Three New Eocene Cycads from Eastern Australia

Aust. J . Bot., 1980, 28, 105-22

Three New Eocene Cycads from Eastern Australia

Robert S . Hill Botany Department, University of Adelaide, Adelaide, S.A. 5001; present address: Botany Department, School of Biological Sciences, James Cook University of North Queensland, Qld 48 11.

Abstract

The descriptions of Pterostoma R.S.Hil1, a new cycad genus with two species ( P . zamiodes R.S.Hil1 and P. anastomosans R.S.Hil1) and Lepidozamia foveolata R.S.Hill add significantly to our knowledge of the floristic components of the Australian Tertiary. Along with other reports of Tertiary cycads, the discovery of these new species supports the theory that cycads had a greater geographical range and diversity in Australia during the Tertiary than at present. The evolutionary position of Pterostoma in relation to other fossil and extant cycads is considered and it is concluded that Pterostoma represents an extinct line. The possibility that fossil cycads could be of value in biostratigraphy and in estimating palaeoclimate is discussed.

Introduction

The early history of Australian Tertiary cycad megafossils was reviewed by Cookson (1953). She concluded that no fossils previously described as cycads were sufficiently substantiated since they consisted of frond impressions with no cuticular remains. Two reports which Cookson did not mention should be considered. Deane (1902) described the single specimen of Eucalyptuspraecoriacea as consisting of 'branchlets' with 'leaves' attached. Deane placed the specimen in Eucalyptus because he thought it was very similar to the extant species E. coriacea A.Cunn. ex Schau. (now E. pauc$ora Sieber ex Spreng.). A recent examination of the architecture of this fossil showed that it bears very little resemblance to E. pauci$ora and is probably part of a frond of a zamioid cycad. A cuticle preparation was attempted, but the protective coating applied to the specimen by Deane made it impossible. Before a positive identification of this species can be made, material with retrievable cuticle must be collected. The type locality at Balcombe Bay near Mornington Beach south of Melbourne was re-collected early in 1979 but no specimens of E. praecoriacea were found (D. C. Christophel. personal communication). Chapman (1925) concluded that seeds from the Morwell brown coal mine in Victoria were similar to those of Lepidozamia peroffskyana Regel, although less than half the size. However, this conclusion was based only on the external morphology of the seeds, and their identity must be regarded with caution until a formal description is made. Including the review by Cookson (1953), three fossil cycad species have been described from eastern Australia. They are Lepidozamia hopeites (Cookson) L. Johnson (described as Macrozamia hopeites by Cookson 1953) from the Oligocene of Bacchus Marsh, Vic. Bowenia eocenica R.S.HiI1 from the Eocene of Anglesea, Vic. and B. papillosa R S H i l l from the Eocene of Nerriga, N.S.W. (Fig. 1). All three species are

known only from fragments of pinnae but in every case the exceptional cuticular detail establishes their generic position. Those Australian Tertiary megafossils which can clearly be classed as cycads are all members of extant genera. Therefore the description of a new cycad genus with two species from the Eocene of eastern Australia with no direct extant relatives is of importance for our understanding of the evolution of Australian cycads.

Fig. 1. Map of eastern Australia showing the fossil localities and the present distribution of Lepidozamia (shaded areas) and Bowenia (open areas) (after Johnson 1959). ho, Lepidozamia hopel; pe, L. peroffskyana; sp, Bowenia spectabilis; se, B . serrulata. Note the large overlap between ho and sp. A small area of Bowenia serrulata occurs at Tinaroo Dam (Johnson 1963) in this overlap and is marked in black.

Materials and Methods The matrix at both the Nerriga and Anglesea localities is grey mudstone. The Nerriga fossils, Pterostoma anastomosans and Lepidozamia foveolata, were separated from the mudstone by maceration in aqueous hydrogen peroxide. The specimens of Ptevostoma zamioides generally consist of black compressions that cannot be removed from the matrix without fragmenting. Megafossils were photographed with a Leitz Aristophot 9 cm by 12 cm plate camera with reflected light for P . zamioides (Figs 14, 15, 17) and transmitted light for P. anastomosans (Fig. 16). Cuticles were prepared by placing compression fragments in 10% chromic acid until all the internal organic material had dissolved (about 15 min). The cuticles were then rinsed in water, neutralized in 3% w/w aqueous ammonia and stained with 1% aqueous

