The Systematics of Solanum Section Allophyllum (Solanaceae)

Three species of Solanum are included in a new section, Solanum sect. Allophyllum. Two of the species, S. allophyllum and S. mapiriense, have been previously known; the third, S. morellifolium Bohs, is described as new. The combination of characters exhibited by members of this section, such as unbranched hairs, 2-3-leaved sympodial units, simple leaves with decurrent bases, and tapered anthers, is unique to this group and does not agree well with any existing subgenus of Solanum. A group of three species first attracted my attention during monographic work on the solanaceous genus Cyphomandra Sendtn. Solanum allophyllum (Miers) Standl., the most widespread and best-known species of the section, had been placed by various workers in Cyphomandra and in Solanum. Determination of the true affinities of this species entailed an examination of the validity of the characters separating the two genera. An account of the morphology of this species and its resultant exclusion from Cyphomandra is given in a separate paper (Bohs, 1989). Bitter (1914) first noted the similarities between S. allophyllum (his S. ellipsoideibaccatum Bitter) and S. mapiriense Bitter (his S. phytolaccoides (Rusby) Bitter), but he did not create a formal taxonomic category to include them. Child (1984) erected sect. Allophylla to contain these two species and regarded the new section as belonging to Cyphomandra rather than to Solanum. He also surmised -that Cyphomandra chlorantha Rusby belonged to sect. Allophylla, but Bohs (1986) confirmed that it belongs in Cyphomandra and that it has no close relationship to S. allophyllum or S. mapiriense. A recent analysis of S. allophyllum (Bohs, 1989) indicates that this species should be placed in Solanum instead of Cyphomandra. Accordingly, sect. Allophylla is here removed from Cyphomandra and placed in Solanum. All three species of Solanum sect. Allophyllum share the following combination of characters: (1) upright herbs or shrubs (up to 2 m tall); (2) 2-3leaved sympodial units; (3) vegetative parts densely covered with "sand-punctae" (idioblasts containing crystal sand); (4) membranaceous leaf blades, usually with decurrent bases and winged petioles; (5) unbranched, few-flowered, relatively short inflorescences; (6) flowers with usually white or greenish white membranaceous corollas; (7) stamens with glabrous filaments, distally tapered anthers, and very small terminal pores; (8) glabrous, filiform styles with small truncate stigmas; and (9) glabrous, globose to ellipsoidal fruits with obtuse apices. Additional characters may further distinguish the section, but these are as yet imperfectly known for all three species. For instance, S. allophyllum and S. mapiriense have light-colored corollas and anthers both with darker spots at the base, but such spots have not been observed in S. morellifolium. Also, the fruits of S. allophyllum have a distinctive color and shape, being white to light orange marked with dark green or purple longitudinal stripes and being ovoid to ellipsoidal in outline but flattened on opposite sides (Bohs, 1989). The color and shape of mature fruits of S. mapiriense and S. morellifolium are unknown. Most of the character states that define sect. Allophyllum are found in many other sections of Solanum and are probably plesiomorphic. These include (1) herbaceous habit; (2) exclusively simple, unbranched hairs; (3) absence of prickles; (4) presence of sand-punctae; (5) cuneate and decurrent leaf bases; (6) glabrous filaments; (7) glabrous, filiform styles with small truncate stigmas; (8) gla' I thank the herbaria listed for loans of specimens, M. Nee, W. G. D'Arcy, and S. Knapp for helpful discussions and specimens, Julie Mitchell for the illustration of S. morellifolium, and G. McPherson for helping me search, albeit unsuccessfully, for S. allophyllum in Panama. 2 Department of Biology, University of Utah, Salt Lake City, Utah 84112, U.S.A. ANN. MISSOURI BOT. GARD. 77: 398-409. 1990. This content downloaded from 169.237.45.23 on Wed, 02 Mar 2016 18:56:20 UTC All use subject to JSTOR Terms and Conditions Volume 77, Number 2 Bohs 399 1990 Systematics of Solanum sect. Allophyllum brous, obtuse fruits; (9) stone granules in the fruits; and (10) small, flattened seeds. Probable derived character states manifested by members of the section include (1) tapered anthers with small terminal pores; (2) 2-3-leaved sympodial units; and (3) self-compatible breeding system (known only for S. allophyllum). The evolutionary significance of the laterally compressed fruits seen in S. allophyllum (and possibly present in the other two species) is uncertain, but compressed fruits occur in other subgroups of Solanum, such as the