

Stems. The main axis of the plant is a typical Solanum sympodium formed by a succession of lateral axes with alternate leaves arranged in a 1/3 phyllotaxic spiral, and inflorescences are terminal at the end of each sympodial unit (Danert 1958). Danert (1958, 1967, 1970) and Child and Lester (1991) documented sympodial units and anthoclades (patterns of foliar lateral branches and associated inflorescences) in the Solanaceae. The number of leaves per sympodium is very regular and of major taxonomic significance in the entire genus Solanum (see Knapp 2001 for a further discussion).

In the Dulcamaroid clade sympodial units are almost always plurifoliate with many (or an irregular number) of leaves between each inflorescence in contrast to other groups such as the tomatoes (Peralta et al. 2008) or the Geminata clade (Knapp 2008) in which sympodial units are composed of 3 or fewer leaves. Plurifoliate sympodial units have the leaves arranged in a spiral fashion along the stems, and the leaves in members of the dulcamaroids are never geminate (paired, as seen in the Geminata clade, Knapp 2008). Most members of the group have monochasial branching, with a single axillary branch arising from below the inflorescence, giving the stems a zig-zag appearance (Danert 1958). Members of the Solanum nitidum group are either monochasial or some species have dichasial branching (see Knapp 1989); in species with dichasial branching two axillary shoots develop and the branching is dichotomous (e.g., Solanum stenophyllum Dunal). Most of these species have a mixture of the two branching types but are predominantly one or the other.

Leaves. The leaves of members of the Dulcamaroid clade are extremely variable, both across the group and within individual plants (see Figure 2 and individual species illustrations). Leaves on reproductive stems are either simple or variously pinnatifid/pinnate. There is always a wing of leaf tissue along the midrib between the lobes connecting all the dissections that is more pronounced than that found in the tomatoes and potatoes (see Peralta et al. 2008 for discussion of tomato leaf morphology), but some species such as Solanum angustifidum and Solanum viscosissimum or some forms of Solanum lyratum have such deeply divided leaves that they are best characterised as pinnate as is commonly done in other such taxa (see Peralta et al. 2008).

Leaf polymorphism is common in the group and in many species individual stems bear both simple and pinnatifid leaves (e.g., Solanum dulcamara, Solanum lyratum, Solanum salicifolium Phil.). The control of this has not been investigated, but may have to do with hormonal balance due to light or nutrients. In other species, most notably in some plants of Solanum viscosissimum, lateral inflorescence-bearing shoots have simple leaves while main axis leaves are pinnatifid or pinnate. This variability has lead to the many synonyms in some species; different leaf morphologies collected from different parts of the plant have been described as separate taxa. In some taxa there seems to be some genetic control of leaf morphology; variants with deeply dissected leaves of Solanum lyratum all come from a few localities and may represent local populations with fixed genetic differences. A set of specimens of the otherwise consistently simple-leaved Solanum uncinellum from near Iquitos have deeply dissected leaves but are identical in flower and fruit characteristics. Leaf morphology in pinnately leaved species of Solanum is complex, and governed by a set of interacting genes (Holtan and Hake 2003); this variation can make identification difficult in some species.

Leaf divisions are narrowly elliptic, elliptic to broadly elliptic, or obovate; the divisions rarely are petiolulate. The base is usually asymmetric, and varies from truncate or rounded, to cordate or lyrate. The apex (of leaves and leaflets) is rounded, acute or acuminate. The margins are entire, never serrate or crenate and the margins are straight or revolute.

The juvenile leaves of most of the species for which they have been observed are usually pinnate (pinnatifid), but because botanists usually only collect reproductive stems, this is not known for many species. Where juvenile leaves are known I have included this information in species descriptions. A sterile specimen with pinnate leaves from Iquitos (Williams 8481, F) was identified by MacBride (1964) as Solanum dulcamara. In my view this is a specimen of the pinnate-leaved seedling of Solanum uncinellum, not a distinct species, nor is it a southern range extension of Solanum dulcamara.