Three New Australian Cycads

107

safranin 0. Cuticles were mounted in phenol-glycerine jelly for transmitted light microscopy and photographed with Nomarski optics. Unstained cuticles were mounted on aluminium stubs with double-sided adhesive tap: and coated in a high vacuum evaporative coating unit to a thickness of c. 200 A with an 80:20 gold palladium alloy. They were studied with an ETEC Autoscan scanning electron microscope operated at 20 kV. Wax embedding of pinna and rachis fragments of P. zamioides showed no structural preservation. Twenty-five random measurements were taken for all cuticular characters except that five random areas totalling 0 . 8 mm2 were used for stomatal index. For stomatal width the two guard cells were combined and measured as one stomate. Measurements were tzken directly f x m the i!!ustrations far Lepid~zmziah~peites(Cookson 19531, which restricted the number of measurements for most characters. The following material was made available for this study: four specimens of Ctenis covonata Douglas, ~ s 60914 v (holotype), ~ s 60913, v GSV 61369 and GSV 61353 (Victorian Department of Minerals and Energy); the holotype of Eucalyptus praecoviacea Deane (National Museum of Victoria); pinnae of Lepidozamia hopei Regel, CBG 043373 and L . pevoffskyana, CBG 01749 (Herbarium of the Canberra Botanic Garden); a pinna of Stangevia eriopus (Kunze) Nash (National Herbarium of New South Wales). An attempt to locate the holotype of Lepidozamia hopeites (Cookson) L. Johnson in the National Museum of Victoria failed.

Localities

The fossils described in this paper are from two localities. Ptevostoma zamioides was found in the ALCOA Anglesea open-cut brown coal mine (38 " 25 ' S., 144' 11'E.). Christophel(l979) described three fossiliferous lenses in the overburden and reported their age as upper Eocene, based on stratigraphifposition and palynological data. Two of the lenses are separated vertically by about 10 m of coarse fluviatile sand, the lower lens being c. 25 m above the coal. The third, larger lens lies 100 m away, also 25 m above the coal. The stratigraphic position of this lens in relation to the others is unclear, because of the homogeneous nature of the sands which lie between them. Ptevostoma zamioides was found to varying extents within all three lenses. The uppermost of the two smaller lenses has yielded only a few very fragmentary specimens with poorly preserved cuticle. This lens is rich in angiosperm leaves, representing several taxa. It also contains specimens with affinities to Podocavpus and Lygodium (Douglas 1977*) and a flower similar to those of Leptospermurn (Douglas 1978). The lower lens is dominated by two taxa, Ptevostoma zamioides and Casuarina. (D. C. Christophel, personal communication, has shown that the Casuarina megafossils from Anglesea and Nerriga belong to division Gymnostomae which is soon to be elevated to generic standing by L. A. S. Johnson (personal communication).) It was from this lens that the best material of P,zamioides, including the holotype (Fig. 14), was collected. A few angiosperm taxa have been found in this lens along with the cycad Bowenia eocenica (Hill 1978). The third, larger lens contains many specimens of P. zamioides but it is not the dominant taxon. There are also many angiosperm taxa within this lens. Recent collections have shown that P. zamioides also occurs in the upper layers of the coal seam (D. C. Christophel, personal communication).

,

*Douglas, J. G . (1977). A new fossil plant assemblage from the Eastern View Coal Measures. Geol. Surv. Vict. Rep. 1977131 (unpubl.).

R. S. Hill


109

Pterostoma anastornosans and Lepidozarnia foveolata were found at Nerriga, N.S.W. (35'07'S.,l 5OCO5'E.). This locality, described by Owen (1975) and Hill (1978), has been dated as middle Eocene (Owen 1975; J. I. Raine, personal communication). The deposit is dominated by angiosperm leaves, with about 50 taxa collected to date. The cycad Bowenia papillosa (Hill 1978) was collected at Nerriga.

Taxonomic Description Order Cycadales Family Zamiaceae L'P,rl3QZ4MIIA Rege!.

Lepidozamia foveolata R.S.Hil1 sp. nov.

Figs 2, 3, 4 Diagnosis Pinnules hypostomatic. Stomata restricted to areas between veins. Stomatal orientation usually parallel to veins. Stomates with four to seven subsidiary cells. Lower epidermal cells with long axis oblique to veins, cell walls of similar thickness throughout, pitting of walls not developed. Upper epidermal cells in regular longitudinal rows parallel to veins and often more than one cell wide, wall pitting highly developed. Cell wall thickness at row boundaries greater than inside rows. Upper epidermal cells unspecialized over veins. Description of the Specimen Architectuve. The holotype (N0.596) is the only specimen in which the lower cuticle is preserved. It is only a fragment of a pinna with no margin preserved. Evidence from the lower epidermal cell arrangement indicates six veins, one of which branches dichotomously. Since no margins are preserved, this represents the minimum vein number. Lower epidermis. Stomates are restricted to the lower epidermis and occur in rows between the veins (Fig. 2). The long axes of the stomates are more or less parallel to the veins, although there may be up to 15" between the two. Stomatal length averages 47.5 ym (42.5-52.5 ym) and width 36.5 ym ( 3 2 . 5 4 0 . 0 y m ) The subsidiary cell arrangement is haplocheilic with four to seven subsidiary cells per stomate. Epidermal Fig. 2. Lower epidermis of Lepidozamiafoveolata (N0596) showing the edge of a stomatal band. Fig. 3. Upper epidermis of L . foveolata (N0596) without Nomarski optics. Note the heavily pitted cell walls. Fig. 4. Upper epidermis of L . foveolata (N0596). Note the pitted cell walls and the alignment of cells into rows. (Veins run from left to right.) Fig. 5. Upper epidermis of L . foveolata (N0594), probably from the extremity of a pinna. Fig. 6. Lower epidermis of L. hopei (CBG043373) showing the stomatal band. Fig. 7. Upper epidermis of L . hopei (CBG043373). Note the pitted cell walls and oblique orientation of the cells. Fig. 8. Lower epidermis of L . pevoffskyana (CBG01749) showing stomatal and interstomatal bands. 9 ) . the oblique orientation of cells and absence of Fig. 9. Upper epidermis of L.peroffskyana ( ~ ~ ~ 0 1 7 4Note wall pitting. Scale applies to all figures.

