The Morelloid clade, also known as the black nightshades or “Maurella” (Morella), is one of the 10 major clades within Solanum L. The pantropical clade consists of 75 currently recognised non-spiny herbaceous and suffrutescent species with simple or branched hairs with or without glandular tips, with a centre of distribution in the tropical Andes. A secondary centre of diversity is found in Africa, where a set of mainly polyploid taxa occur. A yet smaller set of species is found in Australasia and Europe, including Solanum nigrum L., the type of the genus Solanum. Due to the large number of published synonyms, combined with complex morphological variation, our understanding of species limits and diversity in the Morelloid clade has remained poor despite detailed morphological studies carried out in conjunction with breeding experiments. Here we provide the first taxonomic overview since the 19^th century of the entire group in the Old World, including Africa, Asia, Australia, Europe and islands of the Pacific. Complete synonymy, morphological descriptions, distribution maps and common names and uses are provided for all 19 species occurring outside the Americas (i.e. Africa, Asia, Australia, Europe and islands of the Pacific). We treat 12 species native to the Old World, as well as 7 taxa that are putatively introduced and/or invasive in the region. The current knowledge of the origin of the polyploid species is summarised. A key to all of the species occurring in the Old World is provided, together with line drawings and colour figures to aid identification both in herbaria and in the field. Preliminary conservation assessments are provided for all species.

Africa, Asia, Australia, Eurasia, Europe, Madagascar, Pacific, polyploidy, Solanum nigrum complex, supervegetables, weeds

Solanum L., with approximately 1,400 species, is one of the largest genera of flowering plants (Frodin 2004). The genus poses a taxonomic challenge not only due to its large size, but also due to the 6,931 published names, many of which are associated with the cultivated and widespread species of the genus, including the potato (S. tuberosum L.), tomato (S. lycopersicum L.) and the eggplant (S. melongena L.). Recent taxonomic and molecular systematic efforts (http://www.solanaceaesource.org) have identified major clades within Solanum (e.g. Weese and Bohs 2007), clarified relationships and the monophyly of previously recognised morphological sections (e.g. Stern et al. 2011) and provided taxonomic revisions for major clades with keys for species identification (e.g. Knapp 2013).

The Morelloid clade of Solanum, also known as the black nightshades or “Maurella” (Morella), is amongst the 10 robustly supported major clades within Solanum (Bohs 2005; Weese and Bohs 2007). This group, that includes the type species of the genus, S. nigrum L., has a long taxonomic history dating back to the early herbals (Figures 1–2). The group has traditionally been considered difficult, due in part to the weedy nature of its species and its worldwide distribution. The clade comprises 75 currently recognised non-spiny herbaceous and suffrutescent species with simple or branched hairs with or without glandular tips (glandular or eglandular) and inflorescences usually arising from the internodes (Figure 3; Särkinen et al. 2015a). Ploidy level varies from diploid to hexaploid within the group (e.g. Edmonds and Chweya 1997; Moyetta et al. 2013; see Table 1), in part contributing to the difficulties in its taxonomy. While taxonomic revisions of the smaller New World sections within the Morelloids have recently been published (del Vitto and Petenatti 1999; Barboza 2003; Barboza and Hunziker 2005), the group in its entirety has not been revised since the 19^th century (Dunal 1852).

Figure 1. 

“Solanum Hortense” from Joseph Banks' handcoloured copy of Dioscorides De Materia Medica; the flowers were originally painted white but have turned brown because a lead paint was used. This copy dates from the 15^th century. Copyright The Trustees of the Natural History Museum, London. Reproduced with permission.

Figure 2. 

“Solanum Hortense” as illustrated in early herbals AFuchs (1542)Neu KreuterbuchBMatthioli (1544)Di Pedacio Dioscoride AnazarbeoCDodoens (1563)CrüydeboeckDDodoens (1583)Stirpium historiae pemptades sex. Courtesy of the Library of the Natural History Museum, London.

Figure 3. 

Morphology of the Old World species of the Morelloid clade of Solanum A Many species are cultivated for leaves and fruits (S. scabrum, NIJ994750012) B Most species show preference of disturbed, dry habitats in the wild (S. americanum, Knapp et al. 10210) C Strongly ridged and often purple coloured stems present in many species (S. tarderemotum, NIJ 2010-5) D Glandular hairs present in some species of the Morelloid clade (S. memphiticum, A34750473) E Typical stellate, white flowers with spreading or reflexed corolla lobes (S. tarderemotum, A14750151) F Fleshy round berries that are black, green, yellow, orange or red depending on the species (S. nigrum, A44750150) G Stone cells (also known as sclerotic granules or brachysclerids) are found in the fruits of most species of the Morelloid clade and are round in shape as compared to the tear-drop shaped or ellipsoid seeds (S. umalilaense, A24750133) H Stone cells are often easy to see in herbarium specimens (e.g. S. sarrachoides, Blom s.n. BM001207745). Photos by S. Knapp and G. van der Weerden.

General overviews of Black nightshade taxonomy and systematics have been published (Edmonds 1977, 1978, 1979a; Chakrabarti and Sharma 1994), including geographically focused taxonomic treatments for South America (Edmonds 1972), North America (Schilling 1981), Australia (Henderson 1974), Africa (Edmonds and Chweya 1997; Olet 2004; Manoko 2007; Edmonds 2012), Europe (Wessely 1960) and the Indian subcontinent (Schilling and Andersen 1990) and detailed cytological, molecular and morphological studies (Saarisalo-Taubert 1967; Venkateswarlu and Rao 1972; Edmonds 1982, 1983, 1984a; Ganapathi and Rao 1986a, 1986b, 1986c, 1986d, 1987a, 1987b; Nandhini and Paramaguru 2014). These studies have greatly enhanced our understanding of the complex morphological and ploidy level variation present in the group, but much taxonomic work remains within the Morelloids especially in South America, where more than half of the known Morelloid species are found (Barboza et al. 2013).

