CHAPTER XXVI

.

HYDROPTERIDEAE ╭ Marsiliaceae. ╰ Salviniaceae.

The unsatisfactory and meagre records in regard to the past history of these heterosporous Filicales render superfluous more than a brief reference to the recent species.

Marsiliaceae.

This family is usually spoken of as including the two genera _Marsilia_ and _Pilularia_. Lindman[1202] has however founded a third genus, _Regnellidium_, on a Brazilian plant which is distinguished by some well-defined characters from all species of _Marsilia_. The members of the Marsiliaceae live for the most part in swampy situations. _Marsilia_ is represented in Europe by _M. quadrifoliata_ L. which occurs in Portugal, France, Germany and other parts of the Continent, extending also to Kashmir, Northern China, and Japan. Of the other 53 species, 17 are recorded from different regions in Africa, while others occur in South America, Asia[1203], Australia, and elsewhere.

_Pilularia globulifera_ L. is the only British representative of the Hydropterideae. The remaining four species of the genus occur in South America, California, New Zealand, Australia, and _P. minuta_ Dur. is met with in the South of France, Algeria, and Asia Minor in subtropical or warm temperate regions.

The Marsiliaceae are regarded as more nearly related to the Schizaeaceae than to any other family of homosporous ferns[1204]. Their heterospory, the production of sporangia in closed fruit-like sporocarps, and the anatomical features associated with existence in marshy habitats, tend to obscure the resemblances to the true ferns.

• • • • •

The genus _Marsilidium_ proposed by Schenk[1205] for a piece of an axis, bearing apparently a whorl of six leaflets, from the Wealden of Osterwald, cannot be regarded as satisfactory evidence of the existence of the Marsiliaceae in the Wealden flora of North Germany.

The six leaflets of _Marsilidium speciosum_, having a length of 5 cm., are similar in shape to the four leaflets of recent species of _Marsilia_, but they differ in the repeated dichotomy of the veins from the reticulate venation of the recent forms. It is worthy of note, however, that in Lindman’s Brazilian type _Regnellidium diphyllum_ (fig. 326, A), the leaflets are characterised by dichotomous and not by anastomosing veins.

Hollick[1206] has described some impressions of imperfect orbicular leaves with a “finely flabellate obscurely reticulated(?) venation” from Cretaceous rocks of Long Island as _Marsilia Andersoni_, but these are too fragmentary to be accorded this generic designation. My friend Dr Krasser informs me that he is describing some well-preserved leaves from Cretaceous beds of Grünbach in Lower Austria as _Marsilia Nathorsti_[1207]. He compares these with the recent form _Marsilia elata_, a variety of _M. Drummondi_.

Another Lower Cretaceous species _Marsilia perucensis_ has been figured by Frič and Bayer[1208] as a stalked fruit-like body from Bohemia. This was originally described by Velenovský as _M. cretacea_, but under this name Heer[1209] had previously recorded a supposed sporocarp from Greenland. These fossils have little claim to recognition as examples of Marsiliaceous plants.

The fragment figured by Heer[1210] from Tertiary rocks of Oeningen as _Pilularia pedunculata_ is too small to determine with reasonable accuracy. Other supposed representatives of the family mentioned in palaeobotanical literature are not of sufficient importance to describe.

Salviniaceae.

The two genera of Salviniaceae, _Salvinia_ and _Azolla_, are water plants, and are usually described as annuals which survive the less favourable season in the form of detached sporocarps. Goebel[1211] states that all the tropical species of _Salvinia_ known to him have an unlimited existence.

_Salvinia natans_, Hoffm., the only European species, extends from the South of France to Northern China and the plains of India: the other twelve species are mostly tropical. _Azolla_, represented by four species, occurs in Western and Southern North America, South America, Madagascar, Australia, New Zealand, and is widely spread in tropical Asia and Africa.

Species of _Azolla_ frequently form a considerable proportion of the floating carpet of vegetation on inland waters[1212] growing under conditions which might be supposed favourable for preservation in a fossil state.

The Salviniaceae, though probably rather farther removed than the Marsiliaceae from the homosporous Filicineae, are considered by Bower[1213] to be related to the Gradatae, but modified in consequence of their aquatic habit and the assumption of heterospory.

No undoubted examples of fossil species of _Azolla_ have been described. _Salvinia_, on the other hand, is represented by several Tertiary species, for the most part founded on leaves only, and Hollick[1214], who published a list of fossil Salvinias, has described detached leaves as _Salvinia elliptica_ Newb. from what may be Upper Cretaceous rocks from Carbonado, Washington. Some of the leaves figured as Tertiary Salvinias are of no value as evidence of the former distribution of the genus[1215].

From the Coal-beds of Yen-Bäi (Tonkin), probably of Miocene age, Zeiller[1216] has figured some well-preserved impressions of oval or orbicular leaves, 15 mm. long and 10–20 mm. broad, characterised by reticulate venation and by cordate bases, which he refers to Heer’s Swiss species _Salvinia formosa_[1217].

