Palaeogeography, Palaeoclimatology, Palaeoecology l 19 (1995) 385-394
The palaeoclimate of the Karoo: evidence from plant fossils R.J. Rayner B.P.I. Palaeontology, University of the Witwatersrand, P.O. Box 179, Wits2050, South Africa
Received 13 December 1994; revised and accepted 17 March 1995
The common and conspicuous Karoo fossil floras flourished in what appears to be climatically disadvantageous conditions, as southern Gondwana lay within the Antarctic Circle throughout much of the Permian. Most previous climatic reconstructions suggest a cold-temperate climate. Plant structures are sensitive to environmental conditions, particularly climate, and may be used to predict these conditions. Large Glossopteris leaves dominate fossil plant assemblages, and considerable deposits of coal were laid down at this time. Quantitative analyses of fossil leaves from three coeval localities indicate that the climate prevailing at the time of the burial of these Glossopteris floras was seasonal, with summer conditions extremely favourable for plant growth, comparable to tropical regions today, and winters that were also correspondingly warmer than has previously been suggested.
Attempts to establish an association between vegetation and physical environmental conditions have occupied the attentions of generations of botanists. This has been the case with those with a particular interest in the fossil record. As a result, a formal correlation between types of vegetation and climate has become established, and extended back into the geological past (e.g. Rayner et al., 1993). Indeed, the most commonly accepted delimitation of "tropical" and "subtropical" forests is based on the temperature of 18°C for the coldmonth mean (Wolfe, 1979). Desert, Savanna, Temperate Forest and Tundra vegetation likewise are defined in terms of, and controlled by, their prevailing climates. As a corollary, further detailed observations on vegetation types have shown that, in similar climates, although the taxa of plants may be very 0031-0182/95/$15.00 © 1995 Elsevier Science B.V. All rights reserved SSDI 0031-0182(95)00021-6
different, there are morphological consistencies to be found. In a classic study, Bailey and Sinnott (1915, 1916) measured characteristics of certain living angiosperm leaves, and, using comparative morphology, attempted to provide an index of Cretaceous and Tertiary palaeoclimates. This technique was elaborated by Raunkiaer (1934) who provided a complete index of simple leaf morphology related to climates. Wolfe, in a series of studies (1971, 1977, 1978, 1985), predicted Tertiary climates from various parts of the world, based mostly on leaf morphology. It is the purpose of this paper to apply this technique in an attempt to quantify some climatic parameters in the Permian of Gondwana at the time of the deposition of parts of the South African Karoo Supergroup (sensu SACS, 1980). In this way, previous palaeoclimatic interpretations based on other evidence may be tested against the environmental message from fossil leaves.
R.J. Rayner/Palaeogeography, Palaeoclirnatology, Palaeoecology 119 (1995) 385. 394
2. Karoo palaeoenvironment The Karoo sedimentary sequence comprises a unique record of biological evolution. Considerable thicknesses of clastic sediments were laid down in a gently subsiding basin from the Carboniferous to the Triassic. Some sediments were undoubtedly marine, but most were terrestrial and they contain abundant tetrapod fossils, including mammal-like reptiles (Broom, 1932). Plant fossils and coal deposits are also common, and the Karoo records the evolution and diversification of the Glossopteris floras and their eventual replacement by the Dicroidium floras of the Triassic. An understanding of the environment when these animals and plants were being captured by the fossil record has been the goal of many researchers. Most have concentrated on the rock record, detailing sedimentary environments (e.g. Turner, 1980; Smith, 1987; Rubidge, 1988). Vital to an understanding of the environment, both sedimentary and biogenic, however, is an insight into the climates prevailing at the time. Previous views of the Permian climate of southern Gondwana have been influenced by the prolonged Upper Palaeozoic glaciations and workers have argued for cold temperate conditions. For example, from geological and palaeobotanical evidence, Plumstead (1957) likened the Karoo to Siberia or Canada today, with a strongly seasonal climate including very cold winters and temperate summers supporting Glossopteris vegetation which was restricted to sheltered basins. Yemane (1993) cited a variety of evidence which also indicated a temperate climate. The palaeo-position of the continents confirms the similarity with today's Siberia-Canada analogue. In the early Permian, the area occupied by the Karoo basin lay approximately 70 ° south of the equator, within the then Antarctic Circle; in the late Permian Gondwana had drifted north somewhat and the Karoo was about 65 ° south (Smith et al., 1973). Thus, the extreme southerly positions of Gondwana would have influenced the climate considerably, if only to create seasonal conditions, but the fact that the Karoo lay within the Antarctic Circle reinforces the notion of an overall cold climate. Some studies have produced