Leaf size and texture can also vary enormously depending on the environment or position on the stem. In Solanum triquetrum, which grows in arid environments in the southwestern USA, leaves formed in wet periods are larger, sometimes by several times, than those from dry periods (see Figure 97). Leaf texture varies across the group from very delicate and membranous in some temperate species (e.g., Solanum lyratum) to thick and coriaceous in taxa from high elevations (e.g., Solanum macbridei Hunz. & Lallana). Turreson (1922) grew thick-leaved plants of Solanum dulcamara from seaside habitats in Sweden in common inland gardens with thin-leaved plants from more inland forest areas and found that the coastal forms developed thinner (and less pubescent) leaves away from their natural habitat. Clough et al. (1979), also working with Solanum dulcamara, found that light levels were a primary factor influencing leaf morphology and physiological performance; a major effect was on leaf specific weight, or thickness. Leaf texture and morphology are both important species specific characters in the group, although within species variation in some taxa is very large.

Pubescence. Trichome types and density are very useful for species recognition in Solanum. Previous classifications of the genus (Seithe 1962) have been constructed using primarily trichome types, with groups defined by possession of stellate, branched or simple trichomes. Later research (e.g., Knapp 2002a; Bohs 2005) has shown that trichome types do not define monophyletic groups in Solanum, but that they can be useful at the species level. As with many of the vegetative features of members of the Dulcamaroid clade, trichomes are very variable (see Figure 3 for trichome types). Trichomes of members of the Dulcamaroid clade are either uniseriate (consisting of a single file of cells) and either simple or branched, or multiseriate and echinoid and densely dendritic (Roe 1971; Knapp 1989; Mentz and Stehmann 1992). Species of the Solanum nitidum group (sensu Knapp 1989) have branched or densely branched trichomes; some of these approach the echinoid and tree-like trichomes characteristic of the Brevantherum clade (see Roe 1971) with numerous very short branches clustered at the tip (e.g., Solanum stenophyllum and Solanum storkii).

Simple trichomes are either eglandular or glandular, with the glands usually consisting of a single cell. Solanum kulliwaita S.Knapp has unusual bead-like terminal cells on the trichomes of the inflorescence that all appear to be glandular (Figure 52); this type of trichome has also been observed in Nicotiana L. in Australia (Marks et al. 2011). Glandular trichomes are very consistently present in some species (e.g., Solanum lyratum, Solanum viscosissimum), while in others possession of glandular or eglandular trichomes appears to be polymorphic within species (e.g., Solanum endoadenium, Solanum umbelliferum). Possession of glands has often been used as an infraspecific marker in other groups of Solanum (for example some species of section Solanum, Edmonds 1982) but molecular work with these taxa has shown that the trichome types do not define monophyletic groups within these species but instead are polyphyletic and represent polymorphism (Oyet 2004; Manoko 2007).

Terminal glands also occur on branched trichomes, but these trichomes are more typically eglandular. Several different branched trichome types occur in the Dulcamaroid clade: 1) uniseriate trichomes with few branches (e.g., Solanum sanchez-vegae S.Knapp, Figure 84), 2) uniseriate trichomes with many congested branches (e.g., Solanum aureum, Figure 20), 3) multiseriate or uniseriate trichomes with densely congested very short branches (tree-like or “Tannenbaumartig” sensu Seithe 1962, e.g., Solanum stenophyllum, Figure 94) and 4) multiseriate echinoid trichomes (e.g., Solanum storkii, Figure 95). These latter two types have very short, often unicellular branches and are found only in the Solanum nitidum group.

Inflorescences. The inflorescences of members of the Dulcamaroid clade, as in most Solanaceae, are terminal with a subtending lateral bud. The inflorescence usually occupies a position at the end of branches, but if growth continues from the axillary bud it can appear lateral. In some species the inflorescences are borne on what appear to be short shoots with very reduced leaves (e.g., Solanum inodorum, Solanum valdiviense Dunal); the flowering and non-flowering shoots of these plants are very divergent morphologically and the short shoots often have smaller leaves (see Figures 49, 104). Solanum inodorum is extreme in this regard with the flowering shoots seeming to be axillary in larger leaves. In fact, the inflorescence is terminal on a short shoot with much reduced, almost bract-like, leaves.

