A GLOSSOPTERIS WITH WHORLED LEAVES

H. HAMSHA W THOMAS Botany School, Cambridge

MONG the many valuable contribu­ the mode of attachment of the leaves is tions to palaeobotany made by the different. It seems highly probable that this A late Prof. Birbal Sahni, his description specimen shows a whorl of Glossopteris leaves of the cuticles of Glossopteris angustifolia attached to a slender stem, for in the speci­ Brongn. (SAHNI, 1923, p. 277, PL. 17) will men of G. angustifolia figured by Sahni ( 1923, always be remembered. Leaves of this type PL. 17, FIG. 1 ) the midrib appears to become had been known for a century and had usually very slender at some distance from the leaf been regarded as ferns, but this work showed apex. In the same work Feistmantel (1881, that they had characters now only known to p. 113, PL. 41 A, FIGS. 3, 4) described another exist in the leaves of seed-bearing plants. form as Sagenopteris (?) PolYPhylla founded The purpose of the present communication on two specimens each showing seven stalked is to add one more fact to our knowledge of leaves arising from a common axis. The this same genus by describing the attach­ venation of these leaves resembles that of ment of leaves, somewhat similar to those of Glossopteris retifera Feist. G. angustifolia, to a slender stem in a verti­ Etheridge ( 1894, p. 228 ) described another cillate manner. This is seen in a species specimen from Australia which was believed from the Molteno beds of South Africa. to have come from the Newcastle Coal Measures, showing a small stem or caudex PREVIOUS OBSERVATIONS CONCERNING surmounted by a clump of closely packed THE HABIT OF GLOSSOPTERIS fronds. All traces of organic matter had disappeared but the leaves were stated to The h~bitof the plants bearing the leaves resemble those of Glossopteris clarkei Feist. named Glossopteris has been debated for more The stem was ! in. in width and appeared than a hundred years. A century ago Dana to be covered with transversely elongated (in WILKES, 1849, p. 716, PL. 12, FIG. 13 c) leaf scars arranged alternately. Etheridge described and figured a number of fronds thought that his specimen resembled Feist­ referred to G. browniana attached in a clump mantel's Sagenopteris longifolia. to a fragment of a stem. The figure shows None of the above-mentioned examples a number of petioles radiating from a com­ showed stems of the Vertebraria type, long mon level as though belonging to a whorl. known to be commonly associated with Later the opinion was expressed by Zigno Glossopteris leaves, but at the end of the (BUNBURY, 1861, p. 327) that this species century Zeiller ( 1896, pp. 356-362) described had a compound or digitate frond like that specimens from South Africa which seemed of Sagenopteris. There is, of course, consi­ to show Glossopteris leaves attached to these derable similarity between a leaf of Glossop­ stems. It now seems very uncertain that the teris and a leaflet of Sagenopteris in form and specimens which he figured did show the venation. It may well be due to parallel organic connection between leaves and stems, evolution, but it has led to considerable con­ and there can be little doubt that his inter­ fusion. Thus Feistmantel (1881, p. 113) pretation of the structure of Vertebraria is gave the name of Sagenopteris longifolia to a incorrect. Walton and Wilson's important specimen showing about six leaves or leaflets study of the structure of a Vertebraria stem attached at the same level to a common stalk, from southern Rhodesia shows that this but he remarked that" a single leaflet might organ is entirely different from the structure occasionally pass for a Glossopteris angusti­ envisaged by Zeiller. The paper by these folia Brongt. as regards general form and the authors (WALTON & WILSON, 1932, p. 201, secondary veins but the midrib in this latter TEXT-FIG. 1) includes a new figure of an is quite distinct throughout". The stalk, interesting specimen from Vereeniging, South which he figures but does not describe, is Africa, which had previously been figured quite unlike the petiole of a Sagenopteris and by Seward ( 1910, p. 504, FIG. 339). This 436 THE PAL.J\EOBOTANIST shows a group of four leaves of Glossopteris thickening of the lamina along its crenulate browniana Bgt. at the end of a slightly margin. One of the leaflets showed a small tapering axis, about I em. wide. There can lamina and a very long stalk, a remarkable be little doubt that the leaves were attached feature for a compound leaf of the Sagenop­ to the axis in a whorl, while the stem shows teris type. The way in which the leaves or traces of other transverse nodes suggesting leaflets of the two groups cross each other the production of a series of leaf whorls. was figured but not mentioned; it is very There is no evidence that this stem had the similar to the appearance of the specimen Vertebraria structure. which will be described below. I examined this specimen when I was in Cape Town; it GLOSSOPTERIS IN THE MOLTENO BEDS OF seemed to differ considerably from any SOUTH AFRICA specimen of Sagenopteris that I had seen, and its derivation from a member of the Cayto­ Leaves belonging to several species of niales seemed unlikely. Glossopteris are common in the beds of the New E vidence from the Molteno Black Ecca series; they occur less frequently in the Shales - After examinning du Toit's co]]ec­ Beaufort beds, and have been recorded by tion in Cape Town I was able, through the du Toit in the middle portion of the Molteno kindness of Mr. J. W. 13. MacLean, to spend beds. Among the plants forming the rich some time in collecting at the waterfall and well-preserved flora from the waterfall locality in the valley of the Upper Umkomaas in the valley of the Upper Umkomaas, du in Natal from which the specimens had been Toit found leaves which he named G. brown­ obtained. It seemed most important to iana and G. conspicua Feist. (DU TOIT, ascertain if. in fact, the remains of plants 1927, p. 364-367). He also discovered and referable to the CaytoniaJes did occur in the described a beautiful specimen which he Triassic rocks of South Africa. The plant designated Sagenopteris longicaulis sp. nov. remains at this locality are singularly well ( 1927, p. 325, TEXT-FIG. 4). This requires preserved and often allow the preparation further consideration since it is probably an of good cuticle preparations. Many leaves example of the same plant which will be and reproductive structures referabl.e to the described below. In his account of this Pteridosperms (THOMAS, 1933, pp. 193-265 ) form, du Toit states that the slab of black were discovered, but no trace of any structure shale No. 8670 (South African Museum, belonging to the Caytoniales was found. Cape Town) carries the impressions of eight My collection contains, however, about leaflets seemingly belonging to two verticils. twenty examples of leaves referable to Glos­ The largest leaflet was 7 em. long and 1·5 sopteris and a unique specimen, with its em. wide, lanceolate in shape with an acute counterpart, which explains the nature of apex and a lamina which tapers gradually du Toit's supposed Sagenopteris. towards the base in an asymmetric manner; This specimen, which is shown in the photo­ at the base of each leaflet there was a foot­ graphs in Plate 1, Fig. 1, shows two whorls of stalk more than I em. long which passes leaves of the Glossopteris type springing from upwards into a midrib, prominent at the a slender stem.l The attachment of the lower end but breaking up towards the apex leaves of one whorl to the stem is clearly seen into subparallel or gradually diverging veins ( PLATE 1, FIG. 1, lower part, and FIG. 2) but with occasional cross connections. The the specimen is fractured at the place where secondary veins were coarse, given off very the leaves of the other whorl (the uppermost acutely from the midrib from which they in PLATE I, FIG. 1) join the stalk. gradually curved away, forked and anas­ The stem from which the leaves sprang tomosed, producing a network very like that is 3 mm. wide throughout most of its length of Glossopteris. The specimen did not show but its width increases to 4 mm. at the nodes. the attachment of the leaflets to a stem or Its surface is generally smooth but traces of petiole, but from their converging arrange­ low longitudinal ridges are visible in places, ment it was assumed that they had formed as if due to the presence of lines of harder parts of two leaves. In assign~ingthe speci­ tissue in its internal structure; it shows men to Sagenopteris, du Toit seems to have been influenced by the crenulate margin of 1. In order to give a more correct representation of the specimen the whorl which was originally some of the leaflets, the character of the the lower one is shown in the upper part of the midrib, and by the appearance of a slight photograph. THOMAS - A GLOSSOPTERIS WITH WHOI