R. S . Hill

cell length averages 76.8 pm (35.0-167.5 pm) and width 25.0 pm (17.5-30 pm). Their long axes are usually oblique to vein direction. All lower epidermal cells are similar in wall thickness and contrast with the slightly thicker-walled subsidiary cells (Fig. 2). There is no evidence of pitting in the cell walls. Upper epidermis. The upper epidermal cells occur in regular rows (Figs 3, 4, 10) which can be 30 or more cells long and two or more cells wide. Whereas the rows have their sides parallel to the vein direction, the epidermal cells have their long axes oblique to the vein direction (Fig. 10). Toward the ends the rows become narrow and taper to a point. Cell walls between rows of cells are thicker than the cell walls within the rows. Epidermal cells are not specialized above the veins. Cell walls are heavily pitted, which makes observation of wall outlines difficult (Fig. 3). Holotype. Specimen number N0596, housed in the South Australian State Herbarium (AD). Type locality. Nerriga, N.S.W. (35"07'S.,15Oo05'E.). Collector. R. S. Hill. Date. 22/5/1978. Comparison of Fossil and Extant Species Cookson (1953) and Johnson (1959) realized the importance of cuticular characters in Lepidozamia. Apart from several macromorphological features which separate Lepidozamia from Macrozamia Miq., one cuticular character is diagnostic. The long axes of the epidermal cells are oblique or transverse to the long axis of the pinna in Lepidozamia, in contrast to the longitudinal orientation in Macrozamia (Figs 6-9). Lepidozamia foveolata has oblique orientation of epidermal cells developed to an extreme degree, particularly on the upper epidermis (Figs 3, 4). All other cuticular characters correspond with those found in Lepidozamia, and the fossil was therefore placed in that genus. Table 1. Measurements of cuticular characters for the Lepidozamia species Upper ep~dermal cell length (iirn) L. hopei L, pero/jsk.pana

L. hopeites L./oreohta

Upper ep~derrnal ceil width (iim)

Lower eplderrnal cell length (w)

Lower eplderrnal cell width (iim)

Stornatal length (pm)

Stornatal wldth iiim)

No. of subsidiary

cells

Stornatal mdex

111.2 (78.4-148.4) 84.8 (56.0-137.2) 92.0 (45.0-135.0) 63.6 (37.5-1 17.5)

Cuticles of the extant species of Lepidozamia were prepared for comparison (Figs 6-9). Table 1 gives the values for several cuticular characters. In every size character L. foveolata is considerably smaller than the other species but has the highest stomata1 index. Cookson (1953) reported that the thin-walled upper epidermal cells of L , hopei and L . hopeites are pitted whereas those of L.peroffskyana are not. The current study has shown that in L, peroffskyana limited pitting does occur, whereas in L . foveolata this pitting of upper epidermal cells is highly developed (Fig. 3). One specimen of L . foveolata (N0594) has much shorter upper epidermal cells (Figs 5, 11). Pant and Nautiyal(1963) have shown how cell dimensions in cycads can vary

111


considerably along pinnae and it is possible that NO594 is from the extremity of a pinna, whereas the holotype (N0596) was situated nearer the centre. Values in Table 1 for L. pevoffskyana are much closer to those for L. hopei and L. hopeites than those measured by Cookson (1953), and it appears that this species is not as distinct in its epidermal characters as Cookson believed. The cuticle of L.foveolata is very similar to

Fig. 10. Upper epidermis of Lepidozamia foveolata (N0596) showing cells arranged in rows. Note the thicker walls between rows. (Veins run from left to right.) Fig. 11. Upper epidermis of L. foveolata (N0594). Note the shorter cells and less well developed rows.

other Lepidozamia species, but the distinguishing features of smaller cell size, higher stomata1 index, extreme oblique orientation and heavy pitting of the walls of upper epidermal cells justify the new specific status. Order Cycadales Family Zamiaceae Ptevostoma R.S.Hil1 gen, nov. Generic Diagnosis Fronds large, simply pinnate. Rachis swollen basally into an abscission layer. Numerous elongate, narrow lanceolate to linear pinnae attached to upper surface of rachis. Pinna margins entire, tapering to an acute apex. Pinna bases contracted near