The basic inflorescence, as in all other species of Solanum, is a scorpoid cyme with a variety of branching patterns. In the Dulcamaroid clade inflorescences are usually very large and highly branched. Turrill (1935) made crosses between a form of Solanum dulcamara with 1-flowered inflorescences he called “uniflora” and the more typical many-branched form, and suggested that the branches of the complex inflorescence were dichotomous and that subtending bracts were suppressed. Lippman et al. (2008) found that the highly branched inflorescence of Solanum crispum was developmentally similar to compound inflorescence (s) mutants of tomato; each shoot inflorescence meristem produces multiple flowers in a proliferating manner. They suggest that delays in floral termination (perhaps mediated by S, or the genetic pathway that S defines) coupled with sequential expression of AN (anantha, another tomato inflorescence mutant) provide a developmental framework for the modulation of sympodial branching in the entire family. In the Dulcamaroid clade flower number varies from 2-3 (Solanum salicifolium, Figure 82) to more than 100 (Solanum sanchez-vegae, Figure 84); within a plant inflorescence size can vary greatly with inflorescence age as each inflorescence continues to branch and produce flowers throughout its “life”.

In any given inflorescence only a few flowers are open at a time (up to about 40-50 in species with large inflorescences like Solanum sanchez-vegae). Flowers are generally more congested near the inflorescence branch tips; in Solanum aligerum and Solanum pubigerum the flowers are grouped onto tiny “platforms”, or highly congested groups of 3-5 flowers (see Knapp 1989: Figure 2b; Figures 10, 74). The pedicel scars are irregularly and relatively widely spaced, and the space between scars is generally greater more basally in the inflorescence. Differentiation between the peduncle and the end of the shoot is difficult to determine in most species, but in some species (e.g., Solanum kulliwaita), the pubescence of the inflorescence is distinct from that of the shoot.

Pedicels. The pedicels of species in the Dulcamaroid clade are distinctive in being inserted into a tiny sleeve of tissue that appears to arise from the inflorescence rhachis (see Figure 2c in Knapp 1989); this is one of the few morphological synapomorphies for the clade. This is most pronounced in the species of the Solanum nitidum group, in which the sleeve is sometimes up to 2 mm long and forms a small cup in which the base of the pedicel sits (Figure 2C). In most species the pedicel is articulated slightly above the base, leaving a small peg or cup on the inflorescence axis; this accentuates the zig-zag appearance of the inflorescence architecture that results from sympodial growth. The pedicel articulation in Solanum boldoense Dunal & A.DC. of Cuba is unusual in being in the distal third of the pedicel, often directly beneath the calyx tube (see Figure 22). The colour of the pedicel sleeve is sometimes (e.g., Solanum dulcamara) different from the pedicel (see Figure 4D), further supporting the idea that the sleeve is anatomically part of the inflorescence rhachis.

Pedicel orientation in the species of the group is generally deflexed or slightly nodding at anthesis. In fruit, those species with pendent inflorescences generally have deflexed pedicels, but shrubby species such as those of the Solanum nitidum group have erect pedicels in fruit (e.g., Solanum coalitum S.Knapp).

Calyces. The calyx is typically sympetalous and 5-merous with the tube and lobes more or less equal in size. The shape of the calyx lobes varies considerably and is a useful character for species identification. Lobes vary from mere undulations of the calyx rim (e.g., Solanum boldoense, Solanum uncinellum) to deltate (many species), quadrate (e.g., Solanum aligerum, Solanum pyrifolium Lam.) or lanceolate (e.g. Solanum imbaburense S.Knapp); apices are usually acute to acuminate, but are occasionally rounded (e.g., Solanum angustifidum) or elongate (e.g., Solanum viscosissimum). Pubescence of the calyx tube and lobes usually parallels that of the pedicels and inflorescence rhachis, but is generally sparser.

