South African Journal of Botany 2004, 70(4): 646–653Printed in South Africa — All rights reserved

Copyright © NISC Pty LtdSOUTH AFRICAN JOURNAL

OF BOTANYISSN 0254–6299

The earliest occurrence of Angiosperms in southern Africa

MS Zavada

Department of Biology, Providence College, 549 River Avenue, Providence, Rhode Island 02918-0001, United States ofAmericae-mail: mzavada@providence.edu

Received 22 December 2003, accepted in revised form 4 June 2004

Three outcrop localities and 28 onshore and offshorecores and wells were sampled from the east coast,Zululand and Durban Basins, the south coast onshoreAlgoa and the offshore Outeniqua Basins, and the westcoast Orange Basin. Samples ranging in age from theLate Jurassic to the Late Cretaceous were examined formegafossils, microfossils, and fossilised wood to deter-mine the first occurrence of angiosperms in southernAfrica. In all three regions angiosperms first appear inthe Aptian and they are represented by monosulcate,reticulate pollen genera, e.g. Clavatipollenites.

Beginning in the Albian, tricolpate and tricolporatepollen types occur in low frequency with the monosul-cate angiosperm pollen. This first occurrence ofangiosperms in South Africa is more than 25 millionyears later than the earliest global appearance ofangiosperms in the Middle East and north Africa. Theconcurrent presence of monosulcate and tri-aperturatepollen suggests that angiosperms migrated to SouthAfrica via dispersal corridors and that angiosperms didnot originate in southern Africa.

Angiosperms are a monophyletic group (Judd et al. 2002).Although the time and the place of their origin is equivocal,their monophyly necessitates derivation from a commonancestor in a specific geographic location. It can be arguedthat with increasing distance from the geographic origin,angiosperms should appear proportionally later in the geo-logic time.

South Africa occupied a central position in Gondwana,having direct crustal connection to South America,Antarctica, and Madagascar/India (Smith et al. 1994). EastGondwana (Australia, Antarctica, Madagascar, India) startedseparation from West Gondwana (Africa and South America,and the Falkland Plate) in the mid-Jurassic approximately170My before the final separation of South America andAfrica in the Cenomanian (100My). The separation and sub-sequent breakup of Gondwana began in the late Jurassicand continued into the Cretaceous severing the continentaldispersal corridors with the adjacent land masses (Brown etal. 1995). The earliest unequivocal global record ofangiosperms is from the earliest Cretaceous (Berriasian–Valanginian) of Israel (Brenner 1996). There are earlierreports of angiosperm-like pollen from the Triassic of easternNorth America (Cornet 1989, also see Litwin 1985, Doyleand Hotton 1991) and the Late Jurassic of France (Cornetand Habib 1992), however, the taxonomic position of theseremains are equivocal.

The purpose of this study is to determine the earliestoccurrence of angiosperms in southern Africa and discuss

what bearing this may have on the subsequent developmentof angiosperm floras in this region.

Material and Methods

Sample localities

Onshore and offshore Cretaceous sediments in South Africawere sampled for megafossils, microfossils and fossilisedwood (Table 1, Figure 1). The easternmost Cretaceousbasin (Mozambique Basin) outcrops along the southeasterncoast of South Africa (Flores 1973, Dingle et al. 1983, Brownet al. 1995, Figure 1). These Early Cretaceous sedimentsextend and thicken northward into Mozambique. The lowerCretaceous sediments are primarily comprised of theMakatini Formation, which ranges in age from the Barremianto the late Aptian (McLachlan and McMillian 1979). There isan outlier of the Makatini Fm. in the vicinity of the UmfolosiRiver, south of Mtubatuba in Zululand (Umkwelane Outlier,Frankel 1960) that is believed to represent the more wide-spread occurrence of the Makatini Fm. in the past. Two out-crop localities were sampled in this basin, one locality in theMkuze Game Reserve, and the Umkwelane Outlayer (Table1, Figure 1). Four onshore cores (Figure 1(1–3 and 7), Table1), and three offshore wells (Figure 1(4–6), Table 1) weresampled for palynomorphs. The plant fossils recovered fromall of the localities in this basin include impression fossils,fossilised wood, and palynomorphs.