R. S. Hill

frond base but expanding towards apex. Numerous dichotomosing or simple veins, sometimes anastornosing. Pinnae hypostomatic. Stomata scattered with random orientation in broad zones between veins. Guard cells sunken in a cutinized pit formed by a ring of halocheilic subsidiary cells, but thinly cutinized at poles. Cuticular ledge present between guard cells and subisidary cells. Epidermal cells conspicuously sinuous and with thick cuticular ridges on stornatal surface. Stomata present on both surfaces of rachis. Type species. Pterostoma zamioides from the upper Eocene ALCOA Anglesea open-cut brown coal mine (38 " Z ' S . , 144" 11 'E.). Etymology. From Pteros (wing) and stoma (mouth). Named to illustrate the cuticular 'wings' between guard ceiis, and subsidiary cells. Pterostoma is neuter and specific epithets should match accordingly. Ptevostoma zamioides R.S.HiI1 sp. nov. Fig. 14

Diagnosis Fronds over 40 cm long with at least 26 pinnae. Rachis width about 0 . 5 cm, expanded at base to 2 . 5 cm. Basal pinnae 20 cm long, 2 cm wide; apical pinnae 15 cm long, 1 . 5 cm wide. Pinna bases contracted at base of frond, expanding towards apex until equal to maximum width of pinna. Pinnae evenly spaced along rachis. About 15 veins per cm across pinna with a maximum number of 26 veins per pinna. Veins terminating at margin. Upper epidermal cell walls sinuous with decreasing amplitude over veins and toward pinna extremities. Veins marked by narrower cells. Epidermal cells in non-venous areas averaging 68 ym (51-83 ym) long, 47 ,urn (33-58 ym) wide; venous epidermal cells averaging 69 ym (53-80 pm) long, 24 pm (16-30 pm) wide. Two- to three-celled hair bases over veins occur rarely near pinna bases. No outer surface sculpturing. Lower epidermis with randomly orientated stomata between veins. Epidermal cells sinuous, with decreasing amplitude over veins and toward pinna extremities. Veins marked by longer, narrower cells. Epidermal cells in non-venous areas averaging 73 pm (50-91 pm) long, 46 ym (35-57 ym) wide; venous epidermal cells averaging 90 ym (74-106 ym) long, 24 ym (18-29 ym) wide. No hair bases occur on the lower epidermis. Outer surface with heavy cuticular ridges. Stomates average 43 ym (39-47 pm) long, 21 ym (19-24 ym) wide. Four to seven haplocheilic subsidiary cells. Rachis with straight-walled epidermal cells and longitudinally aligned stomata. Description of the Specimen Architecture. The holotype (P 150055) consists of the distal part of a large frond (Fig. 14). A study of the size of individual pinnae on less complete fronds suggests that a hypothesized complete length of at least 1 metre with 40 pinnae attached is not unreasonable. The pinnae attach to the upper surface of the rachis. At the apex of the frond, the width of the attachment of the pinnae to the rachis is equal to the maximum width of the pinnae (Fig. 17). The width of the attachment decreases toward the base of the frond (Fig. 14). Frond bases of P. zamioides are commonly found. The rachis enlarges at the base and forms a smooth surface which has the appearance of an abscission zone (Fig. 15). This, along with the expanded state of the fronds, suggests that P. zamioides shed its fronds freely. Delevoryas and Hope (1976) noted that in all living cycads the fronds do not drop off while still expanded, but shrivel up before separating from the stem.


113

In the pinnae the veins are either simple or dichotomously branched and apart from an occasional anastomosis they appear to terminate at the margin (Fig. 12). In P. anastomosans the lower epidermal cell shape and orientation shows that there was a network of fine, freely anastomosing veins between the main veins (Fig. 13). P. zamioides does not show the same specialization of epidermal cells. Occasionally some evidence of alignment of cells between major veins can be seen and it could be that P. zamioides had a similar network of veins, but they did not affect the shape and orientation of the epidermal cells.

Fig. 12. Venation pattern of a pinna of Pterostoma zamioides showing one anastomosis.

Lowev epidermis. Stomates occur in bands between veins. The stomatal complex appeared to be orientated at random to the vein direction (Fig. 18). To test this hypothesis, the angle of the long axis of the stomatal complex to the nearest vein was measured for 200 stomates. A xZ test was carried out, giving a value of 8.69 (8 degrees of freedom), making 0.30 < P < 0.50. The null hypothesis of random stomatal orientation was accepted.

Fig. 13. Network of fine anastomosing veins between two major veins of P. anastomosans as reflected by the cuticle.

