304 S.-Afr.Tydskr. Plantk. , 1989,55(3): 304-309 An SEM study of the external pollen morphology in and some related genera in the subtribe Senecioninae (: )

P.L.D. Vincent* and F.M. Getliffe Norris Charles E. Moss Herbarium, Department of Botany, University of the Witwatersrand, P.O. Wits, Johannesburg, 2050 Republic of South and National Botanic Gardens, Kirstenbosch, Clrivate Bag X7, Claremont, 7735 Republic of South Africa

Accepted 7 February 1989

The external pollen morphology of 95 species of the genus Senecio, predominantly from southern Africa, and 11 species from related genera of the subtribe Senecioninae, was investigated using the scanning electron microscope. The pollen of all the species studied is spheroidal, uniformly tricolporate and echinate. The distal ends of the spinules are entire and their apices are acute. Only the frequency of the spinules and the length: basal width ratio of the spinules varied amongst the species studied.

Die eksterne stuifmeelmorfologie van 95 Senecio-spesies, oorwegend van Suider-Afrika, en 11 spesies van verwante genusse van die subtribus Senecioninae, is met behulp van die aftaselektronmikroskoop ondersoek. Die stuifmeelkorrels van al die ondersoekte spesies is sfero'idaal, eenvormig trikolporaat en stekelrig. Die distale onvertakte punte van die stekels is skerppuntig. Slegs die frekwensie van die stekels en die lengte: basalebreedte-verhouding van die stekels varieer by die ondersoekte spesies.

Keywords: Pollen morphology, SEM, Senecio, Senecioninae

*To whom correspondence should be addressed

Introduction degree of elaboration of the foot layer of the ectoexine' Unlike many other families, knowledge of the pollen (Jeffrey et al. 1978) . morphology of the Asteraceae has been derived predom­ inantly through light microscope studies (e.g. Stix 1960) Materials and Methods and transmission electron microscope (TEM) studies The external morphology of the pollen from each of the (e.g. Skvarla & Turner 1966a, b), there being relatively species investigated in this study (Table 1) was investiga­ little SEM data. The SEM data of Asteraceae pollen has ted using SEM. In most instances one specimen of each been provided mostly by Jeffrey et al. (1978), Skvarla et species was sampled. at. (1977) and Blackmore (1984). Pollen was obtained from unopened florets of herb­ In studies on the generic limits of Senecio, partially arium specimens, so avoiding the possibility of stimulated by the research of Jeffrey et al. (1978) and contamination with pollen from unknown species. Jeffrey (1980), it was decided to investigate the external Unacetolysed pollen grains were mounted directly onto pollen morphology of 95 Senecio species and 11 species stubs using double-sided sellotape, sputter-coated with from related genera (Table 1) in the subtribe Senecion­ gold/palladium (Polaron SEM coating unit E5100) and inae. The purpose of this investigation was to ascertain viewed using a Jeol JSM-T200 scanning electron whether there were characters of the external pollen microscope. morphology which would facilitate the elucidation of the Data on the following characters were recorded: generic limits of Senecio. Studies aimed at elucidating (1) Shape of pollen grain, using the terminology of the generic limits of Senecio, using micromorphological Erdtman (1969) . features, are already under way (Vincent & Getliffe (2) Spinule frequency. 1989). (3) Length:basal width ratio of the spinules. Included in the taxonomic scheme for investigating the (4) Prominence of the colpi. Senecio complex on a world-wide basis (Jeffrey et al. The spinule frequency was determined from the aver­ 1978), was the appearance of the pollen grain surface age of three measurements of the interbasal distance and of the wall stratification. The initial SEM studies by between spinules of the intercolpoid region, using SEM Jeffrey et al. (1978) of some species within the Senecio contact prints of similar-sized pollen grains magnified to complex, revealed that the pollen grains of Senecio are the same degree. An interbasal distance of 1,7 fLm or 'uniformly tricolporate and echinate and vary only in the more constitutes the category 'relatively low frequency', number, density and length:breadth ratio of the spines, while an interbasal distance of 1,0 fLm or less constitutes and in the degree of prominence of the colpi.' Only the the category 'relatively high frequency'. aforementioned variable features noted by Jeffrey et al. The length:basal width ratio of the spinules was deter­ (1978) were included in the present study. Other studies, mined from the average of three measurements of using TEM, have shown that the pollen grain walls 'vary spinules of the intercolpoid region , using SEM contact in thickness and structure, particularly in development prints of similar-sized pollen grains magnified to the and length of the columellae and in the thickness and same degree. S.Afr.J. Bot., 1989,55(3) 305

