Iridaceae Subfamily Research Articles

New species of Gladiolus L. and Moraea Mill. (Iridaceae) from Tanzania and Mozambique

Summary. Gladiolus clivorum and G. exalatus are new species of this large African and Eurasian genus of Iridaceae subfamily Crocoideae, both from Tanzania. These additions bring the total in the genus in tropical Africa to 84 species and the entire genus to 270 species. Moraea niassensis (subg. Grandiflorae Goldblatt) is a new, montane species of northern Niassa Province, Mozambique. The Africa and western Eurasian Moraea now includes some 225 species, 25 of which occur in tropical Africa. An identification key to the large-flowered species of subg. Grandiflorae in East Africa and Mozambique is provided. The new species are documented with a map, photographs and a conservation assessment.

FLORAL BIOLOGY OF BABIANA (IRIDACEAE: CROCOIDEAE): ADAPTIVE FLORAL RADIATION AND POLLINATION1

Field observations, floral dissections, floral nectar, and pollen load analyses of captured insects of 53 species of Babiana Ker Gawl. show that flowers of this geophytic genus of some 90 species of Iridaceae subfamily Crocoideae, predominantly of the southern African winter-rainfall zone, are cross pollinated by a wide range of animals. These include passerine birds and insects of four different orders, Hymenoptera (mainly Apidae), Diptera (mainly Nemestrinidae), Coleoptera (Scarabaeidae), and Lepidoptera (mainly Noctuidae). Apid pollination involves two discrete systems—passive pollination by anthophorines and native Apis mellifera foraging for nectar and active pollen gathering by A. mellifera and other Apoidea foraging for pollen. From what is known about relationships within Babiana and Crocoideae, it seems likely that passive pollination by anthophorines and honeybees, with nectar secreted in zygomorphic, bilabiate flowers as a reward, is the ancestral condition; it is also the most common, demonstrated for 18 and inferred for 35 more species from all three taxonomic sections of the genus. Active pollination by honeybees and female Apoidea foraging for pollen was demonstrated in one species and is inferred for four more, all of which have radially symmetric flowers and prominent anthers. Pollination by long-proboscid nemestrinids, mostly species of Prosoeca, is recorded for 13 species and inferred for five more, while moth pollination is recorded for one species and inferred for another two. Pollination exclusively by hopliine scarab beetles, known for six species, is associated with development of radial symmetry of the flower. Passerine bird pollination, associated with the classic syndrome of a wide floral tube, red floral pigmentation, and rigid, well-exserted stamens, occurs in two species, and is inferred for one more. Species with a bimodal system in which bees and beetles both visit and accomplish pollen transfer is known for three species. Comparing pollination systems with what is known about species relationships in Babiana, we infer that long-proboscid fly pollination evolved at least four times and moth pollination three times. Active pollination by pollen-collecting bees and hopliine beetle pollination also probably evolved three times each and bird pollination twice. Pollination systems are labile, and we postulate that there has been a minimum of 14 shifts in pollination system, approximately one shift for every six species. Lastly, Babiana species show the same correlation of morphology and floral presentation with particular sets of pollinators, described for several other genera of Iridaceae, e.g., Gladiolus L., Hesperantha Ker Gawl., and Lapeirousia Pourr., as well as Geraniaceae and Orchidaceae. This increases our confidence in predicting pollinators on the basis of floral presentation in other species and genera in which pollinators have not been established.

Taxonomic notes and new species of the southern African genus <i>Babiana</i> (Iridaceae: Crocoideae)

