KEW BULLETIN VOL. 67: 401 Y 411 (2012) ISSN: 0075-5974 (print) ISSN: 1874-933X (electronic) Solanum incanum s.l. (Solanaceae): taxonomic relationships between S. incanum, S. campylacanthum, S. panduriforme and S. lichtensteinii John Samuels1 Summary. In a study designed to elucidate the taxonomy of Solanum incanum sensu lato, S. incanum L., S. campylacanthum Hochst. ex A. Rich., S. panduriforme E. Mey. ex Dunal and S. lichtensteinii Willd. from Africa and SW Asia were investigated using crossability and morphometric techniques. It is proposed that S. panduriforme is, in fact, a subspecies of S. campylacanthum (S. campylacanthum subsp. panduriforme) and that S. incanum and S. lichtensteinii are distinct species. Other information suggests that S. campylacanthum is more closely related to a common ancestor of S. incanum s.l. S. campylacanthum subsp. panduriforme and S. incanum are believed to have diverged away from S. campylacanthum-type predecessors in tropical E Africa, moving southwards or towards the Middle East, respectively. S. lichtensteinii probably evolved from an even earlier ancestor in its migration towards southern Africa. Key Words. Biosystematics, Georg Bitter, interfertility, numerical taxonomy, series Incaniformia, species concept. Introduction Taxonomy of African solanums The last adequately detailed work on African solanums was undertaken by Georg Bitter in the first quarter of the 20th century, when he formulated a revision of the genus Solanum in Africa, based largely on herbarium material (Bitter 1913, 1917, 1921, 1923). More recently, Jaeger & Hepper (1986) reviewed the development of our knowledge of the genus and provided a conspectus of native African species, totalling around 110, belonging to 16 sections in four subgenera. Levin et al. (2006) showed that the African solanums formed part of the distinctive and monophyletic “Old World clade.” The present paper concerns a few of the species in series Incaniformia Bitter in sect. Melongena (Mill.) Dunal of subgen. Leptostemonum (Dunal) Bitter. Series Incaniformia Bitter Bitter (1923) grouped several well-known species such as Solanum incanum L., S. campylacanthum Hochst. ex A. Rich., S marginatum L. f. and other closely allied species, along with S. melongena L. into the series Incaniformia Bitter. Bitter also attempted to expand upon the particularly narrow species concept prevalent at the time, reducing many of Dammer’s (1905, 1906, 1912, 1915) species to infraspecific taxa or synonyms. Bitter’s series Incaniformia nevertheless remained a large group containing 28 species found throughout much of Africa to SW Asia (Samuels 1996). Members of series Incaniformia are typical examples of the subgenus Leptostemonum (the “spiny solanums”) which comprises 450 species (Levin et al. 2006), around 80 of which occur in Africa. They are characterised by stellate pubescence, attenuate anthers and strong armature. Wherever ground is disturbed, members of series Incaniformia may successfully establish themselves and become persistent weeds; they are well-known ruderals and adventives across much of Africa and parts of SW Asia and are found growing at altitudes of 250 – 3000 m. Solanum incanum s.l. Solanum incanum and its allies have been the subject of plant breeding and research based on genetic improvement of the brinjal eggplant, S. melongena L. (e.g. Daunay et al. 1991, 1998, 2001; Lester 1998; Mohammad et al. 1994; Sakata & Lester 1994). An accurate understanding of the taxonomy of this group therefore has important commercial implications. Previous taxonomic treatments of Solanum incanum s.l. (Bitter 1923; Jaeger 1985; Lester & Hasan 1991; Whalen 1984) have provided us with surveys based on informal species groups; they are useful as overviews but do not commit to detailed taxonomic judgements (see Table 1). The main taxa representing the Solanum incanum species complex are: S. campylacanthum (group A in Lester & Hasan 1991), S. panduriforme E. Mey. ex Dunal (group B), S. incanum L. s.str (group C) and S. lichtensteinii Willd. (group D). Their close relationship has been confirmed by Mace et al. (1999) in their AFLP Accepted for publication May 2012. Published online 29 June 2012 1 Trezelah Barn, Trezelah, Gulval, Penzance, Cornwall TR20 8XD, UK. e-mail: john.samuels@virgin.net © The Board of Trustees of the Royal Botanic Gardens, Kew, 2012 E. S. cerasiferum group (S. cerasiferum Dunal & allies) D. S. campylacanthum group (S. campylacanthum & allies) D. S. lichtensteinii & allied taxa C. S. incanum L. sensu stricto only A. (S. bojeri Dunal, S. campylacanthum, S. delagoense Dammer & many other spp. S. incanum sensu lato groups: Lester & Hasan 1991 Jaeger 1985 “Solanum incanum” group: “S. incanum L. agg.” — 5 species groups: Whalen 1984 A. S. incanum group (S. incanum & allies) B. S. lichtensteinii group (S. lichtensteinii & S. subexarmatum Dammer) B. S. panduriforme only C. S. panduriforme group (S. panduriforme only) DNA study, by Lester & Hasan (1991) in their biosystematic investigations, and by Lester & Daunay (2003) in their survey of African vegetable solanums. 