Within the Old World, Africa is a centre of species diversity with a total of 12 species (10 of them native), with fewer species found in Australasia (Figure 4; Tables 2–6). Many of the Old World species have spread to the New World with humans, such as S. nigrum, S. scabrum and S. villosum, while five species currently established in the Old World are widespread weeds originally native to South America but have been taken in the reverse direction, often as contaminants of wool; S. chenopodioides, S. nitidibaccatum, S. pygmaeum, S. sarrachoides and S. triflorum (Tables 2–6; Särkinen et al. 2015a).

Figure 4. 

Distribution of the Morelloid clade of Solanum, showing number of species per 3 degree grid cells (ca. 300 km²) based on native and non-native records.

An overview of the entire section in the Old World, including a revision of the nomenclature and typification of the more than 370 names associated with these taxa, has never been done but is needed in order to provide identification tools for these difficult and morphologically very similar species. Here we provide a taxonomic revision of all 19 species of the Morelloid clade (black nightshades) occurring in the Old World based on a detailed morphological study; we include native and introduced taxa. This is part of our molecular systematic and taxonomic work focusing on producing a monographic treatment of the entire Morelloid clade which has to date focused on understanding species diversity and delimitation in the New World (e.g. Barboza et al. 2013; Särkinen et al. 2013, 2015a, b).

The European species of black nightshades were known to the ancient Greeks and Romans, who used them medicinally. In the 1^st century AD, Pliny the Elder (Riley 1855) called the plant “cucubalus, strumus or strychnon” [translated from the Greek] and documented its use against stings, wounds and lumbago. Dioscorides, whose De Materia Medica was the primary source for much of medicine in Europe until after the Middle Ages, illustrated the plant (Figure 1) and advocated it for its astringent and cooling properties. Later European herbals, building upon the work of Dioscorides (e.g. Fuchs 1542, 1543; Matthioli 1544; Dodoens 1554, 1563) all treated S. nigrum L. as “Solanum Hortense” and recognised it as a native plant and distinct from others which they considered similar plants such as “Halicacabum [Solanum Halicacabum]” (Alkekengi officinarum Moench), “Solanum somniferum” (Withania somnifera (L) Pauq.) or “Solanum furiosum” (probably Atropa bella-donna L.). Fuchs (1542) listed French (Morelle) and German (Nachtschatten) common names, along with the name used in medicine “Solatrum”, stating that the name “Nigrum” given to this plant came from the berries and not from the leaves. It is clear that authors in continental Europe recognised the different fruit colours as the same plant; Matthioli (1544) describes the fruit as “fructu rotundo, uiridi, qui post maturitatem nigricat, aut fuluefcit” (round green fruit, black or yellowish-brown at maturity). Most hand-coloured editions of these early works (e.g. Fuchs 1543; Dodoens 1554) have the berries painted black and the plants in the woodcuts are identifiable as S. nigrum with the sepals more or less appressed to the berries (see species treatments). Dodoens (1583) described the fruit more clearly (“per initia virentes, maturae veròe, aut nigricantes, aut rubentes, aut luteo coloris” [at the beginning green, when mature black, or red, or yellow coloured]) and his illustration is clearly of S. villosum with its characteristic strongly reflexed and longer sepals in fruit (Figure 2).

The first herbal written in English was by John Gerard. He (Gerard 1597: 338–340) extensively treated the Garden Nightshade (as “Solanum Hortense” = S. nigrum) and documented its uses (“…Physicians do worthily use it, and that seldom as a nourishment, but always as a medicine”) and alluded to the confusion of nightshades by earlier authors (“Apuleis, amongst the confused names of Nightshade, who comprehending all the kinds of Nightshade together in one chapter, being so many, hath strangely and absurdly confounded their names.”). He suggested that it was safe and not poisonous – “Neither the juice hereof nor any other part is usually given inwardly, yet it may without any danger.”

By the mid-15^th century, black nightshades were already becoming confused with the more toxic deadly nightshade (Atropa bella-donna). In his Complete herbal, first published in 1653, Nicholas Culpeper called his “Common nightshade” a “cold Saturnine plant” and warned “Have a care you mistake not the deadly nightshade for this [black nightshade]; if you know it not, then you may let them both alone” (Culpeper 1814). He described its weedy nature well; “It groweth wild in this kingdom, and in rubbish, the common paths and sides of hedges, in fields, and also in gardens without any planting.” Black nightshades were given the common name “Petty Morel” (from the French petite morelle) and Atropa was called “Great Morel” (Grieve 1931). Nightshades’ reputation for being poisonous developed over the succeeding centuries in Europe, despite the use of these plants in local cultures across their range (see Uses below).

Linnaeus (1753) broadly circumscribed S. nigrum and divided it into six infraspecific taxa, many of which were based on the plates in Dillenius’s Hortus Elthamensis (Dillenius 1732). He recognised the European S. nigrum (as var. vulgare), S. villosum (as S. nigrum var. villosum), S. americanum (as S. nigrum var. patulum) and included the African cultivated species S. scabrum (as S. nigrum var. guineenense); but he had not seen material of other species treated here (see individual species treatments for details). He clearly recognised all these taxa as very similar and as variants of a worldwide species; his diagnosis reads “Habitat in Orbis totius, cultis” [Habitat in all the world, cultivated]. He also noted many of these looked like mixtures (“Tot varietates β, γ, δ, ε, ζ videntur esse hybridae proles”). In the Species plantarum (Linnaeus 1753), he did not cite many of the works based on non-European plants (e.g. Piso 1648; Rheede von Draakestein 1689), which he had previously cited in Hortus cliffortianus (Linnaeus 1737) and, in the Clifford herbarium (BM), only specimens of S. nigrum, S. scabrum and S. villosum are preserved.

In the sixth edition of his Gardener’s dictionary, Philip Miller used Linnaean binomials for the first time (see Stearn 1974). In this work, he (Miller 1768) described seven members of the Morelloid clade, five of these as new names (S. villosum Mill., S. luteum Mill., S. rubrum Mill., S. americanum Mill., S. scabrum Mill.). He did not recognise infraspecific taxa, but also did not indicate he was raising Linnaeus’s varieties to species level.