A member of Iridaceae subfamily Crocoideae. Babiana Ker Gawl. comprises some 80 species from southern Africa. Field studies have shown the need for several taxonomic and nomenclatural changes, while a number of new species have been discovered. The type of B flabellifolia Harv. ex Klatt is a short-tubed plant that matches B. truncata GJ.Lewis. and that name falls into synonymy. The name B. flabellifolia sensu GJ.Lewis (1959) has been misapplied to long-tubed plants from the western Karoo which are now renamed B. praemorsa sp. nov. In addition. B. truncata, as originally circumscribed, included two species, one short-tubed and a second with a longer tube, which we describe here as B. cuneata sp. nov. The type off? hypogaea Burch, has also been misinterpreted and matches the species described as B.flavida. which thus falls into the synonymy of B hypogaea. A second species. B.falcata GJ.Lewis. closely matches this species and is also reduced to synonymy. The widespread southern African species long known as B hypogaea (hypogea sensu GJ.Lewis) matches the type ol B. bainesii Baker and must now be known by that name. Babiana stricta var. erectifolia (GJ.Lewis) GJ.Lewis is appropriately included in typical B. stricta (Aiton) Ker Gawl. However, var. regia GJ.Lewis is a very different plant and is treated as a separate species. B. regia comb, et stat. nov.. as is long-tubed var. grandiflora GJ.Lewis. which is described as the new species, B. longiflora sp. nov. Plants included in B. stricta var. sulphurea sensu GJ.Lewis are also included in var. stricta. We are unable to match the type of Gladiolus sulphureus Jacq.. basionym of var. sulphurea (Jacq.) Baker, with any known species and the name is thus excluded. Lastly, the name B. disticha Ker Gaw l., type of the genus, is superfluous for Gladiolus fragrans Jacq. and the new combination B fragrans comb. nov. is made, reducing B. disticha to synonymy. Babiana fragrans Eckl., which was thought to prevent the transfer of G. fragrans. is a nomen nudum and thus invalid and cannot be taken into account in considerations of nomenclatural priority.

Reduction of Schizostylis (Iridaceae: Ixioideae) in Hesperantha

Schizostylis has the floral synapomorphies of the African genus Hesperantha, including style branches dividing at the mnouth of the floral tube, long laxly spreading style branches, and articulated anthers twisted to face inward. The basic chromosome number of Schizostylis, x = 13, is shared in Ixioideae by Hesperantha, Geissorhiza, and a few species of Romulea. Schizostylis then differs from Hesperantha only in its red flowers (occasionally pink) and, more importantly, in its rootstock, a short rhizome. Most species of Hesperantha have white or pink flowers, occasionally yellow or purple, and all have corms with hard woody tunics. We suggest that Schizostylis, which is semi-aquatic and grows along streams in seasonally or permanently waterlogged ground, has acquired a rhizome secondarily by loss of the corm. Neither the rhizome nor the red flower is sufficient grounds to warrant generic separation from Hesperantha. Schizostylis, a monotypic, or at best ditypic, genus of southern African Iridaceae subfamily Ixioideae, extends from the Drakensberg Escarpment of Eastern Cape Province, South Africa, northward through Lesotho, Swaziland, and Mpumalanga Province to eastern Zimbabwe. A plant of habitats that are waterlogged during its spring and summer growing season, Schizostylis grows along streams, wet seeps, and the edges of marshes (Goldblatt, 1991, 1993). Except in its rhizomatous rootstock and red (sometimes pink) flowers, Schizostylis closely resembles the large African genus Hesperantha (ca. 65 spp.) morphologically, and it has all of the critical floral features of that genus. These include green, soft-textured bracts; a well-developed perianth tube; a style dividing at the top of the perianth tube; long, spreading filiform style branches; and anthers twisted to face inward. The last three features are synapomorphies for Hesperantha. Schizostylis also shares a derived basic chromosome number, x = 13, with Hesperantha, Geissorhiza, and some species of Romulea (Goldblatt, 1971). Here we address the hypothesis that Schizostylis is not only immediately related to Hesperantha but is also closely related to species within that genus, thus nested in Hesperantha.