3 species collectivae forming part of series Incaniformia Bitter: 1. S. campylacanthum 2. S. bojeri (Dunal) sensu 3. S. incanum (L.) sensu ampliore (Hochst.) sensu ampliore ampliore Bitter (S. panduriforme Bitter (S. incanum L. & S. Bitter (S. campylacanthum Dunal & allies) incanum L. var. lichtensteinii & allies) (Willd.) Bitter consists of 12 (un-named) spp. of S. incanum and allies; largely based on Bitter’s series Incaniformia with majority of species reduced to synonymy KEW BULLETIN VOL. 67(3) Bitter 1923 Table 1. Previous taxonomic treatments of Solanum incanum s.l. 402 © The Board of Trustees of the Royal Botanic Gardens, Kew, 2012 Objectives The main objectives of this paper are to consider all available information, including that produced by the author, to determine the taxonomic status of groups A – D in Lester & Hasan (1991) and to designate species and other taxa appropriately. Materials and Methods For the present study, almost all of the species included in Bitter’s series Incaniformia are considered under Solanum incanum s.l. However, S. marginatum and S. melongena, which on the basis of DNA characteristics (Levin et al. 2006; Mace et al. 1999; Sakata & Lester 1994) and morphological differences (Jaeger 1985; Lester & Hasan 1991) are readily distinguished from S. incanum and its allies, and must, therefore, be treated elsewhere. S. cerasiferum Dunal and its allies were considered by Hasan (1989) to be distinctive from S. incanum and its near relatives. Although S. cerasiferum is closely related to S. incanum s.l. (Bitter 1923; Samuels 1996) subsequent DNA analysis (Mace et al. 1999) and comparative morphology (Samuels, in prep.) have confirmed Hasan’s (1989) view. In the present work the scheme for describing the taxa in Solanum incanum s.l. provided by Lester & Hasan (1991) has been adopted, but with some modifications. Two of Lester & Hasan’s Middle Eastern accessions have been re-located to group C from group D, making group D a collection of purely southern African accessions. In addition, S. bojeri Dunal and S. delagoense Dammer, believed to be distinct species by Lester & Hasan (1991), are considered to be synonyms of S. campylacanthum and S. panduriforme respectively, in the present study. It is also likely that several other species allied to groups A and D by Lester & Hasan (1991) are synonyms of S. campylacanthum and S. lichtensteinii respectively (Samuels, in prep.). Cultivation of living plants Seeds of 20 accessions originally collected from Africa and the Middle East were obtained from the Birmingham University Solanaceae Collection (Lester et al. 2001) and grown on. Plants were cultivated by the same methods used in other related studies (e.g. Lester & Hasan 1991; Lester & Niakan 1986). Accessions included eight of group A and four each of groups B, C and D (Table 2). Crosses between these accessions produced seed, from which 24 F1 hybrid lines were grown (Table 3). Voucher specimens were initially retained at the University of Birmingham and then transferred to the Radboud University Botanical and Experimental Garden, Netherlands. SOLANUM INCANUM S.L. (SOLANACEAE) 403 Table 2. Parental Accessions of Solanum incanum s.l. and methods of study. Studied by Accession No. Group (Hasan 1989) BIRM/S.0859 0931 1064 1398 b 1512 1518 1692 b 1750 1780 1781 1793 2023 2026 2027 2028 2055 2607 2465 2503 c RNL337/1432 A C A B C D D C A B C A A A A B A D D D a = unprovenanced; was employed. b Locality, Source/Collector and No. NTv Uganda, Kyambogo, Lester s.n. 6 Sept. 1969 Israel, Bot. Gdn. Univ. Tel Aviva Uganda, Kampala, Anne Kenrick s.n. Zimbabwe, N.E. Salisbury, Min. Ag. 14 Sept. 1973 Israel, Vadi Pereas, Univ. Jerusalem s.n. South Africa, Pretoria, P.I. Officer s.n. South Africa, Transkei, Arnold Q s.n. Iran, Bandar-Abbas, Wendelbo s.n. Tanzania, Dar-es-Salaam, Hedberg s.n. Tanzania, Dar-es-Salaam, Hedberg s.n. N.E. Ethiopia, Loutfy Boulos s.n. Kenya, Hepper & Jaeger s.n Kenya, Lake Naivasha, Jane Parish No. 3 Kenya, 1.6°S 36°E, Jane Parish No. 4 Kenya, Kilifi, Jane Parish No 5 Malawi, Domasi Valley, Blackmore 315A Belgium, Bot. Gdn. Univ. Liegea South Africa, TUL, Ngwenyi, Balsinhas 3394 South Africa, Pretoria, Lester s.n. Zimbabwe, IBPGR, TGR 1432 * = re-located to group C from group D; c * * * * XX ST * * * * * * * * * * * * * * * * * * * * * * * * * * * * * = seed normality/germination data only available; * denotes technique NTv numerical taxonomy of vegetative characters; XX crossability tests; ST pollen stainability tests. Interfertility studies 235 cross-pollinations involving accessions representing groups A, B, C and D were performed over four growing seasons. Each unit cross, its reciprocal cross and self-pollinations of parents were attempted and repeated several times wherever possible, according Table 3. Cross-combinations of Solanum incanum s.l. used in crossability study. female parent group BIRM/S.0859 0859 0859 1064 0859 0859 1781 1512 1793 1512 1512 1512 1512 1750 1750 1793 1793 1512 1692 1692 2465 1692 1692 2465 A A A A A A B C C C C C C C C C C C D D D D D D × male parent group BIRM/S.2023 2026 2028 2369 1398 1781 2369 2053 2053 0931 1512 1750 1793 1750 1793 1750 1793 2465 0931 1750 1750 1692 2465 2465 A A A A B B B B B C C C C C C C C D C C C D D D to flower availability. The success or failure of pollinations leading to development of fruit (fruit set), proportion of normal seeds per cross (seed normality), the proportion of successful germinations of seed produced per cross-combination (germination success), and the pollen stainability of F1 hybrids were investigated according to the methods employed in similar studies (e.g. Lester & Hasan 1991; Lester & Niakan 1986). Mean percentage values for each of the crossing programme results were calculated and crossing success values between 1 and 5 were calculated using a combination of fruit set, seed production and pollen stainability data (see Table 4 for details). Morphometric study of groups A and B All the morphological characters studied by Lester & Hasan (1991), Lester & Niakan (1986), Pearce (1975) and others were considered. During a preliminary study, erratic flowering presented difficulties with the availability of flowers for study. As a consequence, a total of 28 purely vegetative characters were chosen for measurement on each plant in accessions from Group A and Group B (Fig. 1; Table 5). Focus was given to these two groups as their distinction is particularly problematic. For each of six plants of five parental accessions, three habit, six stem and 19 leaf characters from each of three leaves, were measured. Each leaf was treated as a separate OTU (operative taxonomic unit), but all three leaves from any one plant had the 3 habit and 6 © The Board of Trustees of the Royal Botanic Gardens, Kew, 2012 404 KEW BULLETIN VOL. 67(3) Table 4. Solanum incanum s.l. crossing programme results. A A 41.4 20.8 B 25.0 73.0 C 0 - D 0 - 69.0 94.4 5 Male Parent 95.5 95.6 5 1 1 a = based on 1 value only; b Female Parent B 50.0 98.0a 13.0a 91.9a 5 25.0 0 1 0 1 0 1 C 41.5 6.1 1/1a b 3 16.7 80.0 10.0a 70.3 4 61.8 88.0 37.4 80.4 5 42.8 12.3 a 54.6a 4.0 4 D 81.8a 0 2 75.0a 0 2 75.5 50.8 42.7 69.6 4 57.1 65.0 20.0 83.1 5 = seedling died. Each block of five values is arranged as:- c.s. c.s. c.s. c.s. c.s. = = = = = 1 2 3 4 5 – – – – – pollinations made, no fruit set fruit set, no normal seed produced normal seed produced, no germination of F1 F1 produced from seed, low pollen stainability (0 – 74 %) F1 produced from seed, high pollen stainability (75 – 100 %) stem characters in common. The morphological data were processed and analysed using the CLUSTAN 2.1 numerical taxonomy package (Wishart 1982) as used in similar analyses (e.g. Lester & Hasan 1991). This generated phenograms by Euclidean Distance Squared and Cluster Analysis (Ward’s Method), and scatter diagrams by Principal Components Analysis (see Figs 2 and 3). Results and Discussion Interfertility Relationships (see Table 4) In the following discussion, the female parent group is the first cited of any pair of groups crossed together, whilst the male parent group (pollen donor) is the second. In terms of crossing success, there is general agreement between the present author’s results and those of Hasan (1989). The present study, however, was based on a more detailed analysis of A × B, B × A and C × A crosses, and examined a wide range of reciprocal crosses, not attempted by the former work. Within-group