models, based on circulation and energy balance, which indicate that the Karoo experienced seasonal extremes in climate during middle to late Permian times (Kutzbach and Gallimore, 1989; Crowley et al., 1989). Daytime summer temperatures of central to north Gondwana were, according to Crowley et al., higher than 35°C, and may even have reached 45°C. The area of the Karoo lay between their 15° to 20°C January and - 2 0 ° to - 2 5 ° C July isotherms. There is a mixed message therefore. According to some simulations, the end of the early Palaeozoic period of glaciation was marked, in southern Gondwana, by continued cold, even extremely cold conditions in the winters and cool temperate conditions in the summer (Plumstead, 1957). Or, according to others, the summers may have been hot, but there was a 40-50°C annual range in temperatures and winters were correspondingly cold (Crowley et al., 1989). The effect on the vegetation was that a "cold tolerant" Glossopteris flora emerged and, with its attendant herbivorous fauna, persisted over much of Gondwana throughout the Permian.
3. Leaf morphology Plants can be reliable palaeoenvironmental indicators, for they are sessile and share a fundamental trophic homogeneity, in that they all basically do the same things (Knoll, 1984). The number of adaptations to climatic conditions is, therefore limited, and adaptive indicators may be preserved in the fossil record. The most obvious and common plant environmental indicator is leaf morphology. Leaves are the only means by which energy can be harnessed for metabolic activity for almost all vascular plants. In addition, the structural and functional constraints on such an organ are such that there are only a limited number of solutions and possible leaf designs. For example, in his text, The Tropical Rain Forest, Richards (1952, p. 83) observed: "It is remarkable how the rain-forest environment seems, as it were, to mould the foliage of all species coming under its influence to one particular form. This is well seen when the leaves of tropical
R.J. Rayner/Palaeogeography, Palaeoclimatology, Palaeoecology 119 (1995) 385 394
species are compared with those of near relatives in temperate climates." This evolutionary or ecological convergence was exploited first by Bailey and Sinnott (1915, 1916), but developed and refined by Wolfe and Hopkins (1967), to determine palaeoclimates. The latter authors advanced the hypothesis that foliar physiognomy is a highly sensitive index of the environment in which the assemblage lived, and it has been backed up by mathematical models, which predict that leaf diameter will most closely approximate physical conditions (Parkhurst and Loucks, 1972; Givnish and Vermeij, 1976). The characteristics of leaves most sensitive to climate are: ( 1) type of margin, (2) texture, (3) type of apex, (4) size. In areas of high mean annual temperature and precipitation, leaves typically have entire margins (i.e. lacking lobes or teeth); they are large, evergreen, with a leathery texture and a high proportion of them have attenuated apices ("drip tips"); and a moderate number have cordate (heartshaped) bases associated with palmate venation and joints ("pulvini") in the petiole (Richards, 1952; Wolfe, 1978). Richards quantified this entire-margin character and measured many leaves from two localities in the Rain forest ("Wet Evergreen Forest") of Nigeria, and found that 80% of the species had entire-margin leaves, and of those with non-entire margins only 4% were lobed or incised. He therefore characterised the average rain forest leaf: "Typically it is of a deep, sombre green and from oblong-lanceolate to elliptical in shape; the margin is entire or finely serrate, and there is often a long and distinct acumen, commonly forming a pronounced "drip tip". The texture is hard, often more or less leathery, the upper surface glabrous and often highly polished, any tomentum or hairiness being confined to the lower surface." (p. 80). In more temperate regions, with lower temperatures and rainfall, and greater seasonal ranges of temperature, leaves are commonly lobed and/or serrated; they are also thin and deciduous. The tropical attenuated tips are normally absent on
deciduous leaves (Upchurch and Wolfe, 1987). Leaves in subhumid climates have, in many instances, rounded to emarginate apices.