Introduction

South African Journal of Botany 2004, 70: 646–653 647

The Outeniqua and associated onshore basins occuralong the south coast of Africa. A single outcrop locality wassampled in the Kirkwood and Sundays River Formationsnear Dumbrodie (Table 1, Figure 1). These formations rangein age from Berriasian to early Aptian (Dingle et al. 1983).Three onshore wells (Figure 1(18–20), Table 1) and ten off-shore wells were sampled for palynomorphs. (Figure1(8–17), Table 1). Compression and impression fossils, fos-silised wood and microfossils were recovered from thisbasin.

The Orange Basin occurs along the West coast, and eightoffshore wells were examined for palynomorphs (Figure1(21–28), Table 1). Among the eight offshore wells exam-ined, Deep Sea Drilling Project 361 located off the south-west coast of South Africa has been previously studied(Figure 1(21), Table 1) (McLachlan and Pieterse 1978).Sample interval 48-23 of DSDP-361 was re-examined in thisstudy. This interval ranges in age from Aptian toCenomanian (MacLachlan and Pieterse 1978, Davey 1978).

Both the onshore and offshore localities from all threeregions collectively range in age from the Late Jurassic tothe Late Cretaceous, and yielded a variety of plant fossilmaterial. Plant fossils from some of the localities have beenpreviously studied, e.g. the Makatini Fm, Zululand Basin(Frankel 1960, Flores 1973, Anderson and Anderson 1985),the Kirkwood and Sundays River Formations, Algoa andOuteniqua Basins (Scott 1976, Stapleton and Beer 1977,Anderson and Anderson 1985, Bamford 1986) and OrangeBasin (DSDP-361, McLachlan and Pieterse 1978, Bamfordand Corbett 1994, 1995).

The fossils collected from the outcrops during this studyare part of the permanent fossil collections of the BernardPrice Institute of Palaeontology, University of theWitwatersrand, Johannesburg, RSA. The collections arecurrently housed in the Department of Biology, ProvidenceCollege, Providence, RI, USA. All residues and slides pre-pared from the cores, side wall cores, and cuttings aredeposited at the Petroleum Agency, Parow, South Africa.

Table 1: First occurrence of angiosperms in outcrops and wells from the major depositional basins along the east, south and west coasts ofsouthern Africa. The table gives the depth of the first occurrence and the age of the first occurrence based on foraminifera (Brown et al. 1995).Unless noted all wells are offshore (OS = Onshore). ‘Absent’ indicates that no angiosperms were recovered from that interval in that particu-lar outcrop or well. SWC = Side Wall Core

Well type Depth (m) Age of first occurrence Sample typeEast Coast (Mozambique Basin)1) ZC (OS) 1 462 Albian Core2) B (OS) 1 283 Albian Core3) NZA (OS) Below 310 Albian Core4) JC-C1 2 205 Aptian SWC5) JC-A1 N/A Absent – Barremian – Aptian SWC6) JC-B1 N/A Absent – Barremian SWC7) ZU 1/77 (OS) N/A Absent – Barremian CoreMkuze Game Reserve N/A Absent – Barremian – Aptian OutcropUmkwelane N/A Absent – Barremian – Aptian Outcrop

South Coast (Outeniqua and associated basins)8) E-1 1 102–1 105 Aptian SWC9) E-AA1 2 523 L. Aptian – Albian SWC10) E-AD1 N/A Absent – Barremian – Aptian SWC11) E-B1 1 205–1 210 Aptian Cutting12) F-E1 1 702 Aptian SWC13) PB-A1 N/A Absent – L. Jurassic – Barremian SWC14) GA-B1 1 036–1 097 Aptian Cutting15) G(b)Gemsbok-1 1 145–1 150 Albian Cutting16) G(b)Springbok-1 1 481.7 Albian SWC17) Hb-A1 910–920 Albian Cutting18) AL 1/69 (OS) N/A Absent – Valanginian – Hauterivian SWC19) BR 1/71 (OS) N/A Absent – Portlandian – Hauterivian SWC20) BT 1/74 (OS) N/A Absent – Valanginian – Hauterivian SWCDunbrodie N/A Absent – Berriasian – Aptian Outcrop