Dr Zeiller[1218] in the most recently published part of his series of valuable résumés of palaeobotanical literature refers to a description by Brabenec of specimens of this species from Bohemian Tertiary beds showing both microspores and megaspores.

One of the most complete specimens so far discovered has recently been described by Fritel[1219] from Eocene beds of the Paris Basin as _Salvinia Zeilleri_. This species, founded on portions of stems bearing floating leaves, submerged root-like leaves, and sporocarps, is compared with a recent tropical American species _S. auriculata_.

It is noteworthy that no authentic records of Hydropterideae have been discovered in Palaeozoic rocks[1220]. Comparisons have been made in the case of the genera _Traquairia_ Carr. and _Sporocarpon_ Will. with the reproductive organs of _Azolla_[1221], but these rest on a wholly insufficient basis.

Dawson[1222] proposed the generic name _Protosalvinia_ for some spores of Devonian age, which he regarded on inadequate grounds as evidence of Palaeozoic Hydropterideae.

Zeiller[1223], in discussing the possible relationships of the problematical type _Chorionopteris gleichenioides_ Cord., suggests a possible alliance with the Hydropterideae. Corda founded the genus _Chorionopteris_[1224] on some small fragments of pinnules, 6–7 mm. long, found in the Carboniferous rocks of Radnitz in Bohemia.

The lobes of the pinnules are incurved distally to form a capsule, containing four sporangia, which apparently opened on dehiscence into four valves; the spores are of one size. The material is however insufficient for accurate determination.

There is no evidence contributed by fossil records which indicates a high antiquity for the Hydropterideae. It is unsafe to base any conclusion on the absence of undoubted Palaeozoic representatives of this group; but the almost complete absence of records in pre-Tertiary strata is a fact which may be allowed some weight in regard to the possible evolution of the heterosporous filicales at a comparatively late period in the earth’s history.

A description of the Mesozoic genus _Sagenopteris_ may be conveniently included in this chapter, though as in many other instances the inclusion of a genus under the heading of a recent family name does not by any means imply that the position of the extinct type is regarded as settled.

Sagenopteris.

This generic name was applied by Presl[1225] to small fronds composed of four or rarely two palmately disposed leaflets with a more or less distinct midrib and anastomosing secondary veins. Schimper[1226] compared _Sagenopteris_ with _Marsilia_, but did not regard the resemblance as evidence of relationship. Nathorst[1227] expressed the opinion that certain fruit-like bodies obtained from the Rhaetic beds of Scania are of the nature of sporocarps and were borne by _Sagenopteris_, with the leaves of which they were associated. He published a drawing of part of a fruit showing on its partially flattened surface some raised oval bodies which are considered to be spores. Dr Nathorst kindly placed at my disposal the drawings reproduced in fig. 325 made from some of his specimens found at Bjuf in Scania.

In contour and superficial features, e.g. the veining on the wall, these bodies bear a fairly close resemblance to the sporocarps of recent species of _Marsilia_. They were found in association with the leaves of _Sagenopteris undulata_ Nath., an abundant Scania type similar in form to the English Jurassic species _S. Phillipsi_ (figs. 327, 328). Heer was independently led by an examination of some examples of the Swedish “fruits” to compare them with the sporocarps of _Marsilia_. A small spherical body is figured by Zigno[1228] close to a leaf of his species _S. angustifolia_, which may be a sporocarp. In a recent paper, Salfeld[1229] says that he found fructification on the lower face of the leaflets of _S. Nilssoniana_ Brongn. from German Jurassic rocks, but he brings forward no evidence in support of this statement. The systematic position of _Sagenopteris_ is by no means settled. In a previous account of the genus I expressed the view that it is probably a member of the true ferns[1230], but the resemblance of Dr Nathorst’s drawings to the Marsilian sporocarps influences me in favour of his opinion that _Sagenopteris_ may belong to the Hydropterideae. The evidence, as Solms-Laubach[1231] states, is not wholly satisfactory: Schenk points out that the frequent occurrence of detached _Sagenopteris_ leaflets suggests that they easily fell off the petiole, whereas in _Marsilia_ the leaflets do not fall off independently. The discovery of a new type of Marsiliaceae in Brazil, which Lindman has described as _Regnellidium diphyllum_[1232] (fig. 326, A), affords an additional piece of evidence bearing on the comparison of _Sagenopteris_ with members of this family. In _Regnellidium_ the leaves differ from those of _Marsilia_ in bearing two instead of four leaflets, and in the former the veins are repeatedly forked, and do not anastomose as in _Marsilia_. In the possession of only two leaflets _Regnellidium_ agrees with some forms of _Sagenopteris_ (fig. 328).

[Illustration: FIG. 325. Sporocarp-like bodies found in association with the leaves of _Sagenopteris_. (Nat. size. From drawings supplied by Dr Nathorst.)]

[Illustration: FIG. 326.

A. _Regnellidium diphyllum_ Lind. Single leaf and stalked sporocarp. (⅞ nat. size. After Lindman.) B. Cuticle of _Sagenopteris rhoifolia_. (After Schenk.)]