crosses Within-group crosses. Crossing success for crosses A × A, C × C and D × D were high, as expected; however, B × B cross-combinations yielded no viable seed, which was surprising, especially as parental group B accessions showed high pollen © The Board of Trustees of the Royal Botanic Gardens, Kew, 2012 fertility (Samuels 1994). Some B × B cross-combinations were self-pollinations, and the possibility of selfincompatibility arises here. This is rare in subgenus Leptostemonum (Whalen 1984), and neither Hasan (1989) working on related taxa, nor the present author (Samuels 1994) found any evidence for this phenomenon. Within-group crosses did not consistently demonstrate better fruit set than between-group crosses, which is surprising, since greater success might be expected between more closely related taxa. Furthermore, Daunay et al. (1999) also noted low fruit set in crosses within groups A and B and erratic fruit set in related species of Solanum. Wider Crosses Wider Crosses. Out of 32 attempts only one seed was produced (seed normality of 6.1%) by C × A cross-pollinations, but the resultant seedling died. Such very low seed normality values are probably an indication of the more distant taxonomic relationship between group C and group A plants. The seed normality value compares closely with 6.9% noted by Hasan (1989). Lester & Kang (1998) suggested that incongruity between the two parental genomes in C × A crosses leads to the breakdown of endosperm and sporophyte tissues. Hasan’s (1989) pollen stainability results for the C × A combination (77%) are surprising, since groups C SOLANUM INCANUM S.L. (SOLANACEAE) 405 Fig. 1. Leaf morphology characters used for morphometric study. and A are believed to be taxonomically distinct. Furthermore, any interfertility between the two groups is at best only one way, as a total of 16 attempts at crosspollination in the present study (and Hasan’s (1989) results) showed that the reciprocal cross failed to produce any fruit. Although some other wide cross-combinations (C × B, C × D and D × C) provided substantially fertile pollen (although less than 75% stainability) there may well be genomic incongruities (which parallel taxo- nomic disparity) whereby the fertility and viability of subsequent generations could be seriously impaired. In addition, cross-combinations involving allopatric groups (such as those above) are unlikely to take place in nature because of geographical isolation. Certain cross-combinations resulted in no fruit being set, whereas the reciprocal crosses did. E.g. no fruits were produced when flowers of groups A or B were used as female parents to cross with groups C or D as the male parents, whereas many of the reciprocal © The Board of Trustees of the Royal Botanic Gardens, Kew, 2012 406 KEW BULLETIN VOL. 67(3) Table 5. Definitions and abbreviations of gross vegetative morphology characters employed for numerical taxonomy (see Fig. 2). No. Character definitions (a) Plant habit characters 1. Plant height at highest point, cm 2. Plant breadth at broadest point, cm 3. Plant habit, height/breadth, ratio × 100 (b) Stem characters 4. Stem width at 10 cm from ground, mm × 10 5. Prickle number at half way up, all round the stem for 10 cm 6. Length of longest prickle, at half way up stem for 10 cm, mm × 10 7. Width of longest prickle, at half way up stem for 10 cm, mm × 10 8. Prickle length/prickle width, ratio × 100 9. Stem prickle shape (coded: 1 straight; 2 curved; 3 hooked) (c) General leaf lamina characters 10. Leaf blade attitude (coded: 1 erect; 3 semi-erect; 5 horizontal; 7 semi-pendent; 9 pendent) 11. Leaf lamina shape (coded: 1 elliptic; 3 elliptic ovate; 5 ovate; 7 ovate lanceolate; 9 lanceolate) (d) Individual petiole characters 12. Petiole length, mm 13. Petiole width, in middle, mm × 10 14. Petiole length/petiole width, ratio × 10 (e) Individual leaf lamina characters 15. Total number of prickles on the petiole 16. Leaf lamina length, cm × 10 17. Leaf width at widest part, cm × 10 18. Leaf lamina length/width, ratio × 100 19. Distance from the widest part of the leaf to the tip of the leaf, cm × 10 20. Widest to the tip/length, ratio × 100 21. Leaf blade tip angle 22. Leaf base angle 23. Total leaf lobe number 24. Length of the greatest leaf lobe, mm 25. Width of the base of the greatest leaf lobe, mm 26. Distance from the greatest leaf lobe to the tip, cm × 10 27. Prickle number on the upper lamina surface 28. Leaf undulation, height at highest undulation, mm Lf. undu crosses did set fruit. Hasan (1989) and Lester & Hasan (1991) also noted the complete failure of fruit set of group A flowers pollinated by groups C or D pollen donors. These results may be explained by a one-way prezygotic barrier. The hermaphrodite flowers of groups A and B have long, robust styles and large stigmas, whereas those of groups C and D tend to have thinner, shorter styles, with smaller stigmas. Pollen from group A or group B pollen is more likely to produce pollen tubes that penetrate the stylar tissue of the less robust, shorter styles of groups C and D, than the reverse. Such unilateral fertility barriers are of considerable significance in Solanum incanum and its allies. Baksh & Iqbal (1979, cited in Daunay et al. 1991) reported that many crosses between S. incanum and S. melongena were only successful with S. incanum as the female parent. Olet & Bukenya-Ziraba (2001) noted similar unidirectional success in crosses between S. incanum from Uganda (probably S. campylacanthum) and S. cerasiferum, in which fruit was produced only in crosses in which S. incanum was the female parent. Similar unilateral fertility barriers were noted by Daunay et al. (1991) and Daunay et al. (1999) in crosses between © The Board of Trustees of the Royal Botanic Gardens, Kew, 2012 Abbreviations Pl. ht. Pl. br. Pl. ht./Pl. br. St .wid. St. pric. St. pric. St. pric. St. pric. St. pric. no. len. wid. l/w sh. Lf. att. Lf. shape Pet. len. Pet. wid. Pet. len./wid. Pet. pric. no. Lf. len. Lf. wid. Lf. len./lf. wid. Lf. w-tip. Lf. w-tip./Lf. len. Lf. tip. ang. Lf. base ang. Lf. lobe no. Lf. lobe len. Lf. lobe wid. Lf. lobe-tip Lf. pric. no. Lf. undul. S. melongena and other Solanum species. Furthermore, this phenomenon is well-known and relatively common in several genera in the Solanaceae, including Solanum, Nicotiana and Petunia (Onus & Pickersgill 2004). Interfertility of groups A and B InterfertilityofgroupsAandB. Prolonged seed dormancy, erratic seed germination, unilateral pre-zygotic and complex postzygotic fertility barriers seem to have a strong influence on fertility relationships in Solanum incanum s.l. Therefore, two groups must show full two-way crossing success (i.e. crossing success score of “5”) to be safely regarded as fully interfertile and thereby members of the same biological species. On this basis, only groups A and B show full interfertility and are conspecific. A study performed by Sakata et al. (1991) involving the analysis of chloroplast DNA (cpDNA) from plants in accessions of groups A and B, showed that these groups were closely related. This was later confirmed by Mace et al. (1999) using AFLP analysis. This close genetic relationship tallies with the results of the interfertility study. SOLANUM INCANUM S.L. (SOLANACEAE) 407 Fig. 2. Dendrogram for analysis of 28 vegetative characters of 3 leaves of 6 plants in 5 parental accessions of Solanum incanum s.l., groups A and B. (Accessions are as follows: 0859, 2026, 2028 — group A; 1398, 1781 — group B). Morphological Variation in Solanum incanum s.l Differentiation of groups A and B DifferentiationofgroupsAandB. In Fig. 2 the various individual accessions were shown to be more or less distinctive. However, separation between group A accessions on the one hand, and group B on the other, was unclear, indicating that their morphological distinction is unclear. For example, group B accession S.1781 showed considerable phenetic affinity with group A accession S.2028. A similar situation in Fig. 3 shows that OTUs from individual accessions were © The Board of Trustees of the Royal Botanic Gardens, Kew, 2012 408 KEW BULLETIN VOL. 67(3) Fig. 3. Principal Components Analysis of 28 vegetative characters of 3 leaves of 6 plants in 5 parental accessions of Solanum incanum s.l. more or less tightly clustered, but that group A and group B accessions overlap, with S.2028 and S.1781 on the one hand, and S.2026 and S.1398 on the other, partly intermerging. Again, this suggests a lack of morphological distinction between the respective accessions. Lester & Hasan (1991) also found that groups A and B were morphologically very close, and that several group A and group B accessions showed considerable morphological similarity. © The Board of Trustees of the Royal Botanic Gardens, Kew, 2012 Variation in Group A Variation in Group A. Fig. 3 shows that group A OTUs (S.0859, S.2026 and S.2028) were generally quite dispersed within their clusters, indicating greater variability, whereas group B OTUs (S.1398 and S.1781) clustered more tightly, indicating