Corollas. The corolla of members of the Dulcamaroid clade is sympetalous, 5-merous and regular, in common with the vast majority of Solanum species. Color is either white or varying shades of purple; many species have populations of both color forms, and in Solanum dulcamara these have been described many times at the infraspecific level. Solanum laxum growing in full sun often has pale violet rather than the more commonly encountered white flowers, but the degree to which flower color is environmentally or genetically determined is not clear. At the base of the corolla tube is often a ring or irregular area of differently colored tissue refered to as the eye. In the group of species found in the temperate zone (e.g., Solanum dulcamara, Solanum lyratum, Solanum umbelliferum) at the base of each corolla lobe is a small pair of shiny green dots often ringed with darker green or white tissue. These have been variously interpreted as “pseudo-nectaries” functioning in pollinator attraction (Buchmann et al. 1977), but specific field studies have not been undertaken to test this. This character can be difficult to see in herbarium specimens. The center part of Solanum flowers is usually UV absorbent and seen by bees as a dark guide to the prominent anther cone (Buchmann et al. 1977). Solanum endoadenium has a prominent green eye, but the tissue is not shiny, perhaps indicating it is not homologous with the similar structure in Solanum dulcamara (see Figure 5).

Corolla shape varies from deeply stellate (Solanum uncinellum, Figures 5, 101) to pentagonal (Solanum wallacei, Figure 108). The lobes are triangular and usually either spreading (Solanum dulcamaroides, Figures 5, 38) or strongly reflexed (Solanum lyratum, Figures 5, 61) at anthesis. Solanum macbridei is unusual in having campanulate corollas (Figure 63). The tips are usually somewhat cucullate (e.g., Figure 14D) and are papillose to densely pubescent. In those species with dense overall pubescence, the abaxial surface of the corolla lobes is also usually densely pubescent. Corolla diameter varies from approximately 1 cm (Solanum endoadenium) to almost 5 cm (Solanum wallacei); most species have corollas that range from 1.5–3 cm in diameter.

Anthers. Anthers of members of the Dulcamaroid clade conform to the typical poricidal morphology of all other species of “non-spiny” Solanum (see Knapp 2001, 2002a). In most taxa the pore usually “unzips” during anther dehiscence to form a tear-drop shaped slit (Figure 5 and species illustrations), from which pollen is shed during vibratile pollination (Buchmann 1986). Some species of the Dulcamaroid clade have pores that do not elongate at all (e.g., Solanum dulcamaroides, Solanum seaforthianum), or only elongate slightly with age (e.g., Solanum valdiviense). Anthers in the group are generally ellipsoid, but those of Solanum dulcamara and Solanum uncinellum are elongate and tapering and those of Solanum dulcamaroides and Solanum boldoense are almost spheroidal. Anthers of members of most species are loosely connivent, while those of Solanum dulcamara are tightly connivent, and are held together physically with a sticky “glue” (Glover et al. 2004) such that the pores act as a single opening to a visiting bee. The anthers of Solanum aureum occasionally appear pubescent due to the prescence of elongate papillae on the dorsal surface (see Figure 20).

All five anthers in most of the species of the group are morphologically identical; some taxa, however, have filaments of different length, making the flowers zygomorphic. This filament length difference ranges from quite subtle (Solanum seaforthianum, Figure 86) to very conspicuous (Solanum uncinellum, Figure 5F, 101). Flowers of Solanum flaccidum Vell. that have just opened have filaments of equal length; the long filament apparently elongates over the course of anthesis. This has been observed in other species of Solanum (Solanum turneroides Bitter and Solanum evolvuloides Giacomin & Stehmann, members of section Gonatotrichum Bitter of the Brevantherum clade, see Stern et al. in press) and in the genus Lycianthes Hassl. (Dean 2001).

Pollen of members of the group is similar in size and shape to that of other species of Solanum (Buchmann 1986); it is tricolporate with a relatively smooth exine and held in yellow “pollen shamming” anthers. Some plants of Solanum lyratum have dark purple anthers; this appears not to be related to light or exposure but may be genetic. I have not seen dark anthers in any other member of the group, but anther color is variable in the Cyphomandra clade (section Pachyphylla Dunal, Bohs 2004). Nitrogen and protein content of pollen grains is the same as for other species of Solanum (Buchmann 1986), and is typical for a buzz-pollinated plant.

