Solanum terminale Forssk.

Solanum terminale is the most widespread and common species of the ANS clade on the African mainland, and is often recognised as three distinct entities; S. nakurense, S. terminale s.s. and S. welwitschii (e.g., Edmonds 2012). Bitter (1917, 1922) recognised 20 species and 23 infraspecific taxa, while Heine (1960) amalgamated all these into a single species with three infraspecific taxa in western Africa, and suggested that a similar course of action needed to be taken throughout the continent. We are recognising these populations as a single species because the characters used to distinguish them are continuously variable and field observations indicate that habitat plays a significant role in overall morphology (see below).

Solanum terminale is not broadly sympatric with any other species of the ANS clade, although in eastern South Africa S. africanum also occurs, but not strictly sympatrically; S. runsoriense also occurs in the same regions as S. terminale but usually at higher elevations. Solanum terminale is easily distinguished from others in the clade by its flowers with densely papillate abaxial corolla lobes, flowers clustered at the ends of inflorescence branches (of varying length) and bright red berries. Populations from western Africa tend to have shorter inflorescence branches (see Fig. 25D), tightly connate anthers and elongate berries and have been recognised as S. welwitschii, while those from eastern Africa have more open inflorescences (see Fig. 25B, C), more loosely connate anthers and globose berries and have been called S. terminale.  Small plants with few-flowered inflorescences primarily from eastern Africa have been called S. nakurense (Fig. 25A). The leaves of S. terminale are usually pubescent to some degree, but are occasionally completely glabrous (e.g., the type of S. welwitschii). Dense tufts of trichomes in the vein axils (domatia) do not occur in S. terminale, but the pubescence is often denser near the midvein.

In order to determine if there was any geographical component to the morphogical variation seen in S. terminale, we measured the key characters used to distinguish S. terminale from S. welwitschii in a range of specimens from across continental Africa. Long thin buds that open to flowers with fused anthers, and spicate inflorescences with short to non-existent branches are thought to be typical of S. welwitschii (e.g., Edmonds 2012). Neither character segregated clearly into two classes in our simple analysis (Figure 27). Bud shape is similarly distributed across longitude, but the longest, thinnest buds occur in western Africa (see point at top left-hand corner of Fig. 27A). Spicate inflorescences are somewhat more common in the wet forests of western Africa (Fig. 27B). The type of Solanum welwitschii var. laxepaniculatum (Mildbraed 5265 - http://plants.jstor.org/stable/viewer/10.5555/al.ap.specimen.hbg511510), however, is a good example of a “terminale” type inflorescence in western Africa. Anthers on specimens from wet forests of western Africa (red dots in Fig. 27) were more often fused than those from drier forests of eastern Africa. Anther connation may be a way of protecting pollen in damp, humid environments, but this remains to be tested. The characters used to distinguish the taxa we here recognise as the single species S. terminale appear to be the extremes of continuous variation, and perhaps represent fixation of particular morphologies in certain populations. Field study of these plants coupled with molecular analysis at the population level will be essential to untangle the complex morphological variation with S. terminale.

In their catalogue of Forsskål’s plant collections Hepper and Friis (1994) cited both specimens in C as “type” of S. terminale; of these two sheets Herb. Forsskål 419 (C10003105) has the larger inflorescence with more flowers and we select it here as the lectotype. They considered Herb. Forsskål 406 (C10003106) a duplicate (Hepper and Friis 1994); we concur and consider it an isolectotype.

Both Udo Dammer and Georg Bitter described many infraspecific taxa in what we now recognise as the widely distributed and polymorphic S. terminale. Both botanists worked in Berlin before the Second World War and it is probable that much of the material upon which they based their descisions is no longer extanct (Vorontsova and Knapp 2010). Bitter (1917) acknowledged his lack of ability to travel to other herbaria due to military service and war as a short-coming of his treatment of this polymorphic complex. We have searched extensively for duplicates of this material throughout the course of the PBI Solanum project, and in some cases (e.g., see S. plousianthemum var. subtusbarbellatum above) have encountered duplicates, while in many, if not most, others we have not been successful. Rather than neotypifying these many infraspecific names we cite type collections in the hope that duplicates will eventually be encountered in the collections we have not yet been able to visit. Dammer (1906, 1912, 1915) never cited particular herbaria in association with specimens he used in describing new taxa (see Vorontsova and Knapp 2016), Bitter (1917, 1922) occasionally did so. Where he cited herbaria in association with type specimens we have indicated this, otherwise no herbarium was cited and we chose lectotypes from duplicates we have seen. Our choices of lectotypes are based on the quality of the particular sheet selected. When syntypes were cited, our choice has been made based on the number of herbaria in which particular syntype collections now are found; where more duplicates of a collection are extant, we selected that as the lectotype.
In his treatment of the species of the ANS clade in Solana Africana Bitter (1917) re-circumscribed many of Dammer’s (1906, 1915) species and excluded some of the specimens cited by Dammer as types of new infraspecific taxa. He explicitly cited types for Dammer’s names, something that later authors did not realise, resulting in some superfluous later lectotypifications (e.g., Edmonds 2012). Solanum leucanthum was first described by Dammer (1906) then explicitly redescribed as “S. leucanthum Bitter & Dammer n.sp.” by Bitter (1917).  Dammer (1906) originally cited two syntypes from Cameroon, Mildbraed 894 and Mildbraed 1431a, while Bitter restricted his concept to only Mildbraed 894. Bitter (1917) recognised Mildbraed 1431a as a distinct species S. ruandae. We have found no duplicates of either of these collections that were probably in Berlin and destroyed.