4. Material and methods
There are substantial collections of plant remains from three localities held in the BPI Palaeontology and associated institutions: Hammanskraal, Vereeniging and Lawley. The Hammanskraal plants were collected from refractory clay quarries in the vicinity of the Hammanskraal township, c. 50 km North of Pretoria; the Vereeniging fossils were also found in refractory clays in the Leeukuil Quarries approximately 80 km S of Johannesburg; Lawley I is a brick quarry 30 km SW of Johannesburg. Sediments from all three localities are similar in lithology and are associated with coal deposits. Clay deposition has been suggested to have taken place under pH controlled conditions: small proximal lakes with high pH induced the flocculation of kaolinite rich clays derived from nearby granites (Bredell, 1983). The plants are preserved in these kaolinite mudrocks, the latter two as impressions, the former as coalified compressions (Schopf, 1975). The floras have been described fully (Plumstead, 1952, 1956; Rayner, 1985, 1986; Rayner and Coventry, 1985; Anderson and Anderson, 1985; Smithes, 1978; Kov~ics-Endr6dy, 1991). Although there is some dispute of their ages, the three sets of sediments are probably coeval, and a part of the upper Ecca Group (sensu SACS, 1980). Smithes (1978) and Anderson and Anderson (1985) have suggested a Lower Permian age, but Rayner and Coventry (1985) and Kovfics-Endr6dy (1991) favour the Upper Permian (for a more complete discussion of this age detemination, see KovficsEndr6dy, 1991 ). The collections were examined and appropriate parameters from each glossopterid species and the type material measured. The particular dimensions measured included: the size of the glossopterid leaves (length × width) and calculated surface area (length × width × 0.66); the character of the leaf apex; the percentage of species making up the floras that have entire and non-entire margined leaves.
5. Architecture of
R.J. Rayner/Palaeogeography, Palaeoclimatology, Palaeoecology 119 (1995) 385 394
The common Gondwana tongue-shaped leaves, with their distinctive midrib and secondary reticulate venation was established as the form genus Glossopteris by Brongniart (1828). It has since been expanded to include more than 80 species (Banerjee and Ghosh, 1972); however, their relationship to each other and other plant groups is, in many cases, still uncertain. Nevertheless, they are so common that they have given their name to the entire Gondwanan Permo-Carboniferous flora. The characteristic Glossopteris outline (Fig. 1), which is preserved in two dimensions on the sediment surface, starts as a narrow petiole, from which the lamina expands regularly showing bilateral symmetry to a widest part between a half and two thirds up the blade at which point the margins narrow to a rounded or pointed apex. The secondary venation arcs away from the central midrib as it divides and anastomoses making its way to the leaf margin. The overall size and abundance of Glossopteris leaves, together with substantial deposits of fossil wood, suggest strongly that the leaves were borne on large trees and that the Karoo was forested. Indeed, the latest reconstruction shows Glossopteris to be a fairly substantial tree (Gould and Delevoryas, 1977).
6.1. Leaf margins Glossopteris leaves dominate floral assemblages, typically throughout Gondwana, but particularly at these three localities, and they all have entire margins. However, there are some other elements to the flora. A full species list from each locality is provided (Table 1), and is further divided into those with entire and non-entire margins. The ratios of those with entire margins to non-entire margins at the different localities is: Hammanskraal 28/4, Vereeniging 38/2 and Lawley 17/5 (87.5%, 95% and 77%, respectively). The nonentire species of plants include articulates and ferns, both groups almost certainly forming the understorey of the Glossopteris forests, and small
Fig. 1. Typical Glossopteris leaf showing outline, pointed apex, midrib, detail of secondary reticulate venation and folding. Scale x 0.9.
arborescent lycopods, presumably occurring in the wetter areas (Rayner, 1985, 1986).
6.2. Texture Although a difficult character to be precise about, the available evidence points to the Glossopteris leaves being fairly substantial. The surface area of the two-dimensional compressions (see below), which is in many instances consider-
R.J. Rayner/Palaeogeography, Palaeoclimatology, Palaeoecology 119 (1995) 385-394 Table 1 Species lists. H M K = Hammanskraal; VER = Vereeniging; LWY = Lawley; X = present; O = absent; References: 1 = Rayner (1985); 2 = R a y n e r (1986); 3 = R a y n e r and Coventry ( 1985 ); 4 = Anderson and Anderson ( 1985); 5 = Smithes ( 1978); 6 = Kovfics-EndrOdy ( 1991 ) H M K VER Entire leaves Azaniadendron fertile X Cyclodendron leslii X Unknown lycophyte O Sphenophyllum hammanskraalensis X S. ~peciosum 0 Annularia hammanskraalensis X Raniganjia rayneri O R~ lanceolata O Phyllotheca lawleyensis O Liknopetalon enigmata X Glossopteris obovata 0 G. clarkeana X G. leptoneura X G. ampla X G. stenoneura X G. browniana X G. deeipiens X G. communis X G. elliptica X G. flabellivenosa X G. divergens X G. indica X G. duocaudata X G. parallela 0 G. tortuosa 0 G. insueata X G. maccoyi X G. andreanskyi X G. angustifolia X G. stricta X G. clarkei X G. linearis X G. ferrugistratum X G. intermittens 0 G. taeniopteroides X G. claramarginata O Palaeovittaria kurzii 0 Ottokaria hammanskraalensis X O. transvaalensis 0 Scutum rubidgeum 0 S. draperium 0 Hirsutum dutoitides 0 Arberia leeukuilensis 0 A. allweyensis O Lidgettonia afrieana O Eritmonia natalensis O Scale leaf X Scale leaf A O
0 X O 0 0 O O O O 0 X X X X 0 O 0 X X 0 0 O X X X 0 X X X 0 0 X X X X O X 0 X X X X X O O O O X
0 O X 0 X O X X X 0 0 0 0 X 0 X 0 X 0 0 0 X 0 0 0 0 O 0 0 0 0 0 O 0 X X 0 0 0 0 0 0 0 X X X X O
1 2,5 3 4,5 3 4,5 3,4 3,4 3,4 5 6 6 6 3,4,6 6 3,6 6 3,6 6 6 6 3,6 6 6 6 6 4,6 6 6 6 6 6 4,6 6 3,4,6 3,4 4 4 4 4 4 4 4 3,4 3,4 3,4 3,5 4
Table 1 (continued) H M K VER
SL B SL C SL D Noeggerathiopsis hislopii N. elongata Sphenobaiera eccaensis Metreophyllum lerouxii Ginkgophyllum kidstonii G. spatulifolia Ginkgophyllum sp. A FlabelloJolium sp. A Flabellofolium sp. B Walkomiella transvaalensis Cyparissidium sp. A Taeniopteris gemmina Totals
O O O O 0 0 0 0 0 O O O 0 O 0 28
X X X X X X X X X X X O X X X 38
O O O X 0 0 0 0 0 O O X 0 O 0 17
4 4 4 3,4 4 4 4 4 4 4 4 4 4 4 4
Non-entire leaves Sphenophyllurn mesoeccaense Botrychiopsis valida Asterotheca hammanskraalensis A. leeuluilensis Ginkgoites sp. Sphenopteris alata Neomariopteris polymorpha N. hughesii N. lobifolia Neomariopteris sp. Maithya lobifolia Totals
X 0 X 0 O 0 X 0 0 O X 4
O X O X O 0 0 0 0 O 0 2
O 0 O 0 X X 0 X X X 0 5
4,5 4 4,5 4 3 3 5 3 3 3 5
able, and the fact that these leaves are vascularised throughout, with a considerable length and thickness of vascular tissue in the primary and secondary venation, indicates abundant biomass. The common occurrence of folding on the leaves also supports this hypothesis (Fig. 1). In addition, the size of many Glossopteris leaves necessitates something like a leathery texture to withstand physical damage from wind and the weight of rainwater.
6.3. Type of apex A significant number of leaves in the Glossopteris floras have pointed apices (e.g. Fig. 1). These apices are not attenuated into true drip tips, and a number of Glossopteris leaves are rounded. However, the character of many of the apices, taken together with the attitude and large surface
R.J. Rayner/Palaeogeography, Palaeoclimatology, Palaeoecology 119 (1995) 385 394
area of the leaves suggests a mechanism to remove standing water (see below).
6.4. Size Metabolic activities are generally limited by surface areas rather than body volumes. Leaf size, in terms of surface area, is therefore a highly significant environmental indicator. The sizes of the types are presented, along with their surface areas, in Table 2. I have further divided them, on the basis of surface area, into Raunkiaer's (1934) size-classes (Fig. 2).
There are several assumptions made concerning the nature of the fossil collections: (1) that they represent a mature stable forest community, (2) that the transporting agent or fossilization process has not introduced a bias, and (3) that the collector has not introduced a similar bias. The fact that similar leaf assemblages have been collected from many localities in Gondwana, seems to suggest strongly that this type of assemblage is a true reflection of the vegetation growing close to the basins of sedimentation, and that these are mature stable communities. Further, if a size bias has been introduced by transport or even fossilization processes, this would tend to favour small leaves over the larger ones. However, a "natural" collecting bias has been noticed. Most collectors unwittingly favour complete specimens over incomplete ones (complete large leaves are almost impossible to collect, but complete small leaves may be common), and, such is the preponderance of Glossopteris leaves, collectors tend to favour the non-Glossopteris rarities (Rayner and Coventry, 1985). The palaeoclimatic signal may, therefore, be an underestimate of the true temperature and rainfall. A whole suite of characters affect leaf size and shape--heredity obviously has a strong influence, but so also does the physical environment during ontogeny. In such a case as the Gondwanan
Table 2 A p p r o x i m a t e sizes a n d s u r f a c e a r e a s o f t y p e s o r a r e p r e s e n t a t i v e e x a m p l e o f the G l o s s o p t e r i d a l e s (sensu A n d e r s o n a n d A n d e r s o n , 1985) Species
Area ( m m 2)
Glossopteris obovata G. clarkeana G. leptoneura G. amplu G. stebineura G. browniana G. decipiens G. communis G. elliptica G. flabellivenosa G. divergens G. indica G. duocaudata G. parallela G. tortuosa G. insueta G. rnaccoyi G. andreanszkyi G. angustifolia G. stricta G. clarkei G. linearis G. ferrugistratum G. intermittens G. taeniopteroides G. claramarginata Palaeovittaria kurzii Ottokaria transvaalensis O. hammanskraalensis Scutum rubidgeum S. draperium A rberia leeukuilensis A. allweyensis Lidgettonia africana
166 x 39 70 x 42 155 x 15 500 x 160 107 x 24 155 x 40 80 x 20 275 x 65 130 x 69 120 x 22 200 x 100 300 x 70 110 x 30 250 x 50 225 x 25 90 x 25 250 x 25 160 x 55 90 x 16 225 x 28 110 x 25 76 x 20 250 x 100 100 x 27 270 x 55 200 x 25 205 x 37 313 x 144 162 x 19 247 x 33 235 x 46 172 x 40 250 x 76 170 x 21 n = 34
4300 2000 1530 52,800 1700 4100 1060 11,800 5920 1740 13,200 13,860 2180 8250 3712 1485 4125 5810 950 4158 1820 1000 16,500 1782 9800 3300 5010 29,750 2030 5380 7140 4540 12,540 2350 .? = 7350
P e r m i a n - - a large continental land mass with no apparent catastrophic climatic changes for a considerable time--the evolution of such functional structures as leaves would be expected to exhibit trends, and even show strong evolutionary convergence. Glossopteris, therefore, may not be a biologically coherent evolutionary taxon but still may show morphological consistency. Schopf (1976) thought that Glossopteris was a natural group, but the assemblage of reproductive struc-
R.J. Rayner/Palaeogeography, Palaeoclimatology, Palaeoecology 119 (1995) 385-394
1oo 9O 8O
~_ 511 0
.o 10 r-
Lea£ Area mm2 (Raunkiaer's Jg34 c]assi£icotion)
Fig. 2. The specimensfrom the three localities placed in Raunkiaer's (1934) size-classes.
