Fossil plants, Vol. 2

The genera and species described in this Chapter are founded on sterile leaves or portions of leaves, and in the great majority of cases the reproductive organs are either imperfectly known or have still to be discovered. Some of the genera, the smaller number, are no doubt true ferns, while most of them may safely be regarded as plants which will ultimately be shown to belong to some other group, in most cases that of the Pteridosperms. It is possible that a few of the types may be members of the Cycadophyta rather than of the Pteridospermeae, but evidence as to systematic position is for the most part of a negative kind or too incomplete to lead to any definite expression of opinion as to the cycadean or pteridosperm nature of the imperfectly known Palaeozoic or Mesozoic species. Many of the genera are of little botanical interest, though even the most problematical are of importance as criteria of geological age. Genera which there is good reason for including in the Pteridosperms are dealt with in this section, in order that the Chapter in Volume III. devoted to this important group may be limited to more completely known types.

In most text-books it is customary to employ family names for sterile fern-like fronds which possess similar venation features or have in common certain vegetative characters, the value of which it is impossible to estimate. In the following account family or group names are not adopted, on the ground that such slight utility as they may have is more than counterbalanced by the risk attending a grouping under one name of plants which may agree only in unessential characters. The practice of classifying fossil plants has been carried to excess. Grouping together genera as a matter of convenience unavoidably creates a prejudice in favour of actual relationship, which may or may not exist.

Taeniopteris.

This generic name was instituted by Brongniart[1254] for simple linear or broadly linear leaves with a prominent midrib from which secondary veins, simple or dichotomously branched, are given off at right angles or obliquely. The frond of the type-species Taeniopteris vittata (fig. 332), characteristic of Jurassic floras, was compared by Brongniart with the pinnules of Danaea and Angiopteris. Among recent ferns the Taeniopteris form of frond and venation is represented by Oleandra neriiformis, Asplenium nidus, and many other species. Though usually applied to fronds which there is good reason for regarding as simple leaves, the generic designation Taeniopteris has been extended to include pinnate fronds, e.g. the Upper Palaeozoic species T. jejunata Grand’Eury, and T. Carnoti Ren. and Zeill. (fig. 330, A). The compound fronds from the Lower Coal-Measures of Missouri described by Dr White[1255] as T. missouriensis are characterised by decurrent and confluent Taeniopteroid pinnules. In a later reference[1256] to this plant White pertinently adds, “perhaps it belongs more properly in Alethopteris.”

Leaves of the Taeniopteris type are described by several authors as species of Oleandridium, Angiopteridium, Danaeites, Marattia, and other genera. In such species of Taeniopteroid leaves as have been dealt with in a former Chapter, the occurrence of sori justifies the substitution of a name denoting a close relationship to existing members of the Marattiaceae, but in the absence of fertile specimens the provisional designation Taeniopteris should be retained. It is often difficult to decide between Taeniopteris and Nilssonia as the more suitable name to apply to fragments of fossil leaves of Mesozoic age. Taeniopteris is, however, distinguished from the Cycadean genus by the greater prominence of the rachis, also by the dichotomous branching of the secondary veins, usually close to their origin and at varying distances between the axis of the frond and the edge of the lamina. The genus Taeniopteris, though most abundant in Rhaetic and Jurassic strata, occurs also in Upper Carboniferous and Lower Permian rocks. The generic name Macrotaeniopteris instituted by Schimper[1257] has been used for leaves differing only in size from the usual type of Taeniopteris, but there is no adequate reason for its retention.

The species included in Taeniopteris afford no satisfactory evidence as to their systematic position. It is obviously unwise to adopt such generic titles as Oleandridium, Marattiopsis, etc., merely because of resemblance in the venation of sterile fragments to Oleandra or Marattiaceous ferns.

Some specimens of Taeniopteris fronds described by Mr Sellards[1258] from Permian rocks of Kansas, which are referred to later, have furnished unconvincing evidence of reproductive organs.

Taeniopteris multinervis, Weiss. Fig. 329, A, B.