Gynoecium. The gynoecium is typically bicarpellate; the carpels are fused in a superior ovary with axillary placentation. The ovary is usually conical to globose or slightly ellipsoid and usually glabrous; if pubescent; the trichomes are only sparsely present at the apex (e.g., Solanum cutervanum). The flowers lack nectaries, as do all other species of Solanum. The style is simple, straight or slightly curved, glabrous or variously pubescent, and is exserted beyond the anthers in all but Solanum luculentum S.Knapp (see below). The stigma is capitate (e.g., Solanum valdiviense, Solanum umbelliferum) to clavate (e.g., Solanum amygdalifolium) and is sometimes distinctly bilobed (e.g., Solanum uncinellum). The ovules are anatropous and non-arillate.

Fruits. As with all species of Solanum, the fruit is a bicarpellate berry. Fruits of members of this group are usually brightly colored and juicy (see Figure 6 for respresentative photographs). Many species change fruit color through maturation; for example, Solanum nitidum has green immature fruits that pass through a stage of being bright red before turning dark purple at full maturity. This makes it difficult to use fruit color as an identification aid, and in Solanum uncinellum fruit color has been variously recorded on labels as blue, dark purple, red and white. Without in-depth field studies it is unclear whether these represent populational or regional differences or maturational changes. Species here regarded as synonyms of Solanum umbelliferum in California have sometimes been distinguished using “mature” fruit color (either green or black; Nee 1993b), but such color polymorphisms are found in other Solanum species (most notably Solanum nigrum L. in Europe, see Manoko 2007). Fruit surface morphology is usually shiny, but in some species the pericarp is distinctly dull and non-shiny (e.g., Solanum sanchez-vegae). Solanum endoadenium has bright orange fruits that appear to dry on the plant and release the seeds (see Figure 6D); this has not been studied in detail. Fruit diameter varies from less than 1 cm (Solanum angustifidum, Solanum viscosissimum, Solanum pittosporifolium Hemsl.) to approximately 4 cm (Solanum wallacei).

Seeds. Seed morphology has proved very useful in Solanum taxonomy. Enzymatic digestion of the outer testal walls reveals great variety in the morphology and structure of the lateral testal cell walls, varying both between and within groups (Whalen 1979; Lester and Durrands 1984; Knapp and Helgason 1997). Souèges (1907) provided an early study of Solanum seeds, including Solanum dulcamara. He described the development of the integument and recognised outer and inner layers of the seeds. Seeds of members of the Dulcamaroid clade are oval, obovate or kidney-shaped in outline and flattened laterally. The cells of the outer epidermal layer in some species (e.g., Solanum dulcamaroides, Solanum lyratum, Solanum uncinellum, Solanum umbelliferum) develop radial wall thickenings that form as “hair-like outgrowths” or “pseudohairs” in mature seeds (Lester and Durrands 1984; Lester 1991). These hair-like outgrowths often greatly enlarge the outer layer of the integument and the seed coat appears pubescent; seed measurements here include these projections. In seeds from mature fruits the pseudohairs are translucent, connate or fused laterally to each other and tightly adpressed to the epidermis giving a silky appearance to the seed surface, or if long and distinct produce a hairy and shaggy seed surface. In some species (e.g., Solanum dulcamaroides) the pseudohairs form a prominent wing around the margin of the seed. The testal cells form a reticulate or honeycomb pattern with cell outlines at the basal portions deeply sinuous and irregular (e.g., Solanum boldoense), pentagonal/rectangular (e.g., Solanum seaforthianum, Solanum pubigerum) or somewhat intermediate (e.g., Solanum triquetrum).

Seed number per berry varies from few (fewer than 10, Solanum leiophyllum Benth., Solanum triquetrum) to many (more than 100, Solanum wallacei), but most species fall in the range of 20-50 seeds per berry. Seed size varies from 1.5 mm length and 1 mm width to 8 mm length and 6 mm width. Color varies from yellow or pale brown to dark brown.