The stomates have distinctive cuticular thickenings. There are prominent wings of cuticle between the guard cells and subsidiary cells and, when viewed from the inner surface (Fig. 20), there are depressions where the poles of the guard cells terminated. These depressions clearly indicate that the poles of the guard cells were relatively thinly cutinized. On the outer surface there is the characteristic gymnospermous ring of cuticle around the stomatal pore (Fig. 24). There are normally four or five subsidiary cells, but observation of them is difficult because the cuticle often does not penetrate between their cell walls.

R. S. Hill

Fig. 14. Holotype of Pterostoma zamioides (P 150055). Fig. 15. Frond base of P. zamio~des(P 150057) showing swcdlen abscission zone. Fig. 16. Holotype of P. anastomosans (N0157) showing twc) anastomoses between major veins. Fig. 17. Apex of a frond of P. zamioides (P 150058) showirig the very broad attachment of the apical pinnae.


115

The epidermal cells have sinuous walls with decreasing amplitude over the veins and toward pinna extremities. Epidermal cells over the veins are longer and narrower than non-venous epidermal cells. On the outer surface-there are well-developed cuticular ridges (Fig. 22), which decrease in prominence over the veins. No hair bases occur on the lower epidermis. Upper epidermis. Epidermal cells have sinuous walls which decrease in amplitude over the veins (Fig. 26). Over the veins the epidermal cells are longer and narrower than in non-venous areas. Two- or three-celled hair bases rarely occur over the veins near the pinna bases (Fig. 28). The outer surface of the cuticle is featureless, with hair bases and cuticular ridges absent. Rachis. The rachis has longitudinally aligned stomata which are structurally similar to the stomata of the lower epidermis. The epidermal cells are straight-walled. Holotype. Specimen numbers P 150055 and P 150056 (counterpart), housed in the National Museum of Victoria. Type locality. Anglesea, Vic. (38 "25'S.,144" 11 'E.). Collector. R. S. Hill. Date. 15/6/1977. Ptevostoma anastomosans R.S.Hil1 sp. nov.

Fig. 16 Diagnosis Major veins simple or dichotomously branching with a network of finer veins between them exhibiting frequent anastomoses. About 27 veins per cm across pinna with a maximum of at least 14 veins per pinna. Upper epidermal cell walls sinuous, unspecialized over veins. Upper epidermal cells averaging 71 pm (58-86 pm) long, 50 pm (41-58 pm) wide. Two- to three-celled hair bases occur randomly on pinnae. No outer surface sculpturing on upper epidermis. Lower epidermis with randomly orientated stomata between veins. Lower epidermal cell walls sinuous, amplitude decreasing over veins. Veins marked by longer, narrower cells. Epidermal cells in nonvenous areas averaging 63 pm (56-71 pm) long, 39 pm ( 3 4 4 4 pm) wide; venous epidermal cells averaging 76 pm (68-87 pm) long, 28 pm (24-33 pm) wide. Two- to three-celled hair bases common, usually over veins. Outer surface with heavy cuticular ridges, which radiate from hair bases. Stomates averaging 42 pm ( 3 9 4 5 pm) long, 25 pm (23-28 pm) wide. Four to seven haplocheilic subsidiary cells. Description of the Specimen Architecture. The holotype (N0157) consists of the basal part of a pinna with one margin preserved along the whole length and the other preserved toward the base (Fig. 16). Harris (1964) found that it was sometimes difficult to be sure of the true vein course because the interior of the leaves may have liquefied at an early stage of preservation, causing the veins to shift laterally and give a false appearance of anastomosis. Despite this, the effect that veins often have on shape and orientation of epidermal cells usually means that the true vein course can be found by a study of the cuticle. In P. anastomosans the cuticle shows that major veins, which have been preserved in their true course (Fig. 16), branched dichotomously and occasionally anastomosed. A network of finer veins, which have not been preserved, frequently anastomosed (Fig. 13). Such a network of finer veins running between major veins is not unknown in cycads. Hi11 (1978) described them in both fossil and extant species of Bowenia, although they did not anastomose to the extent of those in P. anastomosans.