Lamarck (1794) recognised seven taxa, including some not known to either Miller or Linnaeus, such as S. radicans L.f. and S. corymbosum Jacq. (members of the clade belonging to the Radicans group, see Särkinen et al. 2015a). He additionally described S. chenopodioides Lam., from material said to be from “île de France” (Mauritius, but see species description) and S. triangulare Lam. based in part on an illustration from Rumphius (1750, = S. scabrum). Some of these early authors re-used epithets (e.g. villosum used by both Miller and Lamarck), but it is not clear whether they were referring to earlier names or not; the principle of priority had not yet become established for botanical naming (see Knapp et al. 2004).

The name for the Morelloid clade is derived from Dunal’s (1813: 119) un-ranked group “Maurella” that included herbaceous or subherbaceous species with entire leaves. He included 15 species, all of which are still considered members of the clade. In his Solanorum synopsis (Dunal 1816), he maintained this group, adding to it taxa described by himself and others, most of which (with the exception of S. quadrangulare Thunb. = S. africanum Mill., a member of the Dulcamaroid clade, see Knapp 2013) are still considered related. Dumortier (1827) used this group for his conspectus of the Belgian species, but with Dunal’s spelling changed to “Morella”. In general, the concept of Morella was narrow and included only those species later recognised as members of Solanum sect. Solanum, but did not include species now recognised as part of the more broadly defined group (Bohs 2005; Weese and Bohs 2007; Särkinen et al. 2013, 2015a). In the Prodromus, Dunal (1852) paid little attention to earlier names and erected an entirely new framework for Solanum mostly composed of gradi ambigui (names of ambiguous rank). Morella, however, was one of the names he continued to use. Within it, Dunal (1852) recognised two groups based on the inflorescence position, “Morellae spuriae” (six spp.) and “Morellae verae” (54 spp.). Circumscription of “Morella” remained obscure and loose during most of the 19^th and 20^th centuries, with many herbaceous non-spiny taxa treated as members of the group, resulting in many names associated with the Morelloid clade. Many of these names do not belong to the clade as now recognised based on phylogenetic data (Bohs 2005; Weese and Bohs 2007). An extreme example of this is S. stipuloideum Rusby from the morphologically distinct Potato clade (Spooner and Knapp 2013), which was assigned to Solanum sect. Solanum in the original description (Rusby 1907).

The very numerous treatments of these species in European floristic works (see species descriptions) usually did not specifically treat these taxa as belonging to infrageneric groups. Often these species were the only solanums treated in these floristic works, and they were often called “black nightshades” (e.g. Opiz 1843). Von Wettstein (1895) followed Dunal’s scheme in his treatment of the European species, as did Dammer (1906) and Bitter (1917, 1919) for Africa and Asia. Bitter (1917) was the first to explicitly recognise the group at the sectional level. Following the rules on use of autonyms, Seithe (1962) was the first to re-name the group containing the type species of the genus (S. nigrum) Solanum section Solanum; she also recognised sects. Campanulisolanum Bitter, Chamaesarachidium Bitter and Episarcophyllum Bitter (all groups confined to the New World, see Barboza 2003; Barboza and Hunziker 2005), now considered part of the larger Morelloid clade (Särkinen et al. 2015a). Her sectional names were followed by Danert (1967, 1970) with little change. D’Arcy (1972) lectotypified the infrageneric groupings in Solanum and provides an overview of the history of these infrageneric names.

Within the Morelloids, four well-supported clades have been recognised based on a detailed molecular phylogenetic study (Särkinen et al. 2015a). Prior to the use of molecular characters in phylogenetic analysis (e.g. Bohs 2005; Weese and Bohs 2007), the Morelloid clade thus defined was not recognised as a natural group. These clades loosely correspond to the previously recognised morphological sections: (1) the Radicans clade which comprises four of the species (but does not include the type species, S. triflorum Nutt.) of Solanum sect. Parasolanum A.Child (Bohs 2005); (2) the Episarcophyllum clade that includes most species of Solanum section Episarcophyllum Bitter; (3) the Chamaesarachidium clade that includes two species of Solanum section Chamaesarachidium Bitter; and finally the largest group (4) the Black nightshade clade, that includes all species of the traditional Solanum sect. Solanum. The first three clades are restricted to the New World, while most species of the Black nightshade clade occur in the New World but with a secondary centre of diversity in the Old World that is treated in this monograph.

The number of taxa included in this clade has not been clear, in part because it contains many widespread and morphologically variable species and has always been considered difficult. Although Edmonds (1972) suggested estimates of species richness had been exaggerated in the group and provided relatively low estimates of species numbers, our ongoing work in the group shows the clade includes ca. 75 species that are mostly restricted to South America (Särkinen et al. 2015a). Nineteen black nightshade species are found in the Old World, of which twelve are native there (four narrowly endemic) and not found in the Americas (Table 2; Särkinen et al. 2015a); the other seven are introductions from the New World.

Numerical taxonomic studies have been undertaken in order to resolve species relationships, parental origin of polyploids and species delimitation in these species (Soria and Heiser 1961; Heiser et al. 1965; Edmonds 1978), but the power of these methods has remained limited due to the complex and often overlapping morphological variation between the closely related species. Species of black nightshades show large amounts of morphological variation, especially in growth form, leaf morphology and indumentum.

Modern regional taxonomic treatments of the Old World members of the Morelloid clade have been done for Europe (Hawkes and Edmonds 1972), India (Ganapathi and Rao 1986a, 1988; Schilling and Andersen 1990), Africa and Madagascar (Bitter 1912a, 1912b, 1913a, 1921, 1923; Jaeger 1985; Bukenya and Hall 1988; Bukenya 1993; Bukenya and Carasco 1995; Edmonds and Chweya 1997) and Australia (Henderson 1974). Edmonds and Chweya (1997) remains the most thorough account of the black nightshades in the Old World, but includes only some of the species known to occur in Africa and does not treat species from Asia or those known from Australasia and the Pacific. Regional floristic treatments for regions of Africa have recently been published with detailed descriptions of local morphological variation (Edmonds 2005, 2006a, b, 2012). Schilling and Andersen (1990) clarified some of the nomenclatural problems and issues relating to misapplication of names in the Indian subcontinent in relation to the most commonly occurring species in the region.