The late Dr Weiss[1259] instituted this species (which he designated Taeniopteris multinervia, though the specific name multinervis is constantly used) for a fragment of a leaf from the Lower Permian of Lebach characterised by numerous forked veins given off at right angles from a prominent rachis (fig. 329, B). This type of frond is recorded from the Permian of Trienbach (Alsace) by Zeiller[1260], by Renault[1261] and Zeiller[1262] from the Upper Carboniferous of Autun, and from other localities. The lamina of the simple leaf reaches a breadth of 6 cm. and a length of 40 cm. (fig. 329, A); the numerous secondary veins (25–36 per cm. of lamina) are either at right angles to the rachis or given off at an acute angle. The mesophyll consists of polygonal cells some of which are elongated at right angles to the surface of the lamina. A very similar form is described by Fontaine and White from the Permian of Virginia as T. Lescuriana[1263].

It is futile to expect to be able to separate the numerous Taeniopteris leaves into well-defined species: all we can do is to group the specimens under different names, using as artificial distinctions such characters as the shape of the leaf, the number of veins per centimetre, and the prominence of the rachis. Another Virginian species of Permian age described by Fontaine and White[1264], T. Newberriana, is said to bear sori, but no satisfactory information is given as to the nature of these organs. Specimens referred with some hesitation to this species and to a similar species, T. coriacea, have been described by Sellards[1265] from material obtained from Permian beds in Kansas. The lamina of the simple linear fronds is characterised by the occurrence of small oval bodies half immersed in the substance of the leaf between the secondary veins (figs. 330, D, E). One of these bodies is represented in an apparently dehisced condition in fig. 330, D. Sellards suggests the possibility that these bodies are sporangia, but, as he points out, they afford no indication of cellular structure nor are they in direct connexion with the veins.

Taeniopteris jejunata, Grand’Eury[1266].

This species differs from T. multinervis in its bipinnate fronds; the linear or oval-linear pinnae are attached by a short stalk to the primary rachis and reach a length of 25 cm.; the veins are less crowded, 12–15 per centimetre.

T. jejunata is recorded from the Coal-fields of the Loire and Commentry[1267] in France, from the Lower Permian of Thuringia[1268], and elsewhere.

Taeniopteris Carnoti, Ren. and Zeiller[1269]. Fig. 330, A.

This species, founded on portions of pinnate fronds from the Coal-field of Commentry, is characterised by rather broader (25–30 mm.) pinnules, with short pedicels and a cordate base, reaching a length of 25–30 cm. The secondary forked veins are more numerous than in T. jejunata. In T. multinervis the pinnules are still broader and have a stronger midrib.

Several species of Taeniopteris have been described from Triasso-Rhaetic rocks in Europe, India, Tonkin and elsewhere. In some cases it is practically impossible to recognise clear specific distinctions between Rhaetic and Jurassic types.

From the Damuda and Panchet series of India (Triasso-Rhaetic) Feistmantel has described large sterile fronds as Macrotaeniopteris Feddeni[1270] which reach a breadth of 20 cm.: these may be compared with the Indian species Taeniopteris lata Oldham[1271], and to T. gigantea from the Rhaetic of Franconia[1272] and Scania. A specimen of this species figured by Nathorst[1273] from Scania has a lamina 33 cm. broad. Other examples are afforded by M. Wianamattae Feist.[1274] from rocks of the same age in Australia and by Taeniopteris superba Sap.[1275] from Lower Rhaetic rocks near Autun.

From the Rhaetic of Tonkin, Zeiller records several species, among which may be mentioned T. Jourdyi Zeill.[1276] and T. spatulata MacClelland (fig. 330, B, C). Both have simple fronds. Those of T. Jourdyi reach a length of 10–40 cm. and a breadth of 10–70 mm.; the rachis is characterised by crowded and discontinuous transverse folds, and the secondary veins (35–50 per cm.) are usually at right angles to the rachis. This Tonkin species is compared by Zeiller with the European Rhaetic species T. tenuinervis Brauns.

The polymorphism of the fronds is a striking feature: in one case described by Zeiller the lamina appears to be divided into segments like those characteristic of the leaf of the Cycadean genus Anomozamites. It is obviously difficult in many instances to distinguish between detached Taeniopteroid pinnae of a compound frond and complete simple leaves. In some compound fern fronds, as in the recent Polypodiaceous genus Didymochlaena, the pinnules are deciduous, and the same feature undoubtedly characterised the fronds of many extinct species. A specimen figured by Zeiller which shows several petioles of T. Jourdyi attached to a thick stem[1277] demonstrates the simple nature of the leaves. In other cases, e.g. T. vittata, specimens occur in which the slightly enlarged petiole-base has a clean-cut surface indicating abscission from a rhizome (fig. 332).

The fronds described by Zeiller as T. spatulata[1278] (fig. 330, B, C) closely resemble Jurassic leaves from Victoria referred to Taeniopteris Daintreei McCoy[1279].

Whether specifically identical or not, these leaves represent a type distinguished from the other species of the genus by the small breadth of the linear-lanceolate or linear-spathulate lamina, which may be 6–15 cm. in length and 3–12 mm. broad. The lamina is often characterised by transverse folds (fig. 330, C).

Taeniopteris Carruthersi. Fig. 331.

The simple fronds included under this specific name are characterised by a strong midrib from which numerous simple or forked secondary veins are given off at a right angle or slightly inclined. The breadth of the lamina decreases gradually towards the petiole. The Australian species named by McCoy Taeniopteris Daintreei, to which Carruthers referred the Queensland fossils, has a much narrower and more linear form of frond, and for this reason Tenison-Woods instituted a new specific name. T. Carruthersi represents a form of leaf met with in Rhaetic, or possibly Upper Triassic, rocks in S. Africa[1280] and Australia. A very similar, perhaps an identical type, was described from Argentina by Geinitz[1281] as T. mareyiaca: among many other examples of this form of frond may be mentioned T. immersa[1282] Nath. from the Rhaetic rocks of Scania and T. virgulata from the Rhaetic of Tonkin[1283].

A comparison of Taeniopteris Carruthersi or various other “species” of Rhaetic fronds with the Jurassic species T. vittata illustrates the slight and unimportant differences on which specific separation is based. It is hopeless to attempt to draw a satisfactory distinction between the numerous Taeniopteris fronds from Upper Triassic and Jurassic rocks.

Taeniopteris vittata, Brongniart. Fig. 332.

The simple leaves to which Brongniart applied this name are characteristic of the Inferior Oolite flora of England, and examples of the same type are recorded from Jurassic rocks of India, Poland, the Arctic regions, Japan, China, Australia and other countries[1284].

Leaf linear-lanceolate, reaching a length of more than 20 cm. and a breadth of 3 cm. The lamina increases gradually in breadth from the base and tapers towards the apex. Numerous secondary veins are given off at right angles from a broad midrib: the lateral veins may be simple or forked close to their origin, near the margin, or in the intermediate portion, of the lamina.

It is exceedingly difficult to use Taeniopteris leaves of this form as evidence in regard to the Jurassic or Rhaetic age of plant-bearing strata. The species T. tenuinervis Brauns, as figured by Schenk[1285] from the Rhaetic rocks of Germany and Persia, and recorded from several other regions, presents a close agreement with T. vittata. Oleandridium lentriculiforme Etheridge[1286] from the Hawkesbury series of Australia is another similar leaf. The species T. vittata from the Yorkshire coast, represented in fig. 332, shows a well-preserved petiole with a clean-cut base like that of the petioles of Oleandra neriiformis and other recent ferns which are detached from the rhizome by the action of an absciss-layer.

A broader form of frond with similar venation was described by Lindley and Hutton[1287] as Taeniopteris major. An examination of the type-specimen from the Inferior Oolite of Yorkshire, now in the Manchester Museum, led me to doubt the necessity of specific separation from T. vittata[1288].

A smaller frond of the same general type as T. vittata is recorded from Wealden strata of North Germany and England under the name T. Beyrichii[1289].