R. S . Hill


117

Veins that have been preserved vary greatly in thickness (Fig. 16). This is not an artifact of preservation since the thickest veins are those that are least branched. Whenever a vein branches in this species it gives rise to two veins of approximately half the original thickness. Judging from the thickness of the veins at the base, there may have been about 25 veins in the widest part of the pinna. Lower epidermis. Stomates occur in bands between the major veins although there are none directly above the minor veins. The stomates appear to be randomly orientated with respect to the major veins (Fig. 19). The presence of the minor veins makes it difficult to determine whether stomatal orientation is affected by the course of the veins. CuticuIar thickenings around the stomates of P. anastomosans (Figs 21, 25) . . are very s;=!ar ta these ef P. zamicides, Subsidiary cd!s Ere difficl~!tto ehserve, as in P. zamioides, but there are normally four or five. Epidermal cells have sinuous walls with decreasing amplitude over the veins and toward the pinna base. Epidermal cells over the veins are longer and narrower than in non-venous areas. Two- or three-celled hair bases occur frequently over veins (Fig. 29). On the outer surface hair bases show as a circular scar where the basal hair cell was once attached (Fig. 23). The outer surface of the cuticle also exhibits highly developed cuticular ridges which occur all over the epidermis and radiate from the hair bases (Fig. 23). Upper epidermis. The sinuous walls of the epidermal cells are unspecialized over the veins (Fig. 27). Two- or three-celled hair bases occur, but not as frequently as on the lower epidermis. These hair bases are probably over the veins, but owing to the nonspecialization of venous epidermal cells this is unconfirmed. The outer cuticular surface is featureless, with no cuticular ridges or hair bases. Holotype. Specimen number NO157 w ~ t htwo cuticle slides, housed in the South Australian State Herbarium (AD). Type locality. Nerriga, N.S.W. (35 "07'S.,150°05'E.). Collector. R. S. Hill. Date. 22/5/1978. Discussion The two known species of Pterostoma share several distinctive characters. They have sinuous epidermal cell walls on both surfaces, random orientation of stomata, similar cuticular ridges present only on the lower surface and similar cuticular thickenings of the stomatal complex (Figs 20,21,24,25). These features form a unique combination among Tertiary cycads, and there can be little doubt that the specimens from the two localities belong to the same genus. There are two distinct differences between the Nerriga and Anglesea specimens that warrant their separation into two species. P. anastomosans has very frequent vein anastomoses which are reflected by the lower epidermal cell pattern (Fig. 13).

Figs 18, 19. Inner surface of the lower cuticle of Pterostoma zamioides (Fig. 18) and P . anastomosans (Fig. 19). A vein runs from top to bottom down the left-hand side. 100 pm scale applies to Figs 18, 19,22, 23.

Figs 20,21. Inner surface of a stomate of P. zamioides(Fig. 20) and P. anasromosans(Fig. 21). 10 pm scale applies to Figs 20, 21. Figs 22,23. Outer surface of the lower cuticle of P. zamioides (Fig. 22) and P . anastomosans (Fig. 23). A vein runs from top t o bottom down the left-hand side.

118

R. S. Hill


119

Although P. zamioides may have similar vein anastomoses they are not reflected by the lower epidermal cell shape and orientation. Hair bases are common on the lower epidermis of P. anastomosans and also occur, but less frequently, on the upper epidermis. In P. zamioides hair bases are absent from the lower epidermis and are rarely found on the upper epidermis. The two species of Pterostoma show some features that are very unusual in cycads. The first is the highly sinuous epidermal cell walls (Figs 18, 19, 26, 27). This is commonly found in Bennettitales but, with the exception of Stangeria T. Moore and two species of Ctenis Lindley and Hutton ( C . stewartiana Harris and C . coronata Douglas), it is unknown among cycads. The second feature is the random orientation of stormt. (Figs 18, 19). In the Bennettita!es the stomates tend to be trancverse to vein direction or more or less at random (T. M. Harris, personal communication), whereas in cycads the stomates are usually aligned parallel to the veins. The third feature is the presence of 'wings' of cuticle between the guard cells and subsidiary cells. This is a characteristic of bennettitalean cuticles, but it also occurs in some cycads. These 'wings' are well developed in Pterostoma (Figs 20,21) but they do not compare with the development often found in Bennettitales. Despite these characteristics, Pterostoma does not have the bennettitalean arrangement of two subsidiary cells parallel to the two guard cells. This syndetocheilic arrangement is the one feature that unequivocally separates the bennettitaleans from other gymnosperms (Harris 1976). If this is universal then Pterostoma, with its haplocheilic stomata, must be regarded as a cycad. The placement of Pterostoma in the Zamiaceae is tentative. In frond form the fossils fit closely to that family but reproductive structures are as yet unknown. Comparison of Fossil and Extant Species

The frond architecture of Pterostoma is superficially similar to that of some extant species of Macrozamia and Lepidozamia. However, the cuticle shows fundamental differences from the conditions in those genera. Tertiary cycads usually resemble extant genera and where cuticle has been preserved (Cookson 1953; Hill 1978) it is virtually identical with that of some extant genus. Therefore it is unlikely that Pterostoma is closely related to Macrozamia or Lepidozamia. Cuticle of Stangeria eriopus was compared to Pterostoma, since Stangeria is the only extant genus with sinuous epidermal cell walls. This proved to be the only cuticular similarity between the two genera, as was expected considering their vast architectural and geographic differences. Pterostoma therefore has no direct extant relatives, which suggests a postEocene extinction. Interesting comparisons can be made between Pterostoma and cycads which occurred earlier in the fossil record. Douglas (1969);described Ctenis coronata from the Cretaceous of Victoria. The holotype consists solely of dispersed cuticle, but both surfaces are preserved and there are many similarities with Pterostoma. For example,

Figs 24, 25. Outer surface of a stomate of P. zamioides (Fig. 24) and P. anastomosans (Fig. 25). Scale applies also to Fig. 24. Figs 26, 27. Inner surface of upper cuticle of P. zamioides (Fig. 26) and P. anastomosans (Fig. 27). Scale applies also to Fig. 26. Figs 28,29. Hair bases of P. zamioides (Fig. 28) and P. anastomosans (Fig. 29). Scale applies also to Fig. 28.