The Old World species of the Morelloid clade do not form a monophyletic group. Phylogenetic analysis using plastid DNA sequence data (Särkinen et al. 2015a; T. Särkinen et al. in prep.) indicates that the native Old World polyploid species (S. hirtulum, S. memphiticum, S. nigrum, S. opacum, S. retroflexum, S. scabrum, S. tarderemotum, S. villosum, S. umalilaense) form a distinct group, to which S. alpinum does not belong. Polyploids from Africa are not each other’s closest relatives. Other species occurring in the Old World are members of distinct South American clades.

The Morelloid clade is one of only a few cases of extensive ploidy level variation in Solanum. Potatoes and the Archaesolanum clade are the only other clades of Solanum in which ploidy level variation is extensive (Poczai et al. 2011; Spooner et al. 2014); due to the economic importance of the species in the Potato clade as contributors of genes for cultivated potato improvement, they have been much more extensively investigated than the morelloids (see Spooner et al. 2014). Potato species are also diploid, tetraploid or hexaploid, but the cultivated potato itself (S. tuberosum L.) comprises both diploid and tetraploid populations (see Spooner et al. 2007); this variation of ploidy level occurs in some New World morelloids, but we have not observed it in the species treated here. The morelloids have been the focus of only a few previous studies (e.g. Stebbins and Paddock 1949; Edmonds 1977, 1979a) and, of the ca. 75 species in the clade, only 44 species have verified and vouchered chromosome counts; the counts show that 30 species are diploid (2n=2x=24), nine are tetraploid (2n=4x=48) and five hexaploid (2n=6x=72). Octoploid individuals have only been recorded three times, all of these in South American plants and are thought to be the result of spontaneous chromosome doubling of tetraploid species (Heiser 1963; Edmonds 1977) and three species (S. interandinum Bitter, S. macrotonum Bitter, S. pygmaeum) appear to have multiple ploidy levels. Eight of the Old World native morelloids are polyploid (six tetraploids and two hexaploids) and sequence data indicate that S. hirtulum is also a polyploid based on the presence of multiple double peaks in low copy nuclear sequence reads similar to known polyploid species (e.g. S. nigrum), but this needs to be confirmed with a vouchered chromosome count (see Table 1; Särkinen et al. 2015a). The ploidy level of S. alpinum remains unknown. Natural hybridisation between sympatric species (Ganapathi and Rao 1985, 1986b) has been suggested as the reason for the large number of polyploids in the Morelloid clade, especially in the Old World, but evidence for this hypothesis is weak at best.

It is thought that most of the polyploids in the Morelloid clade are of alloploid, rather than autoploid origin, because bivalents show regular pairing at meiosis (Edmonds 1977, 1979a; Ganapathi and Rao 1987a; Ojiewo et al. 2007). The putative parental origin of these species has been investigated using traditional crossability and hybridisation techniques (Edmonds 1977; Ganapathi and Rao 1985, 1986b, 1986c, 1986d, 1987a, 1987b) and, more recently, using molecular markers (Manoko 2007; Poczai and Hyvönen 2011). Chromosome painting or GISH (genomic in situ hybridisation) techniques have not been used in determining progenitor genomes in the morelloids as they have with great success in Nicotiana (e.g. Chase et al. 2003; Clarkson et al. 2005) and in the potatoes (e.g. Pendinen et al. 2012).

Edmonds (1977) undertook an extensive crossing programme involving morelloid species from both the New and Old Worlds. She found that hybridisation was relatively common within ploidy level, while successful hybridisations between ploidy levels were rare. Successful crosses between tetraploids and hexaploids were postulated to have resulted due to possession of a common parental genome. She suggested that S. sarrachoides (as S. sarachoides) was a potential diploid parent of S. nigrum, because almost half of the crosses involving it and S. nigrum were successful; she also stated that S. americanum was the “undisputed [diploid] progenitor” of S. nigrum, probably based on the work of Nakamura (1935, 1937). Regular bivalent pairing in hybrids derived from S. nigrum × S. scabrum crosses indicate homology of parental genomes of the two hexaploids (Ganapathi and Rao 1987a). Evidence from other crossing studies suggests that the tetraploid S. villosum, S. retroflexum or one of their close relatives, has also been involved in the parentage of S. nigrum (Ganapathi and Rao 1986b, 1986c) and that S. villosum is one of the parents of S. scabrum (Ganapathi and Rao 1985, 1987a, 1987b). As only a limited set of the diploid and tetraploid species have been used in these crossing studies, results have to be interpreted with care in that the species tested, or any of its close relatives, could have been involved in the parentage. Detailed molecular phylogenetic studies could help to reveal closely related species to the ones used in these crossing studies to further understand potential parents that could then be tested.