In view of the more complete information CONCLUSIONS now available about the present form, it would seem desirable to use a distinct specific The evidence that has slowly accumulated name for the specimens that have been over a hundred years suggests that the described. The most appropriate specific leaves of Glossopteris were borne in whorls on name is the one given by du Toit to the relatively slender stems. specimen which came from the same locality, A specimen collected by the author from and which he called Sagenopteris longicaulis. the black shales in the valley of the Upper I feel sure that his specimen belonged to the Umkomass, Natal, South Africa, shows two same species of plant as the form now des­ whorls of leaves springing from a stem 3 mm. cribed. Such differences as may be observed in diameter. There were probably nine between the two specimens may well be due to leaves in each whorl. a difference of position on the original plant. The shape and venation of the leaves Du Toit's leaves were smaller, tapered more differ somewhat from that seen in other rapidly towards the base, and probably species and the specimen is named Glossopteris sprang from a shoot with shorter internodes. longicaulis (du Toit ) comb. nov. The only real problem is the use of the I t is suggested that the forms from India generic name of Glossopteris for these speci­ and Natal, previously described by Feist­ mens in view of the relatively few anastomos­ mantel and du Toit as species of Sagenopteris, ing branches among the secondary veins. were really examples of Glossopteris with But since all the leaves do show a consider­ whorls of leaves. able number of closed meshes, it would seem The view that Glossopteris leaves sprang unnecessary to create a new genus. It has from stems of the Vertebraria type is regarded been shown above that there are many as unreliable, but such stems may have been indications in the literature that the leaves of of the nature of rhizomes bearing leafy other species of Glossopteris were borne in shoots. whorls, though no specimen seems to have The specimen described and figured comes shown this habit so clearly as the present from the Middle Molteno beds and is of example. Triassic age.

