Fossil plants, Vol. 2

The filiform appendages on the petioles of the recent fern Hemitelia capensis (fig. 235, p. 304) have often been compared with the aphlebioid leaflets of fossil fronds.

Potonié who has discussed the nature of Aphlebiae regards them as vestiges of a once continuous lamina, which formed a winged border to the branched axes of more primitive forms of fronds. It is possible that the pinnules between the pinnae on the rachis of Archaeopteris and the Cyclopteroid leaflets of Neuropteris and Odontopteris may have the morphological significance attributed to them by Potonié. In some cases it is probable that the Aphlebiae, whether vestiges or not, served the purpose of protecting either the whole frond or individual pinnae. Aphlebiae, though especially characteristic of Palaeozoic leaves, are occasionally met with in the form of modified pinnules at the base of the primary pinnae on Mesozoic ferns, e.g. in Coniopteris hymenophylloides[1401].

In some fern fronds the lowest pinnule of each pinna differs in shape or size from the normal ultimate segments, but it would be almost affectation to extend the use of the term Aphlebia to such pinnules. The Jurassic species Cladophlebis lobifolia (Phill.) is a case in point[1402]. In this fern, which some authors speak of, without sufficient reason, as Dicksonia lobifolia[1403], the lowest pinnule is large and different in shape from the others.

Sphenopteris.

Sphenopteris is one of the many generic names which we owe to Brongniart[1404]. It is the generic designation used for a great number of Palaeozoic and later fronds, most of which are those of true ferns while some Palaeozoic species are undoubted Pteridosperms. The genus, which is purely provisional, includes members of widely different families possessing pinnules of the same general type, such as is represented in some recent species of Davallia, Asplenium, and other ferns.

The fronds of Sphenopteris may be bipinnate, tripinnate, or quadripinnate; the rachis may be dichotomously branched or the branching may be of the pinnate type characteristic of most recent ferns. The pinnules are small; they vary considerably in shape even in a single frond, but the chief characteristics are: the lobed lamina, contracted and often wedge-shaped at the base (fig. 352), the dichotomously branched veins radiating from the base or given off from a median rib at an acute angle. The lamina may be divided into a few bluntly rounded lobes (fig. 352, C) or deeply dissected into linear or cuneate segments (fig. 352, A, B, E).

Examples of Sphenopteroid leaves have already been described under the genera Coniopteris, Onychiopsis, Ruffordia, etc. Among the numerous examples of Sphenopteris species from the Carboniferous rocks mention may be made of Sphenopteris obtusiloba Brogn.[1405] (fig. 352, A–C), which occurs in the Middle and Lower Coal-Measures of Britain[1406]. This type is characterised by the almost orbicular, oval or triangular pinnules which may reach a length of 15 mm.; they are occasionally entire, but more usually divided into 3 to 5 rounded lobes. The forked veins radiate from the base of the pinnule. The rachis may be dichotomously branched. Fructification unknown.

The species S. furcata Brongn.[1407], characteristic of the Middle and Lower Coal-Measures of Britain (fig. 352, E), is referred to under Stur’s genus Diplotmema[1408] in which it is included by some authors solely because of the dichotomous habit of branching of the pinnae.

The pinna represented in fig. 353 illustrates a similar type of pinnule. This species, which is very common in the Calciferous Sandstone of Scotland, was described by Lindley and Hutton as Sphenopteris affinis[1409].

The fronds of Sphenopteris affinis were discovered by Mr Peach[1410] in a fertile condition, but he regarded the reproductive organs as those of a plant parasitic on the Sphenopteris fronds. Kidston[1411] substituted Stur’s genus Calymmatotheca for Sphenopteris on the ground that the sporangia figured by Peach under the name Staphylopteris Peachii bear a close resemblance to the organs which Stur described as valves of an indusium in his species Calymmatotheca Stangeri[1412]. An examination of Stur’s specimens by Miss Benson[1413] and by Prof. Oliver and Dr Scott has confirmed Stur’s interpretation of the appendages at the tips of the fertile pinnae as valves of an indusial or cupular structure. The superficially similar bodies on the fertile pinnae of S. affinis are however true sporangia, and cannot legitimately be included in the genus Calymmatotheca as described by Stur. For this reason Miss Benson institutes a new genus Telangium, the type-species of which, T. Scotti from the Lower Coal-Measures of Lancashire, is based on petrified material. The Scotch species Sphenopteris affinis (= Calymmatotheca affinis of Kidston) is also transferred to Telangium; the sporangia are considered by Miss Benson to be microsporangia. This with other species is no doubt correctly included in the Pteridosperms. A complete frond of Sphenopteris affinis, showing a regular dichotomy of the main axes, is represented by an admirable drawing in Hugh Miller’s Testimony of the Rocks[1414].