Manoko (2007) used AFLP (Amplified Fragment Length Polymorphism) molecular markers to examine genetic variation within and between African diploids and polyploids. As AFLP analyses are based on fragment sizes, results from these analyses only allow us to understand genetic variation but do not directly link to phylogenetic relationships. The AFLP study found that tetraploid accessions split into two distinct groups; a set of tetraploids (group A; S. retroflexum, S. tarderemotum (as S. florulentum), S. umalilaense and S. villosum) clustered with hexaploid taxa (S. nigrum, S. scabrum) and another set (group B; S. memphiticum, S. tarderemotum and “S. patens”) clustered with diploid species used in the analysis. It was suggested that, because the tetraploids split into two clear groups in a Neighbour Joining (NJ) analysis, they were unlikely to share diploid progenitors (Manoko 2007), but as highlighted above, AFLP studies allow the inspection of genetic variation but not phylogenetic relationships. Poczai and Hyvönen (2011) used ISSR (Inter Simple Sequence Repeat) and SCoT (Start Codon Targeted) markers in a network analysis to investigate the origin of S. nigrum. They used a limited number of species that had previously been suggested as having common origins; S. americanum, S. chenopodioides, S. nigrum, S. opacum, S. nitidibaccatum (as S. physalifolium Rusby) S. retroflexum, S. scabrum and S. villosum. Their results supported the close relationship between S. villosum and the two hexaploids (S. nigrum and S. scabrum) and they suggested that these species share a diploid progenitor. They also showed that S. villosum is most likely an autotetraploid derived from S. americanum, confirming an earlier chromosome banding study (Sultana and Alam 2007) and hypothesised that the progenitors of S. nigrum are S. villosum and S. americanum and that the species evolved via amphiploidy of sterile triploid progeny (Poczai and Hyvönen 2011). Solanum opacum and S. retroflexum were associated with S. chenopodioides and S. nitidibaccatum (as S. physalifolium) and not with S. americanum, contrary to suggestions by Edmonds 1977, who thought S. americanum was involved in the origins of S. retroflexum based on crossability, suggesting their diploid progenitors are not shared with the other polyploids in the analysis. Results of previous work attempting to establish genome origins for morelloid polyploids is summarised in Table 8.

The species of black nightshades are predominantly self-compatible (Edmonds 1979a; Schilling and Heiser 1979; Olet 2004). Interspecific crosses can be produced between most species within ploidy levels, although fertility is not always high and some breeding barriers exist (Edmonds 1977, 1979a; Olet 2004). Putative hybridisation in the wild has been reported between co-occurring diploid species in North America and Australia (Stebbins and Paddock 1949; D’Arcy 1974b; Henderson 1974). Schilling and Heiser (1979), however, suggested that the ability to cross was not a useful taxonomic character in the morelloids. Natural hybrids have also been reported between diploids and polyploids and between polyploids themselves (Henderson 1974; Leslie 1978; Venkateswarlu and Rao 1972), although Edmonds (1977) showed that within ploidy level crossability success was much higher than that between ploidy levels. Hybridisation, followed by backcrossing to parental species and/or polyploidisation, has been thought to explain some of the complex morphological variation found within the group (Stebbins and Paddock 1949; Edmonds 1979a). This assumption of widespread hybridisation being responsible for taxonomic complexity was also the case for the potatoes, but detailed studies with modern methods have shown that much of the variation is not due to hybridisation but instead to extensive within-population variability (Spooner et al. 2014). Schilling and Heiser (1979) pointed out that the hybrids created in the greenhouse were unlikely to occur in nature and expressed views that the use of the Biological Species Concept (BSC; e.g. Mayr 1982) in this group where most species are interfertile at some level is not useful (see Knapp 2008 for a discussion of the BSC and plant species recognition).

There is considerable confusion regarding the toxicity of species of black nightshades (see Schilling et al. 1992; Olet et al. 2005). Many manuals or floras list black nightshade berries as toxic (e.g. Kingsbury 1964; Edmonds 2005). The European members of the group are often confused with the deadly nightshade Atropa bella-donna L., whose berries are similarly black and shiny, but contain highly toxic tropane alkaloids such as atropine and scopolamine (Parr et al. 1990; Ashtiana and Sefidkonb 2011). No species of Solanum has been shown to contain tropane alkaloids such as atropine or scopolamine (Tétényi 1987; Pigatto et al. 2015), but members of Solanum, including the black nightshades, do contain the nortropane alkaloids known as calystegines (Dräger et al. 1994; Pigatto et al. 2015). In Solanaceae, these compounds are synthesised along the same pathway as the tropanes, but are not as toxic when tested on laboratory rats (Stegelmeier et al. 2008); the finding that they occur in many commonly eaten foods suggests their activity differs from that of the tropanes such as atropine, scopolamine and nicotine (Asano et al. 1997; Welch et al. 2012).

Most species of Solanum contain a wide range of steroidal alkaloids (also known as glycoalkaloids; Bradley et al. 1979) that, while potentially toxic in very large quantities, are not dangerous to humans. Chemical surveys have shown that only unripe fruits contain these potentially somewhat toxic glycoalkaloids (Cipollini et al. 2002), while leaves and ripe fruits usually lack these compounds (Carle 1981; Voss et al. 1993). Many of the studies that have studied the chemical composition of morelloid species have mixed samples of immature and mature berries, leading to conclusions that berries are poisonous (Eldridge and Hockridge 1983; Mathé et al. 1980; Sharma et al. 1983; Arnold 1985). Such findings are important in agricultural trade where species of morelloids occur as weeds and their immature and mature berries are eaten by livestock and can contaminate harvests of a variety of grain and vegetable crops (Ogg et al. 1981; Rogers and Ogg 1981). Toxicity of unripe berries, however, appears to be reduced if forage is fermented as silage (Vogel and Gutzwiller 1993) and does not apply when ripe berries are harvested for human consumption.

Species of black nightshades are used as leafy vegetables or fruits throughout the world, sometimes referred to as ‘supervegetables’ for their high protein and iron content (FAO 1988; Yang and Ojiewo 2013; Cernansky 2015; Ronoh et al. 2017; Fokon and Domugang 1989). Species of the group are wild-harvested or cultivated for their juicy berries in both the Old and the New Worlds, in both the tropics and temperate zones (e.g. Heiser 1969). Their use, however, has been particularly studied in Africa, where they are important components of indigenous cultivation systems (Edmonds and Chweya 1997; Schippers 2000; Keding et al. 2007; Pasquini et al. 2009) and are the focus of plant breeding strategies aimed at improving the nutrition and income generation of women and children in urban and peri-urban areas (Dinssa et al. 2016; Manoko and van der Weerden 2004a, 2004b).