Table 1 The list of the species and varieties of Table 1 Continued Senecio and related genera included in this study. Also provided are the data obtained from SEM observations Score ratio of the pollen of all the species investigated, with respect to the four pollen characters studied. Note that subse­ Character 2 Character 3 quent to the completion of this study, the names of a S. haygarthii Hilliard 0.83 number of species have changed. The currently S. heliopsis Hilliard & Burtt 1.00 accepted names are provided in parentheses. S. helminthioides (Sch. Bip.) Hilliard 0.92 Characters studied: character 1: shape of the pollen S. hieracioides DC. 1.11 grains; character 2: spinule frequency; character 3: S. hirsutilobus Hilliard 0.92 length:basal width ratio of the spinules; character 4: S. hypochoerideus DC. 1.09 prominence of the colpi. Note that characters 1 and 4 S. inaequidens DC. 0.82 are invariable and consequently are not reflected in the S. ingeliensis Hilliard 0.92 Table. Character 1: all pollen grains spheroidal; charac­ S. inornatus DC. 0.93 ter 4: all pollen grains with prominent colpi. S. juniperinus L. f. 0.91 Character state scores: character 2: 1 relatively = S. latifolius DC. 1.00 high; 2 relatively low (see text for definition) = S. lydenburgensis Hutchinson & Burtt Davy 2 1.00 S. macrocephalus DC. 1.20 Score ratio S. macrospermus DC. 1.00

Character 2 Character 3 S. madagascariensis Poi ret 0.73 S. mauricei Hilliard & Burtt 1.00 Natal S. medley-woodii Hutchinson 0.71 S. achilleifolius DC. 0.90 (inc. sed., fide Jeffrey 1986: 934) S. affinis DC. 1.00 S. microglossus DC. 0.82 S. albanensis DC. vaT. doroniciflorus S. mikanioides Otto ex Harvey 0.77 (DC.) Harvey 1.00 (Now Delairea odorata Lem. , fide S. anomalochrous Hilliard 0.93 Jeffrey 1986: 933) S. arabidifolius O. Hoffmann 0.80 S. mooreanus Hutchinson 1.08 S. asperulus DC. 1.00 S. natalicola Hilliard 1.22 S. barbatus DC. 1.00 S. ngoyanus Hilliard 0.90 S. barbertonicus Klatt 0.73 S. othonniflorus DC. 0.88 S. brachypodus DC. 1.00 S. oxyriifolius DC. 1.10 S. brevidentatus M.D. Henderson 1 0.91 S. paludaffinis Hilliard 0.92 S. brevilorus Hilliard 2 0.88 S. panduriformis Hilliard 0.60 S. bupleuroides DC. 0.92 S. pleistocephalus Spencer Moore 1.00 S. cathcartensis O. Hoffmann 0.75 S. polyanthemoides Sch. Bip. 2 1.00 S. caudatus DC. 1.00 S. polyodon DC. S. chrysocoma Meerburgh 1.00 vaT. polyodon 1.00 S. cissampelinus (DC.) Sch. Bip. 1.17 vaT. subglaber (0. Kuntze) (Now Mikaniopsis cissampelina (DC.) Hilliard & Burtt 1.00 C. Jeffrey, fide Jeffrey 1986: 879) S. poseideonis Hilliard & Burtt S. consanguineus DC. 0.89 radiate form 1.10 S. coronatus (Thunberg) Harvey 1.25 discoid form 0.90 S. deltoideus Lessing 1.09 S. sp. aff. S. poseideonis 1.00 S. discodregeanus Hilliard & Burtt 0.86 S. praeteritus Killick 1.18 S. dregeanus DC. 1.50 S. pterophorus DC. 0.92 S. erubescens Aiton S. purpureus L. 0.75 vaT. erubescens 1.00 S. radicans (L.f.) Sch. Bip. 0.67 vaT. crepidifolius DC. 0.79 S. retrorsus DC. 0.93 vaT. incisus DC. 0.64 S. rhomboideus Harvey 0.77 vaT. dichotomus DC. 1 0.90 S. rhyncholaenus DC. 0.83 S. fulgens (J.D. Hooker) Nicholson 2 1.00 S. sandersonii Harvey 1.18 (Now fulgens Hook. f. , fide S. saniensis Hilliard & Burtt 0.77 Jeffrey 1986: 935) S. scitus Hutchinson & Burtt Davy 0.92 S. gerrardii Harvey 1.00 S. seminiveus Wood & Evans 0.83 S. glaberrimus DC. 1.08 S. serratuloides DC. 0.71 S. glanduloso-lanosus Thellung 0.90 S. skirrhodon DC. 0.75 S. glanduloso-pilosus Vol kens & Muschler 0.82 S. speciosus Willdenow 1.13 S. harveianus MacOwan 0.60 S. sp. aff. S. speciosus 1.33 S. hastatus L. 2 0.85 S. subcoriaceus Schlechter 0.69 306 S.-Afr.Tydskr. Plantk. , 1989,55(3)

