2. Solanum campylacanthum A.Rich, Tent. Fl. Abyss. 2: 102. 1850.

Solanum bojeri Dunal, Prodr. [A.P. de Candolle] 13(1): 344. 1852.

Solanum panduriforme Drège ex Dunal, Prodr. [A.P. de Candolle] 13(1): 370. 1852, as ‘‘ panduraeforme ’’.

Solanum delagoense Dunal, Prodr. [A.P. de Candolle] 13(1): 349. 1852.

Distribution. Ubiquitous weed of low altitudes in Southern and Eastern Africa: roadsides, abandoned cultivation, savanna, bushland, dunes, forest edges etc.; usually 0–2000 m, but has been recorded up to 2300 m elevation.

Solanum campylacanthum is extremely widespread and variable (75 heterotypic synonyms [25]), particularly with respect to leaf morphology (Fig. 2) but flowers are relatively uniform throughout its range (Fig. 1A). The vast majority of wild egglant relatives collected in Africa belong to this species, which is commonly and incorrectly called ‘‘ Solanum incanum’’ (see discussion of S. incanum below). Our concept of this species corresponds to ‘‘ Solanum incanum group A’’ and ‘‘ Solanum incanum group B’’ of Daunay et al. [23]; ‘‘group B’’ comprises those plants with narrower leaves from the southern part of the species distribution that have been recognised by some as S. delagoense, S. panduriforme or as infraspecific taxa based on those epithets [31]. From our examination of many herbarium specimens throughout Africa we conclude that this variation represents a north-south cline with leaf shape narrower in more southern populations. The variation is continuous and we do not think it warrants taxonomic recognition at either the specific or infraspecific level. Solanum campylacanthum can form dense stands of monomorphic plants through vegetative reproduction by underground stems; this can lead to the impression that variation is at a population rather than an individual level. Samuels [31] showed that ‘‘A’’ and ‘‘B’’ were fully interfertile, and thus classified them as subspecies.

doi:10.1371/journal.pone.0057039.t002

doi:10.1371/journal.pone.0057039.t003

Crossability of S. campylacanthum with other members of the group has proved difficult [40] and one-way pre-zygotic barriers have been suggested as the reason for this failure of fruit set in crosses with S. incanum and S. lichtensteinii [31]. It is possible that some of these difficulties could be due to ploidy differences within S. campylacanthum. Anaso and Uzo [41,42] reported tetraploidy in S. campylacanthum from Nigeria (reported as S. incanum); their study illustrates the problems with inconsistent application of names in this group, they compared wild tetraploid ‘‘S. incanum’’ (= S. campylacanthum) with cultivated diploid ‘‘S. incanum’’ (= S. aethiopicum L., the unrelated scarlet eggplant). Fortunately they illustrated the plants used in the studies and identification of their material is clear even in the absence of vouchers. There has been an assumption that all relatives are, like the eggplant itself, diploid and earlier cytogenetic studies have not explicitly cited vouchers for counts of ‘‘S. incanum’’ so verification of identities of taxa counted is difficult. Ploidy level variation in Solanum is most common in the potatoes, where the cultivated potato has a number of ploidy forms and wild species vary from diploid to hexaploid [12], and in the Morelloid clade (S. nigrum L. and its relatives; see [43]), but it also occurs in some species of the Leptostemonum clade, particularly in widespread weedy species such as S. elaeagnifolium Cav. [44].