West Coast (Orange Basin)21) DSDP-361 1 050 Albian Core22) K-E1 3 752 Albian SWC23) K-B1 3 900 Albian SWC24) K-D1 3 174 Albian SWC25) A-C1 2 612 Albian SWC26) BA-A1 1 052–955 Albian Cutting27) A-A1 2 452 E. Aptian – Albian SWC28) A-D1 2 515 E. Aptian – Albian SWC

Zavada648

Sample preparation

Fossil pollen in this study was macerated from the sedi-ments using a modified version of Barss and Williams(1973). Pollen from the samples was examined using light,scanning electron (SEM), and transmission electronmicroscopy (TEM). Light microscopic slides were preparedby dehydration of the pollen residue in an alcohol series(10%, 50%, 95%, 100%) followed by acetone, then xylene.Once in xylene, the sample was mixed with the polystyrenemounting medium ProTexx and permanent slides weremade for examination on a Zeiss Axiophot light microscope.The isolation of individual pollen grains from the pollenresidues for SEM and TEM followed the method of Zavada(2003).

Examination of the impression and compression fossilswas done on a Zeiss SV-11 Stereomicroscope equipped forphotomicroscopy. Petrographic thin sections were made of across, radial and tangential section of the petrified woodrecovered from various localities.

Results

The collections of compression, impression, and petrifactionmegafossils from the outcrop localities produced no speci-mens that can unequivocally be considered angiosperm.None of the palynological samples taken from outcrop local-ities produced angiosperm pollen. All records of firstangiosperm occurrence are based on pollen recovered fromthe onshore and offshore wells.

The pollen floras described in this study include taxa thatare typical of early Cretaceous floras from South America,

southern Africa and India (Singh 1972, Herngreen andChlonova 1981 and references therein, e.g. Figures 2–8).The earliest occurrence of angiosperms in southern Africa ispollen from the Aptian interval of the offshore and onshorewells (Table 1). All of the Aptian occurrences of angiospermpollen are derived from sidewall cores or cuttings, and aresusceptible to down hole contamination, and may be derivedfrom younger sediments (possibly Albian or younger).Angiosperms may have been present in southern Africa bythe Aptian, however, they were not widespread, eleven ofthe thirty-one localities produced no record of angiospermremains in intervals that ranged as high as Aptian. However,in all of the offshore and onshore wells angiosperms arepresent in Albian and younger sediments.

The most commonly recorded angiosperm pollen fromAptian aged sediments in the offshore and onshore wells areClavatipollenites, Retimonocolpites, and/or Liliacidites(Figures 9–16) all occur in low frequency, and are often sin-gle occurrences (McLachlan and Pieterse 1978).Multiaperturate pollen (tricolpate and tricolporate types,Figures 17–26) is restricted to Albian and younger age sed-iments in all of the offshore and onshore wells examined inthis study.

Systematic palynology

Pollen of possible Gnetalean affinity

The occurrence of Equisetosporites and Steevesipollenites,in association with Concavisporites sp., Cicatricosporitesaustraliensis (Cookson) Potonie, Monosulcites sp., and anabundance of Classopollis sp. relate the South Africa floras(Scott 1976, Mclachlian and Pieterse 1978) to those of earlyCretaceous floras of South America (Doyle et al. 1977,Herngreen and Chlonova 1981), and support the hypothesisthat a strong South American-African exchange of bothplants and animals existed prior to the break up ofGondwana (Amedegnato 1993, Storch 1993, Buffetaut andRage 1993). In addition, Cyclusphaera, a distinctive LowerCretaceous pollen type recovered from Argentina, is alsopresent in South African wells (Volkheimer et al. 1977,Zavada 1987, 1992, Taylor et al. 1987), but is absent by thelate Early Cretaceous, indicating that terrestrial dispersalcorridors were severed by this time.

Equisetosporites concinnus

Singh (1964) Figure 2: this pollen type is widespread inLower Cretaceous sediments throughout the world and iscommon in the Lower Cretaceous section in southern Africa.It may have gnetalean affinities (Osborn et al. 1993).

Steevesipollenites sp.