Phylogeny of the African Genera Anomatheca and Freesia (Iridaceae: Ixioideae), and a New Genus Xenoscapa

Anomatheca and Freesia, African genera of Iridaceae subfamily Ixioideae, are currently thought to be closely related, but until now the exact nature of their alliance has been difficult to determine. Cladistic analysis has provided one means of assessing their relationship. Using 25 characters found to be useful for phylogenetic analysis, including traditional morphological features as well as original data from pollen morphology, leaf anatomy, embryology, and seed development, Freesia appears to be monophyletic but nested in Anomatheca sect. Anomatheca, one of two sections of the genus. A second section, Xenoscapa, which includes one named species and a second one described here, comprises a well-supported clade that is probably only distantly related to Freesia plus sect. Anomatheca. We herein unite Anomatheca and Freesia, but under the later name Freesia. Because Freesia is already a conserved name and because it is a well-known and horticulturally important genus, Anomatheca has been formally rejected in favor of Freesia. We raise sect. Xenoscapa to generic rank and transfer A. fistulosa to it, describe the new species, X. uliginosa, and provide new combinations in Freesia for the species until now included in Anomatheca. Anomatheca Ker-Gawl. (1805), a tropical and southern African genus currently including five species, and Freesia Klatt (1863), a South African genus of 11 species, are closely related members of Ixieae, one of three tribes of the Old World and largely African subfamily Ixioideae (Goldblatt 1990, 1991) of the Iridaceae. Anomatheca was later treated as a subgenus of Lapeirousia Pouret by Baker (1896, 1898). Lewis (1954), however, disagreed with Baker's circumscription of Lapeirousia and suggested that Anomatheca might be better regarded as a separate genus, which she believed was most closely allied to Freesia. Lewis placed the two genera on the same branch of a phylogenetic tree some distance from Lapeirousia in subtribe Watsoniinae. Goldblatt (1971, 1972) also recognized Anomatheca and removed both it and Freesia from Watsoniinae and treated them as the sole genera of a subtribe Freesiinae. In this paper we explore the relationships of Freesia and Anomatheca and, as a result, propose some changes to the taxonomy of Ixieae. MATERIALS AND METHODS Taxa and Characters. Characters selected for inclusion in the analysis and their states and polarization are discussed in Appendix 1. Data for Anomatheca were largely taken from a study by Goldblatt (1972) and for Freesia from a revision by Goldblatt (1982). Data on leaf anatomy were obtained from work by Rudall and Goldblatt (1991) and supplemented by additional observations for species not known anatomically. Pollen characters, taken from a study by Goldblatt et al. (1990), were also supplemented by original observations presented in Table 1. Chromosome numbers were determined for the new species, confirmed for A. grandiflora Bak. and A. fistulosa (Spreng. ex Klatt) Goldblatt by M. Takei, and will be reported fully elsewhere (Goldblatt and Takei, in mss). Seed characters are taken from ongoing studies (Manning and Goldblatt unpubl. data) of seed development in Ixioideae.

New Species of the Southern African Genus Geissorhiza (Iridaceae: Ixioideae)