greater homogeneity. Observations by Gianoli & Hannunen (2001), Olet & Bukenya-Ziraba (2001) and Robinson (1993) substantiate the view that morphological variation in group A plants is considerable and covers a broad range of vegetative and reproductive characters. SOLANUM INCANUM S.L. (SOLANACEAE) The wider variation of one of the group A accessions, S.0859, is well demonstrated in Fig. 2, as the majority of the S.0859 OTUs were clustered together at D2 = 12.4, compared with much tighter clustering in other accessions (D2 = 8 or less). This accession also seems to be quite dissimilar to other group A (and B) accessions, as the majority of its component plants only linked up with other accessions at D2 = 29.0. Furthermore, one plant (S.0859-1) was so dissimilar that it clustered with different accessions. The cluster analysis in Fig. 3 confirmed that accession S.0859 is quite distinctive, whereby it formed a cluster that is quite removed from the other four accessions. This particular accession displays a range of morphological characteristics that lies near the boundaries of typical group A plants. Sakata & Lester (1994) in their study of cpDNA reported the Ugandan accession S.0859 and several other east African group A accessions to be distinct from 409 the main mass of other group A (and B, C and D) accessions. This diversity in cpDNA suggests that group A is a more ancient group than groups B, C, or D (Sakata & Lester 1994) which all have homogeneous cpDNA profiles. Group A genotypes may have become more diverse through the development of more numerous mutations over a greater period of time. This greater genetic diversity in group A parallels its greater morphological diversity. Differentiation of groups C and D Groups C and D accessions were investigated using morphometric techniques by Jayawickrama (1990). His study revealed a clear distinction between groups C and D; this was later confirmed by the findings of Lester & Hasan (1991) and Samuels (1996). Furthermore, the distinction evident between group D chloroplast DNA and that from group C was demonstrated by Sakata & Lester (1994). Key to the species and subspecies 1. Shrubs or sub-shrubs, less than 2 m high; branches robust, up to 7 mm diam., densely tomentose with stellate hairs; always armed on shoots, leaves, inflorescence axes, and calyces and pedicels of hermaphrodite flowers; leaf lamina ovate; corolla violet, purple, or white 2. Leaf lamina narrowly ovate, margin repand; inflorescence 1 – 5-flowered; corolla white (rarely violet), 2.5 – 3 cm across; fruiting calyx manifestly robust, heavily armed, lobes strongly reflexed; berry 3.5 – 4.5 cm diam . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . S. lichtensteinii 2. Leaf lamina broadly ovate, margin subentire to repand; inflorescence 1 – 15-flowered; corolla violet to purple, 2.5 – 3 cm across; fruiting calyx enlarged, ± armed, lobes slightly reflexed; berry 3 – 3.5 cm diam . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . S. incanum 1. Shrubs, sub-shrubs or herbaceous perennials up to 2 m or more high; branches approx. 4 mm diam., sparsely tomentose with stellate hairs; armed or unarmed; leaf lamina lanceolate to elliptic; corolla violet or purple 3. Leaf lamina ovate-lanceolate or lanceolate, margin subentire to lobed; inflorescence 3 – 15 (– 50)flowered, 1 – 5 (–15) lowest flowers hermaphrodite; corolla violet or purple, 2 – 3.5 cm across; berry 2.5 – 3.5 cm diam . . . . . . . . . . . . . . . . . . . . S. campylacanthum subsp. campylacanthum 3. Leaf lamina elliptic, margin entire to subentire; inflorescence 3 – 12-flowered, lowest flower only (more rarely 1 – 3 lowest flowers) hermaphrodite; corolla violet, 1.5 – 3 cm across; berry 2 – 2.5 cm diam. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . S. campylacanthum subsp. panduriforme Range of variation and distribution of the taxa http://www.ipni.org/urn:lsid:ipni.org:names:77120311-1 Solanum campylacanthum Hochst. ex A. Rich. subsp. campylacanthum Solanum panduriforme Drège ex Dunal in DC., Prodromus 13 (1): 370 (de Candolle 1852). Highly polymorphic group of more or less tomentose, more or less armed shrubs, up to 2 m or more high, with ovate-lanceolate to lanceolate, more-or-less lobed leaves. Flowers up to 50 in each inflorescence, violet or purple, often with several to many hermaphrodite flowers. Infructescence of several to many fruits, up to 3.5 cm diam. Finely tomentose, sparsely armed or unarmed shrubs, sub-shrubs or herbaceous perennials; up to 2 m or more