Some of the Palaeozoic species of Sphenopteris probably represent the fronds of true ferns, but others are known to have been borne by Pteridosperms. S. Hoeninghausi (fig. 290, C, p. 399) is the foliage of Lyginodendron, and Scott[1415] speaks of three species, S. dissecta, S. elegans, and S. Linkii as the leaves of Heterangium. Grand’Eury[1416] has recorded the occurrence in French Coal-Measures of seeds in association with other Sphenopteroid fronds.

Mariopteris, Diplotmema, Palmatopteris.

The discovery of sporangia on the fronds of several Palaeozoic species of Sphenopteris and Pecopteris has led to the institution of new generic names, which indicate an advance in knowledge beyond the stage implied by the use of those provisional designations based solely on the form and venation of the pinnules. Other names have been created by authors in place of Sphenopteris and Pecopteris on the ground that a striking feature in the mode of branching of fronds is sufficiently important to justify generic recognition even in the absence of fertile specimens. As examples of designations based primarily on the branch-system of compound leaves, the genera Mariopteris, Diplotmema, and Palmatopteris may be briefly considered (fig. 354 A–C). Dr Kidston[1417] is of opinion that the creation of new genera for purely vegetative characters of fronds is of no real advantage, and he prefers to retain the older provisional names for species known only in the sterile condition. On the other hand, if we are sufficiently familiar with specimens large enough to enable us to recognise a well-defined morphological character, it may serve a useful purpose to employ a generic designation for features which may have a phylogenetic value. A comparative examination of Palaeozoic, Mesozoic, and recent compound fronds, including both Pteridosperms and true ferns, brings to light certain distinguishing features characteristic of the older types which, as Potonié maintains[1418], point to the derivation of the pinnate habit from a primitive dichotomous system of branching. For a more complete discussion of this question reference should be made to Potonié’s suggestive papers. Among recent ferns Matonia and Dipteris, two survivals from the past, afford instances of fronds with a branching system of the dichotomous type.

Similarly, in Gleichenia, Lygodium, and more rarely in species of Polypodiaceae (e.g. Davallia aculeata, fig. 232) dichotomy is a striking feature of the fronds. In the great majority of recent ferns the fronds have assumed a pinnate habit. Among Palaeozoic fern-like fronds dichotomous branching of the main rachis and of the pinnae is much more common. Potonié draws attention to several other features which distinguish Palaeozoic fronds from the majority of later species: the frequent occurrence of pinnules borne directly on the main rachis (fig. 354, D), and of modified pinnules or Aphlebiae on the rachis and petiole, are characters to which he attributes an evolutionary significance. The main point is that a comparative examination of leaf-form affords evidence in favour of the view that the modern type of frond, with its naked rachis bearing two rows of pinnae, has been derived from a less specialised type in which the distinction between the parts of the leaf is much less evident. The primitive leaf was probably a dichotomously branched axis provided with a continuous lamina which eventually became broken up into separate lobes or pinnules.

As the dichotomy of the frond became less regular, a pinnate habit was acquired, as is clearly seen in many Palaeozoic types which constitute connecting links between forked and pinnate fronds (fig. 354, D). The Aphlebiae may be remnants of the once-continuous lamina on the petiole, and the normal pinnules borne on the rachis may be regarded as the attributes of fronds in which the division of physiological labour had not reached the stage which characterises the leaves of recent ferns.

Mariopteris.