Three new species are recognized in the southern African Geissorhiza, a genus now including 84 species. Geissorhiza silenoides and G. rupicola are members of subgenus Geissorhiza and occur in the Ceres District of the southwestern Cape, South Africa. Both have pale pink flowers with unusually long perianth tubes for the genus. Geissorhiza silenoides is additionally distinguished by having unequal stamens, each filament being of a different length, scabrid-puberulent stems, and leaves with heavily thickened margins and midribs. Geissorhiza rupicola has unusual soft-textured corm tunics and, like G. silenoides, has leaves with strongly thickened margins and midribs. The third new species, G. uliginosa, is restricted to the Swartberg Mountains of the southern Cape and grows on wet cliffs and under waterfalls. It stands out in the genus in having the flowers solitary on the stems, and very soft-textured corm tunics that rapidly decay, sometimes leaving the corms apparently without tunics. All three new species occur in the southern and southwestern part of southern Africa, a region known for the richness and diversity of its flora, particularly of geophytes. More than half of the estimated 1750-1800 species of the petaloid monocot family, Iridaceae, occur in southern Africa (Goldblatt, 1994). The region is well known botanically, yet despite over 250 years of botanical exploration, novelties continue to be discovered there. Although the family is a numerically significant element of the flora of a large part of southern Africa, Iridaceae are most diverse in the southern and western coastal belt and adjacent mountains of the subcontinent, a region of mediterranean-type climate with winter rainfall and summer drought. Three new species of Geissorhiza, a genus virtually restricted to the winter rainfall zone of southern Africa, are described here. Some 81 species were admitted to the genus when it was last revised (Goldblatt, 1985), and the new species described here thus bring the total to 84. Geissorhiza is a member of Iridaceae subfamily Ixioideae, the largest of the four subfamilies currently recognized in Iridaceae (Goldblatt, 1990, 1991). It is one of the largest genera of the subfamily and one of the larger genera in the southern African flora. One of the new species, G. uliginosa, belongs in subgenus Weihea, which is characterized by concentric corm tunics, and G. silenoides and G. rupicola are members of subgenus Geissorhiza, species of which have the characteristic imbricate corm tunics that inspired the generic name, a Greek word meaning tiled root. 1. Geissorhiza silenoides Goldblatt & J. C. Manning, sp. nov. TYPE: South Africa. Western Cape: Ceres District, lower slopes of Gydo Pass, 20 Sep. 1993, Goldblatt & Manning 9739 (holotype, NBG; isotypes, K, MO, PRE, WAG). Figure 1A. Plantae 20-30 cm altae, cormo 5-7 mm diametro tunicis duris imbricatis, foiis li nearibus 1.2-1.5 mm latis marginibus costisque incrassatis, spica 4-7 florum, floribus pallide roseis, tubo perianthii cylindrico 15-17 mm longis, tepalis subaequalibus 12.5-14 x 6-7 mm, filamentis inaequalibus 4, 7 et 9 mm longis, antheris ca. 4.5 mm longis, stylo eccentrico. Plants 20-30 cm high. Cormn + ovoid, 5-7 mm diam., flattened slightly below on one side, the tunics woody, blackish, imbricate, the layers breaking below into triangular points. Cataphylls membranous, the upper light purple above the ground. Leaves 4, the lower 2 basal and longest, reaching to at least the middle of the stem, sometimes to the base of the spike, straight and erect or trailing, the blades linear, 1.2-1.5 mm wide, oval in section with 2 narrow grooves on each surface, the midrib much thickened, the margins raised and arched over the surface, the upper leaves cauline, sheathing for at least half their length, the blades like the basal, the uppermost sometimes only 9-12 mm long and enNovoN 5: 156-161. 1995. This content downloaded from 157.55.39.78 on Sun, 19 Jun 2016 06:24:39 UTC All use subject to http://about.jstor.org/terms Volume 5, Number 2 Goldblatt & Manning 157 1995 Geissorhiza from Southern Africa