R. S. Hill

Douglas (1969) observed that Ctenis coronata has sinuous cell walls, transverse or randomly orientated stomata, a high frequency of hair bases on the lower cuticle, vein anastomoses and 'crowns' at the poles of the stomates where the guard cells are thinly cutinized. There are also several features which readily distinguish Ctenis coronata from the two species of Pterostoma. Ctenis coronata is closely related to Pterostoma and may even belong in that genus. However, Douglas (1969) was correct in his assertion that the cuticle shows no feature to exclude it from the genus Ctenis and until a more complete specimen is located it should remain there. Pterostoma zamioides agrees very closely with the generic description of Pseudoctenis Seward given by Harris (1964). The only differences are that the pinnae . I ~ Pinserted G:: the upper surfzce of the rachis rather than iateraiiy and the epiaermai cell walls are sinuous rather than straight. Pterostoma anastomosans fits more closely to the diagnosis of Ctenis because of the frequent vein anastomoses. However, the anastomoses in Ctenis are between major veins, whereas in P. anastomosans they are in the network of finer veins. Taking this into account, P. anastomosans also fits the generic description of Pseudoctenis as its major veins anastomose only occasionally. Both Ctenis and Pseudoctenis contain predominantly pre-Cretaceous fossils. Harris (1964) observed that the fronds of more than one genus of recent cycads could be placed in Pseudoctenis, and Delevoryas and Hope (1971) believed that the various species of Pseudoctenis could belong to different genera of plants. To place Australian Tertiary cycads in Pseudoctenis would be misleading because this genus is best reserved for vegetative material from earlier periods. For this reason and because neither species fitted the generic diagnosis of Pseudoctenis exactly, the new genus Pterostoma was described.

.

Conclusion Since the review of Australian fossil cycads by Cookson (1953), three deposits on the east coast of Australia have yielded six Tertiary cycads. They are Lepidozamia hopeites from the Oligocene of Bacchus Marsh, Vic., L. foveolata, Bowenia papillosa and Pterostoma anastomosans from the Eocene of Nerriga, N.S.W. and Bowenia eocenica and Pterostoma zamioides from the Eocene of Anglesea, Vic. Preliminary physiognomic analyses of the Anglesea and Nerriga floras performed by Christophel (1979) with the system of Webb (1959) suggest that they were subtropical. The extant species of Bowenia and Lepidozamia are found in north-eastern Australia (Fig. 1). Lepidozamia peroffskyana has the southernmost distribution, extending south to the Manning River district in New South Wales (Johnson 1959). The extant Australian cycads with greatest representation, Macrozamia and Cycas, have yet to be found as Tertiary fossils in Australia. These genera do not commonly grow in rainforests and are currently found in areas where fossilization is not likely. Thus the small number of cycad megafossils already described which belong to extant genera may prove to be useful indicators of palaeoclimate. This is because the extant species of those genera are found in climates similar to those predicted by the physiognomic analyses. The presence of the extinct genus Pterostoma at Nerriga and Anglesea suggests that Eocene cycads not only had an extensive distribution throughout south-east Australia but also they may have had a large diversity as well. Pterostoma exhibits some features more commonly found in Bennettitales than in cycads, namely sinuous cell walls, random orientation of stomata and to a lesser extent cuticular 'wings' between the guard cells and subsidiary cells. Pterostoma may possibly have evolved from Ctenis or