This name, which is due to Zeiller[1419], is applied by him to Palaeozoic fronds characterised by a double bifurcation of the rachis of the primary pinnae. Mariopteris muricata (= Pecopteris muricata Schloth.) may be taken as the type of the genus. This species is common in the Lower and Middle Coal-Measures of Britain and rare in the Upper Coal-Measures[1420]. It is described by Kidston[1421] as one of the most polymorphic and widely distributed Coal-Measure species. The pinnules as seen in fig. 364, B, are of the Sphenopteroid type. No fertile specimens are known, but it is significant that Grand’Eury[1422] has recorded the association of Mariopteris muricata and seeds.

The main rachis gives off alternate naked branches, each of which bifurcates at its apex into two short naked axes, and these are again forked, the ultimate branches having the form of bipinnate pinnae provided with large Sphenopteroid pinnules (fig. 354, B). Zeiller includes in Mariopteris some species which Stur[1423] referred to his genus Diplotmema. Possibly some of the Palaeozoic fronds with a zigzag rachis may have been climbers like Lygodium.

Diplotmema.

This generic name is employed by Zeiller[1424] and other authors in a more restricted sense than that in which it was originally used by Stur. The Upper Carboniferous species Sphenopteris furcata Brongn. (fig. 352, E) may serve as the type. This species occurs in the Middle and Lower Coal-Measures of Britain[1425]. The main rachis gives off branches as in Mariopteris, but in Diplotmema each naked lateral branch is forked at its apex into two opposite pinnae bearing deeply dissected Sphenopteroid pinnules. Zeiller[1426] and Stur have recorded fertile specimens of Diplotmema, but in no case have actual sporangia been discovered. In the species Diplotmema Zeilleri Stur (fig. 354, C, C′) two Aphlebiae occur at the base of each secondary axis[1427]. It has been pointed out by Potonié that in Diplotmema furcatum the equal dichotomy of the lateral branches is not characteristic of the frond as a whole. In the case of branches higher on the rachis the dichotomy becomes unequal and the forked axis is gradually replaced by a simple pinna (fig. 354, A). For this type of frond, Potonié proposed the generic name Palmatopteris in place of Diplotmema, which he discards. The long comparatively slender rachis of P. furcata suggests comparison with the liane species of Lygodium[1428].

Cephalotheca.

This genus was proposed by Nathorst[1429] for some peculiar bipinnate fertile fronds from the Upper Devonian rocks of Bear Island. The pinnae bear slender forked ultimate segments represented by a few detached fragments (fig. 355, B), associated with the rachises. The fertile pinnae are given off in opposite pairs from the main axis over which they are concrescent (fig. 355, A). A mop-like cluster of sporangia is borne on the lower surface and close to the base of a fertile pinna: the exannulate sporangia are compared with those of Scolecopteris. Nathorst compares Cephalotheca with a Belgian species of Upper Devonian age described by Crépin[1430] as Rhacophyton condrusorum and by Gilkinet[1431] as Sphenopteris condrusorum. A similar fossil is also described by Baily[1432] as Filicites lineatus from the Kitorkan Grits of Ireland.

The position of Cephalotheca cannot be definitely determined from the available data, but it is more probable that it was a seed-bearing Pteridosperm and not a true fern. Zeiller[1433] has recently expressed the same opinion.

Thinnfeldia.

The genus Thinnfeldia, founded by Ettingshausen in 1852[1434] on some Hungarian Liassic specimens, though frequently included in the Filicales, cannot be said to occupy that position by virtue of any well-authenticated filicinean features. It is by no means improbable that many of the species referred to this genus are closely allied to Palaeozoic Pteridosperms.

Thinnfeldia may be briefly defined as follows:

Fronds simple and pinnatifid, pinnate or bipinnate: rachis broad and occasionally dichotomously branched. Pinnules often fleshy or coriaceous; broadly linear, entire or lobed, provided with a midrib from which simple or forked secondary veins are given off at an acute angle: or the laminae may be short and broad without a midrib and traversed by several slightly divergent and forked veins.

No satisfactory evidence of reproductive organs has so far been adduced.