In Africa, species are cultivated either for use as leaf vegetables (e.g. S. americanum, S. memphiticum, S. scabrum, S. tarderemotum, S. umalilaense, S. villosum) or for their fruits (e.g. S. americanum, S. retroflexum, S. scabrum, S. villosum). Leaves of African nightshades are high in flavonoids, vitamin C, folates, iron and antioxidants (Akubugwo et al. 2007; Keding et al. 2007; Yang and Ojiewo 2013; Ronoh et al. 2017) and the leaves are often cooked in milk to make them less bitter. Medicinal uses in Africa include antiseptic for eyes and skin, treatment of diarrhoea (Essou and Hermans 2006) and as a general tonic for health (Yang and Ojiewo 2013). Seeds have been found to be rich in lipids and a good source of essential fatty acids used in some areas in Africa (Nzikou et al. 2007).

In Asia, including the Indian subcontinent, S. nigrum is also used as a leaf vegetable (Arora 1981; Jain and Borthakur 1986), for its fruits (Abraham 1981; Vartak 1981) and in medicine (Ammaan and Subramanian 2017). Berries of several species are eaten (e.g. S. americanum, S. retroflexum, S. scabrum, S. villosum; see also individual species treatments). Solanum nigrum is mentioned in several different contexts in the 13^th century Chinese treatise Jinhuang Bencao which was the first of many Chinese works on uses of plants and introduced many plants into the diets of modern Chinese people (Zhu et al. 2015). In that work, it is mentioned both as a food (leaves and berries) and as a medicine. Today, leaves of morelloid species are commonly eaten, particularly in southern China; one of the authors (S. Knapp, pers. obs.) has also observed cultivation of S. americanum for its fruits there.

Disaggregating uses for these species is difficult because many authors treated all morelloids as a single, highly variable taxon; Burkhill (2000) treated all morelloids in western Africa as S. americanum, while they were most certainly several species. The cultivated status of many of these plants went unnoticed by many early plant collectors, who often noted them as “weeds of field margins”. Black nightshades are known by a great many indigenous names, especially across Africa and care must be taken with names applied to species in literature, given the complex taxonomy of the group and difficulties in transliteration of indigenous languages by early collectors. In Table 9, we list the more widely encountered common names by region and language and, in each individual species treatment, list those that we have verified for each species from either herbarium specimens or from literature where we are certain of the species identification.

Solanum scabrum is by far the most commonly cultivated of the Old World species and is cultivated throughout Africa (Berinyuy et al. 2002; Defelice 2003; Bukenya and Carasco 1995; Olet et al. 2006), especially in western Africa. It is also known from the Caribbean where it was most likely brought by enslaved peoples from western Africa. In the United States, S. scabrum was marketed under the name of “Garden Huckleberry” (Soria and Heiser 1959), where it was introduced with great fanfare by the vegetable breeder Luther Burbank (Heiser 1969). Burbank also introduced what he called a new species, the “Sunberry”, later marketed as the “Wonderberry”, which turned out to be the southern African species S. retroflexum (Soria and Heiser 1959; Heiser 1969); Burbank claimed he had created it from a cross between S. scabrum and what he called S. villosum (probably = S. sarrachoides). The story of this and the accusations that Burbank was marketing a poisonous plant are well-documented in Heiser (1969).

Solanum tarderemotum and S. villosum are more commonly cultivated in eastern Africa and many specimen labels note that the fruits of S. villosum are particularly prized by children (also see Keding et al. 2007). Solanum umalilaense is cultivated locally as a leaf vegetable in the Umalila region of Tanzania (Schippers 2000; Manoko and van der Weerden 2004a, 2004b) and its highly branching morphology is perhaps due to human selection. In Asia, S. villosum, S. americanum and S. nigrum are all used for both their berries and leaves, as well as for medicine (Potawale et al. 2008; Jagatheeswari et al. 2013). Details of cultivars and more specific uses are given in the species treatments.

Solanum “nigrum” has a long history of use in Ayurvedic medicine in India, but it is clear that the concept medicinally does not distinguish between S. nigrum s.s. and S. villosum (see Warrier et al. 1996; Jagatheeswari et al. 2013; Ved et al. 2016). The plants are noted for their antiseptic and antidysenteric properties and, in common with other members of the family, are considered to be febrifuges and anti-inflammatories (see Dunal 1813). They have been used in the treatment of many different diseases and complaints in many areas of the Old World (e.g. Dunal 1813; Grieve 1931; Huang 1993; Ghazanfar 1994; Edmonds and Chweya 1997; Burkhill 2000). Early European uses are documented in the herbals (see above) and in the many pharmacopoeia of different countries (summarised in Grieve 1931); Dunal (1813) records S. nigrum as being the most used of European species of Solanum, and that with the most ancient use (dating back to Hippocrates, Dioscorides and Theophrastus). Common properties appear to be antiseptic, anti-inflammatory, expectorant and diuretic and these plants are often cited for the treatment of eye and stomach complaints (e.g. Ignacimuthu et al. 2006).

In Europe, seed remains of S. nigrum have been excavated from the Viking layers at York (Kenward and Hall 1995) and at ten other sites in Denmark, Germany, Ireland, Russia and Sweden (A. Poulsen and A. Kool, pers. comm., manuscript in preparation). Solanum villosum has been found in Israel in Acheulian site in Gesher Benot Ya‘aqov, amongst other food plants that were used by early humans (Melamed et al. 2016). The excavated seeds that represent the black nightshade species still in use today demonstrate that their use has been widespread across different cultures and areas dating back to early humans during their migration out of Africa in mid-Pleistocene (Melamed et al. 2016).

In the Pacific, uses for black nightshades (S. americanum and S. opacum, without distinction) are recorded from Hawaii (Degener 1945). As with other species of the group, both leaves and berries are eaten and used medicinally. Medicinal uses recorded are as an antiseptic, laxative (the juice of the plant) and as a tonic (Degener 1945); bruised leaves are used topically for stomach ache.