tures associated with the leaves argues differently (Rayner and Coventry, 1985). Its component species may have inherited a leaf plan, which was strongly adapted to the environment, from their common ancestor, or they may have converged on this leaf form due to prolonged environmental factors. Either way there is a strong environmental signal preserved in the leaf architecture. Because of the elastic/leathery nature of vascular tissue, the curving of the secondary venation would tend to give the leaf a double arched transverse cross section in the living position (this has been lost by compression, but some of the creasing on well-preserved leaves indicates that this arching occurred). In addition, the length of the leaves would have produced a tendency to arch longitudinally. This is structurally a favourable orientation and would enable the leaf to support its own (considerable) mass and an additional mass of rainwater. The resistance to wind damage would also be enhanced by such an orientation. This three dimensional leaf would, consequently, be ideal to prevent water build up and the inherent dangers of structural collapse. It is suggested that pointed apices would have functioned rather as a drainage mechanism, equivalent in deciduous leaves to tropical drip-tips. If so, this indicates
reasonable or even high rainfall, at least in the growing season. Wolfe (1971) included a table (his Table 1, p. 34) of the percentage of species with entire margined leaves against mean annual temperature. [The leaf margin data came from earlier studies (Bailey and Sinnott, 1916; Brown, 1919)]. The graph (Fig. 3) quantifies the strong relationship, noticed earlier by Richards (1952) and others, between the percentage of species with entire margined leaves and mean annual temperature. The mean annual ranges are also plotted, assuming they are symmetrical about the mean annual temperature. These annual ranges are minimal in tropical areas and increase away from the tropics towards the strongly seasonal, deciduous, temperate forests. Placing the three fossil localities on the regression line predicts that they enjoyed annual temperature means of 25, 28 and 30°C, considerably warmer than the cool temperate climate suggested by Plumstead (1957) and others, and anomalous when compared to other deciduous floras. The Karoo floras, however, were undoubtedly deciduous (this is known from growth rings on fossil wood and the character of leaf accumulations which indicate autumn leaf fall), and Permian temperature ranges must have been con-
R.J. Rayner/Palaeogeography, Palaeoclimatology, Palaeoecology 119 (1995) 385 394
Nean annual t.mpcrat.u'-, (°C)
] 40 ] 60 [80 Perc~nt~ of sfmcies vith entirq ior~jin leaves
Fig. 3. Graph showing mean annual temperatures and mean annual ranges for the localities in Wolfe ( 1971 ), including the three fossil localities.
siderably greater, and summer temperatures higher than the tropical localities. A degree of caution must be applied, however, to climatic comparisons between a group of deciduous gymnosperms, extinct virtually since the Palaeozoic, and modern woody dicots. Glossopteris may have inherited a leaf plan which could not be elaborated by variation and selection into toothed or lobed forms. This does seem unlikely when considering the reticulate venation pattern, similar to angiosperms, the fact that the same physical conditions were presumably acting on the leaves, and, most of all, put against the range in variation of reproductive structures shown by Glossopteris remains. However, all Glossopteris leaves had entire margins, and there are no signs of toothed or lobed margins (other than the very occasional lobed apex or lamina base). It has not been demonstrated that Glossopteris exhibited a morphological cline from toothed or serrated margins to entire margins with climate. Indeed, it may be impossible to demonstrate this or any such relationship between climate and morphology without falling into a circular argument, since the modern analogue, available in the case of living angio-
sperms, is lacking for these plants. What is certain, however, is that trees today only grow such large leaves in favourable (and almost always nonseasonal/tropical) conditions, and those leaves tend to be entire margined. Size may therefore be a more reliable indicator of climate. Glossopterid leaves may be categorised by size (Fig. 1), an approach first adopted by Raunkiaer (1934). Species' leaf surface areas from the three individual localities and a combination of all three are presented. All show the same trend, with a preponderance of the mesophyll class (2025 18,225mm 2) up to 80% in the case of Lawley, and a significant percentage of macrophylls (18,225-164,025mm2), up to 20% at Lawley. This pattern and overall percentages are comparable to the leaves of tropical rain forests sampled by Raunkiaer (1934) and Richards (1952). The large surface area, the texture and shape of the Glossopteris leaves therefore all indicate highly favourable growing conditions in the K a r o o summer. Put simply, trees do not produce that amount of foliage (and some leaves are up to 1 m in length) just to discard it at the end of the season
R.J. Rayner/Palaeogeography. Palaeoclimatology, Palaeoecology 119 (1995) 385 394
unless physical factors are favourable. Further, the position of the three localities on the graph (Fig. 3) suggest that the temperatures were high. The summer climate was, therefore, undoubtedly humid, with high rainfall, the equivalent of a tropical forest today. Winter and summer day length estimates for a position between 65 and 70 ° S of the equator may be calculated. For more than four months in the summer the forests would have experienced more than 16 hours daylight, rising up to 24 hours for several days either side of the summer solstice. The 180 days of the probable growing season would have received an average of 18 hours of daylight-- ideal growing conditions when accompanied by high rainfall and temperatures. The winters were correspondingly lacking in daylight, and the deciduous habit may have been induced merely by shorter day length rather than attendant severe cold temperatures. There is an absence of cold climatically sensitive sediments and permafrost structures which would otherwise indicate cold winters, and Yemane (1993) further suggested that the conspicuous growth rings and periodic leaf falls indicate frost-free growth patterns. Growing conditions in the summers therefore were extremely favourable, with average daily temperatures 30°C or more, and rainfall in the sixmonth season similar to the annual amount in tropical areas today, in excess of 350 cm. It seems likely that these were also suitable climatic conditions to accumulate the massive coal reserves of southern Gondwana, which were laid down at this time. In addition, the fuller understanding of the preferred habitat of this enigmatic group of plants that the leaf morphology provides, invites climatic change as a reason for its sudden disappearance.