Gladiolus somalensis (Iridaceae), a New Species from Northeastern Somalia

Gladiolus somalensis is a new species restricted to the limestone escarpments of the Sanaag and Bari Regions of northeastern Somalia. Although Gladiolus is the largest genus of Iridaceae subfamily Ixioideae and has some 83 species in tropical Africa, only 4 species occur in Somalia and only G. somalensis is endemic there. The affinities of this new species are uncertain, but the relatively small flowers with a short perianth tube appear to place the species in subgenus Gladiolus, a taxon that is most diverse and speciose in southern Africa. The Old World genus Gladiolus, a member of Iridaceae subfamily Ixioideae and one of the larger, if not the largest genus of the family, consists of some 250 species (Goldblatt, in press; Goldblatt & Manning, ms.). The center of Gladiolus in terms of species numbers and taxonomic diversity is temperate southern Africa, more particularly the winter rainfall region of southern Africa (Lewis et al., 1972; Goldblatt, 1991). The genus is, however, well represented in tropical Africa, where some 83 species occur south of the Sahara and north of the borders of Namibia, Botswana, and South Africa (Goldblatt, in press), 75 of them endemic. At least 150 species occur in southern Africa (Goldblatt & Manning, in prep.), another 8 in Madagascar (Goldblatt, 1989), and perhaps 10 more in Europe, North Africa, and the Middle East. Here we describe a new species of Gladiolus restricted to northeastern Somalia, one of only four species of the genus recorded from that country, and the only one endemic there (Goldblatt, 1995). Although first collected in 1956, Gladiolus somalensis remained until now too poorly known to be described or even assigned with confidence to genus. The type collection made in January 1995, however, is well preserved, and is accompanied by photographs and spirit material. This has made it possible to draw up a description, formally name the species, and have an illustration made. Gladiolus somaleusis Goldblatt & Thulin, sp. nov. TYPE: Somalia. Sanaag Region, escarpment S of Laasqoray, near Ragad, 1100'N, 48029'E, evergreen bush on limestone, 16 Jan. 1995, Thulin, Dahir & Hassan 9079 (holotype, UPS). Figure 1. Plantae (7-)12-30 cm altae, cormo ca. 12 mm in diametro, foliis 4-5 linearibus (1-)2-4 mm latis, spica 2-10 florum, bracteis externis viridibus (7-)12-17(-23) mm longis internis minoribus, flos aurantiacus, tepalis lateralibus infernis infra flavis (raro tepalo inferiore infra flavo), tubo perianthii infundibuliformi 6-8 mm longo, tepalis lanceolatis tribus supernis majoribus 16-18 X ca. 8 mm, infernis 12-15 X 5-5.5 mm, filamentis 8-10 mm longis, antheris 3-5 mm longis, ramis styli filiformibus ca. 2.5 mm longis. Plants (7-)12-30 cm high. Corm obconic, ca. 12 mm diam., the tunics of softly textured layers, these decaying with age into fine netted fibers. Leaves four or five, the lower three basal and longest, reaching at least to the base of the spike and one or more often slightly exceeding the spike, the blades + linear, (1-)2-4 mm wide, the upper one or two leaves inserted on the lower half of the stem, smaller than the basal leaves. Stem erect, simple or with one or two branches, ca. 1.2 mm diam. below the base of the spike. Spike lightly flexuose, 210-flowered; bracts green and soft-textured, the outer (7-)12-17(-23) mm long, the inner about two-thirds as long as the outer. Flowers zygomorphic, orange, the lower lateral tepals (rarely the lower median tepal) bright yellow in the lower half; perianth tube funnel-shaped, 6-8 mm long; tepals unequal, lanceolate, the upper three larger than the lower, the dorsal inclined over the stamens, 16-18 X 8 mm, the upper laterals about as long, the lower tepals ? parallel to the ground, the lower lateral NovoN 5: 325-328. 1995. This content downloaded from 207.46.13.114 on Thu, 26 May 2016 06:12:37 UTC All use subject to http://about.jstor.org/terms

New Taxa and Revisions to the Taxonomy of Southern African Lapeirousia Subgenus Lapeirousia (Iridaceae Subfamily Ixioideae)