(McLachlan and Pieterse 1978), Figures 3–5. Severalspecies of this genus are present in the Triassic–Jurassic ofNorth America, South America and Africa (Cornet 1989).Gross morphology of this pollen type suggests an affinitywith the Gnetales (Osborn et al. 1993). The pollen wall ultra-structure of the form recovered from DSDP-361 also sug-

Figure 1: Map showing the approximate locations of the collectionsites and the onshore and offshore wells. The black shaded areasare the approximate extent of Cretaceous sediments within thesedepositional areas, the numbers correspond to the numbered wellslisted in Table 1

����

��

��

����

��

��

���

��

��

������

��

� ���

���

���������������������

�����������������

�� �!"#�

$��%&�'(�)*'

+�%$,'-'

-'�)+)' ��.'�+)/�0

���

-

gests a gnetalean affinity. The robust ribs (Figure 3) possessa thin tectum, a granular infrastructure and a thin continuousbasal layer (Figures 4–5). Steevesipollenites has a pollenmorphology similar to that found in Ephedra and Welwitchia(Osborn et al. 1993), except for the thin basal layer, which isthick in Ephedra and Welwitchia.

Pollen of possible Angiosperm affinity

cf. Cycadopites sp. (Scott 1976, McLachlan and Pieterse1978), Figures 6–8. The specimens recovered from DSDP-361 have a gross morphology similar to Cycadopites, whichis characterised by a monosulcate aperture, circular to ovalshape, and a wall structure presumed to be alveolar or gran-ular. All of these characteristics are indicative of pollen ofcycadalean or bennettitalean affinity. The form recoveredfrom DSDP-361 (Figures 6–8), however, possesses a thicktectum that is traversed by minute channels (i.e. the tectumis microperforate) (Figures 7–8), an infrastructural layer ofthin cylindrical shaped columellae that rest on a thin basallayer (Figures 7–8), a suite of characteristics more indicativeof angiosperms. This pollen type can only be distinguishedfrom monosulcate, round to oval, psilate, cycad- or bennet-titalean pollen with an alveolor or granular homogeneouswall structure by pollen wall ultrastructure. Only one speci-men recovered from DSDP-361 was examined with TEM,hence the stratigraphic range and distribution of this col-umellate pollen type in South Africa is unknown.

Clavatipollenites hughesii

Couper (1958), Figures 9–11: this species is characterisedby being oval in outline, monosulcate with reticulate exinesculpturing (Figures 9–10). The muri of the reticulum appearsomewhat angular (Figures 9–10). The pollen wall structureis tectate columellate, the tectum is thin, however, the basallayers appears thick and massive (Figure 11). It is believedto have chloranthalean affinities. Similar forms have beenrecovered from fossil floral remains of the Chloranthaceae(Archangelsky and Taylor 1993, Eklund et al. 1997, Friis etal. 1997).

Clavatipollenites minutus

Brenner (1963), Figures 12–16: this species is characterisedby oval shape, monosulcate aperture and the reticulateexine sculpturing (Figures 12–14). The perforations of C.minutus are smaller then those of C. hughesi (compareFigure 9 and Figure 12). The pollen wall structure differsfrom that of C. hughesi in having a thick tectum, thin stoutcolumellae (Figure 16), and a thin basal layer (Figure 15). C.minutus is widespread in Lower Cretaceous sediments ofSouth Africa.

Tri-aperturate Angiosperm pollen

Beginning in the Albian a low frequency of tricolpate or tri-

Figures 2–8: (2) SEM of Equisteosporites concinnus, scale bar = 5µm; (3–5) Steevesipollenites sp.; (3) SEM showing the robust ribs, scalebar = 5µm; (4) TEM showing two ribs, the thin footlayer, granular infrastructural layer and relatively thin tectum, scale bar = 2.5µm; (5)Longitudinal TEM section of a rib showing the continuous granular infrastructure, scale bar = 1.25µm; (6–8) Cycadopites sp.; (6) SEM, scalebar = 5µm; (7) TEM showing the columellate infrastructure. Note as the section becomes more tangential the cross sections of the columel-lae are circular (arrow) scale bar = 1µm; (8) TEM showing the thick tectum with small perforations (arrow), the thin cylindrical columellae, andthin footlayer, scale bar = 0.5µm

South African Journal of Botany 2004, 70: 646–653 649

Zavada650

colporate pollen occurs concurrently with the monosulcate,angiosperm pollen. This suggests that the migration of theearliest angiosperms into the South African region was soonfollowed by the more derived eudicots. Below is a represen-tative suite of tri-aperturate pollen types recovered fromDSDP-361.Fraxinoipollenites venustus: Singh (1971), Figure 17;Tricolpites sp. 4 (McLachlan and Pieterse 1978), Figure 18;Tricolpites sp. 9 (McLachlan and Pieterse 1978), Figure 19;Gemmatricolpites scabratus Van Hoeken-Klinkenberg 1964,Figures 20–21;Tricolpites sp. (McLachlan and Pieterse 1978), Figure 22;Tricolporopollenites sp. (Similar to Plate 5, Figure 25;McLachlan and Pieterse 1978), Figures 23–26.