subgenera Potatoe (G. Don) D'Arcy and Bassovia (Aubl.) Bitt. The distribution of flattened fruits among other solanaceous genera is unknown to me. Perhaps this character is relatively insignificant taxonomically and may be related to the enlargement of the ovary wall and septum during fruit development. In summary, sect. Allophyllum seems to combine a host of primitive features with a few seemingly derived character states, most notably tapered anthers. PLACEMENT OF THE SECTION WITHIN SOLANUM Solanum is one of the largest genera of flowering plants, containing an estimated 1,500 species (D'Arcy, 1979; Hunziker, 1979). Substantial taxonomic problems still exist with regard to Solanum and related genera. At present, Solanum is loosely defined and consists of plants with poricidally dehiscent anthers that lack the specialized features of associated segregate genera, such as enlarged anther connectives in Cyphomandra Sendtn., sterile anther beaks in Lycopersicon Mill., and distinctive calyx morphology in Lycianthes (Dun.) Hassl. Although each of these segregate genera may be monophyletic, monophyly has not been established for Solanum. Poricidal anther dehiscence may have evolved more than once in Solanum and its allies, and may thus not be a synapomorphy uniting these taxa. Certainly our understanding of phylogenetic relationships within Solanum is hampered by the lack of basic taxonomic and morphological information for many subgenera and sections within the genus. It seems inevitable that Solanum will be broken up into smaller monophyletic genera, but taking such a step now would be premature based on our current state of knowledge. The infrageneric classification of Solanum and the boundaries between Solanum and its related genera have not been firmly established, and the taxonomy of this group is likely to remain in a state of flux for many years to come. The placement of sect. Allophyllum in Solanum is enigmatic. No existing group in Solanum has the combination of tapered anthers, simple hairs, upright habit, simple leaves, and spinelessness seen in sect. Allophyllum. In fact, the anthers of this group resemble those of some species of Lycianthes, but sect. Allophyllum does not have the truncate calyx with subapical lobes, reduced inflorescences, and large stone cell aggregates in the fruits that distinguish Lycianthes (D'Arcy, 1986). Section Allophyllum may represent a new genus with characters intermediate between Solanum and Lycianthes, but no synapomorphies are known that distinguish sect. Allophyllum from Solanum. Given the possible polyphyly, or at least paraphyly, of Solanum, sect. Allophyllum, differentiated on the basis of a combination of characters but with no obvious synapomorphies, must be placed in Solanum. Further work on generic limits between Solanum and its allies will undoubtedly shed light on the proper generic placement of this section. Additional problems surface when the affinities of sect. Allophyllum within Solanum are considered, because the infrageneric classification scheme in Solanum leaves much to be desired. Several attempts have been made to divide Solanum into subgenera and sections (e.g., Dunal, 1852; Seithe, 1962; Gilli, 1970; Danert, 1970; D'Arcy, 1972). The most widely used recent scheme is that of D'Arcy (1972), who accepted seven subgenera of Solanum, in a classification with basic elements derived from earlier literature. I will refer exclusively to his classification in the discussion below. Although many well-defined sections can be discerned within Solanum, the circumscription of subgenera is more problematic. Most of the subgenera consist of one or several well-defined or well-studied sections along with a number of sections whose affinities are less certain. Thus, D'Arcy's subgenera are often clearly defined with respect to one or a few component sections, but the boundaries between the sugenera are less clear when all the associated sections are considered. The large number of species in Solanum and the relative paucity of adequate taxonomic treatments for many infrageneric groups hampers understanding the genus. Because I lack the facilities for an exhaustive survey of morphology and variation in the entire genus, it is possible that I have overlooked existing groups in Solanum that might be allied with sect. Allophyllum. Nevertheless, I will attempt to enumerate the possibilities for placement of sect. Allophyllum within the existing subgenera of Solanum as I understand them. Section Allophyllum probably does not belong to the subgenera Archaesolanum Marz. (aneuThis content downloaded from 169.237.45.23 on Wed, 02 Mar 2016 18:56:20 UTC All use subject to JSTOR Terms and Conditions 400 Annals of the Missouri Botanical Garden E o) * D I)~~~~~~~~~: