zyxwvutsrqpo zyxwvutsrq zyxwvutsrqpon zyxwv Botanical ~7ournalofthe Linnean Sociely (1992), 108: 55-81. With 5 figures Elucidative studies on the generic concept of Senecio (Asteraceae) zy P. LESZEK D. VINCENT* Department o f Botany, University o f the Witwatersrand, PIBag 3, Wits, 2050, Republic o f South Africa AND FIONA M. GETLIFFE, F.L.S. zyxwvutsrqp zyxwvutsr zyxwvutsr zyxw Herbarium Pac$cum, Department o f Botany, Bernice P. Bishop Museum, P.O. Box 19000-R, Honolulu, Hawaii 96819, U.S.A. Receiued j’iily 1987, reuised and accepted,for publication Ocloher 1990 VINCENT, 1’. L. D. & GEILIFFE, F. M., 1992. Elucidative studies on the generic concept of Senecio (Asteraceae).This paper presents the results of studies of the generic concept of Senecio Aenm slricto. ‘l’he sample of taxa studied consisted of 93 Natal senecios (including seven varieties and two forms), fivr Cape heterochromous senecios and nine non-southern African senecios, including the type of thc gcnus, S. uulgaris L. Also included in thc study were six species from taxonomically closely related genera in the tribe Senecioneae and one Senecio of uncertain taxonomic position. T h e phenotype of these taxa was investigated with respect to a large number (122) of morphological and micromorphological characters. Six characters were selcrtcd as being taxonomically important with respect to elucidating the generic concept of Senecio sensu Jlrictu. T h e generic concept of Senecio has hcen provisionally re-circumscribed and the generic status of rach of the senrcios and non-senrcios studied has been tested according to this concept of Senecio sensu slricto sensu Vincent. In the light of this concept of Senecio, the following species are recommended for exclusion from Senecio sensu strict0 sensn Vincent: S. ciaampelinus, S. transuaalensis, S. syringiflius and S.hockii. The following species are considered to be peripheral to Senecio sensu slrzcto sensu Vincent: S. tanacetopsis, S. semzniueus, S. medley-woodii, S. lamoides, S.helrninthioides, S. barbertonicus, S. breuilorus, S. uiminalts, S. radicans and S.fu1gens. Before any taxonomir changes are made to the current composition of Senecio, the concept of Senecio sensu striclo sensu Vincmt, is being tested on a worldwide sample of thr genus. ADDITIONAL KEY WORDS:-Compositae - numerical taxonomy. CONTENTS Introduction . . . Material and methods . Results and observations Discussion. . . . Conclusion . . . Acknowledgements . Rcferences . . . Appendix . . . . . . . . . . . . 56 . . . . . . . . . . . . . . . . . . . . . 57 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 0024-4074/92/010055 + 27 $03.00/0 55 . . . . . . . . . . 76 74 74 zyxwv . . *Author for correspondence. 63 73 . . . . 75 0 1992 The Linnean Society of London 56 zyxwvutsrqp zyxwvuts zy P. L. D. VINCENT AND F. M. GETLIFFE INTRODUCTION The genus Senecio was formally circumscribed by Linnaeus in Genera Plantarum (1754), although a number of species with this name appeared in Species Plantarum (Linnaeus, 1753) including the type Senecio vulgaris L. (Farr, Leussink & Stafleu, 1979). Since then, the genus has been described from many parts of the world and the generic concept has been subject to varied interpretations. Thus, Senecio now includes species which exhibit a range of character variation that overlaps, and even exceeds, the combined ranges of species currently referred to several other genera (Jeffrey et al., 1977). This expended concept of Senecio results in an unwieldy and patently paraphyletic genus and, in the spirit of the ‘New Synantherology’ (King & Robinson, 1970), several studies in the Senecioneae (Robinson & Brettell, 1973a-d, 1974; Robinson, 1974, 1975; Jeffrey et al., 1977; Nordenstam, 1978; Jeffrey, 1979; Wetter, 1983) have attempted to improve generic concepts in many groups within the Asteraceae. I n particular, Jeffrey (1979) advocated the application of “uniform criteria” to studies of the Senecio complex, throughout its geographic range. O n this basis, it was suggested, systematic conclusions on generic and sectional limits might be more valid. After preliminary surveys, Jeffrey et al. (1977) concluded that: 1. The species currently placed in Senecio can be subdivided into 16 groups, only one of which (group IX) should be called Senecio, since it contains the type species. 2. The remaining groups might be distinct genera, possibly new or possibly congeneric with other widely recognized genera in the family. Jeffrey (1979), re-evaluated group IX in the light of evidence from studies of the pappus (Drury, 1967), chemical characters (Robins, 1977a, b; Bohlmann et al., 1979) and succulent species (Jeffrey, 1979). As a result, group I X was further subdivided into two sections: “eusenecionoids” which includes the type species, and “gynuroids”. It is estimated that Senecio consists of “perhaps 1000 species” (Jeffrey & Chen, 1984), which is considerably fewer than earlier estimates of c. 1500 species (Nordenstam, 1978) or the c. 3000 species of the earlier estimates by Jeffrey Ueffrey et al., 1977). Thus the application of the “uniform criteria” appears to result in a narrower concept of the genus which may be further refined, as attention is drawn to the fact that the “generic limits in many areas are uncertain” (Jeffrey et al., 1977). I n contrast, Barkley (1985), argued for the retention of the broad concept of Senecio, but this has not yet influenced the direction of research on this genus. Unfortunately, Jeffrey & Chen (1984) do not comment explicitly on their concept of Senecio sensu strict0 but Nordenstam (1978) did mention characteristics that he perceived as being characteristic of the genus in its narrower sense. These are ‘senecioid’ filament collars, ‘radial’ endothecial tissue, discrete stigmatic areas in a ‘banded’ configuration, truncate style branches and a haploid chromosome number of 10. The present study was undertaken with all but the last of these features in mind, and with the intention of contributing to what must finally be a zyxwvu zy zyxwv GENERIC CONCEPT OF SENECIO 57 worldwide revision of the generic concept of Senecio, based on the concept of “uniform criteria” as proposed by Jeffrey et al. (1977). MATERIAL AND METHODS The number and selection o f taxa Eighty-four species of the c. 130 species of the genus Senecio in Natal, together with five species of Senecio from the Cape, nine Senecio species from outside South Africa and six species from other genera in the tribe Senecioneae (Table 1) were investigated. Most of the material came from NU, the remainder from BOL, K , N H and PRE. The Natal senecios studied included a11 the purple senecios described by Hilliard (1977) and a large number of yellow senecios. In the selection of the latter, both putatively closely allied and distantly allied species (on the basis of similar morphology) were selected. At the start of this study some of these senecios had an uncertain taxonomic position: S. cissampelinus (DC.) Sch. Bip., S.fulgens D. Hooker) Nicholson, S. transvaalensis Bolus & S. viminalis Bremekamp, incertae sedis,jde Hilliard, 1977, 1978 and S. syringiflius 0. Hoffm. incertae sedis, fide Jeffrey personal communication, 1984, Jeffrey, 1986 and S. hookii De Wild. & Muschl., incertae sedis, jide Jeffrey personal communication, 1984; Emilia hockii (De Wild & Muschl.) C. Jeffrey,jide Jeffrey, 1986. The species descriptions in Hilliard ( 1977) were used to select representative specimens of the Natal species. Harvey’s descriptions in Flora Capensis (Harvey, 1865) were used in the selection of the Cape senecios. While we decided which non-southern African senecios and related genera to include in this study, the selection of the representatives of these taxa was done on our behalf by M r C. Jeffrey of The Royal Botanic Gardens, Kew (C. Jeffrey, personal communication, 1984). A complete list of all the taxa investigated is given in Table 1. zyxwvut zyx u. Morphology and micromorphology A minimum of three representatives of each taxon were studied in detail. I n the selection of the representatives an attempt was made to select specimens from widely different localities or habitats, to enable the effect of any environmentally induced variation to be encountered. Preparation of vegetative and joral tissue Samples of a t least four leaves, four bracts, four calyculus bracts and six to eight florets of each specimen, were cleared in c. 85% lactic acid at c. 120°C. The clearing process was carefully monitored to achieve adequate clearing, without excessive softening of the tissue. A portion of cleared leaf tisssue (c. 15 x 15 mm) of each specimen sampled was mounted in lactic acid ( 85%). The florets were dissected such that the corollas, the anthers and attached filaments and the cypselae, usually with the style still attached, were all displayed separated from each other. Light microscope observations were made using Reichert and Olympus compound microscopes (the latter with phase contrast objectives) and a Zeiss dissecting microscope. - 58 zyxwvutsrq zyxwvu zyxwv P. L. D. VINCENT AND F. M. GE'ILIFFE TABLE 1. Complete list of t h e species a n d varieties of Seneciu a n d related genera included i n this study (SUM = 101). zyxwvutsrqp Natal senecios: The numbers in the left-hand margin correspond to the species numbers for Senecio in Hilliard (1977) and are for identification purposes only. 1. S. madagascariensis Poiret 2. S. skirrhodon DC. 3. S. inaequidens DC. 4. S. harueianus MacOwan 5. S. polyanthemoider Schultz Bipontinus 6. S.pterophorus DC. 7 . S.juniperinus L. fil. 9. S. serratuloides DC. 10. S. microglossus DC. 14. S.pandur2frmis Hilliard 15. S. achilleifolius DC. 16. S. tanacetopsis Hilliard 17. S.seminiueus Wood & Evans 18. S. haygarthii Hilliard 20. S. medley-woodii Hutchinson 22. S.purpureus L. 23. S.gerrardii Harvey 24. S. erubescens Aiton (24A) var. erubescens (24B) var. crepidfolius DC. (24C) var. incisus DC. (24D) var. dzchotumus DC. 25. S. sandersonii Harvey 26. S.glanduloso-lanosus Thellung 27. S. uariabilis Schultz Bipontinus 28. S. umgeniensis Thellung 29. S.glanduloso-pilosus Volkens & Muschler 30. S. subcuriaceus Schlechter 3 1. S. cathcartensis 0. Hoffmann 32. S. ngoyanus Hilliard 33. S.polyodon DC. (33A) var. polyodon (33B) var. subglaber (0.Kuntze) Hilliard & Burtt 34. S. aff. speciosus 35. S. speciosus Willdenow 36. S. macrocephalus DC. 37. S. barbatus DC. 38. S.poseideonis Hilliard & Burtt (38A) radiate forms (38B) discoid forms 39. S. aff. poseideonis 40. S. arabidifolius 0. Hoffmann 4 1 . S. rhyncholaenus DC. 42. S. subrubriflorus 0. Hoffmann 43. S. hastatus L. 46. S. hirsutilobus Hilliard 47. S. ingeliensis Hilliard 48. S. natalicola Hilliard 50. S. hieracioides DC. 5 1. S. consanguineus DC. 53. S.chysocoma Meerburgh 55. S.asperulus DC. 56. S. hypochoerideus DC. 57. S. mooreanus Hutchinson 63. S.paludajinis Hilliard 64. S. ajinis DC. 65. S.lydenburgensis Hutchinson & Burtt Davy 67. S. anomalochrous Hilliard 72. S. breuidentatus M. D. Henderson zyxwvu zy zyxwv zyxwvutsr zyxwvuts zyxwvuts zyxwvutsrq zyxwvuts GENERIC C O N C E P I OF SENECIO 59 TABLE 1. zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHG continued ~ 73. 74. 75. 76. 77. 8I . 82. 83. 85. 86. 88. 89. 90. 97. 98. 99. 10 1. 102. 103. 104. 107. 108. 109. 110. 1 1 I. 116. 1 1 7. 118. 119. 120. 12 I . 122. 124. S. praeteritus Killick S. coronatus (Thunberg) Harvey S. macrospermus DC. S. dregeanus DC. S. discodregeanus Hilliard & Burtt S. caudatus DC. S.mauricei Hilliard & Burtt S.saniensis Hilliard & Burtt S. heliopsis Hilliard & Burtt S. albanensis DC. var. doroniciflorus (DC.) Harvey S. inornatus DC. S. oxjri~/olius DC. (referred to Senecio sect. Peltati, incertae sedisjde Jeffrey, 1986: 934) S. rhomboideus Harvey S.bupleuroides DC. S. urophjllus Conrdth S. scitus Hutchinson & Burtt Davy S. glaberrimus DC. S.IatifoliuJ DC. S. retrorsu~DC. S.othonniflorus DC. S. brachypodus DC. S.pleistocephalus Spencer Moore S. mikanioides [Otto ex] Harvey ( = Delairea odorata Lem.,fide Jeffrey, 1986: 933) S. deltoideus Lessing S. tamoides DC. (incertae sedisjde Jeffrey, 1986: 934) S.helminthioides (Schultz Bipontinus) Hilliard S. barbertonicus Klatt S. breuilorus Hilliard S. uiminalis Bremekamp (referable to Kleinia subg. Kleinia, cf. Jeffrey, 1986: 924) S. radicans (L,fil.) Schultz Bipontinus (referred to Senecio sect. Rowleyana,fide Jeffrey, 1986: 935) S. cissampelinus (DC.j Schultz Bipontinus ( = Mikaniopsis cissampelina (DC.) C. Jeffreyjde Jeffrey, 1986: 879) S.,/ulgens (Hooker fil.) Nicholson ( = Kleiniafulgens (Hook. f.) Nich., subg. Notoniafide Jeffrey, 1986: 935) S. transuaalensis Bolus ( = Emilia transuaalensis (Bolus) C. Jeffrey,jde Jeffrey, 1986: 919) The numbers in the left hand margin of all the following species are for identification purposes only when referred to in the text. Cape 200. 20 1. 202. 203. 204. heterochromous non-yellow (purple) senecios S. grandijorus Berg. S. c a ~ ~ ~ e DC. ~~ius S. arenarins Thunb. S. multibracteatus Harv. 5. elegans L. Non-Southern African senecios 210. S. vulgaris L. (U.S.A., U.K., Europe & E. Asia) 212. S. uiscosus L. (U.S.A. & U.K.) 213. S. gallicus Chaix (Spain & Portugal) 214. S. califrnicus DC. (U.S.A.) 215 S. vernalis Waldst. & Kit. (Europe) 216. S. sylvaticus L. (U.S.A. & U.K.) 217. S. desfontainei Druce (Middle East & E. Asia) 2 19. S. erechtitoides Bak. (Madagascar) 220. S. lautus Willd. (Australia) Senecios of uncertain taxonomic position and other genera in the tribe Senecioneae 223. S. syringifolius 0. Hoffm. ( = S. syringifolius 0. Hoffm. incerfue sedis,fide Jeffrey, 1986: 885) 60 zyxwvutsrqpon zyxwvutsr zyxwvuts zyxwvu P. L. D. VINCENT AND F. M. CETLIFFE TABLE 1. --continued 224. S. hockii De Wild. & Muschl. ( = Emifia hockii (De Wild & Muschl.) C. Jeffrey,fide Jeffrey, 1986: 912) 225. Crassocephahm c e n u u m (L.fi1.) Moench 226. Emiliajammea Cass. 227. Gynura divaricala (L.) DC. 228. Notonia grandiflora DC. ( = Kfeinia grandgora (DC.) N. Rani,fide Jeffrey, 1986: 935) 229. Kleinia neriifolia Haw. 232. Cineraria geifalia L. TABLE2. Complete list of all characters investigated in this study of Senecio and related genera Character Vegetative habit Rootstock Plant height (max.) Leaf distribution Leaf length (max.): Radical Cauline Leaf width (max.): Radical Cauline Leaf shape: Radical Cauline Leaf apex: Radical Cauline Leaf margin: Radical Cauline Leaf base: Radical Cauline Leaf venation: Radical Cauline Leaf texture: Radical Ca u I i n e Leaf attachment: Radical Cauline Leaf pubescence: Abaxial-radical Abaxial-cauline Adaxial-radical Adaxial-cauline Leaf trichomes: Base Apex Appendages Glandular Flowering stem: Appearance Pubescence Glandular Peduncles: Appearance Leaf stomata Inflorescence bracts Capitula: Length Diameter Radiate/discoid Homo/heterogamoIUS Arrangement Character no. 008 009 010 01 1 012 013 014 015 016 017 018 019 020 02 1 022 023 024 025 026 027 028 029 030 03 1 032 033 034 035 036 037 038 039 040 04 I 042 043 044 045 046 047 048 zyxwvu zyxwv zyxw 61 GENERIC CONCEPT OF SENECIO TABLE 2.-continued Character Involucre shape Involucral bracts: Number Length Shape Pubescence Trichomes Apex vestiture Length rel. to disc florets Glandular Colour Calyculus bracts: Number Length Position Pubescence Trichomes Apex vestiture Glandular Colour Shape Flower colour: Ray florets Disc florets Floret number Ray florets: Corolla length Corolla shape Limb venation Corolla trichomes Resinous Limb apex Limb shape Limb epid. cell shape Limb orientation Floret number Disc florets: Corolla length Corolla shape Corolla venation Corolla trichomes Resinous Lobe apex Gynoecium (ray): Style-arm length Style-arm apices Stigmatic surface Style base Nectary Ovary vestiture Ovary wall crystals Style-arm length (Disc): Style-arm apices Stigmatic surface Style base Nectary Ovary vestiture Ovary wall crystals Distal epidermal cell shape Disc corolla: Proximal epidermal cell shape Androecium (disc): Anther apex Anther base appearance Anther length Endothecial tissue: Cell shape Cell wall configuration Character no. 049 050 05 1 052 053 054 055 056 05 7 058 059 060 06 1 062 063 064 065 066 06 7 068 069 070 07 1 072 073 074 075 076 077 078 079 080 08 1 082 083 084 085 086 087 088 089 090 09 1 092 093 094 095 096 097 098 099 100 zyxw 101 102 103 104 105 106 107 zyxwv zyxwvuts P. L. D. VINCE NI' AND F. M. GETLIFFE 62 TABLE 2. continued Charactri Length Basal width Shape Cypsela (ray): Lrngth Shape Ribbing Vestiturr Pericarp cells (shapr) (Disc): Length Shapr Ribbing Vestiture Pericarp crlls (shape) Uniforrn/dimorphic Pappus: Pappus setar (ray): Length Appearance Distal rell number Apex Length (Disc): Appearance Distal ccll number Apex Filament collars: ti Character no. 108 109 110 Ill 112 1 I3 I I4 I I5 116 117 I18 119 I20 121 I22 I23 I24 I25 I26 127 128 129 zyxwvutsrq zyx Characters 001 to 007 inclusive were inadmissible characters (Sneath Sokal, 1973: 103). Recording o f characters One hundred and twenty-two characters were recorded for each taxon investigated. A list of these characters is given in Table 2. Since the details of the character states for each of these characters are, in most instances, quite lengthy, these have not been provided, but can be obtained from PLDV. The coding of the qualitative character states was achieved using NOMINAL-LEVEL MEASUREMENT (non-parametric). In this measurement procedure no assumption is made with respect to ordering or distances between character states. Thus, the codes for the character states of the qualitative characters are arbitrary values separating the character states for each respective character. The maximum value for the quantitative characters was recorded rather than the mean because it indicates the maximum genetic potential of each character. Some of these data were obtained from Hilliard (1977). Number of measurementslobservations o f each character At least five measurementslobservations were made of each character from each of between three and five representatives of each taxon. Master data matrix Using these data, a master data matrix ( M D M ) (Appendix B) was developed for the entire sample set. Where character plasticity in the qualitative characters was encountered amongst the representatives of a taxon, either the average state was chosen or a new state was described which encompassed the variation zyxwvu zyxwv GENERIC CONCEPT OF SENECIO 63 encountered. This ‘average state’ or ‘new state’ code was entered into the MDM. Variation in the continuous quantitative characters was accommodated in the MDM after first having been subdivided into size classes. This subdivision into size classes was achieved by subjecting the original data to frequency analyses, with the aid of the programme; Statistical Package for the Social Sciences (SPSS) by Nie et al. (1975). These analyses enabled peaks and troughs to be noted in the distribution of the data for each quantitative character. Care was taken to avoid the danger of dichotomizing the peaks into two or more size classes when one size class was judged to be a more accurate biological representation of the data producing the peak. Naturally the data of meristic quantitative characters was entered directly into the MDM. zyxwvuts Selection of the taxonomica1l;v important charactersfor deJining the generic Limits of Senecio The data in the M D M were subjected to frequency analyses using the SPSS programme (1.c.). These analyses elucidated the distribution of the data for each character for all the taxa investigated. The typology of specimens of S. vulgaris was kept in mind when scrutinizing the character distribution patterns in search of correlations between characters. The search for the aforementioned correlations was assisted by ordination analyses using the SPSS programme (I.c.). T h e correlations with the typology of S. vulgaris led to the recognition of the putatively important characters for the circumscription of the generic concept of Senecio. T h e data in the M D M , for all the taxa studied, can be obtained from PLDV upon request. T h e data of the six characters of generic importance (generic concept data) are provided in Appendix B. zyxwvuts Analysis of generic concept data The techniques used in the analysis of the generic concept were those contained in the package: numerical taxonomy system of multivariate statistical programmes (NTSYS-pc, ver. 1.50) (Rohlf, 1988). The data were standardized (Sneath & Sokal, 1973) using the STAND (default options) subroutine, resulting in a mean of zero and a standard deviation of unity for each character. T h e data were standardized, since, in most instances, nominal-level (non-parametric) coding was used for the six characters. Standardization of the data would enable each character to contribute toward the overall resemblance inversely in proportion to its variability amongst the entire set of OTUs. Correlation and distance matrices were computed using S I M I N T (for interval data). Cluster analysis was performed using the SAHN subroutine on a distance matrix. Ordination of the data, using non-metric multidimensional scaling (MDS), was used to evaluate the data. The results of a principal components analysis (PCA) served as the initial matrix for the MDS. A minimum-length spanning tree (MST), computed using the distance matrix, was superimposed on the final ordination. The M S T helps to detect local distortions-pairs of points which look close together in a plot but actually are far apart if other dimensions are taken into account. RESULTS AND OBSERVAL‘IONS We confirm that three of the morphological characters used by Nordenstam (1978), to define Senecio, hold true for the sample studied. In addition, the 64 zyxwvuts zyxwvutsrqpon zyxwvu zy P. L. D. VINCENT AND F. M. GETLIFFE following characters are also diagnostic at the generic level: the shape of the anther apex, the length of the filament collars and the shape of the cypselae of the disc florets. According to this study, Senecio sensu stricto sensu Vincent is defined on the basis of the moderate to considerable correlation of one or more character states of the following six characters: 1. Style arm apices truncate with a subterminal semicircle of spreading papillae (Character 095, states 9, 11, 13, 14). 2. Anther apices short (Character 103, state 1). 3. Endothecial tissue-inner anticlinal (Character 107, state 6). 4. Filament collars 0.3-0.9 mm long (Character 108, states, 1, 2, 3). 5. Filament collars baluster-form (Character 110, states 1-6, 9). 6. Cypsela (disc) shape-fusiform and cylindrical (Character 117, states 1, 2, 4). Similarly, the character states which are used to identify senecios which are peripheral to Senecio sensu stricto sensu Vincent (Table 3), also show moderate to considerable positive correlation with each other. I t is this correlation amongst these characters (determined via frequency and ordination analyses of the MDM) that was used to assign them generic importance. Each of these characters comprise a number of character states (see Appendix A ) . These six characters are all 'good' taxonomic characters (Davis & Heywood, 1963) for most of the senecios investigated (Table 1) because (a) they are not subject to wide variation; (b) they do not appear to be susceptible to environmental modification; (c) they appear to have a low intrinsic genetic variability. A summary of which character states circumscribe Senecio sensu stricto and those which are used to identify senecios which are peripheral to Senecio sensu stricto is provided in Table 3. zyxwvutsrq Descr$tion of the six characters of generic importance Gynoecium (disc) : style-arm apices (char 095) Twelve distinct character states were recorded for this character. The stylearms of the majority of the senecios investigated (74 species and six varieties) are truncate with a sub-apical semicircle (incomplete ring) of spreading papillae (state 9). There is a small sub-apical region on the inner (upper) surface of each style-arm which is devoid of papillae, but appears to be physiologically receptive. This 'nude' sub-apical region corresponds to the width of the stigmatic surface. Some representatives of some species have convex to conical apices to the style-branches as opposed to truncate apices. The apices in all the specimens examined are distinctly verrucose, this pattern ending abruptly at the semicircle of spreading papillae. A number of (13 species) of the remaining senecios have style-arms which are very similar to those described above but they differ in that the outer surface, proximal to the semicircle of spreading papillae, is also verrucose, the verrucae facing distally at c. 45" to the outer surface (state 1 1 ) . The appearance of these style-arms is distinct from those of character state 9. The majority of the remaining character states occur only once and are therefore irrelevant in defining groups. zyxwvu zyxwvu zyxwv zyxwvuts zyxwvu GENERIC CONCEPT O F SENECIO 65 TABLE 3. An analysis of t h e character states of t h e six taxonomically significant characters of generic importance ( C h a r s 095, 103, 107, 108, 110, 117-see Appendix A ) , with respect to t h e concept of Senecio sensu stricto sensu Vincent Status Character and state(s) zyxwvutsrqpo Senecio sensu stricto Gynoecium (disc): style-arm apices (Char 095) Character states: 9, 1 I , 13 & 14 Yes No Character states: 10, 15, 16, 17, 19 Character states 20, 21 and 22 are outside the concept of Senecio sensu Vincent Androecium (disc): anther apex (Char 103j Character state: 1 Yes No Character state: 3 Character states 4 and 5 are outside the concept of Senecio sensu Vincent. Endothecial tissue: cell wall configuration (Char 107) Character state: 6 Yes Character state 10 and 11 are outside the concept of Senecio sensu Vincent. Senecio (peripheral) No Yes No Yes zyxwvutsr Filament collars: length (Char 108) Character states: 1, 2, 3 Character states: 4, 5 Yes No Filament collars: shape (Char 110) Yes Character states: 1 , 2, 3, 4, 5, 6 Character states 7, 8 and 9 are outside the concept of Senecio sensu Vincent. Cypsela (disc): shape (Char 117) Character states: 1 , 2 & 4 Character state: 5 Character state 6 is outside the concept of Senecio sensu Vincent. Yes No Yes No Yes Yes Yes Yes zyxw Androecium (disc) : anther apex (char 103) This character refers to the shape of the apical appendage (the sterile connective extension) of each anther. Four distinct character states were recorded for this character, three of these (states 1, 3, 4) being found amongst the senecios and some of the related genera studied. T h e fourth character state (state 5) is restricted to Gynura diuaricata. The majority (85 species and six varieties) of the senecios produce relatively short anther apices (state 1). A further ten senecios produce anther apices that have the same length: width ratio as those of character state 1, but the anther apices are approximately twice the size. Endothecial tissue: cell wall conjguration (char 107) As has been discussed in a previous paper (Vincent & Getliffe, 1988), the endothecium of all but two of the taxa studied, exhibited what we have termed “inner anticlinal” endothecial thickening (state 6 ) . This variant of “radial 66 zyxwvutsr zyxwvutsrqp P. L. D. VINCENT AND F. M. GE‘TLIFFE zyxwvutsrqp zyxwvutsrqp Figure 1. Inner anticlinal configuration of endothecium; view of adaxial surface of anther; longitudinal axis of anther vertical; connective down right hand side. Sample from Senecio speciosus Willdenow (Hiiliard 1061 N U ) . zyxwvu thickening” (Dormer, 1962), has the ribs of thickening restricted to the inner anticlinal wall of each endothecial cell (i.e. the anticlinal wall nearest the connective (Fig. 1). This configuration appears to have been illustrated in Wetter’s account of the endothecium (Wetter, 1983), though he did not distinguish it from “radial” thickening. Likewise, we believe that Nordenstam (1978) also illustrated this variant of “radial thickening” without comment. The exceptions were S. cissumpelinus, in which polarized thickening was found, and Gynura diuaricata which was conspicuously different, having cells with a distinct imperforate base plate with rib-like extensions from the edge of the plate. This configuration appears very similar to that noted in Valeriana capensis Thunb. (Valerianaceae) (Noel, 1983). Filament collars: length (char 108) The filament collar (Koyama, 1967; Drury, 1973, 1975; Nordenstam, 1978) alternatively called the anther collar (Robinson & Brettell, 1973d; Jeffrey et al., 1977) is a “downward extension of connective tissue which forms an incomplete monostromatic collar of thick-walled cells” (Drury, 1975). zyxwvu zyxwv zy GENERIC CONCEPT OF SENECIO Figure 2. A ‘cacalioid’ (cylindrical) filament collar. Sample from Senecio medley-woodiz Hutchinson (Hilliard s.n. NU). 67 zyx The majority of the senecios investigated have filament collars in the size range 0.3-0.6 mm, the second most frequent size range being 0.7-0.9 mm. Nordenstam (1978) has drawn attention to the occurrence of “very elongated” filament collars in Notonia DC., Notoniopsis B. Nord. and Kleinia Mill. These “very elongated” filament collars are quite possibly in the size range 1.O-1.4 mm, for the specimens of Kleinia grandzjora and Kleinia nerizfolia included in the present study both have filament collars in this size range (state 4). However, both S. tamoides and S. barbertonicus have filament collars which exceed considerably the lengths of the filament collars of those genera examined by Nordenstam (1978). Filament collars: shape (char 110) The fine structure of the filament collar has only recently attracted the attention of synantherologists. Koyama (1966, 1967), drawing on the earlier observations of Kitamura (1937), commented on differences in the filament collar shape found in senecioid genera and within the sections of Senecio itself. Drury (1973, 1975), described two types of filament collar: ‘cylindrical’, when the diameter of the filament collar and the filament proper are the same and the 68 zyxwvutsrqpo P. L. D. VINCENT AND F. M. GETLIFFE zyxwv zyxwv zyxwvu zyx Figure 3. A ‘senecioid’ (balusterform) filament collar. Sample from Senecio speciosus (Hilliard 1061 NU). zyxwvut cells of the collar are all the same size (Fig. 2); and ‘balusterform’, when “the basal cells of the filament collar are enlarged so that the diameter of the collar a t the proximal end is greater than the filament proper” (Drury, 1975) (Fig. 3 ) . Nordenstam ( 1978) noted that “cylindrical” filament collars generally characterize cacalioid genera, but that they were also found in several genera traditionally associated with Senecio (e.g. Brachyglottis J.R. & G. Forst., Nemosenecio (Kitam.) B. Nord., Sinosenecio B. Nord., Tephroseris Reichenb. & Urostemon B. Nord.) Nordenstam (1978) also drew attention to the term ‘cylindrical’ pointing out that ‘cylindrical’ filament collars are, “strictly speaking, not cylindrical since they are neither hollow nor circular in transect. Instead they tend to be somewhat flattened, often with adaxially involute margins, and more or less reniform in transect”. Since there is a strong correlation between the ‘cylindrical’ filament collar and several other characters typical of cacalioid genera, this shape of filament collar is called ‘cacalioid’ (Nordenstam, 1978) and is the term adopted in the present study. Basally dilated filament collars “are found in Senecio sensu strict0 and in many more or less senecioid genera, such as Crassocephalum Moench, Culcitium Humb. & Bonpl., Dendrosenecio (Human ex Hedb.) B. Nord., Dorobaea Cass., Emilia Cass., zyxwvu zy zyxwvuts GENERIC CONCEPT OF SENECIO 69 Euryops Cass., Jacmaia B. Nord., Odontocline B. Nord., Othonna B. Nord., Othonna L., Stilpnogyne DC. and Xyridopsis B. Nord.” (Nordenstam, 1978). Since the variation in the degree of dilation of these filament collars amongst members of the Senecioneae, falls under the descriptive term ‘balusterform’, the nondescriptive term ‘senecioid’ is used by Nordenstam (1978) for all filament collars which are basally distended. ‘Senecioid’ and ‘cacalioid’ filament collars are produced by the taxa investigated in this study (Table 1). All the ‘senecioid’ filament collars have a distinctly balusterform shape. However, within this balusterform shape, both narrow and conspicuously broad shapes are found (states 1-6, 9). Two distinct narrow balusterform filament collars occur (states 2, 9). I n the first form (state 2), the cells of the monostromatic collar are virtually isodiametric and increase in size three to five times towards the proximal end of the filament collar. In the second form (state 9), the cells of the monostromatic collar are noticeably elongate and somewhat irregularly shaped, increasing to approximately twice the size in the proximal end of the filament collar. The conspicuously broad balusterform filament collars were distinctly distended into a basal bulge (state 3 ) . The two cacalioid forms of filament collar recorded (states 7, 8 ) , differ only in that in the one form (state 7), the width of the filament collar and the filament proper are the same, while in the second form (state 8), the width of the filament collar is consistently slightly greater than that of the filament proper. While most of the taxa investigated possessed one of the several forms of filament collar in a consistent manner, a few senecios exhibited intraspecific variation in that they produced senecioid filament collars varying between narrowly balusterform and noticeably broadly balusterform. This intraspecific variation is recorded in character states 4-6. Nordenstam (1978) observed that the cacalioid form of filament collar was nearly always associated with ‘polarized’ endotheciai tissue and the senecioid form with ‘radial’ endothecial tissue. A notable exception was the genus Graphistylis B. Nord. (Nordenstam, 1978). This association between the senecioid form of filament collar and inner anticlinal endothecial tissue was also noted amongst the majority of the taxa studied. However, we recorded nine more exceptions to the aforementioned association. Senecio cissampelinus possessed senecioid filament collars associated with an intermediate form of endothecial tissue which is predominantly ‘polarized’ with a limited amount of inner anticlinal endothecial tissue. T h e remaining eight taxa had cacalioid filament collars associated with inner anticlinal endothecial tissue. They are: S. medley-woodii, S. tamoides, S. viminalis, S.fulgens, S. sylvaticus, Gynura divaricata, Kleinia grandgora and Kleinia nerizflia. Cypsela (disc): shape (char 117) The predominant shape of the cypsela of the disc florets amongst all the senecios investigated is ‘fusiform and cylindrical’ (state 2). Four character states represent the variation in this character. Infraspecific variation was noted with respect to character states 1 and 2. Where either of these two character states is used, the choice represents the more frequent occurrence of the two shapes (states) for the particular taxon concerned. zy 70 zyxwvuts zyxwv zyxwvutsrqpon P. L. D. VINCENT AND F. M. GETLIFFE Groupings of the species The application of the aforementioned generic concept of Senecio to the senecios investigated, results in the formation of 11 groups and several ungrouped species of Senecio (Appendix C). Outside this concept are ten species we regard as peripheral to the genus sensu strict0 (Appendix C ) and the remaining four species (S. cissampelinus, S. transvaalensis, S. syringzfolius and S. hockii) proved unequivocally to be excluded from Senecio. Group I consists of several morpho-biological series namely: A, leafy-stemmed herbs or shrubs with fpinnately divided, not strongly glandular leaves and yellow or heterochromous capitula (species 1, 3, 4, 7, 200-204, 210, 212, 213, 2 15, 2 17, 2 19, 220); B, subcuspose herbs with pinnately lobed strongly glandular leaves and yellow or more often whitish, dull yellow, mauve or purple or blue homochromous capitula (species var./form 22, 24A, 27, 29, 30, 32, 33B, 34, 35, 38B, 48, 50, 51); C, subscapose herbs with unlobed crenate-serrate fibrous, not markedly glandular leaves and yellow or purple homochromous capitula (species 63, 64, 65, 67, 73, 77, 81, 88); D, peltate-leaved non-glandular subsucculent herbs (species 89); E, rhomboid-leaved, non-glandular subsucculent herbs (species 90). However, these series are often repeated in other groups which may be analysed as follows: Group 11: series A (species 15); series F, succulent scandent shrubs (species 107); series G, subsucculent scandent yellow discoid herbs with cordate leaves (species 109). [In another phenetic study (Drury, 1967), species 109 clustered with species 11 1 and a species close to 1 16. Species 1 1 1 and 116 are here grouped among the peripheral senecios.] Group 111: series H, leafy-stemmed, non-glandular, entire-leaved, yellowheaded homochromous herbs (species 97-99, 102, 103). Group IV: series B (species 25 & 28), series C (species 85). Group V: series B (species var. 31, 33A, 42); series C (species 76); series H (species 104). Group VI: series A (species 9, 53, 214); series B (species 37, 39); series C (species 14, 74, 82); series H (species 101). Group VII: series A (species 2 ) ; series B (species var./form 23, 24B, 24C, 24D, 36, 38A, 40, 56, 57). Group VIII: series B (species 41, 43, 46). Group IX: series B (species 47); series H (species 75). Group X: series B (species 26, 83); series C (species 86). Group XI: series B (species 55); series C (species 72). zyxwvu Ungrouped species Series A (species 5, 6, 10, 18, 216) [species 216 is known to be capable of hybridization with species 210 of Group I (Hegi, 1954)l; series F (species 108); series I scandent non-succulent herbs with cordate leaves (species 110). Peripheral species Series A (species 16, 17, 1 18); series J, perennial, succulent, yellow-flowered, tomentose herbs (species 20); series K, scandent subsucculent cordate-leaved, yellow, radiate herbs (species 11 1, 116); series F or near it (species 11 7); whiteflowered succulent shrubs referable to Kleinia subgenus Kleinia (species 1 19); white-flowered succulent herbs currently referred to Senecio section Rowleyana zyxwvut zyxwvu zyxwv zyxwvuts GENERIC CONCEPT OF SENECIO 71 Ueffrey, 1986) (species 120); and orange-flowered succulent herbs currently referred to Kleinia subgenus Notonia (Jeffrey, 1986) (species 122). The