Based on pollen size, shape, sculpturing and aperture typea variety of eudicots are present in South Africa beginning

in the Albian. Few of these pollen types from South Africahave been investigated with SEM and TEM. These tech-niques can potentially increase the number of taxonomiccharacters to better understand the botanical affinities of tri-aperturate pollen. Tricolporopollenites sp., recovered fromDSDP-361, is tricolporate, triangular in shape, oblate, thecolpi extending nearly to the poles (Figure 23). Exine sculp-turing is slightly rugulate (Figure 23). In non-apertural areasthe pollen is atectate, homogeneous, and endexineappears to be present (a less electron dense inner layerthan the outer atectate, homogeneous ectexine) (Figure24). Adjacent to the apertural region the tectum is thick andis underlain by a granular to columellate infrastructurewhich rest on a irregular basal layer (Figures 25–26),endexine appears to be present and thickens toward theaperture (Figures 25–26).

Figures 9–15: (9–11) Clavatipollenites hughesii: (9) SEM, scale bar = 5µm; (10) SEM, scale bar = 5µm; (11) TEM of the same grain in Figure10 showing massive footlayer, columellate infrastructure and thin tectum, scale bar = 5µm; (12–16) Clavatipollenites minutus: (12) SEM, scalebar = 5µm; (13) SEM of a grain similar to those in Figures 12 and 14, scale bar = 5µm; (14) SEM, scale bar = 5µm; (15) TEM of the samegrain in Figure 12 showing thin footlayer, the columellate infrastructure and thick tectum, scale bar = 3.5µm; (16) TEM of the same grain inFigures 12 and 15 showing columellate infrastructure and thick tectum, scale bar = 0.75µm

Figures 17–26: (17) SEM of Fraxinoipollenites venustus, scale bar = 10µm; (18) SEM of Tricolpites sp. 4 (McLachlan and Pieterse 1978),scale bar = 5µm; (19) SEM of Tricolpites sp. 9 (McLachlan and Pieterse 1978), scale bar = 10µm; (20–21) Gemmatricolpites scabratus: (20)SEM, scale bar = 5µm; (21) SEM of the flip side of the grain illustrated in Figure 20, scale bar = 5µm; (22) SEM of Tricolpites sp., scale bar= 5µm; (23–26) Tricolporopollenites sp.; (23) SEM, scale bar = 5µm; (24) TEM of the apertural region of the same grain in Figure 23 show-ing the thick tectum, the infrastructural layer with irregular shaped columellae and the thickened inner layer, scale bar = 3µm; (25) TEM of thesame grain in figure 23 in the region of the aperture showing the columellate-like structures in the the infrastructural region and what appearsto be a less dense endexine (E), scale bar = 1.5µm; (26) TEM of the same grain in Figure 23 showing the homogeneous non-apertural wall,note the ragged, less dense inner layer which may represent endexine, scale bar = 1.5µm

South African Journal of Botany 2004, 70: 646–653 651

Zavada652

Discussion

South Africa occupied a central position in Gondwanathroughout the Triassic and the Early–Mid Jurassic (Brownet al. 1995). The South African region had land connectionsto southern South America, Antarctica, Madagascar/India,and Australia, and shared common faunal and floristic ele-ments during the Triassic through the lowermost Cretaceous(Srivastava 1978, 1981, Herngreen and Chlonova 1981,Dettmann 1992, Doyle 1999). The breakup of the southerncontinents began in the Late Jurassic and by the Barremianthe South African region was separated from South America,Antarctica and Madagascar/India. The rifting that took placeduring these tectonic events led to the formation of deposi-tional basins that now occupy the east, south, and the westcoast of South Africa (Dingle et al. 1983). Palynological evi-dence from the Middle East and North Africa suggests thatangiosperm originated during the earliest Cretaceous(Berriasian) (Brenner 1996), or at some time prior to the ear-liest Cretaceous, and the diversification of the angiospermswas centered in the North African–South American region(Doyle 1999). From the first unequivocal record ofangiosperms in Middle East and North Africa during theBerriasian (Brenner 1996) (145My), by the Barremian(115My) early angiosperm pollen records were consistentlypresent in North America, Europe, South America, northAfrica and Asia and made up a small, but distinctive per-centage of palynofloras from these regions (Traverse 1988,Hughes 1994, and references therein).