Glycoalkaloids from S. americanum, S. nigrum and S. opacum were often used as a steroid base for human contraceptives until the industry moved to production of synthetic compounds in the 1980s (Bradley et al. 1979; Carle 1981). Aqueous extract from S. villosum (as S. alatum) and S. nigrum have been shown to exhibit anti-inflammatory, hepato-protective, anti-cancer effects and potential to control and prevent gastric ulcers (Lin et al. 1995; Heo and Lim 2005). Leaf extracts of several morelloid species have been shown to have a molluscicidal effect against Bulinus (Ndamukong et al. 2006) and Biomphalaria (El-Sherbini et al. 2009) snails that act as hosts of Schistosoma parasites that cause the tropical disease bilharzia (schistosomiasis) and against the parasites themselves (Obare et al. 2016). On the Indian subcontinent (Singh et al. 2001), tests of leaf extracts of S. nigrum have shown promise as larvicides against mosquitoes that are important vectors of human disease such as the malaria vector Anopheles culicifecies Giles and the filariasis vectors Culex quinquefasciatus Say and Aedes aegypti (Linnaeus in Hasselquist). Extracts from the berries of S. villosum offer promise as biocontrol agents for the mosquito dengue fever vector A. aegypti (as Stegomyia aegypti) since resistance to commonly used insecticides is on the increase in many parts of the developing world (Chowdhury et al. 2008). The steroid alkaloid solanine from S. americanum has been used as an agricultural pesticide (Sumner 2000) and produces a proteinase inhibitor (PI II) that has shown to confer insect resistance in transgenic plants (Xu 2001).

Old World species of the Morelloid clade are also sources of resistance genes used in agriculture. Several species of black nightshades, including S. nigrum, S. americanum and S. villosum, are known to carry resistance genes against late blight (Phytophthora infestans (Mont.) de Bary), a major fungal pest in potato, tomato and eggplant (Kamoun et al. 1999; Campos et al. 2002; Flier et al. 2003; Lebecka 2008, 2009; Śliwka et al. 2012; Witek et al. 2016). Although resistance traits have been transferred to potato (S. tuberosum L.) in some experiments (Colon et al. 1993; Eijlander and Stiekema 1994; Horsman et al. 1997; Zimnoch-Guzowska et al. 2003), black nightshades have not been widely used in traditional plant breeding, but new cloning techniques coupled with the large number of resistance genes found in S. americanum suggest these plants hold promise for biotechnology (Witek et al. 2016). Solanum nigrum has also been used as a model system with which to study herbivore defence responses (Schmidt et al. 2004), especially since response chemistry seems to differ significantly from that seen in tomato (Schmidt and Baldwin 2006).

Our goal for the treatment of the Old World species of the Morelloid clade has been to provide circumscriptions for the members of this morphologically variable group of species, while clearly highlighting areas, taxa and populations where further in-depth research is still needed to confirm current taxon concepts. 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 and are therefore of little utility in a practical sense (see Knapp 2008). 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). Our decisions relied on clear morphological discontinuities to define easily distinguished species. Specific characters used for recognition are detailed with each species description and in the key. Some potential reasons for variability and intergradation are recent divergence, hybridisation and environmental influence on morphology. In this revision, we have tried to emphasise similarities between populations instead of differences, which so often reflect incomplete collecting or local variation. We have not recognised subspecies or varieties, but have described and documented variation where present, rather than formalise such variability with a name which then encumbers the literature. We have been conservative in our approach, recognising as distinct entities those population systems (sets of specimens) that differ in several morphological characteristics. Many of the species in the group (and of Morelloids in general) are extremely widespread and variable; variation exists in certain characters, but the pattern of variation is such that no reliable units can be consistently extracted, nor is geography a completely reliable predictor of character states. Here variability within and between populations seems more important than the variations of the extremes that other taxonomists have recognised as distinct. We describe this variation realising that others may wish to interpret it differently.

Although infraspecific taxa have been recognised by others within the group, we do not recognise any here due to the complex morphological variation observed within each species, where the inspection of a larger number of specimens quickly reveals no apparent natural breaks in variation but rather a mixing between highly morphologically variable populations of widespread species. Results from seed protein studies (Edmonds and Glidewell 1977), as well as from more recent population genetic studies (e.g. Manoko 2007; Manoko et al. 2008) support our morphological observations in finding no obvious patterns supporting infraspecific structure within particular variable species such as S. americanum, S. villosum and S. nigrum.

Our taxonomic treatment is based on results from recent molecular systematic studies considering the taxonomy of the section, including focused morphological and AFLP studies of the African species by Jacoby (2003: South Africa), Olet (2004) and Manoko (2007), network analysis of the polyploid species and their putative diploid ancestors by Poczai and Hyvönen (2011) and the molecular phylogenetic study of the entire Morelloid clade by Särkinen et al. (2015a). Descriptions are based on field work and examination of >10,000 herbarium specimens from 146 herbaria (A, AK, ALCB, ANG, ASE, AV, AZU, B, BA, BAH, BH, BHCB, BISH, BM, BP, BR, BRI, BSD, B-W, C, CAL, CAS, CEN, CEPEC, CESJ, CGE, CHR, CICY, COI, COL, CORD, CPUN, CTES, DD, DS, DSM, DUKE, DVPR, E, EA, EAC, ECON, EIU, ESA, F, FCQ, FHO, FT, FUEL, FURB, G, GB, G-BOIS, G-DC, GH, GOET, H, HAMAB, HAO, HAS, HBG, BSHC, HOH, HOXA, HST, HUCS, HUEFS, HUEM, H, IAC, ICN, INB, INPA, IPA, JOI, JPB, K, KUFS, L, LE, LG, LIL, LOJA, LP, LPB, LU, M, MA, MAC, MBM, MBML, MEL, MEXU, MHU, MO, MOL, MPU, MPUC, NHT, NIJ, NSW, NY, OTA, OUPR, OXF, P, PACA, PAL, PBL, PH, QAP, QCNE, S, SCA, SGO, SI, SJRP, SMDB, STU, TAN, TCD, TO, TUB, U, UC, UDBC, UEC, UFP, UFRN, UMO, UPCB, UPS, US, USF, USM, UT, VIES, W, WAG, WIS, WRSL, WU, YA, YU, Z). We have extensively compared introduced species across their entire ranges, not just in the Old World; this in part accounts for the large number of herbaria cited here.