peripheral species share some of the character states associated with Senecio sensu stricto sensu Vincent suggesting a dendritic or paraphyletic development of the genus in its broadest sense. Species excluded f r o m Senecio Senecio transvaalensis (species 124) shows considerable affinity with the tropical African genus Emilia Cassini, for it has lilac florets, style-branch apices with a median fascicle of fused papillae and distinctly 5-angled cypselae. These characteristics suggest that it is best placed in Emilia, as suggested by Hilliard (1977, 1978). Since completion of our research, the new combination Emilia transuaalensis (Bolus) C . Jeffrey (basionym: S. transuaalensis) has been published (Jeffrey, 1986). Senecio cissumpelinus (species 121) shows considerable affinity with the tropical African genus Mikaniopsis Milne-Redhead, its prehensile petioles with their thickened bases being a notable feature of the vegetative habit of this genus. However, the genus Mikaniopsis and the Asiatic genus Cissampelo~sis(DC.) Miq., have been under study by Jeffrey (C. Jeffrey, personal communication, 1984) with the view to elucidating the relationship between these two genera. Since the completion of our research, the new combination, Mikaniopsis cissampelina (DC.) C. Jeffrey (syn.: S. cissampelinus (DC.) Sch. Bip.) has been published (Jeffrey, 1986). O n the basis of our analysis, S.syringiflius (species 223) seems to be allied with Crassocephalum Moench, while Jeffrey ( 1986) regards the position of S. syringiflius as incertae sedis. Senecio hockii (species 224) appears to be allied with Emilia Cass., as suggested by Jeffrey et al. (1977). Subsequently, Jeffrey (1986) published the new combination Emilia hockii (De Wild. & Muschl.) C. Jeffrey (basionym: 5’. hockii De Wild. & Muschl.). zyxwv zyxwv zy Numerical analysis The concept of numerical taxonomy has progressed from simple comparisons of similarities and differences to more sophisticated multivariate techniques (for a review, see Sneath & Sokal, 1973). Numerical techniques can be useful in understanding the distribution and amount of variation of morphological characteristics (McNeil, 1979). Methods such as clustering and ordination can assist in developing taxonomic concepts, but they do not themselves define the taxonomic boundaries. Each analysis must be carefully interpreted, and can only serve as a guideline in taxonomic delimitations. A PCA on the non-metric MDS, overlayed with a minimum spanning tree, of these six characters (Figs 4, 5), helps to illustrate how these characters discriminate Senecio sensu stricto from: the peripheral senecios, the senecios which should be excluded from Senecio and the related genera. The three factors of the initial PCA (prior to the non-metric MDS), accounted for 80.6% of the trace (factor 1 : 45.2%, factor 2 : 19.5y0, factor 3 : 15.90/). This indicates that the three components (Figs 4, 5) are a reasonably accurate representation of the six characters analysed. Component I has correlations 72 zyxwvutsrqpon zyxwvuts zyxwvuts P. L. D. VINCENT AND F. M. GETLIFFE zyxwvutsrqponm zyxwvutsrq zyxwvutsrqp zyxwvutsrqpon -1.0 1.5 0.0 2 .o 1.0 / 121 A = 8 = 22 16 I 7 I 3 4 7 2 2 2 4 A 27 29 30 32 338 34 35 388 48 50 5 1 63 6 4 65 67 73 17 81 8 8 8'3 90 20( 201 202 203 204 210 2 1 2 213 215 2 1 7 219 220 C = 9 1 4 1 5 37 39 53 7 4 82 101 107 109 2 1 4 D = 10 91 98 99 102 103 0.7 E F G = H = I = = = 25 28 85 3 1 339 4 2 16 104 2 2 3 2LB 24C 240 36 3 8 1 4 0 56 5 7 4 1 4 3 46 55 72 ... I- H ! m BCO:': 6:? GH - 0.1 I J K rstr. [:\ N O N - M o TAXA 22s PERIPHERAL \ SENECIOS L I -0.8 -1.6 COMPONENT I Figures 4,5. The first three principal component axes (rotated) from analysing all the taxa studied (Table I ) , with respect to the six characters ofgeneric importance (Table 3). Fig. 4. Component I vs Component 11. Fig. 5. Component I vs Component 111. A minimum-length spanning tree (MST) is superimposed over the plot in Fig. 5 for non-coincident taxa. The taxa are identified by their relevant number (Table 1). The boundaries between Senecio sensu stricto sensu Vincent, the peripheral senecios and non-Senecio taxa, according to the concept of Sen& sensu stricto sensu Vincent, are demarcated on these two figures. greater than 0.5 for: STYLE-ARM APICES (CHAR 095) (0.86), ANTHER APEX (CHAR 103) (0.91), FILAMENT COLLARS: LENGTH (CHAR 108) (0.64) and FILAMENT COLLARS: SHAPE (CHAR 110) (0.74). Component I1 has correlations greater than 0.5 for: ENDOTHECIAL TISSUE: CELL WALL CONFIGURATION (CHAR 107) (0.58), FILAMENT COLLARS: LENGTH (CHAR 108) (0.53) and CYPSELA (DISC): SHAPE (CHAR 117) (0.67). Component I11 has correlations greater than 0.5 for: ENDOTHECIAL TISSUE: CELL WALL CONFIGURATION zyxwvu zyxwvu zyxwv zyxwvut GENERIC CONCEPT OF SENECIO -1.0 0.0 1.0 73 2.0 1.5 0.7 zyxwvuts zyxwvutsrqpon - I I- H 31 a 0 6 v -0.1 - 0.8 \ A 8 C D = I = F = = = = G = H = I = J = K = L = zyxwvutsrqpon 16 1 7 1 1 8 120 1 5 55 72 1 0 9 1 3 4 7 22 24A 27 29 30 3 2 338 34 35 388 4 8 50 5 1 6 3 6 4 65 6 7 7 3 7 7 81 88 8 9 90 200 2 0 1 202 203 204 210 2 1 2 213 215 217 2 1 9 220 9 1 4 3 7 39 5 3 14 82 I01 214 2 6 8 3 86 97 9 8 99 102 103 4 7 75 2 5 28 85 2 2 8 229 2 23 248 24C 240 36 38A 40 56 57 41 4 3 46 108 3 1 331 42 76 104 - 1.6 (OM IENT 1 Figure 5 (CHAR 107) (0.61) and CYPSELA (DISC): SHAPE (CHAR 117) (0.72). Since the components of a PCA are arranged in numerical order of importance, the composition of the three components of this PCA indicates, to a degree, the relative importance of each of the six characters with respect to each other. It is noted from Figs 4 and 5 that the boundary between Senecio sensu stricto (sensu Vincent) and the non-Senecio taxa is relatively distinct. While the affinity of some of the peripheral senecios is indistinct and the affinities of the peripheral senecios as a whole appears dendritic. DISCUSSION While Barkley (1983, 1985) argues for the retention of Senecio as a broad concept, the implications of this in the African context are biologically 74 zyxwvutsrqpo zyx P. L. D. VINCENT AND F. M. GETLIFFE untenable. As Jeffrey (1986) points out, the only means of retaining Senecio as a broad concept for the African species in an other than paraphyletic manner would be to include all of the African Senecioneae within Senecio. This would render meaningless the biological and other information associated with such names as Crassocephalum, Gynura and a number of other members of the African Senecioneae. Consequently segregation seems to be the correct course to follow, primarily on biological grounds. An implication of this choice is that Senecio will probably remain a paraphyletic “core genus” as criticized by Funk (1983). Nevertheless, while Funk (1983) contends that the segregates must be defined by synapomorphies, an evolutionary situation can be imagined in which a monophyletic core group of species exists which is defined only by those synapomorphies which are common to the whole subtribe Senecioniinae Ueffrey, 1986). While in a cladistic analysis such a core group would appear paraphyletic, Jeffrey (1986) argues that it would be falsely so even though there would be no way of verifying that falsity. CONCLUSION The generic concept of Senecio sensu Vincent delimits a natural assemblage. While this concept also supports many of the suggestions of Jeffrey et al. (1977) and Jeffrey (1979)) it also indicates that the concept of Senecio held by these authors could still be too broad. T h e concept also supports some of the suggestions by Nordenstam (1978) on characters he has perceived as being characteristic of Senecio sensu stricto and elaborates upon them. It is obvious that the concept needs to