Weichselia.

This generic name was instituted by Stiehler[1290] for impressions of bipinnate sterile fronds, presumably ferns, from Lower Cretaceous rocks near Quedlinburg. The same type of leaf from English Wealden beds had previously been referred by Mantell and other authors to Pecopteris, and by Brongniart to his genus Lonchopteris[1291]. It is, however, advisable to follow Nathorst’s example[1292] and restrict the latter name to Palaeozoic species. As already suggested, it would obviate confusion to substitute a new generic designation for Lonchopteris in the case of Triassic species which are probably members of the Osmundaceae. The type-species of Stiehler, Weichselia Ludowicae[1293], does not differ in any important character from Weichselia Mantelli, the species originally described by Stokes and Webb from the Wealden of England as Pecopteris reticulata.

Weichselia Mantelli (Brongn.)[1294]. Fig. 333.

Frond bipinnate, rachis broad; pinnae very long, of uniform breadth and with prominent axes; pinnules crowded, entire, with obtuse apex, usually oblong but more or less triangular or rounded towards the distal ends of the pinnae. The pinnules, which may reach a length of 9 cm., are characterised by a fleshy lamina attached by the whole breadth of the base; the two rows of segments on each secondary rachis are usually inclined towards one another so that they form with the axis of the pinna a wide-open V instead of lying in one plane (fig. 333, C). From a median rib are given off numerous anastomosing branches (fig. 333, B).

This characteristic Wealden species is recorded from England, Germany, France, Belgium, Austria, Russia, Bornholm, North America, and Japan. It is by no means certain that Weichselia Mantelli is a true fern: no satisfactory evidence of fructification has been adduced.

The broad and strong rachis is comparable with that of a Cycadean leaf and the thick lamina suggests a plant of xerophilous habit. I have retained the specific name Mantelli on the ground of long established usage instead of following Fontaine in his adherence to strict priority.

Glossopteris.

The name Glossopteris was proposed by Brongniart in 1822[1295] for an imperfect leaf-impression which he called Filicites (Glossopteris) dubius, but the specimen so named has since been identified as part of a sporophyll of a Lepidostrobus. The author of the genus afterwards published[1296] a diagnosis, based on well-preserved leaves from Permo-Carboniferous rocks in Australia and India, of the type-species Glossopteris Browniana, the Indian examples being distinguished as G. Browniana var. indica while the Australian form was named G. Browniana var. australasica. Schimper[1297] afterwards raised the Indian fossils to specific rank as G. indica though some authors[1298] have continued to consider the two forms as insufficiently distinct to be regarded as different species.

The genus Glossopteris may be defined as follows:

Leaves simple, varying considerably in size, shape, and venation characters, but almost without exception characterised by repeatedly anastomosing lateral veins. The leaves are of two kinds: (i) foliage leaves; apparently always sterile, usually spathulate, with an obtuse apex, a well-marked midrib which may persist to the apex or die out in the upper half of the lamina, characterised by its slight prominence and comparatively great breadth especially in the basal part of the frond. In most cases the lamina extends as a narrow margin to the leaf-base, but in a few forms there is a short petiole (fig. 334). Though usually spathulate, the frond may be linear-lanceolate, or ovate; the apex is sometimes acute. Leaves vary in length from 3 to 40 cm. and may in larger forms have a breadth of 10 cm. Numerous lateral veins curve upwards and outwards to the margin of the lamina or pursue a straight course almost at right-angles to the midrib. (ii) Scale-leaves[1299] which differ from the foliage-leaves in their much smaller size and in the absence of a midrib; they are deltoid, oval or cordate in shape and generally terminate in an acute apex; the edge of the lamina may be slightly incurved so that the leaf presents a convex upper surface supplied with anastomosing veins. The scale-leaves, which vary in length from about 1 to 6 cm., probably acted as sporophylls. The only evidence as to the nature of the fructification so far obtained is represented by empty sporangium-like organs (1·2–1·5 mm. long by 0·6–0·8 mm. broad) frequently associated with the scale-leaves[1300].