The genus is chiefly characteristic of Upper Triassic, Rhaetic, and Jurassic floras, though it was in all probability represented in Permian floras. Several species, many of which are valueless, are recorded also from Cretaceous and Tertiary formations. Search should be made for fertile specimens or for evidence as to the association of seeds with Thinnfeldia fronds.

Some Permian fossils from Kansas which Sellards[1435] has made the type of a new genus, Glenopteris, appear to be indistinguishable generically from leaves of Lower Mesozoic age universally recognised as typical examples of Thinnfeldia.

Thinnfeldia odontopteroides (Morris)[1436]. Figs. 356–358.

This is a very variable species as regards the shape and size of the ultimate segments and their venation. It is a type of extended geographical range characteristic of Rhaetic or Upper Triassic rocks in Australia, South Africa, India, South America, and various European localities.

Frond bipinnate; the broad rachis may be dichotomously branched. Pinnules with a thick lamina which may be almost semicircular in form, deltoid, broadly oval or broadly linear, and often confluent at the base. Short and broad pinnules occur on some fronds directly attached to the main rachis between the pinnae. The longer and narrower pinnules (fig. 356, C), resembling those of the Palaeozoic genus Alethopteris, have a well-defined midrib, while the smaller segments are characterised by several slightly divergent veins which spring directly from the rachis (fig. 356, A). Epidermal cells polygonal or, above the veins, rectangular in shape; stomata, which are slightly sunk, occur on both the upper and lower epidermis. Fertile specimens unknown.

The portion of a lobed pinnule shown in fig. 356, B, illustrates a form of segment intermediate between the linear type with a midrib and a row of shorter pinnules without a median vein. Fig. 356, D, represents another instance of variation in the arrangement of the veins in segments of different sizes. Various specific and generic names have been assigned to Thinnfeldia fronds of Rhaetic age on the ground of the occurrence of pinnules longer and narrower than those usually associated with T. odontopteroides; but in view of the range of variation met with in a single leaf it is advisable to extend rather than to restrict the boundary of what we are pleased to regard as a specific type.

The name Thinnfeldia lancifolia has been applied by Morris to fossils from Australia which may be identified with T. odontopteroides, and the same designation is employed by Szajnocha and by Solms-Laubach[1437] for Rhaetic specimens from South America. Similar fronds are described by Geinitz[1438] as Thinnfeldia tenuinervis from Argentine Rhaetic strata. Odontopteris macrophylla Curran, T. falcata Ten.-Woods, Gleichenia lineata Ten.-Woods, and Cardiopteris Zuberi Szaj. afford other examples of what are probably closely allied forms[1439].

Some exceptionally large examples of T. odontopteroides are figured by Feistmantel[1440] from the Hawkesbury series of New South Wales in which the bipinnate frond has a breadth of 25–30 cm. A specimen from the Molteno beds of South Africa, probably of Rhaetic age, represented in fig. 357, illustrates a smaller leaf with pinnules of the linear type, some of which are partially divided into shorter pinnules with forked veins. The example represented in fig. 358, from Cyphergat (S. Africa), shows two equal branches of a rachis with small contiguous segments.

Some specimens figured by Zeiller[1441] from the Rhaetic strata of Tonkin as Pecopteris (Bernouillia?) sp. may be portions of Thinnfeldia fronds, and the large leaves which he refers to Ctenopteris Sarreni differ but slightly from the Australian specimens described by Feistmantel as T. odontopteroides.

Thinnfeldia rhomboidalis, Ettingshausen. Figs. 359, 360, C.

Under this name Ettingshausen described the type-specimen of the genus from Lower Lias strata at Steierdorf in Hungary. He assigned the plant to the Coniferae on the ground of a resemblance of the pinnules to the phylloclades of Phyllocladus. Thinnfeldia rhomboidalis bears a close resemblance to T. odontopteroides, but the pinnules are usually longer and narrower, as shown in the English specimen from the Lower Lias of Dorsetshire represented in fig. 359. The darker margin of the pinnules shown in fig. 360, C, gives the impression of a revolute lamina, but a microscopical examination points to a thicker cuticle at the edge of the segments.