Studies of the pollination biology of the southern African species of Lapeirousia subg. Lapeirousia have indicated the need for several taxonomic changes in the alliance. Lapeirousia simulans is a new species, until now confused with L. arenicola, a fairly widespread species of the Cape west coast: L. simulans occurs sympatrically with L. arenicola, but subtle differences in flower patterning, capsule and seed morphology, and bract size separate the two. New evidence indicates that varieties spinosa and tenuis of L. divaricata should both be treated as separate species, and they are recognized here as L. spinosa and L. tenuis. The former is an allopatric vicariant of L. divaricata, differing in its compact growth form and larger flowers. Lapeirousia tenuis, however, appears to be more closely related to L. dolomitica, from which it differs in flower size, tepal color, and length of the perianth tube. Examination of living plants matching the type of L. macrospatha indicates that the species was incorrectly treated as conspecific with L. barklyi. The two differ consistently in flower color and markings, the length of the perianth tube, and the shape of the floral bracts, and are most likely not immediately related. Lastly, populations of the widespread L. pyramidalis with a dark blue, purple, or red perianth, a longer perianth tube, and odorless flowers are segregated here as subspecies regalis. Subspecies regalis is most likely adapted to a different pollinator than subspecies pyramidalis, which has a pale blue, pink, or white perianth, a shorter perianth tube, and a strong floral odor. The tropical and southern African genus Lapeirousia Pourret comprises some 40 species currently segregated in two subgenera (Goldblatt & Manning, 1990, 1992). Subgenus Lapeirousia, with 21 species, is centered in the coastal and near interior parts of southwestern Africa but includes two tropical African species (Goldblatt, 1990). Subgenus Paniculata Goldblatt & Manning, with 19 species, comprises the largely tropical African section Paniculata (14 species), with one species in the southwestern part of southern Africa, and section Fastigiata Goldblatt (5 species), restricted to the southwestern Cape region of South Africa (Goldblatt & Manning, 1992). Field studies of the pollination biology of subgenus Lapeirousia, conducted over the past two years, have yielded new information about the genus and indicated the need for substantial taxonomic revision. Sixteen southern African species were included in subgenus Lapeirousia by Goldblatt (1972), who then treated it as section Lapeirousia. One cryptic species, included in L. arenicola Schlechter, is now understood to be a distinct species, here described as L. simulans. Lapeirousia divaricata N. E. Brown, treated by Goldblatt (1972) as comprising three varieties, is now reconstituted. Variety spinosa Goldblatt, raised to species rank as L. spinosa, differs consistently from L. divaricata in its ecology and morphology and merits separate species status. Variety tenuis Goldblatt, now L. tenuis, appears to be more closely related to L. dolomitica Dinter than to L. divaricata. It has broadly conic corms with spiny margins identical to those of L. dolomitica a d can be distinguished from this species largely by its short-tubed flower with narrow tepals and noninflated floral bracts. In the widespread Lapeirousia pyramidalis (Lamarck) Goldblatt, patterns of floral variation suggest the need for the recognition of two subspecies. Typical subspecies pyramidalis has strongly scented white, pale blue, or pale pink flowers with a perianth tube of intermediate length. What we recognize here as subspecies regalis is restricted to the northwestern Cape and has darkly pigmented, red, purple, or v olet flowers with a long perianth tube and no scent. Lastly, it is now apparent that L. barklyi Baker and L. macrospatha Baker were incorrectly united by NOVON 4: 339-346. 1994. This content downloaded from 207.46.13.131 on Sat, 15 Oct 2016 04:26:09 UTC All use subject to http://about.jstor.org/terms

Pollen apertures, exine sculpturing and phylogeny in Iridaceae subfamily Iridoideae

Specialization from the basic monosulcate and semitectate-reticulate pollen grains of Iridaceae subfamily Iridoideae occurs in all of its four tribes. Disulcate apertures characterize Tigridieae subtribe Tigridiinae (4 genera). Zonasulculate grains occur in two derived species of Cipura (Tigridieae subtribe Cipurinae) and Dietes iridoides (Irideae subtribe Iridinae), all three autogamous. Aperture membranes are typically smooth, but are verrucate in Alophia (Cipurinae), a condition evidently apomorphic for the genus. Operculate apertures occur in Hermodactylis and Iris reticulata (subgenus Hermodactyloides) and there appears to be a gradation in the subgenus to a zonasulculate condition as found in I. danfordiae. While a fairly uniformly reticulate exine predominates in Iridoideae, a trend towards microreticulate or reticulo-rugulate exine is evident. This is accomplished in Gelasine by elongation of the muri and narrowing of the lumina, and by thickening of the muri in Ennealophus and Herbertia. Notable increase in lumina size characterizes some species of Moraea, and the species of Cipura and Dietes in which the grains are zonasulculate. Lumina of sharply different sizes are a notable feature of the exine of Isophysis (Isophysidioideae) and the species of Cypella sometimes segregated as Hesperoxiphium. We conclude that variability in both apertures and exine sculpturing is sufficiently great in the larger genera of Tigridieae and Irideae, and in some smaller ones as well, that palynological data should be sought routinely in systematic and phylogenetic studies in the two tribes. A smaller degree of variation in Mariceae makes it seem less useful to include such studies here. Data from this and other palynological investigations of the Iridoideae indicate that at least at the tribal level for Tigridiinae, and at the generic level for Alophia (and possibly Ennealophus and Gelasine), pollen characters provide valuable synapomorphies for taxon delimitation.