Acknowledgements I thank the South African Foundation for Research Development and the University of the Witwatersrand for financial support. Judith Masters, Bernie M o o n and Grigor Aitken read and commented on an early version of the manuscript. J.A. Wolfe and L.A. Frakes provided essential critical advice.
References Anderson, J.M. and Anderson, H.M., 1985. Palaeoflora of Southern Africa: Prodromus of South African Megafloras Devonian to Lower Cretaceous. Balkema, Rotterdam. Bailey, I.W. and Sinnott, E.W., 1915. A botanical index of Cretaceous and Tertiary climates. Science, 41: 831-834. Bailey, I.W. and Sinnott, E.W., 1916. The climatic distribution of certain types of angiosperm leaves. Am. J. Bot., 3: 24-39. Banerjee, M. and Ghosh, A.K., 1972. Further studies on the genus Glossopteris and its stratigraphic significance. Proc. 2nd Gondwana Symp., South Africa, 1970, pp. 561-570. Bredell, J.H., 1983. Refractory clays in the Transvaal: Exploration, genesis and property interrelationships. Thesis. Univ. Pretoria (unpublished). Brongniart, A., 1828. Prodr6me d'une histoire des v6g6taux. Dict. Sci. Nat., 57:16 212. Broom, R., 1932. The Mammal-like Reptiles of South Africa. Witherby, London. Crowley, T.J., Hyde, W.T. and Short, D.A., 1989. Seasonal cycle variations on the supercontinent of Pangaea. Geology, 17: 457460. Givnish, T.J. and Vermeij, G.J., 1976. Sizes and shapes of liane leaves. Am. Nat., 110: 743-778. Gould, R.E. and Delevoryas, T., 1977. The biology of Glossopteris: evidence from petrified seed-bearing and poUenbearing organs. Alcheringa, 1: 387-399. Knoll, A.H., 1984. Patterns of extinction in the fossil record of vascular plants. In: M.H. Nitecki (Editor), Extinctions. Univ. Chicago Press, pp. 21-68. Kovfics-EndrOdy, E., 1991. On the Late Permian age of Ecca Glossopteris floras in the Transvaal Province with a key to and description of twenty five Glossopteris species. Mem. Geol. Surv. S. Afr., 77:1 111. Kutzbach, J.E. and Gallimore, R.G., 1989. Pangean climates: megamonsoons of the megacontinent. J. Geophys. Res., 94: 3341 3357. Parkhurst, D.F. and Loucks, O.L., 1972. Optimal leaf size in relation to environment. J. Ecol., 60:505 537. Plumstead, E.P., 1952. Description of two new genera and six new species of fructifications on Glossopteris leaves. Trans. Geol. Soc. S. Afr., 55: 281-328. Plumstead, E.P., 1956. Bisexual fructifications borne on Glossopteris leaves from South Africa. Palaeontographica, B100: 1-25. Plumstead, E.P., 1957. Coal in Southern Africa. Univ. Witwatersrand Press, Johannesburg. Raunkiaer, C., 1934. The Life-Forms of Plants and Statistical Plant Geography. Clarendon Press, Oxford. Rayner, R.J., 1985. The Permian lycopod Cyclodendron leslii from South Africa. Palaeontology, 28:111-120. Rayner, R.J., 1986. Azaniadendron, a new genus of lycopod from South Africa. Rev. Palaeobot. Palynol., 47: 129-143. Rayner, RJ. and Coventry, M.K., 1985. A Glossopteris flora from the Permian of South Africa. S. Afr. J. Sci., 81: 21-32. Rayner, R.J., Moon, B.P. and Masters, J.C., 1993. The