121

some element of the more nebulously conceived Pseudoctenis, but many gaps need to be filled before its true ancestry can be traced. Unfortunately Ptevostoma, like so many megafossils, is known only from vegetative material, although reproductive structures may eventually be found, especially at the vast Anglesea deposit. Although little can be said of the structure of Pterostoma as a complete organism, some idea of its habitat can be gained from a study of the associated floras. As Christophel and Blackburn (1978) noted, leaves, flowers and fruits can be transported only limited distances before decay, mechanical disruption and hydrodynamic sorting have significant effects on them. Pterostonza zamioides is a dominant taxon of one of the three lenses collected at Anglesea. Because of this, and its excellent state of preservation, it is reasonabie to assume that P.zamioides occurred ciose ro r'ne sire of deposition. The co-dominant of that lens is Casuarina (Christophel1979) which is also well preserved. Therefore it is possible that the vegetation was dominated by Casuarina with a Ptevostoma understory. The extant species of Macvozamia usually grow in association with sclerophyllous species (Johnson 1959) and Ptevostoma may have occupied a similar ecological niche. After extensive collections at Nerriga, Ptevostoma anastomosans has been found as only one small pinna fragment. It is therefore possible that it may not have grown in the immediate vicinity of the site of deposition and could have occurred among a different vegetation type to that represented in the deposit. However, Casuavina has also been found at Nerriga and an association similar to that suggested for Anglesea may have occurred there. Ash (1976) listed three criteria that a megafossil must meet to be of use stratigraphically. It must have easily identifiable features, be widely distributed and have a narrow stratigraphic range. When Tertiary deposits, which are usually dominated by angiosperms, are studied cycads easily fulfil the first criterion. Both the architecture of cycad fronds and, more importantly, the cuticular structure separate them easily from angiosperms and usually allow quick identification at least to the generic level. The second and third criteria are as yet largely unknown, although Lepidozamia and Bowenia which are present in the Eocene and still survive today must cover a vast stratigraphic range. While all Tertiary gymnosperms are potentially valuable as stratigraphic indicators, they often occur in such small numbers that they may easily be overlooked in a deposit (Pterostoma zamioides is a rare exception). Therefore if a plant megafossil stratigraphy of the Australian Tertiary is possible it should be based on angiosperm fossils, probably leaves. Acknowledgments This work was carried out with the assistance of an A.R.G.C. grant (E77115532) to D r D. C. Christophel, a Commonwealth Post-graduate Research Award and James Cook University U.R.G. number 597. I would like to thank Mr C. Temple for access to the Nerriga locality and ALCOA of Australia Ltd for access to the Angelsea coal mine during 1977 and 1978. The following institutions provided material essential to this study: The National Museum of Victoria; The Herbarium of the Canberra Botanic Garden; The National Herbarium of New South Wales. I am indebted to the following people for their advice and assistance: Dr D. C. Christophel, Professor T. M. Harris, Professor T. Delevoryas, Dr L. A. S. Johnson, Dr J. G . Douglas, Dr T. H. Rich, Dr M. D. Crisp and Mr P. J. Lang.

R. S. Hill

References Ash, S. R. (1976). The systematic position of Eoginkgoites. Am. J . Bot. 63, 1327-31. Chapman, F. (1925). On some seed-like bodies in the Morwell brown coal. Geol. Surv. Vict. Rec. No. 4, 487-9. Christophel, D. C. (1979). Tertiary megafossil floras of Australia as indicators of floristic association and palaeoclimate. In 'Ecological Biogeography of Australia', ed. A. Keast. (W. Junk: The Hague, Netherlands.) (In press.) Christophel, D . C., and Blackburn, D. T. (1978). Tertiary megafossil flora of Maslin Bay, South Australia: a preliminary report. Alcheringa 2, 31 1-19. Cookson, I. C. (1953). On Macrozamia hopeites-an early Tertiary cycad from Australia. Phytomorphology 3, 306-12. Eeane, II.(1902). Notes on the fossi! flora ofpitfield and Mornington. Geol. Surv. Vici. Rei. No. 1, 15-20. Delevoryas, T., and Hope, R. C. (1971). A new Triassic cycad and its phyletic implications. Postilla 150, 1-21. Delevoryas, T., and Hope, R. C. (1976). More evidence for a slender growth habit in Mesozoic cycadophytes. Rev. Palaeobot. Palynol. 21, 93-100. Douglas, J. G . (1969). The Mesozoic floras of Victoria. Pt. 1. Geol. Surv. Vict. Mem. No. 28. Douglas, J. G . (1978). Victoria's oldest flowers. Vict. Nut. 95, 13740. Harris, T. M. (1964). 'The Yorkshire Jurassic Flora. Pt. 2, Caytoniales, Cycadales and ~teridos~erms.' 186 pp., 8 plates. (Br. Mus. Nat. Hist.: London.) Harris, T. M. (1976). The Mesozoic gymnosperms. Rev. Palaeobot. Palynol. 21, 119-34. Hill, R. S. (1978). Two new species of Bowenia Hook. ex. Hook. f. from the Eocene of eastern Australia. Aust. J. Bot. 26, 837-46. Johnson, L. A. S. (1959). The families ofcycads and the Zamiaceae of Australia. Proc. Linn. Soc. N.S. W . 84, 64-1 17. Johnson, L. A. S. (1963). Cytological and taxonomic notes on Zamiaceae. Contrib. N.S. W. Herb. 3, 23542. Owen, J. A. (1975). Palynology of some Tertiary deposits from New South Wales. Ph.D. Thesis, Australian National Univ. Pant, D. D., and Nautiyal, D. D. (1963). Cuticle and epidermis of recent Cycadales. Leaves, sporangia and seeds. Senckenbergiana Biol. 44, 257-347. Webb, L. J . (1959). A physiognomic classification of Australian rain forests. J. Ecol. 47, 551-70.

Manuscript received 5 July 1979