Notes on Alophia (Iridaceae) and a New Species, A. veracruzana, from Vera Cruz, Mexico

Alophia veracruzana is a new species of this American genus of Iridaceae subfamily Iridoideae. Closely related to the widespread A. drummondii, it differs in its pale to dark lilac-blue perianth with white to yellow markings, secund flower, and more or less erect inner tepals with lightly raised marginal areas bordering a smooth median zone. The generic characters of Alophia lie exclusively in its stamens and style, and the genus is defined by pandurate, nearly latrorse anthers, and deeply, but incompletely divided style branches with slender, ascending-recurved, terminally stigmatic style arms

The Genera of the Sisyrinchium Alliance (Iridaceae: Iridoideae): Phylogeny and Relationships

A phylogenetic analysis of the New World members of Sisyrinchieae, most primitive of the four tribes of Iridaceae subfamily Iridoideae, using leaf anatomical and chromosomal data in conjunction with traditional morphological criteria suggests substantial realignments in the current taxonomy. The New World Sisyrinchieae constitute the majority of the taxa in the tribe; the re- maining Bobartia and Diplarrhena, African and Australian, are not considered in the study. Phaiophleps, Chamelum, and Ona appear to be more closely related to Sisyrinchium sect. Eriphilema than the last does to sects. Sisyrinchium and Echthronema, although all these taxa appear to comprise a monophyletic assemblage based on the characters used in our study. Solenomelus, Tapeinia, and Orthrosanthus are more distantly related to Sisyrinchium and its satellite taxa, but the three genera share some features with the Sisyrinchium clade. We treat sects. Sisyrinchium and Echthronema as subgenera of Sisyrinchium and unite sect. Eriphilema and the genera Phalophleps, Chamelum, and Ona in a single genus, Olsynium (1837). Appropriate changes to the taxonomy of Sisyrinchium are presented and necessary new combinations in Olsynium are made: 0. acaulis, 0. biflorum, 0. lyckholmi, 0. bodenbenderi, 0. frigidum, 0. luteum, 0. filifolium, 0. junceum, 0. philippii, 0. scirpoideum, and 0. obscura. Additional anatomical observations are presented for Tapeinia, Ona, and Sisyrinchium spicatum, which complement extensive anatomical observations already published.

New Species of Galaxia (Iridaceae) and Notes on Cytology and Evolution in the Genus

Galaxia kamiesmontana and G. parva are new species, both of subgenus Eurystigma. Galaxia kamiesmontana is restricted to the Kamiesberg in Namaqualand, where it grows in shallow soils overlying rock, while G. parva is known only from the Mierkraal flats near Bredasdorp in the southern Cape. Chromosome number is n 9 in G. kamiesmontana and n = 6 in G. parva. Both numbers are new for the subgenus, in which n 8, 7, and 17 have been recorded. The chromosome number in G. kamiesmontana links subgenus Eurystigma to the more specialized subgenus Galaxia in which x = 9 is basic. Chromosome evolution in Galaxia, reviewed in the paper, is seen as having proceeded from a base of x = 9 by aneuploid reduction to n = 8, 7 and ultimately 6 by unequal translocation and centric loss. Galaxia kamiesmontana is evidently a primitive relict species retaining the ancestral basic karyotype, while G. parva, an inbreeding, autogamous species with a highly derived karyotype is one of the most specialized. Galaxia is a small genus of Iridaceae subfamily Iridoideae restricted to the winter rainfall area of the southwestern coast and interior of southern Africa. All species of Galaxia are small corm bearing plants with a basal rosette of short bifacial leaves and Crocus-like flowers with an underground ovary and a perianth tube that raises the flower well above the ground. The flowers are fugaceous, lasting for 3-6 hours, and this, together with their small size, makes the plants very inconspicuous. As a result, the genus is relatively poorly collected. In the recent revision of the genus (Goldblatt, 1979a) I recognized 12 species in two subgenera. The two species described here were discovered subsequent to the publication of the 1979 revision, G. kamiesmontana in 1980 and G. parva in 1981. Both are clearly very local endemics and in view of the high level of knowledge of the Cape Flora, it seems unlikely that they occur elsewhere.