The first occurrence of angiosperms in South Africa is inthe Aptian, more than 25 million years after the first occur-rence of angiosperm pollen in Middle East and North Africa(Brenner 1996). In addition, angiosperm pollen records fromthe Aptian in southern Africa are not widespread and makeup an almost negligible percentage of the palynofloras. All ofthe Aptian angiosperm occurrences are from SWC, or cut-tings. These types of samples are susceptible to down holecontamination. The first unequivocal occurrence ofangiosperms is in the Albian, however, if the Aptian occur-rences and distribution of angiosperm pollen are accepted,angiosperms are not widespread floristic elements in SouthAfrica during the Aptian. The Albian occurrences ofangiosperm pollen become more widespread and the occa-sional occurrence of tri-aperturate pollen types (tricolpatesand tricolporates) suggests that by the end of the EarlyCretaceous the frequency and abundance of angiospermswas rising throughout the South African region. This may bedue to the introduction of these elements into this region bydispersal (migration) or by diversification, however, theabundance of angiosperm pollen types is still well below thepercentages recorded from North America, South America,Europe, North Africa, Middle East and Asia (Zavada, work inprogress). It is during the Cenomanian that angiospermsbecome consistently present in South African floras, andcomprise a significant percentage of these palynofloras(McLachlan and Pieterse 1978).

The relatively late arrival and diversification ofangiosperms in South Africa implies that the origin ofangiosperms did not occur in the South African region. Theoccurrence of plant remains from the Triassic through Early

Cretaceous with gnetalean affinitiy and the widespreadGondwana occurrence of corystosperms and peltaspermsduring the early Mesozoic (Anderson and Anderson 1983,1985, Anderson et al. 1999) plant groups considered, at onetime or another, to be possible angiosperm sister groups,suggests that these Gondwana species are not involved inthe origin of angiosperms. Although southern Gondwanamay not have been the geographic centre of the origin of theangiosperms and may have occupied a distant and margin-al relationship to the diversification of angiosperms in theCretaceous, all of the continents that comprised Gondwana,except Antarctica (South America, Southern Africa,Madagascar, Australia and India) are among the most floris-tically diverse regions in the world today. The migration ofangiosperms via fluctuating dispersal corridors from a vari-ety of exotic sources during the Cretaceous into marginalGondwana regions, and the subsequent development ofnew dispersal corridors in the Tertiary due to the changingpositions of the continents, may have contributed to the highfloristic diversity observed in these regions.

Acknowledgements — I wish to thank Dr James Benson of SantosLtd, Adelaide Australia who generated much of the palynologicalwell data while at the Soekor, South Africa, and Dr Ian McLachlan ofthe Petroleum Agency, Parow, South Africa, and the PetroleumAgency who graciously allowed me to view their material and usethe data in Table 1.

References

Amedegnato C (1993) African-American relationships in the acridi-ans (Insecta, Orthoptera). In: George W, Lavocat R (eds) TheAfrican–South American Connection. Oxford Monographs onBiogeography No. 7, Oxford Science Publications, New York, pp59–75

Anderson JM, Anderson HM (1983) Palaeoflora of southern Africa:Molteno Formation (Triassic), Vol. 1: Part 1 — Introduction, Part2 — Dicroidium. AA Balkema, Rotterdam, 227pp

Anderson JM, Anderson HM (1985) Palaeoflora of Southern Africa.Prodromus of South African Megafloras, Devonian to LowerCretaceous. AA Balkema, Rotterdam, 423pp

Anderson JM, Anderson HM, Archangelsky S, Bamford M, ChandraS, Dettmann M, Hill R, Mcloughlin S, Rosler O (1999) Patterns ofGondwana plant colonisation and diversification. Journal ofAfrican Earth Sciences 28: 145–167