The species is recorded from Jurassic rocks of France, Germany, Italy, India, Australia, and elsewhere[1442].

Palaeobotanical literature contains numerous records of Jurassic, Cretaceous and some Tertiary species referred to Thinnfeldia, but many of these are probably not generically identical with T. odontopteroides or T. rhomboidalis. Mr Berry[1443] in a paper on The American species referred to Thinnfeldia concludes that the genus is “a rather indefinite one ... and badly in need of revision.” He regards the Middle and Upper Cretaceous American species as Conifers related to Phyllocladus and probably forming a link between the Podocarpeae and Taxeae: for these forms he proposes the generic name Protophyllocladus. The opinion has been expressed elsewhere[1444] that this “problematical[1445]” genus rests on an unsatisfactory basis; the available data do not justify the use of a name which implies the existence in North American Cretaceous floras of a type related to the New Zealand and Tasmanian Conifer Phyllocladus. We are not in a position to assign a single species of Thinnfeldia to the Filicales or the Gymnosperms.

A leaflet from Jurassic rocks of Poland figured by Raciborski[1446] shows what this author regards as the impression of a circular sorus: no sporangia have been found. A specimen in the British Museum[1447], which is said to come from Rhaetic beds in Queensland, shows a row of contiguous polygonal prominences on each side of the midrib which resemble the sori of a fern; but until sporangia are discovered we cannot determine the precise nature of this apparently fertile frond.

A species described by Fontaine[1448] from the Potomac beds (Wealden-Jurassic) of North America as Thinnfeldia variabilis affords a good example of a plant which cannot be identified with any degree of confidence either as a fern or a seed-bearing type. Mr Berry draws attention to the former application of this name by Velenovský to a distinct Lower Cretaceous Bohemian species and proposes for Fontaine’s plant the name T. Fontainei; he maintains that no one has doubted the fern-nature of the Potomac plant. T. variabilis may indeed be a fern, but the evidence is not such as to preclude legitimate doubts as to the correctness of this suggestion. Solms-Laubach[1449], in referring to Schenk’s view that Thinnfeldia and its allies may represent a group intermediate between Ferns and Gymnosperms, admits that it is a possible supposition; he is, however, inclined to consider Lomatopteris and Cycadopteris, “genera especially comparable with Thinnfeldia” as more probably ferns.

At this point we may conveniently consider a series of genera which occupy an equally uncertain position and bear a very close resemblance to Thinnfeldia.

Lomatopteris.

The generic name Lomatopteris was proposed by Schimper[1450] for some bipinnate fronds originally described by Kurr[1451] from Jurassic rocks of Württemberg as Odontopteris (?) jurensis (fig. 360, A). I have elsewhere expressed the opinion[1452] that this German species may be identical with Thinnfeldia rhomboidalis Ett. Kurr’s type-specimen, a portion of which is reproduced in fig. 360, A, consists of a frond or large pinna characterised by a prominent and broad rachis giving off alternate linear pinnae bearing bluntly rounded, contiguous and basally concrescent pinnules having a thick or revolute border and a central rib. The lateral veins are visible in the ultimate segments of Kurr’s fossil. Saporta[1453] has described several species, which he refers to Schimper’s genus, from French Jurassic strata: it is, however, difficult to recognise some of the examples represented in his illustrations as specifically distinct forms. This author notices the resemblance of Lomatopteris to Thinnfeldia, not only in habit but in the structure of the epidermal cells[1454]. In Lomatopteris and in Thinnfeldia the cells have straight and not sinuous walls and the slightly sunken stomata are surrounded by a ring of epidermal cells. Salfeld[1455] has recently described portions of fronds from Jurassic rocks of South-West Germany, which he identifies as Lomatopteris jurensis. He disagrees with my view that Lomatopteris does not differ sufficiently from Thinnfeldia to be accorded generic autonomy, chiefly on the ground that the folded-over edge of the pinnules is a distinguishing feature of Lomatopteris. There is, however, no difference, in appearance at least, between the leaflets of some species of Thinnfeldia, e.g. T. rhomboidalis from Liassic rocks of England[1456], and those referred to Lomatopteris. In a later paper, Salfeld[1457] describes some Portlandian fragments from North Germany as Lomatopteris Schimperi, identifying them with a Wealden fossil of somewhat doubtful affinity, which Schenk[1458] makes the type of his species. The Portlandian specimens are described as tripinnate, with thick decurrent obtusely terminated pinnules with a revolute edge. The general form of the frond is very similar to that of L. jurensis. Salfeld publishes a photograph of a large specimen which he describes as fertile and a drawing of a piece of a pinna: the latter is reproduced in fig. 360, B. He speaks of sori occurring in two rows, probably attached to lateral veins, in the groove between the midrib and the revolute edge of the lamina. The sporangia are described as “nicht näher bekannt[1459].” An examination of the figures reveals nothing as to the nature of the “sori.” The specimens are considered by Salfeld to afford decisive evidence against the view that Lomatopteris and Thinnfeldia are generically identical. Nothing has so far been published which constitutes a valid argument in favour of retaining Schimper’s generic name.