Table 1 Continued Results The data of the four pollen characters recorded for each Score ratio of the species investigated, are provided in Table 1.

Character 2 Character 3 Shape of pollen grains S. subrubriflorus O. Hoffmann 0.82 All the species studied (Table 1) , have spheroidal pollen S. tamoides DC. 0.92 grains (Erdtman 1969) (Figure 1). (inc. sed. , fide Jeffrey 1986: 934) S. tanacetopsis Hilliard 1.15 Spinule frequency S. transvaalensis Bolus 1.15 The conical-shaped spines forming the echinate appear­ (Now Emilia transvaalensis (Bolus) ance of the intercolpoid regions of the exine of the pollen C. Jeffrey, fide Jeffrey 1986: 919) grains are here termed spinules, since they are always S. umgeniensis Thellung 1.10 shorter than 3 fLm (Erdtman 1969). S. urophyllus Conrath 1.08 The majority of the species studied (Table 1), had a S. variabilis Sch. Bip. 1 0.64 relatively high frequency of spinules (Figure 1), while S. viminalis Bremekamp 2 0.91 the following 10 species had a relatively low spinule Cape heterochromous non-yellow (purple) Senecios frequency: S. polyanthemoides Sch. Bip. (Figure 2), S. S. arenarius Thunb. 1 1.00 hastatus L., S. lydenburgensis Hutchinson & Burtt S. cakilefolius DC. 1.00 Davy, S. brevilorus Hilliard, S. viminalis Bremekamp, ~ . S. elegans L. 1.00 fulgens (Hook. f.) Nicholson, S. lautus Wild., S. S. grandiflorus Berg. 1.00 syringifolius O . Hoffm., Gynura divaricata (L.) DC. and S. multibracteatus Harv. 1.00 Kleinia grandiflora (DC.) N. Rani. Non-southern African Senecios S. californicus DC. 1.17 Length:basal width ratio of the spinules S. desfontainei Druce 0.64 The length:basal width ratio of the spinules ranged from S. erechtitoides Bak. 1.13 0.60 to 1.50 with the majority of the species having ratios S. gallicus Chaix 1.08 in the range 0.80 to 1.00, indicating that, in the majority S. lautus Willd. 2 1.00 of the species, the spinules are very broad-based (Figure S. sylvaticus L. 1.17 5) to moderately broad-based (Figure 3). S. vernalis Waldst. Kit. 0.92 S. viscosus L. 1.10 Prominence of the colpi S. vulgaris L. 1.14 All the species investigated have tricolporate pollen S. hockii De Wild. & Musch!. 1.00 grains (Erdtman 1969) , characterized by noticeably wide (Now Emilia hockii colpi (Figure 1). (De Wild. & Musch!.) C. Jeffrey, fide Jeffrey 1986: 912) Other observations S. syringifolius O .Hoffm. 2 1.00 The exine, of all the species studied, has a rough texture (inc. sed., fide Jeffrey 1986: 885) between the spinules (Figure 4) and in the colpi (Figure Species of other genera in the subtribe Senecioninae 1) . This rough relief usually stops at the bases of the Cineraria geifolia L. 0.93 spinules but in some instances extends a little distance Crassocephalum cernuum (L.f.) Moench above the bases of the spinules. The distal ends of the nom. illeg. 1.00 spinules are entire and smooth and their apices are acute (Now C. rubens (Jacq.) S. Moore, (Figure 4) . The appearance of the spinules is supported Jeffrey 1986: 904) by the unpublished (Skvarla, pers. comm.) TEM micro­ Emilia flammea Casso nom. superfl. illeg. 0.91 graph of a portion of an intercolpoid region of an aceto­ (Now E. javanica (Burm. f.) Merr., lysed pollen grain of S. desfontainei Druce (Figure 6). Jeffrey 1986: 908) Gynura auriculata Casso 2 1.25 Discussion (Now G . divaricata (L.) DC., fide The evidence from this study suggests that the external Jeffrey 1986: 929) pollen morphology of the species studied (Table 1), is Kleinia neriifolia Haw. 1 1.17 reasonably consistent in the genus Senecio and in the Notonia grandiflora DC. 2 1.00 species of the closely related genera in the sub tribe (Now Kleinia grandiflora (DC.) Senecioninae, with very little variation so far observed. N. Rani. fide Jeffrey 1986: 935) These observations concur with most of those of Jeffrey et al. (1978). It is customary to use acetolysed pollen grains in most A list of all the specimens sampled is available from pollen morphological studies (Erdtman 1969). However, the senior author. Blackmore (1984) argues that acetolysis is unnecessary S.Afr.J. Bot., 1989,55(3) 307