Archangelsky S, Taylor TN (1993) The ultrastructure of in situClavatipollenites pollen from the early Cretaceous of Patagonia.American Journal of Botany 80: 879–885

Bamford MK (1986) Aspects of the Palaeoflora of the Kirkwood andSundays River Formations, Algoa Basin, South Africa. MScThesis, University of the Witwatersrand, Johannesburg, 160pp

Bamford MK, Corbett IB (1994) Fossil wood of Cretaceous age fromthe Namaqualand continental shelf, South Africa. PalaeontologiaAfricana 31: 83–95

Bamford MK, Corbett IB (1995) More fossil wood from theNamaqualand coast, South Africa, onshore material.Palaeontologia Africana 32: 67–74

Barss MS, Williams GL (1973) Palynology and nannofossil process-ing techniques. Geological Survey of Canada, Paper 73-26, 25pp

Brenner GJ (1963) The spores and pollen of the Potomac Group ofMaryland. Maryland Department of Geology, Mines and WaterResources, Bulletin 27: 1–215

Brenner GJ (1996) Evidence for the earliest stage of angiospermpollen evolution: a paleoequatorial section from Israel. In: Taylor

DW, Hickey LJ (eds) Flowering Plant Origin, Evolution andPhylogeny. Chapman & Hill, New York, pp 91–115

Brown LF, Benson JM, Brink GJ, Doherty S, Jollands A, JungslagerEHA, Keenan JHG, Muntingh A, Van Wyk NJS (1995) Sequencestratigraphy in offshore South African divergent basins. AmericanAssociation of Petroleum Geologists, Studies in Geology, 41, 184pp

Buffetaut E, Rage J-C (1993) Fossil amphibians and reptiles and theAfrica–South America connection. In: George W, Lavocat R (eds)The African–South American Connection. Oxford Monographs onBiogeography No. 7, Oxford Science Publications, New York, pp87–99

Cornet B (1989) Late Triassic angiosperm-like pollen from theRichmond Rift Basin of Virginia, USA. Palaeontographica B 213:37–87

Cornet B, Habib D (1992) Angiosperm-like pollen from theammonite-dated Oxfordian (Upper Jurassic) of France. Review ofPalaeobotany and Palynology 71: 269–294

Couper RA (1958) British Mesozoic microspores and pollen grains.Palaeontographica B 103: 75–179

Davey RJ (1978) Marine Cretaceous palynology of Site 361, DSDPLeg 40, off southwestern Africa. Initial Reports, Deep Sea DrillingProject 40: 883–913

Dettmann ME (1992) Structure and floristics of Cretaceous vegeta-tion of southern Gondwana: Implications for angiosperm bio-geography. Palaeobotanist 41: 224–233

Dingle RV, Siesser WG, Newton AR (1983) Mesozoic and Tertiarygeology of Southern Africa. AA Balkema, Rotterdam, 375pp

Doyle JA (1999) The rise of angiosperms as seen in the AfricanCretaceous pollen record. In: Scott L, Cadman A, Verhoeven R(eds) Palaeoecology of Africa and the Surrounding Islands. AABalkema, Rotterdam, 26, pp 3–29

Doyle JA, Biens P, Doerenkamp A, Jardin S (1977) Angiospermpollen from the Pre-Albian Lower Cretaceous of Equatorial Africa.Bulletin du Centre Recherches Elf-Aquitaine Exploration-Production 1: 451–473

Doyle JA, Hotton CL (1991) Diversification of early angiospermpollen in a cladistic context. In: Blackmore S, Barnes SH (eds)Pollen and Spores. Systematics Association, Special Volume 44:169–195

Eklund H Friis EM, Pedersen KR (1997) Chloranthaceous floralstructures from the Late Cretaceous of Sweden. PlantSystematics and Evolution 207: 13–42

Flores G (1973) The Cretaceous and Tertiary sedimentary basins ofMozambique and Zululand. In: Blant G (ed) Sedimentary Basinsof the African Coasts. Part 2: South and East Coasts. ParisAssociation of African Geological Surveys, pp 81–111

Frankel JJ (1960) The geology along the Umfolozi River, south ofMtubatuba, Zululand. Transactions of the Geological Society ofSouth Africa 63: 231–252