We have not selected a neotype for Dammer’s (1906) S. suberosum; of the seven syntype collections cited we are certain that duplicates will be found eventually. If not, the description and comments on the corky stem morphology will be easy to match, and a recent and widely distributed collection from Cameroon should be selected as a neotype.

Dammer’s (1906) illegitimate name S. laurentii (not S. laurentii De Wild., 1905) was described using a single specimen collected by “Laurent” from an unspecified locality in “Congo”. This sheet was in all likelihood at B and is no longer extant; the lack of specifics makes duplicates very difficult to trace. Durand and Durand (1909: 394) coined the replacement name S. subcoriaceum, but again did not specify a herbarium, sheet or locality; nor did they specify which Laurent, Émile or Marcel, was the collector of the type. We did not find a sheet annotated with this name in BR. We have selected a neotype in the Brussels herbarium collected by Marcel Laurent in 1905 (BR0000014801273) for S. laurentii Dammer (=S. subcoriaceum De Wild. & Durand) that matches the description in Dammer’s (1906) protologue.
Dammer (1906) described S. plousianthemum citing three collections made by C. Holst in Tanzania (Holst 232, 3731, 8927). Bitter (1917) later cited Holst 232 and Holst 3731 as elements in his redefinition of S. plousianthemum, and effectively lectotypified the name by stating “letztere Nr. der Typus zu Dammer’s originaldiagnose” – the latter number [Holst 3731] the type of Dammer’s original diagnosis). He cited Holst 8927 as the type for his variety conglutinans without citing a herbarium. Edmonds (2012) superfluously lectotypified S. plousianthemum with the Kew duplicate of Holst 8927, none of the other collections cited in Dammer’s protologue have been found. We select the Paris duplicate (P00341605, with flower and fruit) of Holst 8927 as the lectotype of Bitter’s var. conglutinans.

Similarly C.H. Wright (1894, 1906) rarely cited herbaria in new taxon descriptions, but because he worked at the herbarium of Kew, we have designated lectotypes at that institution unless a much higher quality specimen was available elsewhere (e.g., S. welwitschii where of the two syntypes Welwitsch 6081, 6098 the BM sheet of the better documented syntype Welwitsch 6098 has both flowers and fruits).  In some cases (e.g., Wright 1897) the herbarium at Kew was mentioned in the introductory material in the publication, thus Scott-Elliot 6800 at Kew (K000096900) is the holotype of S. nakurense.  However, in the case of S. lykipiense where Wright (1894) states “ the specimen from which this description is drawn is about 6 in. long, but sufficient to show that the species is a very distinct one”, we have lectotyified the name with the K sheet which answers to this description rather than taking it as the holotype because there is no evidence that there were not several sheets with small fragments. Two collections, Schweinfurth 3498 and Mann 274 were cited in the protologue of S. welwitschii var. strictum (Wright 1894: 127); we select the Kew duplicate (K000414054) of Mann 274 as the lectotype because another sheet is in Paris and we have only seen one duplicate of Schweinfurth 3498. For S. phytolaccoides, we have selected Schimper 310 as the lectotype, rather than the other collection (Johnston s.n.) cited, so that can be used as the lectotype of S. plousianthemum var. connmixtum. In herbaria where we have examined duplicates of Schimper 310 two different collection dates have been written on sheets, but the sheet in US (labelled as “ex Berol.”) has both dates; all duplicates of Schimper 310 are here cited as isolectotypes, regardless of collection date.

De Wildeman and Durand (1899), Durand and Durand (1909) and De Wildeman (1922) worked at Brussels, but again did not cite herbaria in their species descriptions. For these names we have selected lectotypes at BR with the best quality duplicate available; for example, the lectotype for S. lianiforme is not the one without the original label, the sheet we selected (BR0000008993489) has label in De Wildeman’s hand with the species name and is a flowering specimen.  The late Richard Lester affixed holotype labels to specimens at BR (e.g., Thonner 96 the type collection of S. symphyostemon), and his “designations” were followed by Edmonds (2012). We have, in general, used these sheets as the lectotypes, unless duplicates in BR are of better quality.

Two collections were cited in the description of S. balboanum (Chiovenda 1935); Balbo 425 and Balbo 579. We have selected Balbo 579 (FT003031) because it has both flowers and fruits.