Cycadopteris.

Zigno[1460] founded the genus Cycadopteris on Italian Jurassic impressions regarded by Schimper as indistinguishable from Lomatopteris. As Solms-Laubach[1461] points out, the supposed sori of Cycadopteris described by Zigno are not convincing. There appear to be no satisfactory reasons for separating Cycadopteris from Lomatopteris, nor do the fronds described under these names exhibit any important differences from Thinnfeldia.

Ptilozamites.

Nathorst[1462] founded this genus on a remarkable series of specimens from the Rhaetic Coal-beds of Scania and assigned it to the Cycadophyta. The species Ptilozamites Heeri may be taken as a representative type. The leaves are linear and simply pinnate. In the example shown on a much reduced scale in fig. 361 the frond is 53 cm. long and 2·1 cm. broad. The upper edge of each pinnule is straight or slightly concave; the lower edge is rounded; the veins are slightly divergent and dichotomously branched (fig. 356, E, p. 539). In some of Nathorst’s specimens the broad rachis is forked as in many Thinnfeldias.

As a comparison of fig. 356, A and E, shows, the pinnules of some specimens of Thinnfeldia odontopteroides are identical with those of Ptilozamites. In the latter genus the rachis is either unbranched or occasionally forked, while in Thinnfeldia the branching may be of the dichotomous or pinnate type. In Ptilozamites the segments appear to be always without a midrib, while a median vein frequently occurs in those of Thinnfeldia. There can be little doubt as to the very close alliance between the Rhaetic species referred to these two genera. The name Ptilozamites should perhaps be retained for such long and narrow fronds as that shown in fig. 361: no species included in Thinnfeldia is known in which the rachis reached so great a length without branching. The habit of Ptilozamites Heeri predisposes one in favour of Nathorst’s opinion that the fronds are Cycadean: we have no information in regard to the nature of the reproductive organs.

Ctenopteris.

This name was instituted by Saporta[1463], at Brongniart’s suggestion, for Liassic species characterised by pinnules like those of Thinnfeldia, but distinguished by the bipinnate habit of the frond. Saporta compares the genus with the Palaeozoic leaves known as Odontopteris, and with Italian Jurassic plants referred by Zigno to his genus Dichopteris.

The name Ctenozamites is applied by Nathorst[1464] to the type of frond which Saporta, Zeiller, and other authors refer to Ctenopteris. Nathorst instituted Ctenozamites for fossils agreeing in the form and venation of the pinnules with his genus Ptilizamites but differing in being bipinnate and not pinnate.

Fronds of Ctenopteris are characteristic of the Jurassic and Rhaetic series; they are known only in the sterile condition. As Zeiller[1465] says, Ctenopteris may be a member of the Cycadofilices, an extinct group founded on Palaeozoic plants combining Cycadean and Filicinean characters, and some of which are now known to be Pteridosperms. It is probable that the genus is not a true fern: it is more likely to be a member of the Cycadophyta or of some generalised extinct group.

Ctenopteris cycadea (Brongniart). Fig. 362.