be tested on further representatives of Senecio from many more areas of its cosmopolitan distribution, before any of the consequences resulting from the application of this concept are put forward in new taxonomic delimitations. Nevertheless, it is encouraging to see that the application of this concept appears to be meeting the need of circumscribing the assemblage of Senecio in Natal into a more homogenous group. Further phenetic studies are currently underway on the worldwide Senecio flora and representatives of related genera in the Senecioneae, to test the concept of Senecio sensu stricto here presented. Perhaps these further studies will indicate that Senecio sensu stricto sensu Vincent is a paraphyletic “core genus” as criticized by Funk (1983) and as suggested by Jeffrey (1986). However, it must be remembered that a taxonomic system in practice will always remain a hypothetical reflection of phylogeny. ACKNOWLEDGEMENTS The authors acknowledge financial support from the CSIR and from the University of Natal, while the senior author was a postgraduate student of the latter. Funding from the University of the Witwatersrand is also gratefully acknowledged. The use of the research facilities of the Department of Botany, University of Natal, Pietermaritzburg and the University of the Witwatersrand is also much appreciated. The r61e played by Prof. 0. M. Hilliard in introducing the senior author to the very exciting field of synantherology, particularly studies in Senecioneae, is very gratefully acknowledged. zyxwvu zyxw zyxwvutsr zyxw zyxwvutsrqp zyxwvut zyxwv zyxwvutsrq GENERIC CONCEPT O F SENECIO 75 REFERENCES BARKLEY, T. M., 1983. Generic concepts in the Senecioneae. American Journal ofBotany, 70f.5): 99. BARKLEY, T. M., 1985. Generic boundaries in the Senecioneae. Taxon, 34: 17-21. BOHLMANN, F.,ZDERO, C., BERGER, D., SUWITA, A., MAHANTA, P. &JEFFREY, C., 1979. Neue Furanoeremophilane und Weitere Inhaltsstoffe aus Sudafrikanischen Senecio-arten.Phytochemistry, 18: 79-93. DAVIS, P. H. & HEYWOOD, V. H., 1963. Principles of Angiosoperm ?axonomy, New York: D. Van Nostrand Company, Inc. DORMER, K. J., 1962. 'The fibrous layer in the anthers of Compositae. New Pbtologist, 61: 150-153. DRURY, D. G., 1967. A taxonomic study ofCompositae with special reference to Senecio. Unpublished Ph.D Thesis of the University of Southampton. DRURY, D. G., 1973. Annotated key to the New Zealand shrubby Senecioneae-Compositae and their wild garden hybrids. N e m zealand Journal of Botany, 11: 731-784. DRURY, D. G., 1975. A comparison of Senecio kirkii (New Zealand) and Senecio insularis (Lord Howe Island) with senecios endemic to the island of St. Helena. N e w zealand Journal of Botany, 13: 769-780. FARR, E. R., LEUSSINK, J. A. & STAFLEU, F. A. (Eds), 1979. Index Nominum Genericorum (Plantarum), 3. The Hague: Dr W. Junk b.v. FUNK, V. A,, 1983. Cladistics and generic concepts in the Cornpositae. American Journal ofBotany, 70(5): 99. HARVEY, W. H., 1865. Senecio. In W. H. Harvey & 0. W. Sonder, Flora Capmsis, 3: 346408. HEGI, G., 1954. Illustrierte Flora von Mitteleuropa, 6: 795. HILLIARD, 0. M., 1977. Compositae in Natal. Pietermaritzburg: University of Natal Press. HILLIARD, 0. M., 1978. The geographical distribution of Compositae native to Natal. Notes. Royal Botanic Garden (Edinburgh), 36: 4077425. JEFFREY, C., 1979. Generic and sectional limits in Senecio (Compositae): 11. Evaluation of some recent studies. Kew Bulletin, 34: 49-58. JEFFREY, C., 1986. The Senecioneae in East Tropical Africa. Notes on Compositae: IV. Kew Bulletin, 41: 873-943. JEFFREY, C. & CHEN, Y. L., 1984. Taxonomic studies on the tribe Senecioneae (Compositae) of Eastern Asia. Kew Bulletin, 39: 205-454. JEFFREY, C., HALLIDAY, P., WILMOT-DEAR, M. &JONES, S. W., 1977. Generic and sectional limits in Senecio (Compositae): I. Progress Report. Kew Bulletin, 32: 47-67. KING, R. M . & ROBINSON, H., 1970. The new synantherology. Taxon, 19: 6-1 1. KITAMURA, S., 1937. Les Senecio du Japon. Acta Phytotaxonomica et Geobotanica, 6: 265-275. KOYAMA, H., 1966. O n the section Nemosenecio of the genus Senecio. Acta Phytotaxonomica et Geobotanica, 22: 15-20. KOYAMA, H., 1967. Taxonomic studies on the tribe Senecioneae of Eastern Asia. I. General part. Memoirs. College of Science. Kyoto University, Series B, 33: 181-209. LINNAEUS, C., 1753. Species Plantarum. Stockholm. LINNAEUS, C., 1754. Genera Plantarum, ed. 5. Stockholm. McNEILL, J., 1979. Purposeful phenetics. Systematic ,zbology, 28: 465-482. NIE, N. H., HADLAI-HULL, C., JENKINS, J. G., BREMMER, K. S. & BENT, D. H., 1975. Statistical Package f o r the Social Sciences, 2nd edn. New York: McGraw Hill. NOEL, A. R . A,, 1983. The endothecium-a neglected criterion in taxonomy and phylogeny? Bothalia, f 4 : 833-838. NORDENSTAM, B., 1978. Taxonomic studies in the tribe Senecioneae (Compositae). Opera Botanica, 44: 1-84. ROBINS, D. J., 1977a. Alkaloids of the Compositae. Compositae Nemsletter, 5: 1-1 1. ROBINS, D. J., 1977b. Senecioneae-chemical review. In V. H. Heywood., J. B. Harborne & B. L. Turner (Eds), The Biology and Chemistv of the Compositae, 2: 831-850. London: Academic Press. ROBINSON, H., 1974. Studies in the Senecioneae (Asteraceae). VI. The genus Arnoglossum. Phytologia, 28: 294295. ROBINSON, H., 1975. Studies in the Senecioneae (Asteraceae). V I I . Additions to the genus Roldana. Phytologia, 32: 33 1-332. ROBINSON, H. & BRETTELL, R. D., 1973a. Studies in the Senecioneae (Asteraceae). I. A new genus, Pittocaulon. Phytologia, 26: 45 1 4 5 3 . ROBINSON, H. & BRETTELL, R . D., 1973b. Studies in the Senecioneae (Asteraceae). 11. A new genus, Nelsonianthus. Phytologia, 27: 53-54. ROBINSON, H. & BRETTELL, R. D., 1973c. Studies in the Senecioneae (Asteraceae). 111. The genus Psacalium. Phytologia, 27: 254-264. ROBINSON, H. & BRETTELL, R . D., 1973d. Studies in the Senecioneae (Asteraceae). IV. The genera Mesadenia, Syneilesis, Miracacalia, Koyamacacalia and Sinacalia. Pbtologia, 27: 265-276. ROBINSON, H. & BRETTELL, R., 1974. Studies in the Senecioneae (Asteraceae). V. The genera Psacaliopsis, Barkltyanthus, 7elanthophora and Roldana. Phytologia, 27: 402-439. ROHLF, F. J., 1988. NTSYS-pc: Numerical 7axonomy and Multivariate Analysis System. New York: Applied Biostatistics Inc. zyx 76 zyxwvutsrqp zyxwvutsrq zyxwvuts zy P. L. D. VINCENT AND F. M. GETLIFFE SNEATH, P, H. A. & SOKAL, R. R., 1973. Numerical Taxonomy. San Francisco: W. H. Freeman & Co. VINCENT, P. L. D. & GETLIFFE, F. M., 1988. The endothecium in Senecio (Asteraceae).Botanical Journal of the Linnean Society, 97: 63-7 1 . WETTER, M. A,, 1983. Micromorphological characters and generic delimitation of some new world Senecioneae (Asteraceae). Briltonia, 35(1): 1-22. APPENDIX A. CHARACTERS AND CHARACTER STATES Characters and character states of the six characters of generic importance. The numbers in the left-hand margin or adjacent to illustrations correspond to the character state codes. The character state codes for each taxon studied (Table l ) , for the six characters of generic importance, are provided in Appendix B. The numerical sequence of the character state codes, within each character, is incidental (unweighted). Gynoecium (disc): style-arm apices (Char 095) 1 zyxwvu zyxwvuts GENERIC CONCEPT OF SENECIO Androecium (disc): anther apex (Char 103) 77 zyxwvutsrqpo Endothecial tissue: cell wall configuration (Char 107) 6. Inner anticlinal configuration-marked peg-like pattern, sometimes with a median ‘T’ shaped process on some of the inner anticlinal ribs (see Vincent & Getliffe, 1988). 10. Polarized endothecial thickening and a little inner anticlinal thickening (ribs) o n some cells. 