R.J. Ravner/Palaeogeography. Palaeoclimatology. Palaeoecology 119 (1995) 385 394
Makapansgat australopithecine environment. J. Hum. Evol., 24: 219-231. Richards, P.W., 1952. The Tropical Rain Forest. Cambridge Univ. Press. Rubidge, B.S., 1988. A palaeontological and palaeoenviromnental synthesis of the Permian Ecca-Beaufort contact in the Southern Karoo between Prince Albert and Rietbron, Cape Province, South Africa. Thesis. Univ. Port Elizabeth (unpublished). SACS (South African Committee for Stratigraphy), 1980. Stratigraphy of South Africa. 1. Lithostratigraphy of the Republic of South Africa, South West Africa/Namibia and the Republics of Bophuthatswana, Transkei and Venda. Handb. Geol. Surv. South Afr., 8 : 1 - 6 9 0 (L.E. Kent, compiler). Schopf, J.M., 1975. Modes of fossil preservation. Rev. Palaeobot. Palynol., 20:27 53. Schopf, J.M., 1976. Morphologic interpretation of fertile structures in glossopterid gymnosperms. Rev. Palaeobot. Palynol., 21:25 64. Smith, A.G., Briden, J.C. and Drewry, G.E., 1973. Phanerozoic World Maps. In: N.F. Hughes (Editor), Organisms and Continents Through Time. Spec. Pap. Palaeontol., 12: 1-42. Smith, R.M.H., 1987. Morphology and depositional history of exhumed Permian point bars in the south-western Karoo, South Africa. J. Sediment. Petrol., 57:19 29. Smithes, S.J., 1978. Studies in a Middle Ecca (Lower Permian) flora from Hammanskraal, Transvaal, South Africa, with
emphasis on the glossopterid fructification Ottokaria Zeiller. Thesis. Univ. Witwatersrand, Johannesburg (unpublished). Turner, B.R., 1980. Palaeohydraulics of an Upper Triassic braided river system in the main Karoo Basin, South Africa. Trans. Geol. Soc. S. Afr., 83: 425-431. Upchurch, G.R. and Wolfe, J.A., 1987. Mid-Cretaceous to Early Tertiary vegetation and climate: evidence from fossil leaves and woods. In: E.M. Friis, W.G. Chaloner and P.R. Crane (Editors), The Origin of Angiosperms and Their Biological Consequences. Cambridge Uniw Press, pp. 75 106.. Wolfe, J.A., 1971. Tertiary climatic fluctuations and methods of analysis of Tertiary floras. Palaeogeogr. Palaeoclimatol. Palaeoecol., 9:27 57. Wolfe, J.A., 1977. Paleogene floras from the Gulf of Alaska region. U.S. Geol. Surv. Prof. Pap., 997:1 108. Wolfe, J.A., 1978. A paleobotanical interpretation of Tertiary climates in the Northern Hemisphere. Am. Sci., 66: 694-703. Wolfe, J.A., 1979. Temperature parameters of humid to mesic forests of Eastern Asia and relation to forests of other regions of the Northern Hemisphere and Australasia. U.S. Geol. Surv. Prof. Pap., 1106:1 37. Wolfe, J.A., t985. Distribution of major vegetational types during the Tertiary. In: E.T. Sundquist and W.E.S. Broecker (Editors), The Carbon Cycle and Atmospheric CO2: Natural Variations Archean to Present (Geophys. Monogr., 32). Am. Geophys. Union, Washington, DC, pp. 357 376. Yemane, K., 1993. Contribution of Late Permian palaeogeography in maintaining a temperate climate in Gondwana. Nature, 3651:51 54.