REFERENCES ARBER, E. A. N. ( 1905). Catalogue of the fossil SEWARD, A. C. (1910). Fossil Plants. 2. Cam­ plants of the Glossopteris Flora in the British bridge. Museum. London. THOMAS, H. H,\MSHAW (1933). On some Pteri­ BUN BURY, C. J. F. ( 1861) Notes on a Collection dospermous Plants from the Mesozoic Rocks of of Fossil Plants from Nagpur, Central India. South Africa. Phil. Trans. Roy. Soc. London. Quart. } ourn. Geol. Soc. London. 17: 325-346. B 222: 193-265. Du TOrT, A. L( 1927). Fossil Flora of the Upper WALKOM, A. B. (1922). Palaeozoic Floras of Karroo Beds. Ann. South African lV!useum. Queensland. Pt. 1. The Glossopteris Flora of 22, Pt. 2: 290-420. the Lower and Upper Bowen Series. Queensland ETHERIDGE, R. J UN. (1894). On the mode of Geo!. Surv. Pub. 270. attachment of the leaves or fronds to the caudex WALTON, J. & WILSON, J. A. R. ( 1932). On the in Glossopteris. Proc. Linn. Soc. New South Structure of Vertebra·ria. Proc. Roy. Soc. Edin­ Wales. Ser. 2. 9: 228-249. burgh. 52 (2): 200-207 FEISTMANTEL, O. (1881). Fossil Flora of the WILKES, C. (1849). U.S. Exploring Expedition. Gondwana System. 3, Pt. 3. The Flora of the 10 (Geology). Philadelphia. Damuda-Panchet Divisions. Palaeontologica In­ ZEILLER, M. R. (1896). Etude sur quelques dica. 12: 78-149. plantes fossiles, en particulier Vertebraria et SAHNI, B. ( 1923 ). On the Structure of the Cuticle Glossopteris, des environs de Johannesburg in Glossopteris angustifolia Brongn. Records ( Transvaal). Bull. Soc. Geo!. France. 3me Sth. Geol Surv. India. 54, Pt. 3: 277-280 24: 356-362

EXPLANATION OF PLATE 1

1. Glossopteris longicaulis (du Toit) Thomas from Shows the attachment of one whorl of leaves, to the valley of the Upper Umkomaas, Natal, South the stem. X 5/2. Africa. The stem bearing the two whorls of leaves 3. Portion of the counterpart with parts of leaves. is seen on the left side of the photograph. X 7/5. Shows the midribs and secondary veins on the lower 2. Counterpart of the specimen shown in Fig. 1. side of the leaves. X 5/2.