high; with elliptic, entire to sub-entire leaves; up to 12 violet flowers in each inflorescence, usually only one flower (more rarely up to 3) hermaphrodite; infructescence of up to 3 fruits, up to 2.5 cm diam. DISTRIBUTION. Centred around tropical eastern Africa DISTRIBUTION. Centred on eastern and south-eastern Africa. and extending across to Madagascar. Solanum incanum L. Solanum campylacanthum subsp. panduriforme (Drège ex Dunal) J. Samuels stat. nov. Densely tomentose, armed perennial shrubs; up to 2 m high; with broadly ovate, sub-entire to repand leaves; up to 15 purple or violet flowers in each simple inflorescence, © The Board of Trustees of the Royal Botanic Gardens, Kew, 2012 410 usually only one flower (more rarely up to 3) hermaphrodite; infructescence of up to 3 fruits, up to 3.5 cm diam. DISTRIBUTION. Across north-eastern Africa, extending southwards into Somalia and Kenya, westwards to Senegal, across the Middle East, and as far eastwards as northern Pakistan and northern India. Solanum lichtensteinii Willd. Densely tomentose, armed shrubs or sub-shrubs, 0.5 – 2 m high; with narrowly ovate, repand leaves; up to 5 white (or more rarely violet) flowers in each simple inflorescence, usually only one flower (more rarely up to 3) hermaphrodite; infructescence of up to 3 fruits, up to 4.5 cm diam. DISTRIBUTION. Across southern East Africa and much of southern Africa. Acknowledgements Firstly, my gratitude for expert advice and guidance given by the late Dr R. Lester and the late Prof. J. Hawkes OBE must be recorded here. My thanks also go to Mr A. Esquilant, Dr S. Hasan, Mr H. Jayawickrama and Dr J. Kang for their practical assistance. I would also like to thank Dr M. Vorontsova for assistance at K, Mr E. Thewlis who typed the original manuscripts and Mr J. Tennant for his encouragement. I am indebted to the many herbaria and plant collectors who provided the dried plant specimens used in this study. Lastly, I am grateful to the Annals of Botany journal for their research fellowship which funded the early part of this study, and also to the National Science Foundation (USA) for funding recent research on the PBI Solanum: a Worldwide Treatment project at K and BM. References Bitter, G. (1913). Solana Africana Part I. Bot. Jahrb. Syst. 49: 560 – 569. ——— (1917). Solana Africana Part II. Bot. Jahrb. Syst. 54: 416 – 507. ——— (1921). Solana Africana Part III. Bot. Jahrb.Syst. 57: 248 – 286. ——— (1923). Solana Africana Part IV. Repert. Spec. Nov. Regni Veg. 16: 1 – 320. Dammer, U. (1905). Solanaceae Africanae. Repert. Spec. Nov. Regni Veg. 38: 57 – 60. ——— (1906). Solanaceae Africanae I. Repert. Spec. Nov. Regni Veg. 38: 176 – 195. ——— (1912). Solanaceae Africanae II. Repert. Spec. Nov. Regni Veg. 48: 236 – 260. ——— (1915). Solanaceae Africanae III. Repert. Spec. Nov. Regni Veg. 53: 325 – 352. © The Board of Trustees of the Royal Botanic Gardens, Kew, 2012 KEW BULLETIN VOL. 67(3) Daunay, M.-C., Dalmon, A. & Lester R. N. (1999). Management of a Collection of Solanum Species for Eggplant (Solanum melongena) Breeding Purposes. In: M. Nee, D. E. Symon, R. N. Lester & J. P. Jessop (eds), Solanaceae IV: Advances in Biology and Utilization, pp. 369 – 383. Royal Botanic Gardens, Kew. ———, Lester, R. N., Dalmon, A., Ferri, M., Kapilima, W., Poveda-Aguilar, M. M. & Jullian, E. (1998). The use of wild genetic resources for eggplant (Solanum melongena) breeding. II: crossability and fertility of interspecific hybrids. In: A. Palloix & M.-C. Daunay (eds), Proceedings of the Xth Eucarpia Meeting on Genetics and Breeding of Capsicum and Eggplant, 7 – 11 Sept 1998, Avignon, France, pp.19 – 24. INRA, Paris. ———, ———, Gebhardt, C., Hennart, J. W., Jahn, M., Frary, A. & Doganlar, S. (2001). Genetic resources of eggplant (Solanum melongena) and allied species: a new challenge for molecular geneticists and eggplant breeders. In: R. G. van den Berg, G. W. M. Barendse, G. M. van der Weerden & C. Mariani (eds), Solanaceae V: Advances in Taxonomy and Utilization, pp. 251 – 274. Nijmegen University Press, Nijmegen. ———, ——— & Laterrot, H. (1991). The use of wild species for the genetic improvement of brinjal (Solanum melongena L.) and tomato (Lycopersicon esculentum Mill.). In: J. G. Hawkes, R. N. Lester, M. Nee & N. Estrada (eds), Solanaceae III: Taxonomy, Chemistry, Evolution, pp. 389 – 412. Royal Botanic Gardens, Kew, de Candolle, A. P. (1852). Prodromus systematis naturalis regni vegetabilis 13. Sumptibus Victoris Masson, Parisis. Gianoli, E. & Hannunen, S. (2001). Plasticity of leaf traits and insect herbivory in Solanum incanum L. (Solanaceae) in Nguruman, S.W. Kenya. Afr. J. Ecol. 38: 183 – 187. Hasan, S. M. Z. (1989). Biosystematic Study of Solanum melongena L. in Asia and Africa. Unpublished Ph.D. Thesis, University of Birmingham. Jayawickrama, H. D. (1990). Study of the Diversity of Solanum incanum agg. Unpublished M.Sc Thesis, University of Birmingham. Jaeger, P.