Measurements were made from dried herbarium material supplemented by measurements from living material. Colours of corollas, fruits etc. are described from living material or from herbarium label data. Specimens with latitude and longitude data on the labels were mapped directly. Some species had few or no georeferenced collections; in these cases, we retrospectively georeferenced the collections using available locality data. Maps were constructed with the points in the centres of degree squares in a 1° square grid. Conservation threat status was assessed following the IUCN Red List Categories and Criteria (IUCN 2016) using the GIS-based method of Moat (2007) as implemented in the online assessment tools in GeoCat (http://geocat.kew.org). The Extent of Occurrence (EOO) measures the range of the species and the Area of Occupancy (AOO) represents the number of occupied points within that range based on the default grid size of 2 km². We present only the EOO in the threat assessments for widespread species; AOO is very sensitive to georeferencing bias. For introduced taxa we also present their conservation status in the putatively native range in each of the species treatments.

Type specimens in this group have proved difficult to trace. Many taxa were described based on types housed in herbaria whose collections were lost during the Second World War, including the Berlin herbarium (B) (see Vorontsova and Knapp 2010). In our searches of many potential repositories for original material, we have been able to trace duplicates for at least some of these. For taxa recognised only as synonyms, we have cited the taxa in synonymy and indicated that duplicates have not been found rather than neotypifying these (often infraspecific) taxa. We have also made use of specimen images available via online databases: BP (for the herbarium of Pál Kitaibel, https://gallery.hungaricana.hu/hu/Herbarium/), G (http://www.ville-ge.ch/cjb/bd.php), P (https://science.mnhn.fr/institution/mnhn/collection/p/item/search/form), W (http://herbarium.univie.ac.at/database/search.php), Z (http://www.herbarien.uzh.ch/index.html), the African Plants Initiative and Global Plants (http://plants.jstor.org).

Many of the species and infraspecific names coined for European plants in the 19^th and early 20^th century floristic and other works have no specimens cited. In some cases, we have been able to trace potential original material, but in others, plants may have been described from living material. We have lectotypified these names where we have been able, but have not selected neotypes for those names that have rarely been used beyond their first description (mostly infraspecific taxa of S. nigrum and S. villosum). Where specific herbaria have not been cited in protologues, we have followed McNeill (2014) and designated lectotypes rather than assuming holotypes exist. We cite page numbers for all previous lectotypifications. In general, we have lectotypified names with the best preserved, or in some cases only, herbarium sheet we have seen; in these cases, we have not outlined our reasoning for the letotypifications. Where there has been difficulty or where the choice may not be obvious, we detail our reasoning at the end of the species discussions.

The amateur Czech botanist, Paul M. Opiz, collected prolifically and described many taxa based on his and other collections, most of these specimens being housed in PR or PRC (Kirschner et al. 2007). We have been unable to visit those collections during the course of this work, so have not lecto- or neotypified these names.

Georg Bitter (e.g., 1913a, 1917, 1921, 1923) described many taxa of Solanum in the course of his monumental work on African solanums and worked widely in Germany in the period between the two World Wars (Weber 1928), including, but not exclusively, at Berlin. His protologues frequently include specific herbarium citations, but not in all cases. We have cited specimens as holotypes only when a single specimen with a single herbarium citation is indicated in the protologue; we have not assumed his types are all in B.

Edmonds (1972, 1979b) cited as “holotype” specimens at single herbaria that are in fact more correctly cited as lectotypes; we have accepted these as effective lectotypifications (Art. 7.10, McNeill et al. 2012) but corrected her citation of “holotype” to lectotype. D’Arcy (1974a) cited “type”, “syntype” or “lectotype” for many of the names treated here in his treatment of Solanum for Flora of Panama. He explicitly cited some of these as “lectotype” and we treat these as validly published lectotypifications because his intention was clear. We also treat his citations as “type” coupled with the citation of a single herbarium as unintentional (“inadvertent”) lectotypifications (e.g. Prado et al. 2015), following the stipulations of Art. 7.10 of the Code (McNeill et al. 2012). We have tried to find previous lectotypifications to the best of our ability, but some done inadvertently in floras may have escaped our notice.

Type specimens are cited with their barcodes in square brackets after each herbarium code, written as they are spelled with barcode readers, either as a continuous string (i.e. [G00104280]) or with a dash (i.e. [MO–1781232]) depending on the style of barcode used in each herbarium. Sheet numbers are cited where barcodes are missing and these are indicated as distinct from barcodes (i.e. MO [acc. # 1037660]). For widespread species, we have not cited geographically representative specimens in the text under each species treatment, but instead listed the countries of occurrence by region (Africa, Asia, Europe, Pacific [incl. Australia]; also see Tables 1–3). Identities of all numbered collections seen for this study are in Supporting Material (Appendices 1–3, including Index to numbered collections and specimens cited in pdf, xls and csv formats). The full searchable specimen details are also available on the Solanaceae Source website (www.solanaceaesource.org) and in the dataset for this study deposited in the Natural History Museum Data Portal (https://doi.org/10.5519/0009648). For narrowly endemic taxa, we have cited all the specimens we have examined. For introduced taxa, we cite only Old World specimens in this treatment, but the data set on the NHM Data Portal (https://doi.org/10.5519/0009648) includes all material seen for these species, including New World collections examined for globally distributed species.

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). Following Knapp (2013), we 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 are thus attributed only to Dunal. For “ex” attributions, we cite only the publishing author, as suggested in the Code (McNeill et al. 2012). Standard forms of author names are according to IPNI (International Plant Names Index, http://www.ipni.org).

Common names and uses given under each species treatment are recorded for verified specimens only. The aim here is to provide an authoritative summary; further details on uses and names can be checked from the literature cited. Similarly, we refer to chromosome counts based on voucher specimens that we have been able to verify or based on studies that describe the voucher material in adequate enough morphological detail that allows us to draw firm conclusions on the identity of the original material. This means that some studies (e.g. Venkateswarlu and Bhiravamurty 1969) have been excluded.