The leaves, in some cases at least, were borne near together on a cylindrical stem or rhizome which produced branched adventitious roots[1301]. The fossils long known as Vertebraria were recognised by Zeiller[1302] and by Oldham[1303] as the stems of Glossopteris.

The systematic position of Glossopteris must for the present be left an open question. Though usually spoken of as a fern, it is noteworthy that despite the enormous abundance of its foliage leaves in the Permo-Carboniferous strata of India, Australia, South Africa, and South America, no single example has been discovered which shows undoubted remains of sori or sporangia. Many authors have described fertile leaves of Glossopteris; but it was not until Arber’s discovery of sporangia in close association with the scale-leaves that any light was thrown on the nature of the reproductive organs.

The probability is that Glossopteris was not a true fern but a member of that large and ever-increasing class, the Pteridosperms. This opinion is based largely on negative evidence. Such sporangia as have been described may have contained microspores and the plant may have been heterosporous. The occurrence of seeds in association with Glossopteris fronds recorded by more than one writer[1304], though by no means decisive and possibly the result of chance association, is favourable to this view. Dr White[1305] has suggested that the small leaves described by Zeiller[1306] as Ottokaria bengalensis from Lower Gondwana (Permo-Carboniferous) rocks of India, and similar fossils recorded by himself from Brazil as O. ovalis, may represent “sporangiferous” organs of Glossopteris or Gangamopteris, “both of which are probably pteridospermic.” There is, however, no conclusive evidence in support of this suggestion.

The genus, whatever its position may be, has a special interest for the geologist and for the student of plant distribution; it is a characteristic member of a Permo-Carboniferous flora which flourished over an enormous area, including India, South Africa,—extending from Cape Colony to Rhodesia and German East Africa[1307],—Australia, and South America[1308]. This flora, known as the Glossopteris flora, differed considerably in its component genera from that which overspread Europe and North America and some more southern regions in the Upper Carboniferous and Permian periods.

The discovery by Amalitzky[1309] of Glossopteris, and other genera characteristic of the Glossopteris flora, in the Upper Permian rocks in Vologda (Russia) demonstrates the existence of a northern outpost of the southern botanical province, and Zeiller’s discovery of the genus in the Rhaetic flora of Tonkin[1310] shows that Glossopteris persisted beyond the limits of the Palaeozoic epoch. Dr David White[1311] has recently proposed to re-christen the Glossopteris flora the Gangamopteris flora on the ground that Gangamopteris is strictly Palaeozoic in its range, whereas Glossopteris persisted into the Mesozoic era; this is perhaps hardly a sufficient reason for giving up so well established a title as the Glossopteris flora. A fuller account of this southern flora must be reserved for another volume.

Glossopteris Browniana, Brongniart[1312]. Figs. 334–36.

The specific name Browniana is now applied to obtusely pointed leaves which sometimes reach a length of 15 cm., but are usually rather shorter. In form and venation they closely resemble the leaves of the recent genus Antrophyum and species of Acrostichum. The comparatively broad midrib may be replaced in its proximal portion by several parallel veins: from it are given off numerous lateral veins which form a reticulum characterised by meshes approximately equal in size and elongated in a direction parallel to the general course of the secondary veins (fig. 334).

The drawings, originally published by Zeiller[1313], reproduced in fig. 335 illustrate the venation and its range of variation; the meshes are usually hexagonal and arranged as shown in figs. A and B, but occasionally (fig. 335, C) they follow a more steeply inclined course.

Small leaves with a more or less distinct midrib, 2–3 cm. in length, supply transitional stages between foliage- and scale-leaves. In the true scale-leaves spreading and occasionally anastomosing veins take the place of the midrib and lateral veins of the ordinary frond. McCoy[1314] in describing some Australian specimens of Glossopteris in 1847 spoke of scale-like appendages of the rhizome which he compared with the large ramenta of Acrostichum and other ferns. It was, however, Zeiller[1315] who first recognised the leaf-nature of these scales and adequately described them; additional figures of scale-leaves have been published by Mr Arber[1316] and by myself[1317]. The importance of these small leaves has been considerably increased by Mr Arber’s discovery of associated sporangia which, as he suggests, were probably borne on their lower concave surface.

The sporangia (fig. 336) are compared by Arber with the microsporangia of recent Cycads and with the Palaeozoic sporangia described by Zeiller as Discopteris Rallii (fig. 256, D); the latter are distinguished by the well-defined group of thicker walled cells representing the annulus of true fern sporangia. We know nothing as to the contents of the Glossopteris sporangia, whether they contained microspores or whether they are the spore-capsules of a homosporous plant.

The rhizome of Glossopteris Browniana has been described in detail by Zeiller, who first demonstrated that the fossils originally assigned by Royle[1318] to the genus Vertebraria represent the stem of this and, as we now know, of some other species of Glossopteris. Vertebraria occurs in abundance in Permo-Carboniferous strata in association with Glossopteris; the differences between Australian, Indian, and South forms, though expressed by specific names, are insignificant. The stems are usually preserved in the form of flattened, single or branched, axes sometimes bearing slender branched roots and characterised by one or two, or less frequently three, longitudinal grooves or ridges (fig. 337) from which lateral grooves or ridges are given off at right angles, dividing the surface into more or less rectangular areas 1 cm. or more in length. The surface of these areas is often slightly convex and in some specimens the outlines of cells may be detected. Mr Oldham has described some interesting examples of Vertebraria from India in which the longitudinal and transverse grooves are occupied by a dark brown ferruginous substance or by the carbonised remains of plant-tissues (fig. 338, C, D). In transverse section, a Vertebraria cast appears to be divided into a number of wedge-shaped segments radiating from a common centre. Prof. Zeiller[1319] has figured specimens of Vertebraria with portions of Glossopteris fronds still attached.

The rhizome of Glossopteris, as represented by the Vertebraria casts, is aptly compared by Zeiller[1320] with that of the recent Polypodiaceous fern Onoclea struthiopteris. Sections of the recent stem (fig. 338, E, F) show that the form is irregularly stellate owing to the presence of prominent wings which anastomose laterally at intervals as shown by the examination of a series of sections. The leaf-traces are derived from the steles of adjacent wings. Fig. 338 (B and A) represents somewhat diagrammatically a longitudinal and transverse view of a Vertebraria; the radiating arms represented in the transverse section (fig. A) are the stem ribs or wings and the segments between them are intrusions of sedimentary material. The rectangular areas characteristic of the surface of a Vertebraria are the intruded segments of rock: these are separated at intervals by transverse grooves, which mark the course of vascular strands given off at each anastomosis of the longitudinal wings to supply the leaves.

Mr Oldham, who discovered the connexion between Glossopteris and Vertebraria independently of Dr Zeiller, does not agree with the interpretation of the structural features of the rhizome which Zeiller bases on a comparison between Vertebraria and Onoclea struthiopteris. Oldham[1321] describes Vertebraria as consisting of a central axis “joined to an outer rind by a series of radial septa,” the spaces between the septa being divided into chambers by transverse partitions. His view is that the rhizome of Glossopteris was a cylindrical organ and not an irregularly winged axis like the stem of Onoclea. Zeiller[1322] has replied in detail to Oldham’s interpretation and adheres to his original view, that the rhizome consisted of a solid axis with radial wings or flanges which at intervals anastomosed transversely in pairs at the nodes. It may, however, be possible that the spaces between the longitudinal and transverse grooves on a Vertebraria axis, which have been filled with the surrounding rock, were originally occupied in part at least by secondary wood, and the transverse strips of carbonaceous material[1323] lying in the grooves may represent medullary-ray tissue and accompanying leaf-traces. The longitudinal striations seen in some specimens of Vertebraria on the areas between the grooves may be the impressions of woody tissue. It is impossible without the aid of more perfectly preserved material to arrive at a satisfactory conception of the structural features of a complete Glossopteris rhizome.