-M. L. (1985). Systematic Studies in the Genus Solanum in Africa. Unpublished Ph.D. Thesis, University of Birmingham. ——— & Hepper, F. N. (1986). A review of the genus Solanum in Africa. In: W. G. D’Arcy (ed.), Solanaceae: Biology and Systematics, pp. 41 – 55. Columbia University Press, New York. Lester, R. N. (1998). Genetic resources of capsicums and eggplants. In: A. Palloix & M.-C. Daunay (eds), Proceedings of the Xth Eucarpia Meeting on Genetics and Breeding of Capsicum and Eggplant, 7 – 11 Sept. 1998, Avignon, France, pp. 25 – 30. INRA, Paris. ——— & Daunay, M.-C. (2003). Diversity of African vegetable Solanum species and its implications for a SOLANUM INCANUM S.L. (SOLANACEAE) better understanding of plant domestication. In: H. Knuppfler & J. Ochsmann (eds), Rudolf Mansfield and Plant Genetic Resources; Schriften zu Genetischen Ressourcen (DEU) Symposium Dedicated to the 100th Birthday of Rudolf Mansfield. Gatersleben (DEU) 22: 137 – 152. ——— & Hasan, S. M. Z. (1991). Origin and domestication of the brinjal eggplant, Solanum melongena from S. incanum in Africa and Asia. In: J. G. Hawkes, R. N. Lester, M. Nee & N. Estrada (eds), Solanaceae III: Taxonomy, Chemistry, Evolution, pp. 369 – 387. Royal Botanic Gardens, Kew. ———, Hawkes, J. G., Daunay, M.-C., van der Weerden, G. M. & Barendse, G. W. M. (2001). The sources, successes and successors of the Birmingham University Solanaceae Collection (1964-2000). In: R. G. van den Berg, G. W. M. Barendse, G. M. van der Weerden & C. Mariani (eds), Solanaceae V: Advances in Taxonomy and Utilization, pp. 391 – 412. Botanical Garden of Nijmegen, Nijmegen University Press. ——— & Kang, J. (1998). Embryo and endosperm function and failure in Solanum species and hybrids. Ann. Bot. 82 (4): 445 – 453. ——— & Niakan, L. (1986). Origin and domestication of the scarlet eggplant, Solanum aethiopicum, from S. anguivi in Africa. In: W. G. D’Arcy (ed.), Solanaceae: Biology and Systematics, pp. 433 – 456. Columbia University Press, New York. Levin, R. A., Myers, N. R. & Bohs, L. (2006). Phylogenetic relationships among the “spiny solanums” (Solanum subgenus Leptostemonum, Solanaceae). Amer. J. Bot. 93 (1): 157 – 169. Mace, E. S., Lester, R. N. & Gebhardt, C. G. (1999). AFLP analysis of genetic relationships among the cultivated eggplant, Solanum melongena L., and wild relatives (Solanaceae). Theor. Appl. Genet. 99: 626 – 633. Mohammad, A., Baksh, S. & Iqbal, M. (1994). Cytogenetic studies on on the F1 hybrid Solanum incanum 411 X S. melongena var. American Wonder. Cytologia 59 (4): 433 – 436. Olet, E. A. & Bukenya-Ziraba, R. (2001). Variation within the Solanum incanum complex in Uganda and its relationship with Solanum cerasiferum. In: R. G. van den Berg, G. W. M. Barendse, G. M. van der Weerden & C. Mariani (eds), Solanaceae V: Advances in Taxonomy and Utilization, pp. 97 –108. Botanical Garden of Nijmegen, Nijmegen University Press. Onus, A. N. & Pickersgill, B. (2004). Unilateral incompatibility in Capsicum (Solanaceae): occurrence and taxonomic distribution. Ann. Bot. 94 (2): 289 – 295. Pearce, K. G. (1975). Solanum melongena L. and Related Species. Ph.D. Thesis, University of Birmingham. Robinson, R. W. (1993). Variability of Solanum incanum in Kenya. Solanaceae Newslett. 3 (3): 4 – 7. Sakata, Y. & Lester, R. N. (1994). Chloroplast DNA diversity in Eggplant (Solanum melongena) and its related species S. incanum and S. marginatum. Euphytica 80: 1 – 4. ———, Nishio, T. & Mathews, P. J. (1991). Chloroplast DNA analysis of Eggplant (Solanum melongena) and related species for their taxonomic affinity. Euphytica 55: 21 – 26. Samuels, B. J. (1994). Solanum incanum sensu lato (Solanaceae): A Taxonomic Survey. M.Sc. Thesis, University of Birmingham. ——— (1996). Solanum incanum sensu lato (Solanaceae): Taxonomy, Phylogeny and Distribution. Ph.D Thesis, University of Birmingham. Whalen, M. D. (1984). Conspectus of species groups in Solanum subgenus Leptostemonum. Gentes Herb. 12: 179 – 282. Wishart, D. (1982). CLUSTAN User Manual. Edinburgh University Library Program Unit, Edinburgh. © The Board of Trustees of the Royal Botanic Gardens, Kew, 2012