Habitats and distribution. Members of the Dulcamaroid clade are both tropical and temperate, and occur in a wide range of habitats from deserts to lowland rainforest. Taxa occurring in the temperate areas of Asia and Europe occupy a wide range of habitats and elevations; Solanum dulcamara, for example, is found on sea coasts and extends into high elevations in the mountains of Italy and the Alps. Genetic work done in the early 20^th century (Turesson 1922) revealed some hereditary component to the immense morphological variation in this species, but habitat is also important (Turrill 1935). Solanum lyratum has been collected from sea level to more than 2000 m and is found in many forest types and even in weedy gardens in cities. Most of the species of the group are not of conservation concern (see Table 3 and species treatments) but some have narrow distributions and are assessed as endangered or vulnerable.

Most species in the group are South American, and in the Americas the species of the group range from southern Oregon (USA) to central Argentina and onto the islands of the Caribbean. In South America highest species diversity occurs in two areas, the forests of Atlantic coastal Brazil and the eastern slope of the Andes from Colombia to Argentina (see Table 2). This circum-Amazonian pattern of species richness is common in other groups of Solanum (Knapp 2002b); endemism follows the same pattern. Argentina has the highest species diversity in the Dulcamaroid clade (with Peru and Ecuador very close behind), due in part to the mixing of species from these two centres of species richness, in addition to several endemics or near endemics along the Andean flanks.

Floral biology and pollination. Most Solanum species have hermaphroditic flowers, but in some groups derived sexual systems are common (Whalen and Costich 1986; Anderson and Symon 1989; Knapp et al. 1998; Martine et al. 2009). Andromonoecy is very common in the Leptostemonum clade (Whalen and Costich 1986; Levin et al. 2006), as is dioecy (Anderson and Symon 1989; Martine et al. 2009). Dioecy is phylogenetically widespread in Solanum, with dioecious species occurring primarily in the spiny solanums (subgenus Leptostemonum Bitter, Anderson and Symon 1989), but also in the non-spiny solanums in the Potato (Anderson and Levine 1982), the Geminata (Knapp et al. 1998) and Brevantherum (S. Knapp, pers. obs., Solanum punctulatum Dunal from Jamaica) clades. Solanum luculentum of the Dulcamaroid clade has all the morphological correlates of a dioecious species, but has not yet been examined in detail to determine whether or not it is indeed truly dioecious. If so, Solanum luculentum represents the independent evolution of dioecy in an otherwise completely hermaphroditic group, rather than within a predominantly andromonoecious clade, as is common in the spiny solanums (Martine et al. 2009). The evolution of dioecy in a group with strictly hermaphroditic flowers has also occurred in the Potato clade (e.g., Solanum appendiculatum Dunal, see Anderson and Levine 1982).

Self-compatibility (SC) is widespread in Solanum (Whalen and Anderson 1981) and is essentially irreversible once acquired from the ancestral self-incompatible (SI) state ( 2003, 2006). Only three species of the Dulcamaroid clade have been assessed for their compatibility status: Solanum dulcamara, Solanum laxum and Solanum wallacei (Goldberg et al. 2010). Of these, only Solanum laxum has been shown to be SI (Whalen and Anderson 1981), but Solanum dulcamara flowers failed to set seed when isolated from visiting insects (Turrill 1935) and Cruden and Lyon (1985) scored Solanum dulcamara from Iowa as SI (“xenogamous”). Other authors, however, have stated that Solanum dulcamara is SC (Eijlander and Stiekema 1995; Vallejo-Marín and O’Brien 2006). In greenhouses and field plots in Nijmegen (Netherlands) self-crosses of Solanum dulcamara proved difficult and set few seed (G. van der Weerden, pers. comm., October 2010). There may be a geographic component to self-compatibility in Solanum dulcamara; those reporting SC have used American collections (naturalised) and those reporting SI have used European accessions. Buchmann et al. (1977) report that bagged flowers of Solanum umbelliferum (as Solanum xanti) failed to set fruit and concluded it was SI; this species, however, has not been included in recent large scale analyses of SI in the family (Goldberg et al. 2010). Solanum crispum grown in European gardens rarely sets fruit, and attempts to self flowers resulted in 0% fruit set (pers. obs.), but these experiments were not done in a controlled manner. That Solanum wallacei, an island endemic with a very narrow range, is SC is not at all surprising; it will be interesting to assess again the compatibility status of its close relative Solanum umbelliferum, which has been thought to be SI (Buchmann et al. 1977). Goldberg et al. (2010) have shown that lineages with self-incompatibility in the Solanaceae are more species-rich, due in part to higher extinction rates in SC lineages. It appears from the limited data available in the Dulcamaroid clade that widespread, clonal species (e.g., Solanum laxum, Solanum dulcamara, Solanum crispum) are SI, supporting the results of Vallejo-Marín and O’Brien (2006).