Figure 1 Pollen grains of Senecio sp. aff. S. speciosus to Figure 3 A portion of a pollen grain of S. polyodon DC. var. exhibit: a. the spheroidal shape of the pollen , which is charac­ polyodon exhibiting the moderately broad-based spinules teristic of all the taxa studied (Table 1); b. the character state which are characteristic of the pollen of many of the taxa - relatively high spinule frequency (see text for definition); c. studied. Scale bar = 2 j-lm (from: Hilliard & Burtt 9478, NU). the noticeably wide col pi , which are characteristic of the pollen of all the taxa studied; d . the rough texture of the colpus membrane. Scale bar = 10 j-lm (from: Hilliard 5236, NU). of preparative techniques now available, combined with the great depth of field and the high resolution of the latest models of scanning electron microscopes, can now for most studies based on SEM alone. The decision to provide almost all the morphological data previously use unacetolysed pollen grains in the present study was only obtainable through the use of transmission electron based on the consideration that the four features to be microscopes. Furthermore, the preparation for SEM observed would be unaffected by the presence of Pollen­ studies is generally a lot simpler than that for TEM. kitt, and avoiding the acetolyis procedure would speed Besides, the three-dimensional results, obtained by up the investigation of the large number of taxa (111) viewing the wall structure of either sectioned pollen investigated. Furthermore, the study by Jeffrey et al. grains or of broken walls of pollen grains, are easier to (1978), in which the same four features were observed, interpret. A simple and very effective method of also used unacetolysed pollen grains. preparing and sectioning pollen for SEM has been It could be argued that any study of pollen morph­ described by Blackmore & Dickinson (1981). Blackmore ology should be based on light microscopy, transmission (1981, 1982) has used this method in a palynological electron microscopy and scanning electron microscopy. However, Blackmore (1984) points out that the variety

Figure 4 A portion of a pollen grain of S. vulgaris L., exhibiting: a. the wide perforations around the bases of the Figure 2 A pollen grain of S. polyanthemoides Schultz spinules and the narrower perforations between the spinules; Bipontinus, exhibiting the character state - relatively low b. the rough texture of the exine between the spinules; c. the spinule frequency (see text for definition). Scale bar = 5 j-lm distal ends of the spinules which are entire and smooth and (from: Hilliard 4026, NU). with acute apices. Scale bar = 2 j-lm (from: Britton s.n., K). 308 S.-Afr.Tydskr. Plantk., 1989,55(3)

generic concept of Senecio. In the light of the observations and comments by Blackmore (1984), further studies of the pollen morph­ ology of Senecio and related genera in the Senecioneae are being initiated using SEM on acetolysed pollen grains. The emphasis is now being placed on the wall structure of sectioned and/or broken walls of pollen grains, as described by Blackmore & Dickinson (1981). The data thereby obtained will be evaluated with respect to elucidating the generic concept of Senecio and other aspects of the systematics of the Senecioneae, such as the 'senecioid' and 'helianthoid' structural patterns of the pollen wall, already observed in a number of the Sene­ cioneae (Skvarla & Turner 1966b; Skvarla et al. 1977).