11. Imperforate base plate with a number of rib-like extensions from both anticlinal edges of the base plate. Filament collars: length (Char 108) 1. 0.2 mm or less 2. 0.3-0.6 mm 3. 0.7-0.9 mm 4. 1.0-1.4 mm 5. 1.6-2.3 mm Filament collaw: shape (Char 110) zyxwvutsrqponm zyxwv PREOOM. TYPE 2 B U T A L S O TYPE 1 0 PREOOM. T Y P E 1 B U T ALSO T Y P E 2 PREOOM. T Y P E 1 B U T ALSO TYPE @ 3 7n zyxwvutsrqpon zyxwvutsrqpo P. L. D. VINCENT AND F. M . GETLIFFE Cypsela (Disc): shape (Char 117) I . Oblong and cylindrical. 2. Fusiform and cylindrical. 4. Narrowly turbinate. 5. Cylindric and 5-angled. 6. Obovate, laterally compressed, margins winged APPENDIX B. MASTER DATA MATRIX ( M D M ) zyxw zyxwvutsrqponmlk zyxwvutsrq zyxwvutsrq Master data of the six characters of generic importance for all the taxa investigated. The numerical values in this matrix correspond to the relevant character states of the character concerned. See Appendix A for a description/illustration of the character states of the six characters of generic importance. 095 103 Characters 107 ion 110 117 I 2 1 2 2 2 2 2 2 2 2 2 2 2 4 2 2 ~~ 1. 2. 3. 4. 5. 6. 7. 9. 10. 14. 15. 16. 17. in. 20. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. S. madagascariensis S. skirrhodon S. inaequidens S.harueianus S.polyanthemoides S. pterophorus S. juniperinus S. serratuloides S. microglossus S. pandurnf rmis S. achilleijolius S. tanacetopsis S.seminiveus S.haygarthii S. medlg-woodii S. purpureus S. gerrardii 9 9 9 9 9 9 9 9 9 9 II 17 17 13 1 1 1 1 1 I 1 1 1 1 1 3 3 1 14 9 9 1 9 1 9 9 9 9 9 9 1 1 1 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 2 2 2 2 6 6 6 6 6 6 6 6 6 6 6 6 2 2 2 2 2 3 2 2 2 2 2 2 6 6 6 6 6 6 2 2 2 2 2 2 1 1 2 2 2 2 2 2 2 2 3 2 2 I 1 1 1 6 1 2 3 2 1 1 1 1 7 1 1 1 S. erubescens (24A) var. erubescens (24B) var. crepidijoliur (24C) var. incisus (24D) var. dichotomus S. sandersonii S. glanduloso-lanosus S. uariabilis S. umgeniensis S.glanduloso-pilosus S. subcoriaceus S. cathcartensis S. ngoyanus S. po!yodon (33A) var. polyodon (33B) var. subglaber S. sp. aff. S. speciosus S.speciosus S. macrocephalus S. barbatus S. poseideonis (38A) radiate forms (38B) discoid forms S. aff. poseideonis S. arabidijolius S. rhyncholaenus S. svbrubrQ7orus S. hastatus 9 1 1 1 1 1 1 9 9 9 9 1 1 1 9 9 9 9 9 9 1 9 9 9 9 9 9 9 I 1 1 1 1 1 1 2 2 2 2 2 2 2 1 1 I 1 4 2 1 4 1 1 5 1 5 1 1 1 1 2 1 2 1 1 1 2 2 2 2 2 2 2 2 2 2 2 2 1 2 1 zyxwvutsr 1 1 1 1 I 1 1 2 1 2 5 2 2 2 1 1 2 1 zyxwvu zyxwv zyxwvutsrqp zyxwvutsrq zyxwvutsr GENERIC CONCEPT OF SENECIO 79 APPENDIX B.-continued Charactrrs 095 103 107 108 I10 2 2 2 1 117 ~~ 46. 47. 48. 50. 51. 53. 55. 56. 57. 63. 64. 65. 67. 72. 73. 74. 75. 76. 77. 81. 82. 83. 85. 86. 88. 89. 90. 97. 98. 99. 101. 102. 103. 104. 107. 108. 109. 110. 111. 116. 117. 118. 119. 120. 121. 122. 124. 200. 201. 202. 203. 204. 210. 212. 213. 214. 215. 216. S.hirsutilobus S. ingeliensis S. natalicola S.hieraciaides S. consanguineus S. chrysacoma S. asperulus S.hypochoerideus S. mooreanus S.paludajinis S.ajinis S. lydenburgensis S.anomalochrous S. breuidentatus S.praelerihs S. coronatus S. macrospermus S. dregeanus S. discodregeanus S.caudutus S.mauricei S. saniensis S. heliopsis S. albanensis (var. doranic$ororus) S. inornalus S. oxyriijolius S. rhomboideus S. bupleuroides S. urophyllus S.scifus S.glaberrimus S. latlfolius S. retrorsus S.olhonniflorus 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 11 I1 11 9 11 1 1 1 I I 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 I 1 1 1 1 1 1 1 1 1 1 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 10 6 6 6 6 6 6 6 6 6 6 6 6 6 3 2 2 2 2 3 2 2 2 2 2 2 1 1 2 I 1 1 1 3 2 1 1 1 1 1 2 2 3 2 5 2 2 2 2 3 2 1 I 2 2 4 3 2 2 2 2 2 2 2 2 2 2 2 2 2 2 5 2 1 1 1 2 2 2 2 2 2 5 1 4 2 2 2 i 1 2 2 2 2 2 1 1 2 2 2 2 2 2 2 1 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 1 2 2 2 2 2 2 2 2 2 1 zyxwvutsr zyxwvutsrq S. brachypadus S. pleistocephalus S. mikanioides S. deltoideus S. tamoides S.helminlhioides S. barbertonicus S.brevilorus S. uiminalis S. radicans S.cissampelinus S. fulgens S. transuaalensis S. grandiJlorus S. cakilefolius S. arenarius S. multibracteatus S.elegans S. uulgaris S. uiscosus S. gallicus S. c a l i j h c u s S. uernalis S.syluaticus 11 9 I1 II 11 14 I1 11 I1 10 15 9 11 16 19 9 9 9 9 9 9 9 9 9 9 9 1 I 1 1 1 3 3 3 3 3 3 3 3 4 1 1 1 1 1 I I 1 1 1 1 5 3 4 4 2 4 2 2 2 2 2 2 2 2 2 2 2 2 1 1 1 n n 2 1 8 3 1 1 I 1 1 1 I 1 2 1 8 5 2 2 2 2 2 2 2 2 2 2 2 zyxwvutsrq 80 zyxwvutsrqpon zyxwvut zyxwvu P. L. D. VINCENT AND F. M. GETLIFFE zyx zyxw zyxwvutsr APPENDIX B.-continued ~ 095 21 7. 219. 220. 223. 224. 225. 226. 227. 228. 229. 232. S.desfontainei S.erechtitoides S. lautus 103 9 9 9 14 19 20 14 21 22 22 9 S. hockii S.syringifolius Crassocephalum cernuum Emilia Jammea Gynura diuaricata Notonia grandiJora Kleinia neriijolia Cineraria geijolia 1 1 1 3 3 4 4 5 3 3 1 Characters 107 108 6 6 6 6 6 6 6 11 6 6 6 2 2 2 3 2 3 2 2 4 4 2 110 117 1 1 1 2 2 2 2 2 2 2 2 1 1 6 2 1 9 2 8 8 8 1 APPENDIX C Complete list of the species and varieties of Senecio included in this study (Table 1) arranged according to the new concept of Serrecio wzsu slricto sensu Vincent, together with an indication of which species are excluded from Senecio sensu stricto Jensu Vincent. T h e I 1 groups and ungrouped species of Senecio, together with the peripheral and excluded senecios are distinguished. ’lhe listing of the character states after each group follows the order of the characters in Table 3. Taxa are identified by their respective number used in Table 1. Senecio sensu stricto senm Vincent Group 1: Taxa: Character states: 9, 1, 6, 2, I , 2 I , 3, 4, 7, 22, 24A, 27, 29, 30, 32, 33B, 34, 35, 38B, 48, 50, 51, 63, 64, 65, 67, 73, 77, 81, 88, 89, 90, 200, 201, 202, 203, 204, 210, 212, 213, 215, 217, 219, 220. Group 11: Taxa: Character states: 11, 1, 6, 2, 1, 2 15, 107, 109. Group 111: Taxa: Character states: 11, 1, 6, 2, 2, 2 97, 98, 99, 102, 103. Group IV: Taxa: Character states: 9, 1, 6, 2, 4, 2 25, 28, 85. Group V: Taxa: Character states: 9, 1, 6, 2, 5, 2 31, 33A, 42, 76, 104. Group V1: Taxa: Character states: 9, I , 6, 2, 2, 2 9, 14, 37, 39, 53, 74, 82, 101: 214 Group VII: ‘Taxa: Character states: 9, 1, 6, 2, 1, 1 2, 23, 24(A-D), 36, 38A, 40, 56, 57 Group V I I I : Character states: 9, 1, 6, 2, 2, 1 Taxa: 41, 43, 46. Group IX: Taxa: Character states: 9, 1, 6, 3, 2, I 47, 75. Group X: ‘l’axa: Character states: 9, 1, 6, 3, 2, 2 26, 83, 86. Group XI: I axa: Character states: 9, I , 6, 3, 1, 2 55, 72. r . zyxwvutsrq zyxwvutsrq zyxwvuts zyxwvu GENERIC C O N C E P T OF SENECIO APPENDIX C.-continued Ungrouped species of Senecio sensu stricto sensu Vincent 5. Character states: 9, 1, 6, 1, 1, 2 6. Character states: 9, 1, 6, 1, 6, 2 10. Character states: 9, 1, 6, 2, 3, 2 18. Character states: 13, 1, 6, 2, 1, 4 108. Character states: 11, 1, 6, 2, 1, I 110. Character states: 14, 1, 6, 2, 1, 2 216. Character states: 9, 1, 6, 2, 8, 2 Peripheral senecios 16. Character states: 17. Character states: 20. Character states: 1 11. Character states: 116. Character states: 117. Character states: 118. Character states: 119. Character states: 120. Character states: 122. Character states: Species excluded from 121. Character states: 124. Character states: 223. Character states: 224. Character states: 3, 6, 2, 1, 2 3, 6, 2, 1, 2 1 , 6, 3, 7, 2 3, 6, 5, 8, 2 3, 6, 4, 2, 2 3, 6, 5, 2, 2 3, 6, 3, 2, 2 15, 3, 6, 4, 8, 2 9, 3, 6, 4, 2, 2 16, 3, 6, 4, 8, 1 17, 17, 14, 11, 11, 11, 10, Senecio 11, 3, 19, 4, 14, 3, 19, 3, 10, 2, 1, 2 ( = Mikaniopsis cissampelina (DC.) C. Jeffrey) 6, 2, 3, 5 ( = Emilia tmmzrvaalensis (Bolus) C. Jeffrey) 6, 3, 2, 2 ( = S. syringifolius 0. Hoffm. incertae sedir) 6, 2, 1, 2 ( = Emilia hockii (De Wild. & Muschl.) C. Jeffrey) 81