_Sagenopteris Phillipsi_ (Brongniart)[1233]. Figs. 327, 328.

1828. _Glossopteris Phillipsi_, Brongniart, Hist. vég. foss. p. 225, Pls. LXI. _bis_, LXIII.

1838. _Sagenopteris Phillipsi_, Presl, in Sternberg’s Flor. Vorwelt, vii. p. 69.

[Illustration: FIG. 327. _Sagenopteris Phillipsi._

A. From the type-specimens of Lindley and Hutton (_Glossopteris Phillipsi_). Gristhorpe Bay, Yorkshire. British Museum, No. 39221. Slightly reduced. M.S. B. From a specimen in the British Museum (39222). Nat. size. Figured by Lindley and Hutton as _Glossopteris Phillipsi_.]

The fronds of this common Jurassic species, which is recorded from many European localities, from North America, Australia, the Antarctic regions[1234], and elsewhere, are very variable as regards the form, size, and number of the leaflets.

Frond petiolate, in some forms the petiole bears four linear or oval-lanceolate leaflets having a distinct midrib and oblique anastomosing veins. In others a shorter winged petiole bears one or two shorter and broader, somewhat obcuneate, leaflets without a midrib.

It is probable that Bunbury[1235] was correct in his opinion that the specimen figured by Lindley and Hutton[1236] as _Otopteris cuneata_, characterised by two leaflets (fig. 328), is not specifically distinct from the normal form with four leaflets (fig. 327).

Similarly, such specimens as that represented in Pl. XVIII., fig. 3 of the first part of my _Jurassic Flora_, in which a short stalk bears only one leaflet may, provisionally at least, be included in Brongniart’s species. Yabe[1237] describes a form with two leaflets from Jurassic rocks of Korea as _Sagenopteris bilobata_ which resembles _S. Phillipsi_; and Moeller[1238] records a specimen similar to that represented in fig. 328 from Bornholm as _S. cuneata_ (Lind. and Hutt.).

[Illustration: FIG. 328. _Sagenopteris Phillipsi._ From a specimen in the Manchester University Museum. Nat. size.]

The leaf shown in fig. 327, A, in which the longest segments are 4·5 cm. in length, represents the most abundant form and illustrates the very close agreement between _S. Phillipsi_ and the Rhaetic species _S. rhoifolia_. Fig. 327, B, which is drawn from a specimen figured by Lindley and Hutton[1239], shows a leaf with longer (6·5 cm.) and much narrower segments. Broader leaflets are occasionally met with in which the lamina reaches a length of 11 cm.[1240]

Leaves with leaflets narrower (3 mm. broad) than those represented in fig. 327, B, are described by Zigno[1241] from Jurassic beds of Italy as _S. angustifolia_ and by Moeller[1242] from the Jurassic of Bornholm as _S. Phillipsi_ f. _pusilla_. A coarser type of venation than that of _S. Phillipsi_ is occasionally found in Jurassic examples, as in _S. grandifolia_ Font.[1243] from Oregon and _S. Nathorsti_ Barth. from Bornholm[1244].

• • • • •

_Sagenopteris_ is recorded also from several Rhaetic floras. The best known species, _S. rhoifolia_ Presl[1245], is hardly distinguishable from some forms of _S. Phillipsi_ or from the Italian Jurassic species described by Zigno as _S. Goeppertiana_[1246], though the leaflets are usually rather larger. This species was first described by Brongniart as _Filicites Nilssoniana_[1247], and a few authors[1248] have adopted this specific name because of its priority over Presl’s designation. As Nathorst remarks, to give up the well-known name _S. rhoifolia_ for _S. Nilssoniana_ is “mere pedantry.” The epidermis of _S. rhoifolia_ as figured by Schenk[1249] consists of cells with straight and not undulating walls: stomata occur on the lower surface (fig. 326, B).

Rhaetic leaves of the type represented by _S. rhoifolia_ have a wide geographical distribution.

The specimens described by Feistmantel from the Damuda series of India as _Sagenopteris longifolia_ are no doubt fronds of _Glossopteris longifolia_[1250].

The Wealden species _Sagenopteris Mantelli_ (Dunk.)[1251] agrees closely in habit and in the form of the leaflets with _S. Phillipsi_ and _S. rhoifolia_. It is probable that some of the leaves described by Velenovský[1252] from Lower Cretaceous rocks in Bohemia as _Thinnfeldia variabilis_ are portions of _Sagenopteris_ fronds. _S. Mantelli_ is recorded from several European localities, from California[1253], and elsewhere.

_Sagenopteris_ appears to have been widely distributed during the Rhaetic, Jurassic and Lower Cretaceous floras. The very great similarity between the specimens recorded from these three formations renders the genus an uncertain guide in regard to geological age. Decisive evidence as to its position in the plant kingdom is at present lacking: the inclusion of the genus as a possible member of the Hydropterideae has still to be justified.

##