Flowers of all members of the Dulcamaroid clade are typical for Solanum and are buzz-pollinated by a wide variety of bees. In temperate regions species are visited by various species of bumblebees (Bombus); Solanum dulcamara and cultivated Solanum laxum and Solanum crispum in Europe are both visited by several species of bumblebees (pers. obs.). I have also observed bumblebees visiting several of the montane tropical species of the group (e.g., Solanum nitidum in Peru and Bolivia, Solanum storkii in Costa Rica), and Bombus species were common vibratile pollinators of Solanum umbelliferum (as Solanum xanti) in southern California (Buchmann et al. 1977). Pollen grains left on the anthers or near the pores by vibratile bees were collected by the gleaning bees from Solanum umbelliferum (Andrena spp., Andrenidae; Stenodynerus cochisensis (Viereck), Eumenidae; Buchmann et al. 1977). These bees are not as efficient at effecting pollination, however, as they are not necessarily in contact with the stigma on every visit.

My goal for this revision has been to provide species circumscriptions for the members of this large and morphologically variable clade, while clearly highlighting those taxa where further in-depth research would be useful. Delimitation of species here basically follows what is known as the “morphological cluster” species concept (Mallet 1995): i.e., “assemblages of individuals with morphological features in common and separate from other such assemblages by correlated morphological discontinuities in a number of features” (Davis and Heywood 1963). Biological (Mayr 1982), phylogenetic (Cracraft 1989) and the host of other finely defined species concepts (see Mallet 1995) are almost impossible to apply in practice when dealing with large, complex groups such as the clades of Solanum, and are therefore of little utility in a practical sense. It is important, however, to clearly state the criteria for the delimitation of species, rather than dogmatically follow particular ideological lines (see Luckow 1995; Davis 1997). My decisions relied on clear morphological discontinuities to define the easily distinguished species. Specific characters used for recognition are detailed with each species description and in the keys. Potential reasons for variability and intergradation are recent divergence and hybridization. In this revision I have tried to emphasise similarities between populations instead of differences, which so often reflect incomplete collecting or local variation. I have not recognised subspecies or varieties, as I do not feel these are useful categories, either in a taxonomic or evolutionary sense. The variation is better described and documented, rather than formalised with a name which then encumbers the literature. I have been conservative in my approach, recognising as distinct entities those population systems (sets of specimens) that differ in several morphological characteristics. Minor differences in morphology, distribution, habitat, and ecology are important in some groups, where the common groundplan for the species is very similar. On the other hand, I have recognised several extremely widespread, polymorphic species, e.g., Solanum dulcamara, Solanum lyratum, Solanum umbelliferum, Solanum uncinellum and Solanum viscosissimum. In these species variation exists in certain characters, but the pattern of variation is such that no reliable units can be consistently extracted. In these cases the variability within and between populations seems more important than the variations of the extremes other taxonomists have recognised as distinct. In these cases I have described the variation, realizing that others may wish to interpret it differently.