Acknowledgements Figure 5 A portion of a pollen grain of S. desfontainei Druce, The authors acknowledge financial support from the exhibiting: a. the very broad-based spinules which are charac­ CSIR and from the University of Natal, while the senior teristic of the pollen of many of the taxa studied. Scale bar = 2 author was a postgraduate student of the University of J.lm (from: Abd EI Ghani, s.n., K). Natal. The use of the research facilities of the Depart­ ment of Botany and the Electron Microscopy Unit, University of Natal, Pietermaritzburg is also gratefully survey of the the tribe Lactuceae. In this study the acknowledged. The role played by Prof. O.M. Hilliard 'anthemoid', 'helianthoid' and 'senecioid' patterns of in introducing the senior author to the very exciting field exine described by Skvarla et al. (1977) on the basis of of synantherology, particularly studies in Senecioneae, is TEM, were readily distinguished using SEM. Further­ very gratefully acknowledged. more, in SEM investigations of the complex exines of Asteraceae pollen, the spatial arrangements of the exine References components are much more clearly discernible, this BLACKMORE, S. 1981. Palynology and intergeneric being in part due to the great depth of field available relationships in subtribe Hyoseridinae (Compositae: with SEM (Blackmore 1984). Lactuceae). Bot. J. Linn. Soc. 82: 1- 13. These facts are particularly relevant to future palyn­ BLACKMORE, S. 1982. Palynology of subtribe ological studies in the Senecioneae, for it appears from Scorzonerinae (Compositae: Lactuceae) and its taxonomic the present study that data from the external morphol­ implications. Grana 21: 149-160. ogy will neither contribute much to an understanding of BLACKMORE, S. 1984. Pollen features and the affinities of taxa within the senecioid complex in the systematics. In: Current concepts in plant , eds tribe nor will it contribute much to the elucidation of the Heywood, V.H. & Moore, D.M., pp. 135-154, Academic Press, London, New York. BLACKMORE, S. & DICKINSON, H.G. 1981. A simplc method for sectioning pollen grains. Pollen Spores 23: 281-285. ERDTMAN, G. (1969). Handbook of palynology. Hafner Publishing Co. Inc., New York. JEFFREY, C. 1980. Generic and sectional limits in Senecio (Compositae): II. Evaluation of some recent studies. Kew Bull. 34(1): 49-58. JEFFREY, C. 1986 . The Senecioneae in East Tropical Africa. Notes on Compositae. IV. Kew Bull. 41(4): 873-943. JEFFREY, c., HALLIDAY, P., WILMOT-DEAR, M. & JONES, S.W. 1978. Generic and sectional limits in Senecio (Compositae): 1. Progress Report. Kew Bull. 32(1): 47-67. SKVARLA, J.J. (pers. comm.). Department of Botany and Microbiology, University of Oklahoma, 770 Van Vleet Oval, Norman, Oklahoma, 73019, U.S.A. SKVARLA, J.J. & TURNER, B.L. 1966a. Pollen wall ultrastructure and its bearing on the systematic position of Blennosperma and Crocidium (Compositae). Am. J. Bot. Figure 6 A transmission electron micrograph of a sectioned 53: 555-563. portion of an intercolpoid region of an acetolysed pollen grain SKVARLA, J.J. & TURNER, B.L. 1966b. Systematic from S. desfontainei, illustrating the rel.atively entire and implications from electron microscope studies of smooth appearance of the spinules with their acute apices. Compositae pollen - a review. Annis Miss. Bot. Gdn 53: Scale bar = 2 J.lm (voucher specimen details unknown). 200-256. S.Afr.J.Bot.,1989,55(3) 309

SKVARLA, J.J., TURNER, B.L., PATEL, V.c. & TOMB, STIX, E. 1960. Pollenmorphologische untersuchungen and A.S. 1977. Pollen morphology in the Compositae and in Compositae. Grana Palynologica 2: 41-114. morphologically related families. In: The biology and chemistry of the Compositae, eds Heywood, V.H., VINCENT, P.L.D. & GETLIFFE, F.M. 1989. Elucidative Harborne, J.B. & Turner, B.L., pp. 141-248, Academic studies on the generic concept of Senecio (Asteraceae). 1. Press, London, New York. Linn. Soc. (in press).