A Survey of Seed Surface Morphology in Hesperantha (Iridaceae)

A Survey of Seed Surface Morphology in Hesperantha (Iridaceae)

Notes on Geissorhiza (Iridaceae): The Species in Madagascar

The two Madagascar species regarded as belonging to the southern African genus Geissorhiza, G. bojeri and G. ambongensis are shown to have characteristics incompatible with this genus. Geissorhiza bojeri accords well with Gladiolus and is transferred here as Gladiolus bojeri (Baker) Goldbl. Geissorhiza ambongensis is not sufficiently well known to allow generic placement at this time but is excluded from Geissorhiza. Geissorhiza is a large genus of southern African Iridaceae subfamily Ixioideae. In the only complete modern treatment of the genus, Foster (1941) recognized 55 species in two subgenera. Subgenus Geissorhiza has 51 'species, all from the Cape Flora Region and surrounding areas, and subgenus Ixiopsis has only four, three of which occur in the Cape Region and the fourth, G. bojeri Baker in Madagascar. The second Madagascar Geissorhiza, G. ambongensis, described in 1939 (Perrier, 1939), was treated by Foster among his doubtful and unknown species. Foster saw no specimens, but commented that from the description, it seemed close to G. bojeri, but distinct. Both species were placed in Geissorhiza in the Flore de Madagascar (Perrier, 1946). I am at present revising Geissorhiza and it has become clear to me during this study that the two Madagascar species do not belong in Geissorhiza at all. THE MADAGASCAR SPECIES OF GEISSORHIZA 1. Geissorhiza bojeri. The earlier of the two Madagascar Geissorhiza species, G. bojeri was described in 1876 by J. G. Baker, based on collections by Bojer and Hilsenberg from the mountains of central Madagascar. Since then the species has been collected repeatedly and is evidently common on the central plateau in grassland and rocky situations. Geissorhiza bojeri is a slender plant, variable in height from 15 to 35 cm, with two linear, equitant basal leaves, and one to three more, smaller, cauline, partly sheathing leaves, the uppermost often bractlike. The small corms have tunics of fine reticulate fibres, the character that Baker used to distinguish subgenus Ixiopsis, of which G. bojeri is the type species. The bracts are herbaceous and quite short; and the yellow flower is small, short tubed, and apparently actinomorphic, with lanceolate tepals. The fruits are unusual in being long and slender and the seeds have an entire, membranous wing. The seeds match exactly those of Gladiolus, and the fruits are similar to fruits of some Gladiolus species, although more slender than any other species known to me. The significant taxonomic characters are the fibrous corm tunics, apparently actinomorphic flower, long capsule, and winged seeds and only one of these, the actinomorphic flower, corresponds with Geissorhiza as it is currently understood.

Notes on the Cytology and Distribution of Anapalina, Tritoniopsis, and Sparaxis, Cape Iridaceae

Anapalina and Tritoniopsis are two closely related genera of Iridaceae subfamily Lxioideae, both restricted in distribution to the Cape Province of South Africa and both centered in the Cape Floristic Region. Sparaxis, also a member of subfamily Lxioideae is only distantly related to these two genera, and its range is restricted to the western Cape and adjacent karoo. Recent collecting expeditions to South Africa have enabled me to extend knowledge of the cytology and distribution of Tritoniopsis and of Sparaxis, and this new information is presented here. Chromosome counts were made from root tip mitoses, squashes being prepared by the method outlined in other papers (Goldblatt, 1976; Goldblatt & Gentry, 1979).