Approximately 9000 collections were examined for this monograph from herbaria worldwide; specimens were seen at the following herbaria (all abbreviations from Index Herbariorum, http://sciweb.nybg.org/science2/IndexHerbariorum.asp ): A, AAU, B, B-W, BH, BM, BOLV, BRI, C, CAS, CICY, COL, CORD, DS, E, EA, ECON, EIU, F, FCQ, G, G-DC, GH, GOET, HA, HIB, HITBC, HBR, HUA, JBSD, JEPS, K, L, LAGU, LE, LIL, LOJA, LPB, MA, MBM, MEXU, MO, MOL, NY, OXF, P, PE, PMA, POM, Q, QCA, QCNE, RSA, S, SGO, SI, TI, UC, US, USM, UT, W, WIS, WU. Many of the species of the Dulcamaroid clade are common and widespread (e.g., Solanum dulcamara, Solanum laxum, Solanum uncinellum), so only selected specimens are cited; these represent the entire species range. All numbered collections are cited in the Index to Numbered Collections. Any collections without specific numbers are cited as s.n. (sine numero) in the Specimens examined sections of each species. Collection details for all specimens used for this monograph can be found on the Solanaceae Source website (http://www.solanaceaesource.org ) and in the data file available with this paper (see Supplementary Material). The specimens of Solanum at the New York Botanical Garden (NY) were databased as part of the PBI project; those specimens are best accessed through the NY Virtual Herbarium (http://sciweb.nybg.org/science2/VirtualHerbarium.asp ).

The JStor Plant Science website (http://plants.jstor.org ), established to display the images of type specimens digitised as part of the GPI (Global Plants Initiative) project funded by the Andrew W. Mellon Foundation, was an invaluable resource for images of type specimens, and many of the isotypes identified here were discovered via the website. Many type specimens in European herbaria were photographed in the early 20^th century by J.F. Macbride of the Field Museum (http://fieldmuseum.org/explore/our-collections/berlin-negatives ), including those at B, since destroyed. These photographs are cited here with a F negative number (F neg. #), but herbaria in which I have seen the photographs (usually mounted on sheets) are not cited. Conrad V. Morton of the Smithsonian also took photographs of many type and other nomenclaturally important specimens, these are cited as Morton negative numbers.

With all specimens of types I have cited the barcode or accession number of each sheet if possible. Barcodes are cited with no space between the herbarium acronym and the number (e.g., BM000845345), while for accession numbers I have inserted a hyphen between the acronym and the number (e.g., MO-00678987). Some herbaria such as MA, have barcodes and accession numbers that are the same, these are cited as on the barcode label (e.g., MA-634567); other herbaria such as MO have barcodes and accession numbers that differ, here I have cited the accession number as it is the only element visible without a barcode reader.

In the specimen citations I have selected specimens that represent the species range, and present these citations in alphabetical order by country rather than a geographical order. For those taxa that are introduced and often naturalised (Solanum laxum and Solanum seaforthianum) I have separated the specimen citations into those from countries in the native range (or presumed native range) and those from countries where the species is introduced or naturalised.

Citation of literature follows BPH-2 (Bridson 2004) with alterations implemented in IPNI (International Plant Names Index, http://www.ipni.org ) and Harvard University Index of Botanical Publications (http://kiki.huh.harvard.edu/databases/publication_index.html ). In particular, I have used the square bracket convention for publications in which a species is described by one author in a publication edited or compiled by another – the traditional “in” attributions such as Dunal in DC. for those taxa described by Dunal in Candolle’s Prodromus Systematis Naturalis Regni Vegetabilis; this work is cited here as Prodr. [A.P de Candolle] and the names thus attributed only to Dunal. For “ex” attributions I have cited only the publishing author, as suggested in the Code (McNeill et al. 2012). Author names are abbreviated according to IPNI (International Plant Names Index, http://www.ipni.org ).

Each species treated here has been assessed for a preliminary conservation status using the GIS protocols of Moat (2007) that involve calculation of AOO (Area of Occupancy) and EOO (Extent of Occurrence) from georeferenced specimen data. A grid cell size of 10% of the EOO was used (see Moat 2007); additional analyses using the 2 km² grid size as recommended by IUCN (2001) were done, but specimen collection density made the results difficult to interpret. When threat status as determined by EOO and AOO differ, I have been conservative and selected the higher (more threatened) assessment as the overall assessment of the species (for an example, see Solanum alphonsei), and have used my knowledge of the biology of these species to suggest amendments to these calculated values. These assessments are not complete in the sense of IUCN (2001), but are indicative of the threat status for the taxa treated here. Any species known from fewer than 5 collections has been assessed as Data Deficient (DD); this does not mean these are not of conservation concern, but that the available data are not sufficient for their status to be modelled using the AOO/EOO method.