Friis EM, Crane PR, Pederesen KR (1997). Anacostia, a new basalangiosperm from the Early Cretaceous of North America andPortugal with trichotomocolpate/monocolpate pollen. Grana 36:225–244

Herngreen GFW, Chlonova AF (1981) Cretaceous microfloralprovinces. Pollen et Spores 23: 441–555

Hughes NF (1994) The Enigma of Angiosperm Origins. CambridgeUniversity Press, Cambridge, 303pp

Judd WS, Campbell CS, Kellogg EA, Stevens PF, Donoghue MJ(2002) Plant systematics: a phylogenetic approach (2nd edn).Sinaur Association, Sunderland, MA

Litwin RJ (1985) Fertile organs and in situ spores of ferns from theLate Triassic Chinle Formation of Arizona and New Mexico.Review of Palaeobotany Palynology 44: 101–146

McLachlan IR, McMillian IK (1979) Microfaunal biostratigraphy,chronostratigraphy and history of Mesozoic and Cenozoicdepositson the coastal margin of South Africa. Geokongress 77,Special Publications of the Geological Society of South Africa 6:161–181

McLachlan IR, Pieterse E (1978) Preliminary palynological results:Site 361, Leg 40, Deep Sea Drilling Project. Initial Reports, DeepSea Drilling Project 40: 857–881

Osborn JM, Taylor TN, De Lima MR (1993) The ultrastructure of fos-sil ephedroid pollen with gnetalean affinities from the LowerCretaceous of Brazil. Review of Palaeobotany and Palynology77: 171–184

Scott L (1976) Palynology of Lower Cretaceous deposits from theAlgoa Basin (Republic of South Africa). Pollen et Spores 18:563–609

Singh C (1964) Microflora of the Lower Cretaceous MannvilleGroup, east central Alberta. Resource Council of Alberta Bulletin15: 1–239

Singh C (1971) Lower Cretaceous microfloras of the Peace Riverarea, northwestern Alberta. Resource Council of Alberta Bulletin28: 1–310

Singh HP (1972) Palynology of Lower Cretaceous sediments ofIndia. Seminar Palaeopalynology Indian Stratigraphy, pp 159–166

Smith AG, Smith DG, Funnel BM (1994) Atlas of Mesozoic andCenozoic coastlines. Cambridge University Press, Cambridge,99pp

Stapleton P, Beer EM (1977) Micropalaeontological age determina-tion for the Brenton Beds. In: Papers on BiostratigraphicResearch. Bulletin of the Geological Survey of South Africa 60:1–10

Storch G (1993) ‘Grube Messel’ and African-South American faunalconnections. In: George W, Lavocat R (eds) The African–SouthAmerican Connection. Oxford Monographs on Biogeography No.7, Oxford Science Publications, New York, pp 76–86

Srivastava SK (1978) Cretaceous spore-pollen floras: A global eval-uation. Biological Memoirs, Palaeopalynology Series-5 3: 2–130

Srivastava SK (1981) Evolution of Upper Cretaceous phytogeo-provinces and their pollen flora. Review of Palaeobotany andPalynology 35: 155–173

Taylor TN, Zavada MS, Archangelsky S (1987) The ultrastructure ofCyclusphaera psilata from the Cretaceous of Argentina. Grana26: 74–80

Traverse A (1988) Paleopalynology. Unwin Hyman, Boston, 600ppVan Hoeken-Klinkenberg PMJ (1964) A palynological investigation

of some Upper Cretaceous sediments in Nigeria. Pollen et Spores6: 209–231

Volkheimer W, Caccavari de Filice MA, Sepulveda E (1977) Datospalinologicos de la Formacion Ortiz (Grupo La Amarga),Cretacico inferior de la cuenca neuquina (Republica Argentina).Ameghiniana 14: 59–74

Zavada MS (1987) The occurrence of Cyclusphaera sp. in southernAfrica. VII Simposio Argentino de Paleobotanica y Palinologia,Actas, pp 101–105

Zavada MS (1992) The wall ultrastructure of fossil discoid pollen.Bulletin of the Torrey Botanical Club 119: 44–49

Zavada MS (2003) The ultrastructure of angiosperm pollen from theLower Cenomanian of the Morondova Basin, Madagascar. Grana42: 20–32

Edited by JU Grobbelaar

South African Journal of Botany 2004, 70: 646–653 653